Abstract

Accurate crop yield forecasting plays a critical role in ensuring national food security, guiding agricultural policy, and informing market strategies. India, with its vast agro-ecological diversity and a population exceeding 1.4 billion, faces unique challenges in aligning production with demand. This research presents a comprehensive multi-model machine learning (ML) framework for predicting crop yields across 30 Indian states, explicitly integrating population density and urban–rural composition as demand-related features. Eleven algorithms are evaluated, including Random Forest, XGBoost, LightGBM, CatBoost, Gradient Boosting, Bagging, AdaBoost, Decision Tree, Extra Trees, K-Nearest Neighbors, and Multi-layer Perceptron. The dataset spans 1997–2020 with 19,689 records, incorporating demographic, climatic, and agronomic variables. Results show ensemble methods outperform individual models, with Random Forest achieving the highest performance (R2 = 0.9803, RMSE = 125.79), followed by Bagging (R2 = 0.9793) and XGBoost (R2 = 0.9766). Population features contributed a modest yet consistent improvement of 0.6% in predictive accuracy, with market accessibility and urban–rural ratio being the most influential. LightGBM exhibited the greatest stability (CV = 0.9679 ± 0.0131), while Random Forest offered the best trade-off between interpretability and accuracy. This study highlights the importance of integrating both supply-and demand-side variables for robust agricultural planning and improved food security.

1 Introduction

Global food systems face increasing pressure from climate change, population growth, and resource constraints. For India, agriculture supports the livelihoods of over half the population and contributes significantly to GDP [7]. Accurate yield forecasting is essential to ensure supply meets demand, optimise resource allocation, and stabilise markets. Traditional statistical models, such as regression and time-series approaches, often fail to capture the non-linear, high-dimensional interactions in agricultural data [5]. ML techniques can model these complex relationships, offering improved accuracy [10]. Most prior Indian studies focus on supply-side factors like rainfall, fertiliser use, and cropping patterns, neglecting demand-side influences such as population density and market accessibility. This study bridges that gap by evaluating 11 ML algorithms while integrating demographic features, aiming for balanced supply-demand yield forecasts.

The research addresses several critical gaps in current agricultural forecasting literature. First, while machine learning has been applied to crop yield prediction globally, comprehensive comparative studies in the Indian context remain limited. Second, the integration of demographic and socio-economic factors with traditional agronomic variables represents a novel approach that captures the complex interplay between agricultural production and human settlement patterns. Third, the evaluation of state-of-the-art gradient boosting algorithms (XGBoost, LightGBM, CatBoost) alongside traditional ensemble methods provides insights into the most effective approaches for Indian agricultural data.

The significance of this research extends beyond academic contribution to practical agricultural planning. With India’s population projected to reach 1.5 billion by 2030, understanding how demographic differences — such as variations in population density, market accessibility, and urban–rural composition — influence agricultural demand and production patterns becomes crucial for food security planning. The integration of population density and urbanization patterns into yield forecasting models enables policymakers to anticipate how these differences affect agricultural demand and adjust production strategies accordingly.

2 Literature Review

2.1 Machine Learning in Agricultural Forecasting

Machine learning methods, particularly tree-based ensembles and neural networks, have shown strong predictive capability in agricultural forecasting [11]. LSTM networks excel in modelling temporal dependencies in sequential agricultural data [9], while hybrid models combining process-based and ML approaches improve generalisability.

The evolution of machine learning in agricultural forecasting has followed several distinct phases. Early applications focused on simple regression models and decision trees, which provided interpretable results but limited predictive accuracy. The introduction of ensemble methods, particularly Random Forest, marked a significant advancement by combining multiple decision trees to reduce variance and improve generalization. More recently, gradient boosting algorithms have demonstrated superior performance in various agricultural applications, with XGBoost, LightGBM, and CatBoost emerging as state-of-the-art solutions.

Recent studies have demonstrated the effectiveness of deep learning approaches in agricultural forecasting. Convolutional Neural Networks (CNNs) have been successfully applied to satellite imagery analysis for crop monitoring, while Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks have shown promise in capturing temporal dependencies in yield data. However, these approaches often require large datasets and extensive computational resources, making them less suitable for regions with limited data availability. Additionally, deep learning models are generally less interpretable and explainable compared to traditional machine learning methods, which poses challenges for stakeholder trust and regulatory compliance in agricultural policy applications

2.2 Ensemble and Gradient Boosting Methods

Ensemble methods aggregate predictions from multiple models to improve accuracy and stability. RF, XGB, Light GBM, and CB have proven effective in agricultural applications, handling non-linearities, noise, and high-dimensional datasets [1, 2, 6, 8].

The theoretical foundation of ensemble methods lies in the principle of combining multiple weak learners to create a strong learner. This approach addresses several limitations of individual models, including overfitting, sensitivity to noise, and limited generalization capability. Random Forest, for instance, constructs multiple decision trees on bootstrapped samples of the training data, reducing variance through averaging while maintaining low bias.

Gradient boosting represents a more sophisticated ensemble approach that builds models sequentially, with each subsequent model focusing on the errors of its predecessors. XGBoost extends this concept with advanced regularization techniques, including L1 and L2 regularization, which help prevent overfitting and improve generalization. LightGBM optimizes the training process through leaf-wise tree growth and histogram based algorithms, making it particularly suitable for large datasets.

CatBoost introduces several innovations, including ordered boosting and native handling of categorical features, which addresses common challenges in agricultural data preprocessing. The algorithm’s robust default settings and reduced sensitivity to hyperparameters make it particularly valuable for practitioners with limited tuning expertise.

2.3 Population Density and Agricultural Productivity

Population density influences agricultural productivity through intensification, infrastructure development, and market access [5]. In India, urban proximity affects crop choice and resource allocation [3].

The relationship between population density and agricultural productivity operates through multiple interconnected mechanisms. First, higher population density typically leads to increased demand for agricultural products, driving intensification of production through improved technology adoption, better irrigation systems, and more efficient resource utilization. Second, population density influences infrastructure development, with more densely populated areas typically having better access to agricultural inputs, markets, and extension services.

Urbanization patterns further complicate this relationship. As rural areas become more urbanized, agricultural land use patterns shift, often leading to more intensive production on remaining agricultural land. Additionally, urban proximity affects crop choice, with farmers near urban centers often shifting toward high-value crops that can be sold in urban markets. This phenomenon, known as the “urbanization effect,” has been documented in various developing countries and represents an important consideration for agricultural planning.

Market accessibility, closely related to population density and urbanization, plays a crucial role in determining agricultural productivity. Areas with better market access typically have higher agricultural productivity due to improved input availability, better price information, and reduced transaction costs. The integration of market accessibility metrics into yield forecasting models represents a significant advancement in capturing the full spectrum of factors influencing agricultural productivity.

2.4 Indian Agricultural Context and Challenges

India’s agricultural sector faces unique challenges that make accurate yield forecasting particularly important. The country’s diverse agro-climatic zones, ranging from tropical to temperate regions, create significant variations in crop suitability and productivity. Additionally, India’s agricultural system is characterized by small landholdings, with approximately 86% of farmers operating on less than 2 hectares of land. This fragmentation presents challenges for data collection and analysis, as well as for the implementation of forecasting-based policies.

Climate change poses additional challenges for Indian agriculture, with increasing variability in rainfall patterns, rising temperatures, and more frequent extreme weather events. These changes affect both crop yields and the reliability of historical data for forecasting purposes. The integration of climate variables into yield forecasting models becomes increasingly important as these patterns continue to evolve.

2.5 Related Work Summary Table

These studies were selected based on their direct relevance to our research objectives and methodological approach. Van [10] and Ghar et al. [4] influenced our decision to conduct a comprehensive multi-algorithm comparison, while de et al. [3] specifically guided our 4 integration of demographic features into crop yield forecasting models for the Indian context. The foundational algorithm papers [1, 2, 6, 8] shaped our understanding of ensemble methods and guided our hyperparameter tuning strategies. Sharma et al.’s work [9] on Indian agricultural data provided important benchmarks for expected performance levels and demonstrated the effectiveness of advanced ML techniques in the Indian agricultural context.

3 Methodology

3.1 Data Provenance and Collection

Data were sourced from the Directorate of Economics and Statistics, Ministry of Agriculture and Farmers’ Welfare, Government of India, supplemented with Census demographic data, covering 1997–2020. The primary dataset contains comprehensive information on crop production, including yield, area under cultivation, fertilizer and pesticide usage, and rainfall patterns across 30 Indian states and union territories.

The demographic data was obtained from the Indian Census conducted in 2001, 2011, and projections for other years. Population density calculations were based on total population divided by geographical area, with urban-rural ratios derived from census definitions of urban areas. The integration of these datasets required careful temporal alignment and spatial matching to ensure consistency across different data sources.

Additional data sources included the Indian Meteorological Department for rainfall data, the Department of Fertilizers for input usage statistics, and various state agricultural departments for crop-specific information. The comprehensive nature of the dataset, spanning 24 years and covering multiple dimensions of agricultural production, provides a robust foundation for machine learning analysis.

3.2 Data Preprocessing and Quality Assurance

Cleaning included duplicate removal, unit standardisation, and median imputation for missing numeric values. Missing data constituted approximately 8.3% of the total dataset (1,635 out of 19,689 records), distributed across multiple variables with rainfall data showing the highest missing rate (4.2%) followed by fertilizer usage (2.8%). Missing data patterns were analyzed and determined to be missing at random (MAR) based on Little’s MCAR test (p < 0.001), indicating that missingness was related to observable variables rather than the missing values themselves. Outliers were addressed using IQR-based thresholds, with approximately 3.7% of observations flagged as potential outliers. The preprocessing pipeline was designed to maintain data integrity while ensuring compatibility with machine learning algorithms.

Duplicate removal was performed using multiple criteria, including crop type, state, year, season, and area under cultivation. Unit standardization involved converting all measurements to consistent units (tons for production, hectares for area, millimeters for rainfall). Missing value imputation was performed using median values within crop-state-season combinations to preserve the natural variation in agricultural data.

Outlier detection and treatment followed a systematic approach. Values beyond 1.5 times the interquartile range were flagged as potential outliers. Outliers representing legitimate extreme values (such as exceptional yields due to favorable weather conditions) were retained based on agricultural domain knowledge and data distribution analysis. Otherwise, they were capped at the 95th percentile to prevent undue influence on model training.

3.3 Feature Engineering and Selection

Label encoding was applied to categorical features. Derived metrics included fertiliser-per-area and pesticide-per-area. Market accessibility was derived from urban–rural ratios. The feature engineering process was guided by domain knowledge and statistical analysis to ensure relevance and predictive power.

Categorical variables, including crop type, season, state, and population category, were encoded using label encoding. While one-hot encoding could provide more detailed representation, label encoding was chosen for computational efficiency and to maintain the ordinal relationships present in some categorical variables (such as population density categories).

Derived features were created to capture important ratios and interactions. Fertilizer-per-area and pesticide-per-area ratios provide measures of input intensity that may be more predictive than absolute usage values. Market accessibility was calculated as a function of urban-rural ratio, reflecting the hypothesis that more urbanized areas have better market access and infrastructure.

Feature selection was performed using both statistical methods and domain expertise. Correlation analysis identified highly correlated features that could lead to multicollinearity, while feature importance analysis from preliminary Random Forest models guided the selection of the most predictive variables. The final feature set comprised 13 variables, balancing predictive power with computational efficiency.

3.4 Model Implementation and Architecture

Implemented algorithms: RF, Bagging, AdaBoost, Extra Trees, XGB, LightGBM, CB, GBM, Decision Tree, KNN, MLP, using scikit-learn, XGBoost, LightGBM, and CatBoost libraries. Each algorithm was implemented with careful attention to parameter settings and computational requirements.

Random Forest was implemented with 200 estimators, maximum depth of 15, and minimum samples split of 5. These parameters were chosen based on preliminary experimentation and literature recommendations for agricultural datasets. The algorithm’s ability to handle mixed data types and provide feature importance rankings made it particularly suitable for this application.

Gradient boosting variants (XGBoost, LightGBM, CatBoost) were implemented with 200 estimators, maximum depth of 6, and learning rate of 0.1. These conservative parameter settings were chosen to prevent overfitting while maintaining computational efficiency. The algorithms’ advanced regularization techniques and optimization algorithms provide superior performance for complex datasets.

Traditional machine learning algorithms (Decision Tree, K-Nearest Neighbors, Multilayer Perceptron) were implemented as baseline models for comparison. These algorithms provide important benchmarks for evaluating the effectiveness of ensemble methods and help identify the specific advantages of more sophisticated approaches.

3.5 Hyperparameter Tuning Strategy and Optimization

Random search with five-fold CV optimised hyperparameters (tree depth, estimators, learning rate). The tuning process was designed to balance exploration of the parameter space with computational efficiency, ensuring robust model performance without excessive computational cost.

The hyperparameter search space was defined based on literature recommendations and preliminary experimentation. For tree-based models, key parameters included the number of estimators, maximum depth, minimum samples split, and minimum samples leaf. For gradient boosting models, learning rate, subsample ratio, and column sampling ratios were also considered.

Cross-validation was performed using stratified sampling to ensure representative distribution of crop types and states across folds. This approach provides more reliable estimates of model performance and helps identify models that generalize well to unseen data.

The optimization objective was to maximize R-squared score while maintaining reasonable computational requirements. Models that showed signs of overfitting (high training performance but low validation performance) were penalized in the selection process.

3.6 Evaluation Protocol and Performance Metrics

Train-test split (80-20) with stratified sampling. Metrics: R2, RMSE, MAE, and bias (mean prediction error). CV assessed stability. The evaluation protocol was designed to provide comprehensive assessment of model performance across multiple dimensions, including accuracy, precision, and systematic error patterns.

The train-test split was performed using stratified sampling to ensure representative distribution of crop types and states across both sets. This approach is particularly important for agricultural data, where different crops and regions may have significantly different yield patterns and variability.

Performance metrics were chosen to capture different aspects of model performance. R-squared measures the proportion of variance explained by the model, providing an overall assessment of fit quality. RMSE penalizes larger errors more heavily, making it sensitive to outliers and extreme values. MAE provides a straightforward interpretation of average prediction error, useful for practical applications.

Cross-validation was performed using 5-fold stratified sampling to assess model stability and generalization capability. The standard deviation of cross-validation scores provides important information about model reliability and suitability for production deployment.

4 Results

4.1 Descriptive Statistics and Data Characteristics

Yield values ranged from below 1 t/ha to above 10 t/ha. Input usage varied widely by crop and state. The dataset exhibits significant variation across multiple dimensions, reflecting the diverse nature of Indian agriculture.

The yield distribution shows considerable skewness, with most observations concentrated in the lower range and fewer observations at higher yield levels. This pattern is typical of agricultural data and presents challenges for modeling, as models must accurately predict both typical and extreme yield values.

Input usage patterns reveal significant variation across crops and regions. Fertilizer usage ranges from minimal application in subsistence farming systems to intensive application in commercial agriculture. Pesticide usage shows similar variation, with some crops and regions showing minimal usage while others demonstrate intensive pest management practices.

Population density shows extreme variation across states, from sparsely populated mountainous regions to densely populated urban centers. This variation provides valuable information for understanding the relationship between demographic factors and agricultural productivity.

4.2 Comparative Model Performance Analysis

The comprehensive evaluation of 11 machine learning models reveals significant performance variations across different algorithms. Table 2 presents the complete performance ranking:

Ensemble methods consistently outperform individual models, demonstrating the value of combining multiple decision strategies for agricultural forecasting. Random Forest achieved the highest performance with R² of 0.9803 and RMSE of 125.79, representing the best balance of accuracy and interpretability. The algorithm’s ability to handle high-dimensional data and capture complex feature interactions makes it particularly suitable for agricultural forecasting where multiple factors influence yield simultaneously.

Bagging follows closely with R² of 0.9793 and RMSE of 128.68, demonstrating the effectiveness of bootstrap aggregation in reducing variance. The algorithm’s parallel training capability and stability make it suitable for production environments where consistent performance is crucial.

XGBoost achieves excellent performance with R² of 0.9766 and RMSE of 136.79, showcasing the power of advanced gradient boosting techniques. The algorithm’s built-in regularization and optimization algorithms provide superior performance for complex datasets, though at the cost of increased computational complexity.

4.3 Analysis of High-Performing Models and Algorithm Comparison

RF achieved the highest accuracy; Bagging was close, XGB balanced performance and computational efficiency. The analysis reveals important trade-offs between different algorithms and provides insights into their suitability for various applications.

Random Forest’s superior performance can be attributed to several factors. The algorithm’s ability to handle mixed data types, capture non-linear relationships, and provide robust predictions makes it particularly suitable for agricultural data. Additionally, Random Forest’s feature importance analysis provides valuable insights into the factors driving agricultural productivity.

Bagging’s strong performance demonstrates the effectiveness of bootstrap aggregation in reducing variance and improving generalization. The algorithm’s parallel training capability and stability make it suitable for production environments where consistent performance is crucial.

XGBoost’s performance highlights the advantages of advanced gradient boosting techniques. The algorithm’s built-in regularization, early stopping, and optimization algorithms provide superior performance for complex datasets. However, the increased computational complexity and sensitivity to hyperparameters may limit its suitability for some applications.

4.4 Model Stability and Cross-Validation Analysis

Light GBM had the lowest CV variance, indicating consistent performance. Cross-validation analysis reveals important insights into model stability and generalization capability, providing guidance for model selection in production environments.

LightGBM demonstrates the highest stability with CV mean of 0.9679 and standard deviation of 0.0131, indicating consistent performance across different data subsets. This high stability makes LightGBM particularly suitable for production environments where consistent performance is crucial.

Random Forest shows good stability with CV mean of 0.9563 and standard deviation of 0.0447, providing a good balance between performance and reliability. The algorithm’s robustness to outliers and noise in agricultural data contributes to its consistent performance.

Decision Tree and K-Nearest Neighbors show the lowest stability with high standard deviations, indicating sensitivity to data variations and potential overfitting issues. These algorithms may not be suitable for agricultural forecasting without extensive regularization and feature selection.

4.5 Forecasting Bias Patterns and Error Analysis

Detailed bias analysis reveals systematic patterns in model predictions that provide important insights into model behavior and potential areas for improvement. Figure 1 provides comprehensive diagnostic plots for the top-performing models, while Table 3 presents the bias analysis for the top 5 models:

Figure 1: Model diagnostic plots showing residual analysis, prediction vs actual comparisons, and error distributions for the top-performing models. These plots reveal systematic bias patterns, prediction accuracy across different yield ranges, and model reliability characteristics.

All top models show slight underforecasting tendencies, indicating conservative prediction behavior. Random Forest shows minimal systematic bias with mean error of 0.171 units, indicating well-calibrated predictions. The algorithm’s robust nature and ability to handle outliers contribute to its balanced performance across different yield ranges.

Bagging and XGBoost also show minimal systematic bias, with mean errors of 0.335 and 0.125 units respectively. These algorithms’ ensemble nature and advanced regularization techniques help maintain balanced predictions. CatBoost and LightGBM show some systematic patterns in residuals, particularly underestimating high yields and overestimating low yields. These patterns suggest that these algorithms may benefit from additional tuning or feature engineering to address the bias. The different patterns observed in the diagnostic plots reflect the algorithms’ distinct approaches to handling data complexity: the top three models (Random Forest, Bagging, XGBoost) show more uniform scatter patterns with points closely aligned to the diagonal line, indicating better calibrated predictions. In contrast, CatBoost and LightGBM exhibit more curved or S-shaped patterns in their residual plots, suggesting systematic prediction biases that vary across different yield ranges, likely due to their sequential boosting mechanisms being more sensitive to extreme values in the agricultural dataset.

4.6 Population Feature Impact and Demographic Analysis

Demographic features improved accuracy by 0.6%, supporting their inclusion. The integration of population features provides valuable insights into demand-side factors affecting agricultural productivity and demonstrates the value of comprehensive feature engineering. Table 4 presents the detailed impact of population-related features:

Crop type emerges as the dominant predictor with 84.6% importance, followed by state location (9.8%) and cultivation area (1.4%). Population-related features collectively contribute 0.6% to overall model performance, with market accessibility and urban-rural ratio being the most influential demand factors. This modest but consistent improvement demonstrates the value of incorporating demographic information into agricultural forecasting models.

Market accessibility and urban-rural ratio each contribute 0.30% to prediction accuracy, suggesting that urbanization patterns and market infrastructure significantly influence agricultural productivity. These features likely capture the effects of improved input availability, technology adoption, and market access in urbanized areas.

Population category shows minimal contribution (0.02%), suggesting that absolute population density is less important than urbanization patterns and market accessibility. This finding indicates that the quality of infrastructure and market access is more important than the sheer number of people in determining agricultural productivity.

5 Discussion

5.1 Operational Deployment Feasibility and Implementation

RF and LightGBM can be deployed in agricultural dashboard for near real-time forecasting. The research findings provide important guidance for the operational deployment of machine learning models in agricultural forecasting systems.

Random Forest’s combination of high performance and interpretability makes it particularly suitable for operational deployment. The algorithm’s feature importance analysis provides valuable insights for stakeholders, while its robust performance ensures reliable predictions across different conditions.

LightGBM’s high stability and computational efficiency make it suitable for real-time forecasting applications. The algorithm’s fast training and prediction times enable near-real-time updates, while its consistent performance ensures reliable predictions.

The deployment of these models in agricultural dashboards would provide policymakers, farmers, and market participants with timely and accurate yield forecasts, supporting better decision-making and resource allocation.

5.2 Model Interpretability Considerations and Stakeholder Trust

Tree-based ensembles support feature importance and partial dependence plots for policy transparency. The interpretability of machine learning models is crucial for gaining stakeholdertrustandensuringwidespreadadoptionofforecasting-baseddecision-making.

Random Forest’s feature importance analysis provides clear insights into the factors driving agricultural productivity, supporting evidence-based policy development. The algorithm’s decision tree structure enables the creation of partial dependence plots that show how individual features influence predictions.

The transparency provided by these interpretability tools is particularly important in agricultural contexts, where stakeholders may have limited technical expertise but require confidence in forecasting results. Clear explanations of model predictions and the factors influencing them support better decision-making and policy development.

5.3 Limitations and Assumptions of Current Approach

The current analysis is subject to several limitations that should be considered when interpreting the results and planning future research. These limitations provide important context for understanding the scope and applicability of the current findings.

The models assume that the relationships between features and yields remain constant over time, which may not hold true in the face of significant changes in agricultural practices, climate conditions, or policy environments. This assumption limits the long-term applicability of the models and suggests the need for regular retraining and validation.

The population density data is estimated based on historical trends and may not capture sudden demographic differences or migration patterns. This limitation affects the accuracy of population-related features and suggests the need for more frequent updates of demographic data.

The analysis focuses on Indian agricultural data, limiting the generalizability of the results to other agricultural contexts. While the methodologies and algorithms may be applicable elsewhere, the specific findings and parameter settings may not transfer directly to other regions or agricultural systems.

The dataset lacks crop quality indicators such as protein content, moisture levels, and post-harvest characteristics, which are important factors in determining the economic value of agricultural output. The models focus solely on yield quantity without considering quality attributes that significantly influence market prices and food security outcomes. This limitation affects the comprehensive assessment of agricultural productivity and suggests the need for future research incorporating quality metrics alongside yield predictions.

5.4 Computational Considerations and Scalability

The computational requirements of different algorithms present important considerations for operational deployment and scalability. These considerations affect the choice of algorithms for different applications and the infrastructure requirements for deployment.

Random Forest and Bagging algorithms can be parallelized effectively, making them suitable for deployment on multi-core systems. These algorithms’ parallel nature enables efficient training and prediction on large datasets, supporting real-time forecasting applications.

Gradient boosting algorithms (XGBoost, LightGBM, CatBoost) require more computational resources but provide superior performance. The choice between these algorithms and simpler ensemble methods depends on the specific requirements for accuracy, speed, and computational resources.

The deployment of these models in production environments requires careful consideration of computational infrastructure, including processing power, memory requirements, and storage capacity. These requirements affect the cost and feasibility of operational deployment.

6 Conclusion

This study benchmarks eleven ML algorithms for crop yield forecasting in India, demonstrating ensemble superiority and the benefits of including demographic features. RF was the top performer; LightGBM was the most stable. The comprehensive evaluation provides important insights into the effectiveness of different machine learning approaches for agricultural forecasting.

The superior performance of ensemble methods, particularly Random Forest, demonstrates the value of combining multiple decision strategies in agricultural forecasting. The significant performance gap between ensemble methods and individual models highlights the importance of sophisticated modeling approaches for complex agricultural data.

The integration of population features provides consistent improvements in forecasting accuracy, supporting the inclusion of demographic factors in agricultural forecasting models. While the improvement is modest, it represents a meaningful enhancement that contributes to better agricultural planning and policy development.

The research contributes to improved agricultural forecasting by demonstrating the value of comprehensive model evaluation and ensemble methods in agricultural prediction. The findings support the development of multi-model forecasting systems that can provide more robust and reliable predictions for agricultural planning and policy development.

7 Future Work and Research Directions

Future research should explore crop-specific models, integration of real-time climate and remote sensing data, soil index integration, explainable AI for stakeholder trust, and scenario modelling for climate impact assessment. These directions build on the current findings and address important gaps in agricultural forecasting research.

The development of crop-specific models could significantly improve forecasting accuracy by capturing the unique characteristics and requirements of different crops. The high importance of crop type in the current models suggests that specialized approaches for different crop categories could provide substantial improvements in prediction accuracy.

The integration of real-time climate data and remote sensing information could enhance the models’ ability to capture environmental factors affecting agricultural productivity. These data sources provide more timely and detailed information about growing conditions, potentially improving short-term forecasting accuracy.

The development of explainable AI techniques, including SHAP values and partial dependence plots, could improve stakeholder trust and support better decision-making. These techniques provide clear explanations of model predictions and the factors influencing them, supporting transparency and accountability in agricultural forecasting.

Scenario modeling for climate impact assessment could help policymakers understand the potential effects of climate change on agricultural productivity and develop appropriate adaptation strategies. These models could incorporate various climate change scenarios and assess their impact on crop yields and food security.

References

[1] L. Breiman. Random forests. Machine Learning, 45(1):5–32, 2001.

[2] T. Chen and C. Guestrin. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 785–794, 2016.

[3] D. De Clercq and A. Mahdi. Feasibility of machine learning-based rice yield predic- tion in india at district level. arXiv preprint arXiv:2403.07967, 2024.

[4] N. M. Gharakhanlou. Leveraging ensemble machine learning for enhanced crop yield prediction. Science of The Total Environment, 937:172587, 2024.

[5] D.Headey, P.Hazell, etal. Populationdensityandagriculturalproductivity: Theory and evidence. Agricultural Economics, 33(2):121–134, 2005.

[6] G. Ke, Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T.-Y. Liu. Lightgbm: A highly efficient gradient boosting decision tree. In Advances in Neural Information Processing Systems, volume 30, 2017.

[7] Government of India. Agriculture statistics at a glance, 2023.

[8] L. Prokhorenkova, G. Gusev, A. Vorobev, A. V. Dorogush, and A. Gulin. Catboost: Unbiased boosting with categorical features. In Advances in Neural Information Processing Systems, volume 31, 2018.

[9] S. Sharma, S. Rai, and N. C. Krishnan. Wheat crop yield prediction using deep lstm model. arXiv preprint arXiv:2011.01498, 2020.

[10] T. A. van Klompenburg, A. Kassahun, and C. Catal. Crop yield prediction using machine learning: A systematic literature review. Computers and Electronics in Agriculture, 177:105709, 2020.

[11] Y. Wang, H. Zhang, Q. Li, and Y. Sun. Progress in research on deep learning- based crop yield prediction: Trends, challenges, and future directions. Agronomy, 14(10):2264, 2024.

About the author

Advika Lakshman

Advika is currently pursuing Artificial Intelligence and Data Science at Shiv Nadar University, Chennai. Her academic interests span across machine learning, deep learning, natural language processing, big data analytics, and speech technology. She has worked on diverse projects such as early sepsis prediction using clinical time-series data, sketch-to-face translation with DCGANs, geophysical data inpainting with Masked Autoencoders, and salary prediction using ensemble models. Advika has also interned at the National University of Singapore (Big Data, Deep Learning, Generative AI) and the Spring Lab at IIT Madras, where she developed ASR pipelines using HuBERT and ESPnet for multilingual speech data.

Outside academics, Advika is a professional Bharatanatyam dancer, with over 12 years of training and multiple state and national-level awards, including recognitions from Doordarshan. She also actively contributes to university events and communications through marketing and public relations initiatives.

The post Comprehensive Crop Yield Forecasting in India: A Multi-Model Machine Learning Approach with Population Density Integration for Agricultural Planning appeared first on Exploratio Journal.

Abstract

The long-running socio-economic crisis in Venezuela has severely damaged public healthcare systems, which now face persistent shortages and restricted service availability. This research investigates Venezuelan patients’ healthcare challenges and their strategies for managing these obstacles, with particular attention to gender-specific barriers in service accessibility. The researcher conducted semi-structured interviews with 13 Venezuelan adults from different regions between July 14 and July 16, 2025. Interviews were conducted in Spanish through WhatsApp text and voice notes, then transcribed, translated into English, and pseudonymized. Inductive thematic analysis identified recurring patterns, while frequency counts determined the relative prevalence of each theme. Seven main themes emerged: medication shortages, poor public hospital facilities, insufficient specialist care, reliance on private healthcare services or HCM insurance, non-traditional medical approaches and foreign medication acquisition, preventive self-care measures, and community-based fundraising and support systems. Women described additional challenges accessing reproductive healthcare, managing chronic illnesses, and obtaining medical treatment outside their local areas. Current patient adaptation strategies—including preventive actions, unofficial support networks, and cross-border medication procurement—remain unstable and create unequal outcomes that affect women most severely. Policy priorities should focus on reliable medication supply chains, protected crossborder medical a

Keywords: Venezuela; healthcare access; qualitative interviews; women’s health; resilience; health systems.

1. Introduction

The public health system of Venezuela which used to rank as one of the best in Latin America has experienced a significant decline because of ongoing political instability, economic decline and hyperinflation (World Health Organization [WHO] 2023; Ortega et al. 2020). The healthcare facilities in Venezuela face ongoing problems with medicine, supply shortages, equipment breakdowns and massive healthcare worker departures to find better employment opportunities abroad (Doocy et al. 2019; Pan American Health Organization [PAHO] 2023). The healthcare system operates at reduced capacity because of these disruptions which force wealthier patients to seek private care or foreign medical services while low-income patients receive inadequate treatment (Rodríguez 2020; Hetland 2021).

The research examines how typical Venezuelans handle the structural failures that affect their healthcare system. The research investigates two main questions which are (1) What obstacles do patients face when they try to receive medical care in their daily lives? and (2) What strategies do patients use to overcome these barriers and how do women patients specifically handle these challenges? The research uses patient testimonies from 2025 to show the actual experiences that exist beyond official reports about system failures and it also provides a patient-centered perspective through its focus on their coping mechanisms and emotional responses which enhance existing quantitative and infrastructure-based studies of Venezuela’s healthcare emergency (Doocy et al. 2017; PAHO 2023).

This research has shown that hospitals face medicine shortages, stockouts and administrative breakdowns but it has not fully explored how patients make decisions when resources are scarce. The research connects these findings by showing how Venezuelans handle broken healthcare systems and how gender influences their treatment routes, women experience unique barriers to healthcare access because of their reproductive needs and caregiving duties. Limited mobility actually worsens their unequal healthcare opportunities. The analysis of gendered healthcare system performance under extreme economic conditions provides valuable knowledge about system breakdowns.

Research Questions:

1. What barriers do patients in Venezuela encounter in everyday efforts to obtain healthcare?
2. How do patients, particularly women, adapt to or circumvent these barriers?

2. Literature Review

Multiple research studies from national and international experts have documented the complete breakdown of Venezuela’s healthcare system. Multiple research studies confirm the severe state of infrastructure deterioration through reports about continuous medicine shortages, equipment breakdowns and even healthcare worker departures (Doocy et al., 2019; Ortega et al., 2020; PAHO, 2023; WHO, 2023). The public healthcare system which used to serve as a regional benchmark now operates with permanent shortages that force hospitals to stop vital services while patients must bring their own medical supplies. The healthcare system shows that major urban hospitals continue to operate at reduced capacity but peripheral medical facilities have lost almost all of their operational ability which results in major disparities between urban and rural areas as well as between different social classes (Rodríguez, 2020; Hetland, 2021).

Research on healthcare system collapse has led experts to study how patients experience medical care delivery in deteriorating facilities. The combination of unstable supply chains and currency fluctuations results in unpredictable access to vital medications according to Doocy et al. (2017) and Hetland (2021). The research community agrees that patients directly experience the consequences from both international sanctions and domestic governance problems although they disagree about which factor has the most impact. Households actually manage medicine shortages through three main strategies which include buying from informal markets and using remittances and implementing medicine rationing (Freitez, 2022; International Organization for Migration [IOM], 2022). The survival systems that Venezuelans create operate independently from official healthcare systems through parallel networks or function completely outside of them.

The current research on healthcare system failures provides essential information about large-scale breakdowns yet fails to examine how people and their families make independent decisions when facing scarcity at the individual level. Research using qualitative and feminist methods demonstrates that healthcare emergencies reveal existing gender-based social inequalities (Melo et al., 2023; Rueda-Salazar & García, 2024). The healthcare challenges Venezuelan women encounter stem from multiple factors which include reproductive care restrictions, family care duties as well as limited freedom of movement. The combination of financial constraints and shut-down specialized maternal facilities forces women to handle their pregnancies and also chronic diseases and family health crises without institutional backing. Research conducted in neighboring countries indicates that Venezuelan migrant women encounter equivalent healthcare obstacles because of their immigration status and social discrimination which demonstrates that gender-based healthcare disparities exist across international borders (Melo et al., 2023; Rueda-Salazar & García, 2024).

Academic researchers now employ resilience frameworks to study how people and their communities handle extended crisis situations. According to Norris et al. (2008) community resilience emerges from social networks and economic resources and information access which enable stress absorption. The survival of Venezuelans depends on individual resourcefulness as well as the support systems which include family networks and community structures and international connections. The research on Venezuelan families shows that they use remittances together with informal support networks and preventive self-care techniques to create their survival strategies (Freitez, 2022). The distribution of these survival strategies remains unequal due to people who possess foreign currency and have access to border travel and stable communication networks succeed in adapting better than those who do not.

Research on Venezuelan healthcare infrastructure deterioration has received extensive study but patient decision-making within Venezuela actually remains highly understudied. Most of the research found focuses on institutional breakdowns instead of showing how patients experience these breakdowns in their daily lives. The research combines patient testimonies with structural evaluations to show how women and other individuals handle healthcare access in a system that has stopped operating effectively.

3. Methods

This study used qualitative methods to explore Venezuelan patients’ experiences of healthcare access and adaptation under system collapse. Semi-structured, one-on-one interviews were chosen to capture the nuance of personal narratives and to be able to allow participants to describe their healthcare decisions in their own personal words.

3.1 Participants and Recruitment

The research included thirteen women aged 17 to 65 who participated in individual interview sessions. The researcher picked participants who lived in Venezuelan cities and mainly surrounding areas including Caracas, Guarenas, Maracay, Ciudad Bolívar and Valles del Tuy. The researcher was able to use trusted local contacts to find participants in June 2025 before they expanded the participant pool through WhatsApp using snowball sampling techniques; this method was selected because it helped achieve both geographic and socioeconomic diversity while safeguarding participants from political dangers in the unstable setting.

The study participants were all Venezuelan residents who received healthcare from the national system during the previous years. While the study excluded participants who lacked capacity to give consent or lived outside Venezuela. The study selected women as participants because their input was necessary for conducting gender-based research.

The three-day interview schedule from July 14 to July 16 2025 worked with participants’ work commitments and protected them from dangerous extended online sessions because of unreliable power supply. The researcher understands that snowball sampling could have brought social-network bias because people with restricted smartphone access became less probable to join the study. The Study Limitations section provides further information about these methodological restrictions.

3.2 Data Collection

The research team conducted Spanish-language interviews through WhatsApp text and voice calls from July 14 to July 16 2025 with each session lasting between 30 to 55 minutes. The platform enabled participants to interact at their own pace while ensuring their safety through asynchronous communication. The interview guide asked participants to answer open-ended questions about their care-seeking actions, their experiences with medication access, healthcare facilities, specialist availability and their financial approaches and their coping strategies.

The researcher conducted a pilot test of the interview guide with two Venezuelan contacts to verify both cultural understanding and language precision. The participants gave their consent through WhatsApp messages of all interviews while bilingual reviewers checked the accuracy of the English translation to preserve the original meaning which confirmed their willingness to participate and their ability to leave the study anytime. The transcription researcher substituted all personal information with pseudonyms to protect participant confidentiality.

3.3 Ethical Considerations

The research followed all necessary ethical guidelines for qualitative studies with low risk while the participants received information about digital communication privacy risks in Venezuela while being told to keep hospital names and official identities undisclosed. The researcher stored all data through encrypted files which required password protection and immediately removed identifying information from translated transcripts. The researcher also did not offer any payment to participants because they wanted to protect them from external influences in a region with limited resources.

3.4 Data Analysis

The researcher conducted inductive thematic analysis (Braun & Clarke, 2006) to find recurring patterns and categories. The researcher performed multiple readings of transcripts for familiarization before using line-by-line coding to produce initial codes which later became broader thematic categories. The researcher examined candidate themes throughout the complete dataset to develop them into final themes while creating analytical memos for support.

The researcher documented all coding choices in a very detailed audit trail while performing negative-case searches and by also maintaining reflexive notes about positionality and translation decisions.The study’s exploratory nature required a single coder to analyze data while reflexive documentation replaced traditional intercoder reliability assessment.

4. Results

The research study found seven core themes which demonstrate how Venezuelan patients handle their healthcare needs in a system that lacks stability and has broken down into seven separate parts. The themes show how patients experience multiple forms of scarcity, adaptation and inequality which demonstrate both the complete breakdown of official healthcare services as well as the development of survival methods outside formal care. The following section presents an overview of theme distribution and relationships through Figures 1 and 2 before moving to the detailed analysis.

Figure 1 presents a frequency chart showing how often each of the seven themes appeared across the thirteen interviews. All participants discussed medication unavailability but specialist shortages and substandard public hospital conditions were mentioned by most many participants. The frequency chart in Figure 1. The themes of preventive health behaviors and private insurance use emerged less often but generated intense emotional as well as social responses. The visual data shows that scarcity-related problems took center stage in participant experiences while serving as the foundation for their coping mechanisms.

The binary-coded matrix in Figure 2 shows which seven key themes each of the thirteen participants discussed during their interviews. The left column contains participant identification numbers which match the thematic categories found in the right columns. The maroon filled cells in the matrix indicate participants discussed specific themes while yellow cells show their absence. The visualization shows that participants experienced both common barriers like medication shortages and less common challenges related to community fundraising and also preventive self-care. The chart demonstrates how different participants experienced various structural barriers at different levels which reveals how coping strategies spread unevenly throughout the healthcare system collapse.

4.1 Theme 1→Medication Unavailability

The thirteen participants in the study all described their ongoing struggles to get necessary medications through Venezuela’s public healthcare system. The participants identified medication shortages as their main obstacles which made it difficult to treat hypertension, diabetes, Parkinson’s disease and even asthma.

The woman from Nueva Casarapa explained that her friend received no medical supplies during her pneumonia hospitalization because her family needed to purchase all necessary items including gloves, she stated “the experience made me lose faith in the public healthcare system.” The patient population learned to accept that hospitals would only offer physical care facilities.

A healthcare provider at Guarenas Hospital stated that medications remain the hospital’s top priority but the supplies never reach on schedule. The resident of Ciudad Bolívar described how his grandmother suffered from severe leg pain because the hospital lacked any pain medication. The testimonies demonstrate how the official supply system has failed to deliver basic pain medication which causes emotional distress to patients.

The scarcity situation made people change their daily routines according to multiple participants. The high prices and unavailability of products in Valles del Tuy forced residents to share their medication and reduce their dosage for extended usage. The practice of medicine supply management through informal prescription sharing and dose reduction demonstrates both creative problem-solving as well as very high risk dangerous consequences because families had to stretch their limited medication supplies.The participants chose to purchase medications from outside their country because they had no other options left.

The mother from Maracay described her experience of buying medicine in Cúcuta after her son developed asthma because the hospital refused to provide oxygen, “the experience made me understand that we needed to leave the country.” The situation forced her to migrate to Colombia with her family after she obtained medication in Cúcuta.

The participants from different areas shared similar experiences about medicine shortages which they viewed as both a system failure and a heavy emotional and moral challenge. The participants displayed three main emotional responses to the situation: they worried about their family members’ worsening health and felt responsible for their inability to help and even lost faith in the healthcare system they used to trust. The widespread nature of these accounts demonstrates that medicine shortages represent both a national healthcare emergency and a personal indicator of Venezuela’s failing healthcare system.

4.2 Theme 2 →Substandard Public Hospital Conditions

Nine participants painted a grim picture of the extent to which Venezuelan public hospitals were very overcrowded, dirty, chronically under-resourced and unsanitary. Together, their accounts offer a glimpse into how systemic infrastructure failure has transformed the very meaning of hospital care from a site of treatment to one of uncertainty and survival.

A Guarenas woman described, “In public hospitals it’ s a mess, there are too many patients, no medicine, and sometimes not even water. And you sit there for three hours only to be told there’ s no specialist available.” Her experience is a mirror of the common experience of trying to work one’s way through institutions which can no longer maintain minimum standards of hygiene or even effectiveness.

A third participant from Caracas shared her mother-in-law’s ordeal after falling: “There, at Salud Chacao, they took an X-ray and said there was no traumatologist, we had to carry her across town to El Llanito and there were not even painkillers. Imagine seeing an old person in pain and being told there’ s no basic pain medicine.” This accounting of pain shows how staff shortages and supply deficits are magnifying suffering ushering routine emergencies into the territory of traumatic scourge.

A health-office worker who used to oversee public-sector spending gave an infrastructure-oriented explanation: “I could look at paper and see how little money was given versus what communities needed. This was always a gap between what people needed and what they could get.” And her double identity as administrator and patient bridges personal experience as well as systemic dereliction.

Some also relayed what they had seen of medical personnel being run ragged with unmanageable work duties where one woman said, “You see a doctor alone trying to assist 20-something people with not one glove nor disinfectant, absolutely nothing. After hours of waiting, most leave without being seen.” In the eyes of patients, these images of sick and very exhausted staff serve as visual reminders that the public health system had tragically caved from within.

In all of these stories, public hospitals serve as concrete avatars for Venezuela’s broader institutional collapse as respondents didn’t depict them as isolated failures but rather demonstrated how they were the product of visible outcomes from a chronic underinvestment, staff migration, and bureaucratic rot. Their stories demonstrate how systemic failure undermines trust by forcing patients to rely on private clinics or informal options whenever they can.

4.3 Theme 3→Specialist Shortages and Delays

Most participants described long waiting times and difficulty accessing medical specialists such as cardiologists, neurologists, and oncologists. These delays were not isolated inconveniences but structural outcomes of Venezuela’s shrinking medical workforce and deteriorating hospital capacity as participants consistently framed these shortages as both a medical as well as psychological burden turning treatable conditions into prolonged uncertainties.

A man from Caracas recounted, “My mother had tachycardia. At the emergency room, the EKG machine wasn’t working and there was no cardiologist. We had to take her to a private hospital.” His story reveals how technical failures and missing personnel really forced families to seek private care regardless of their financial hardship.

A former Indigenous Health Office employee offered a similar account from the administrative side: “My father needed a cardiology exam. Every hospital said the same thing, the machine was broken, or the technician wasn’t there. In the end, we had to pay privately, and my family had to pool money just to make it happen.” Her testimony links institutional scarcity to unequal outcomes where access depends on financial capacity rather than the patients medical needs.

Another participant from Caracas recalled the fatal consequences of these delays: “We took my mother to Vargas Hospital for an emergency. They didn’t even run tests, they just said she was fine and sent her home. Two days later, she died. I can’t explain the anger and pain.” Her account conveys the moral, emotional toll of systemic neglect as well as the collapse of clinical accountability.

For others, the absence of specialists created a cycle of deferred care. A participant from Valles del Tuy summarized, “If you need a specialist, you travel to Caracas, wait weeks, and still might not get seen. People just give up.” These accounts show how spatial and temporal barriers combine to make even basic specialist consultations inaccessible.

Together, these testimonies portray specialist scarcity as both a symptom and a driver of broader healthcare inequality. The inability to access specialized treatment deepens existing vulnerabilities, particularly for older adults and those managing chronic illness. Specialist shortages thus stand as one of the clearest indicators of Venezuela’s fractured health infrastructure where time itself has become a form of rationed resource.

4.4 Theme 4→ Reliance on Private Care or HCM Insurance

As public healthcare deteriorated, many participants described shifting toward private clinics or even employer-sponsored insurance programs known as HCM (Hospitalización, Cirugía y Maternidad). This reliance on private coverage emerged not as a preference but as a survival strategy as an adaptation to the state’s withdrawal from healthcare provision. Participants’ narratives show that while private access provided greater reliability it also reinforced very deep financial inequalities and excluded those without stable employment.

A woman from Guarenas explained, “When my neighbor couldn’t get care in the public hospital, she went to a private clinic. It cost her everything she had, but at least she was treated. People say, ‘If you want to live, you have to pay.” Her statement captures both the necessity and resentment surrounding privatized care.

Several participants reported prioritizing jobs that offered HCM insurance as a healthcare worker from Caracas shared, “I took a position mainly because it included private health insurance. Public hospitals don’t have medicine or equipment, so you end up needing HCM for even basic care.” Insurance thus functioned as a form of social capital, one that actually determined not only medical outcomes but also employment choices.

Another participant highlighted the limitations of such plans: “Even with HCM, you still pay a lot out-of-pocket. The coverage runs out fast, and the prices keep going up. It’ s like having a lifeboat with holes in it.” Her metaphor reflects the precariousness of middle-class adaptation under crisis conditions.

For women, these disparities were compounded by reproductive health needs. A participant from Caracas noted, “There’ s almost nowhere left for gynecological care unless you pay privately. Public hospitals cancel appointments all the time, and traveling far alone doesn’t feel safe.” Her account illustrates how gender, safety, and mobility intersect to limit care options even for those with insurance access.

Across these testimonies, private healthcare appears as both refuge and reminder of inequality. While HCM coverage temporarily shields patients from systemic failure it simultaneously deepens divides between those with institutional protection and those without as participants portrayed this duality with ambivalence gratitude for access mixed with anger that survival had become conditional on wealth or employment.

4.5 Theme 5→Alternative Practices and Cross-Border Procurement

When official healthcare systems failed many participants described turning to alternative remedies and informal medication channels to manage illness. These adaptations ranged from home herbal treatments to complex cross-border purchasing networks coordinated through many relatives abroad. The accounts portray a spectrum of creativity and desperation which are strategies that temporarily alleviated suffering but also carried risks of misinformation and even inequity.

A woman from Guarenas shared, “People started using natural medicine as teas, herbs, whatever helped. We learned recipes from neighbors or online because the pharmacies were empty.” Her story highlights how traditional knowledge re-emerged as an informal safety net when biomedical options vanished.

A healthcare worker from Caracas described a similar pattern of cautious substitution: “When medicines disappear, you do what you can use home remedies, eat healthy, try to stretch the little medicine you have left. It’ s survival.” This pragmatic tone resignation without illusion echoed through many interviews.

For others, adaptation required geographic movement. A mother from Maracay who had ultimately migrated to Colombia, explained, “Every few months, someone from our neighborhood crossed the border to buy insulin or antibiotics. We all pitched in money and sent lists. Sometimes they came back empty-handed; sometimes they didn’t make the trip.” Her narrative shows how scarcity created collective networks of cross-border cooperation grounded in trust and necessity.

Participants viewed these alternative strategies ambivalently: they provided temporary control but also underscored dependence on unstable, unregulated systems. A man from Valles del Tuy summarized, “You feel proud that people find ways to survive, but also scared—because it shouldn’t be this way.” His reflection captures the tension between resilience and resignation that runs throughout the data.

Ultimately, these testimonies reveal that informal and transnational health practices have become integral to everyday survival in Venezuela. Yet the uneven access to information, money, and cross-border mobility means that such coping mechanisms often reproduce the very inequalities they are meant to alleviate.

4.6 Theme 6→ Preventive Health Behaviors

Six participants described developing preventive health routines as a way to reduce dependence on Venezuela’s collapsing medical infrastructure. With hospitals being unreliable and medicines scarce many prevention became a deliberate survival strategy as a an attempt to regain control in an unpredictable environment.

Participants framed these lifestyle adjustments not as wellness trends but as pragmatic risk management.A woman from Guarenas explained, “I try to eat well, exercise, and meditate. During the hardest years, when medicine disappeared, that was all we could do.” Her statement reflects how self-care practices became substitutes for unavailable treatments.

A healthcare worker from Caracas expressed a similar sentiment: “You learn to avoid getting sick. I rely on my job’ s small clinic, but mostly I take care of myself because public hospitals are too dangerous.” The combination of precaution and fear shows how prevention emerged from distrust rather than health promotion.

Another participant from Venezuela’s Indigenous Health Office connected this shift to structural awareness: “After seeing how little funding there was, I started saving money for emergencies and taking vitamins. Prevention is cheaper than depending on the system.” Her comment reveals an economic logic to self-care, framing it as an investment in personal resilience.

Across these accounts, preventive health practices carried both empowerment and burden; they offered participants a sense of agency yet simultaneously transferred responsibility from institutions to individuals. A man from Guarenas summarized this trade-off: “It’ s good to be healthy, but it’ s also exhausting because you’re doing the government’ s job.”

These narratives highlight the privatization of risk at the household level and prevention, once a public-health goal, has become an individual coping mechanism as a an act of necessity shaped by structural abandonment.

4.7 Theme 7→ Community Fundraising and Support Systems

When formal health systems broke down the respondents narrated seeking solutions through traditional and informal medicine for their ailing bodies. These adaptations covered different levels, from herbal treatment in the home to intricate transborder shopping strategies passed on through relatives living abroad. The accounts reflect a range of creativity and desperation strategies that offered such temporary relief while also posing the risk of misinformation, expense, and inequality.

One health-care worker in Caracas said that these were part of the gradual process of standing-in: “You do what you can because when medicines vanish from the store shelves, you have to treat yourself at home and eat healthfully and stretch out the last bit of prescription drugs. It’ s survival.” This pragmatic coarseness of resignation without illusion sounded through many interviews.

Others had to move geographically to adapt, “From time to time, someone we knew would go over and buy insulin or antibiotics,” said a mother in Maracay who eventually relocated to Colombia. “We all chipped in money, and we sent lists. Sometimes when they went out in the morning, they didn’t come back with anything; sometimes they never left.” Feign’s story illustrates how the scarcity generated cross-border networks of collective action based on trust and mutual need.

Such alternatives were actually only weakly accepted by the participants; in a way as they gave temporary power but at the same time repeated and underlined an insufficient dependency on unstable, unregulated systems. A man from Valles del Tuy put it like this: “You feel proud that people are finding ways to survive but also scared because people shouldn’t have to do things this way.” His reflection is a poignant tension between that of resilience and resignation that courses through the data.

Ultimately, these testimonies illustrate that informal and transnational health practices are now central to everyday survival in Venezuela but the unequal ability to access information, money, and cross-border mobility means that these coping strategies can themselves replicate the inequalities they are intended to ameliorate.

5. Discussion

5.1 Understanding Patient Adaptation Through Resilience

The research results show that Venezuelan patients’ healthcare collapse experiences extend beyond institutional breakdowns because they demonstrate intricate social and psychological adaptation patterns which align with resilience theory. According to Norris et al. (2008) community resilience exists as the ability of social systems to withstand disturbances while keeping their fundamental operations intact. While according to Ungar (2018) resilience emerges through the continuous interaction between personal initiative and environmental resource availability. The presented stories show Venezuelans performing both individual adaptation and social environment boundary adjustment at the same time.

The coping strategies of participants including preventive health practices and informal medication distribution and private insurance usage and community-based fundraising efforts demonstrate the interconnected systems which Norris and Ungar describe. People use “navigation and negotiation” according to Ungar to find or even establish alternative resource access routes when faced with restricted circumstances. The adaptive strategies people use to cope with their situation demonstrate how digital networks function as a replacement for missing state-based infrastructure. The adaptive capacity shows uneven distribution patterns among the population. The research supports Norris et al. (2008) who state that resilience depends on four essential domains which include economic development and social capital and information and community competence yet these resources remain unevenly accessible in present-day Venezuela.

The research indicates that patients demonstrate remarkable resourcefulness but their adaptive actions take place within systems that maintain significant social inequalities. Women encountered multiple barriers which restricted their ability to join resource-sharing networks because they faced limitations in mobility and safety and reproductive healthcare access. The anthropological concept of “bounded resilience” describes their situation because they showed bravery through adaptation yet their actions remained restricted by structural barriers.

The research evidence contradicts positive views about resilience as a solely empowering force and it also reveals that people must take on institutional responsibilities when public institutions abandon their duties which results in resilience becoming a sign of systemic failure. The Venezuelan patients’ ability to adapt serves as proof of human flexibility yet it also reveals the unacceptable circumstances which force people to endure such hardships.

5.2 Gendered Vulnerabilities and Health Inequality

The research findings from women participants demonstrate that gender plays a fundamental role in determining how Venezuelan women experience the collapse of their healthcare system. Women took on dual responsibilities by providing care to their families while simultaneously acting as medical coordinators who located medications and arranged transportation and handled healthcare costs. The gendered tasks women performed during this time increased their stress levels and financial difficulties which strengthened existing social and economic inequalities.

A woman who lives in Caracas described her multiple responsibilities when she stated, “You need to perform all nursing duties and medical tasks and psychological support and financial management for your family members.” The statement demonstrates the “care work paradox” which feminist scholars describe as women taking on unpaid work to replace absent institutional care according to Hochschild (1995) and Tronto (2013).

The research data indicates that women experienced different levels of risk because of their reproductive and chronic health requirements. Women participants explained that they experienced prolonged delays when trying to obtain gynecological care and birth control methods and prenatal medical assistance. The participant described her experience of seeking gynecological care in Caracas because she needed to travel long distances but doctors frequently canceled her appointments.

Women who are pregnant or have medical issues must handle their health needs independently because they lack proper care. The breakdown of healthcare systems leads to increased health dangers for women and restricts their ability to control their bodily autonomy. Women encountered special risks when seeking medical care because they needed to navigate dangerous areas while dealing with transportation problems and security threats. The participant from Maracay expressed her fear about needing Caracas treatment because bus services were unavailable. The combination of gender with geographical location and security risks produces what scholars call “layered precarity” (Butler, 2016) which restricts women’s ability to make decisions.

Women showed impressive organizational abilities despite facing numerous obstacles in their community such as organized medicine-sharing programs and operated online fundraising campaigns and distributed health information through WhatsApp messaging. Women’s ability to solve problems through resourcefulness continues to perpetuate the societal norm that they should handle institutional breakdowns through emotional work and logistical management. The participant expressed her question about why mothers consistently need to find solutions for every problem.

The research results support feminist anthropological theories which demonstrate that health emergencies reveal and intensify existing social inequalities as the social structure of gender determines how people experience risk exposure and healthcare access. Women in Venezuela experience both the power of their social connections and the weight of enduring a healthcare system that ignores their needs.

5.3 Policy Implications and Structural Change

The healthcare crisis in Venezuela exists beyond resource shortages because it stems from fundamental problems with system organization and unequal distribution of resources. The solution to these problems needs both emergency policy solutions and enduring structural changes to healthcare systems as the healthcare system needs to focus on these essential priorities that participants identified to rebuild trust and minimize gender-based and geographic healthcare inequalities.

The healthcare system needs to create dependable systems for medication distribution across the country. The participants in this research study all mentioned their struggles to get necessary medications which proves that Venezuela needs better medicine procurement systems and improved distribution networks with monitoring capabilities for all regions. The stabilization of pharmaceutical access through international humanitarian organization partnerships should include oversight systems to stop diversion and corruption activities.

The government needs to create official programs that protect patients who need medical treatment outside their home country. The participants needed to use unofficial networks to buy medication outside their country which exposed them to dangerous situations and legal consequences. The government should also establish controlled humanitarian medication import systems through official agreements with Colombia and Brazil. The current system of private and dangerous medication procurement would become unnecessary through these policies.

The healthcare system needs to create immediate specialized outreach programs which focus on treating women and patients who need ongoing medical care. The research shows that women face the most severe health problems which receive insufficient attention during reproductive and preventive care. Mobile health units together with regional telemedicine programs should be implemented to provide better healthcare services in underserved areas while minimizing travel-related dangers. The implemented measures need to adopt gender-sensitive approaches which account for women’s dual responsibilities in caregiving and their limited mobility.

The protection of financial resources for healthcare services must become a priority to stop the growing separation between different social groups in healthcare access. The current healthcare system operates based on wealth because most people depend on private insurance (HCM) and pay medical expenses directly from their pockets. The health system needs more than technical fixes because it requires public trust to be rebuilt as the participants expressed their sense of being left behind and their complete emotional depletion. The recovery of public trust in healthcare institutions depends on transparent governance and professional retention strategies and community involvement in health planning processes. The success of well-designed reforms depends on rebuilding trust between healthcare providers and their patients because without it these reforms will fail to benefit the intended population.

The proposed recommendations follow the principles of resilience frameworks developed by Norris et al. (2008) and Ungar (2018) which state that recovery success depends on building up social and institutional frameworks at the same time. The combination of supply chain improvement with humanitarian channel legalization and gender-focused policies will enhance healthcare results while bringing back the lost sense of national security.

5.4 Study Limitations and Future Research

As with all qualitative inquiry, this study has several limitations that should guide the interpretation of its findings.

First, the sample was relatively small and although it reached regional diversity between age groups, it cannot be generalized to all women in Venezuela. The study’s aim was depth rather than breadth and its findings should not be taken as statistically generalizable but illustrative.

Second, recruitment occurred through WhatsApp networks and snowball sampling leaving out consideration of those without digital access or outside their social network. As a result, the lives of the most marginalized groups, like people in rural areas who do not possess smartphones or are in absolute poverty, may have been unaccounted for.

Third, the 3-day window of data collection precluded follow-up interviews or even longitudinal tracking. Further fieldwork over a longer term might show how coping strategies shift with continued economic or political changes.

Fourth, all interviews were translated from Spanish to English and some nuance may have been lost in translation despite the benefit of careful checking. Future research would gain from bilingual analysis or Venezuelan co-researchers to increase cultural and linguistic precision.

Lastly, the study focuses on gendered experiences and does not provide in-depth analyses of how class, ethnicity, and disability intersect. Future research could explore the interplay of these attributes with access to care and resilience strategies.

Despite these constraints, the study provides a valuable look at how average Venezuelans are dealing with a failing healthcare system. The uniformity of themes by respondents suggests that such issues are not few and far between and could be examined further through mixed-method or cross-regional designs.

6. Conclusion

The research investigates how Venezuelan patients experience the breakdown of their national healthcare system. The research used qualitative interviews with thirteen participants from different areas to document their daily experiences with medication shortages and insufficient hospital facilities and specialist shortages and their need to use private and informal healthcare services and the gender-based differences in healthcare access. The stories present a survival mechanism which people develop because of necessity which researchers term as resilience under constraint.

The participants demonstrate that their ability to cope with the situation exists without choice and has its limits as people and their families have developed survival methods through resourceful behavior, mutual support and personal sacrifices which demonstrate the moral and structural weaknesses of failed institutions. Women take on most of the responsibility for caring for others and managing healthcare services and maintaining household stability. The women’s community support work upholds social structures yet demonstrates how women consistently take on unpaid care duties because of societal expectations about their role.

The study uses resilience theory (Norris et al., 2008; Ungar, 2018) to demonstrate that adaptation depends on the availability of social, economic and informational resources which Venezuelan society distributes unfairly. The creative solutions patients develop show human capability but complete healing needs more than individual perseverance. The path to recovery needs people to work together while institutions must change and public trust needs to be restored.

The research joins multiple academic studies that analyze health emergencies as humanitarian crises while revealing their connection to social inequalities and governmental failures and ethical duties. The testimonies from participants demonstrate that resilience should never serve as a reason for the government to step away from its responsibilities. The recognition of resilience should lead to system reconstruction which will transform survival from improvised measures into an entitlement.

Acknowledgments

This research was made possible by the generosity of thirteen Venezuelan participants who shared their life experiences, including some of their most challenging moments. I am deeply grateful to those who assisted with translating the Spanish interviews into English while preserving the integrity and emotion of each story. I also thank everyone who contributed to transcription and offered valuable feedback throughout the research process.

References

Braun, V ., & Clarke, V . (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa

Butler, J. (2016). Frames of war: When is life grievable? Verso Books.

Doocy, S., Page, K. R., de la Hoz, F., Spiegel, P., & Beyrer, C. (2019). Venezuela’s public health crisis: A regional emergency. The Lancet, 393(10177), 1254–1260. https://doi.org/10.1016/S0140-6736(19)30344-0

Doocy, S., et al. (2017). The humanitarian response to the Venezuelan migration crisis: Needs, coordination, and challenges. Journal of Refugee Studies, 30(3), 1–17.

Freitez, A. (2022). Household strategies and remittance dependence in Venezuela’ s economic crisis. Universidad Católica Andrés Bello.

Hetland, G. (2021). Crisis and inequality in Venezuela: The limits of redistribution. Latin American Perspectives, 48(1), 5–22.

Hochschild, A. R. (1995). The managed heart: Commercialization of human feeling. University of California Press.

International Organization for Migration (IOM). (2022). Venezuelan migration and healthcare access report. https://www.iom.int

Melo, S., Rueda-Salazar, A., & García, P. (2023). Gendered health disparities among Venezuelan migrants in Colombia. Global Public Health, 18(2), 215–230.

Norris, F. H., Stevens, S. P., Pfefferbaum, B., Wyche, K. F., & Pfefferbaum, R. L. (2008). Community resilience as a metaphor, theory, set of capacities, and strategy for disaster readiness. American Journal of Community Psychology, 41(1–2), 127–150. https://doi.org/10.1007/s10464-007-9156-6

Ortega, D., Guerra, J., & Salas, R. (2020). Public health and governance in Venezuela: Between collapse and adaptation. Revista de Salud Pública, 22(4), 501–515.

Pan American Health Organization (PAHO). (2023). Health in the Americas: Venezuela country profile. PAHO. https://www.paho.org

Rodríguez, F. (2020). The macroeconomics of Venezuela’s collapse. Center for Economic and Policy Research.

Rueda-Salazar, A., & García, P. (2024). Venezuelan women’s healthcare under migration and crisis: A comparative perspective. Journal of Migration Studies, 12(1), 88–104.

Tronto, J. C. (2013). Caring democracy: Markets, equality, and justice. New York University Press.

Ungar, M. (2018). Systemic resilience: Principles and processes for a science of change in contexts of adversity. Ecology and Society, 23(4), 34. https://doi.org/10.5751/ES-10385-230434

World Health Organization (WHO). (2023). Venezuela (Bolivarian Republic of): Health profile and key indicators. WHO. https://www.who.int

About the author

Hillary Porco

Hillary Porco is a senior researcher at NSU University School whose work centers on global health access, qualitative research, and health policy in Latin America, particularly Venezuela. She completed this project under the mentorship of Dr. Reed Jordan (NYU Public Policy) where her research interests include healthcare inequities, patient narratives, and community-based health systems. She intends to pursue a career in medicine and global public health.

The post Barriers to Healthcare Access in Venezuela: A Qualitative Interview Study of Patient Experiences appeared first on Exploratio Journal.

Abstract

As global demand for sustainable energy solutions grows, nanogenerators have emerged as a promising technology for the ability to convert various forms of energy into electricity at small scales. This review focuses on two particularly unexplored but rapidly advancing subfields: Moisture-sorption-based Energy Harvesting (MSEH) and Transpiration-driven Electrokinetic Power Generation (TEPG). Both technologies convert moisture and water flow into usable electricity without the need for external power sources. MSEH leverages the ion migration resulting from water molecule sorption and desorption on hygroscopic sorbents, while TEPG generates electricity through transpiration-like capillary-driven flow and electrokinetic interactions within charged nanochannels. Through an in-depth analysis of recent eight most highly cited publications in TEPG, this paper examines the impact of sorbent composition, form factor, and water environments on device performance. Researches yielded high results when sorbent were cotton-based systems, particularly those enhanced with conductive additives like Carbon Black and MXene, and in ionized water environments. Furthermore, the recent trends in the number of MSEH and TEPG publications simulates the early-growth trajectories of triboelectric and piezoelectric nanogenerators, highlighting their rapid growth and significant potential. Although the current power outputs are insufficient for commercial electronics, continued advancements in sorbent material and hybrid integration promise a strong future.

Keywords: moisture-induced energy harvesting (MSEH), transpiration-driven electrokinetic power generation (TEPG), nanogenerators, energy harvesting, sorbent material.

1. Introduction

As the global demand for sustainable energy solutions grow, harvesting energy from ambient environments, such as heat, light, and humidity, has opened a vital research frontier. Among these, Moisture-Sorption-Based Energy Harvesting (MSEH) offers a unique advantage: continuous operation under low-resource conditions. Through MSEH, energy can be generated through just a solvent and moisture: when water vapor sticks to a sorbent, it forms bonds and releases heat; when water evaporates off a sorbent, it absorbs heat from surroundings, which produces a cooling effect. During this process, constantly adsorbed and desorbed, water molecules interact with hydrophilic groups on sorbents, leading to ion migration, which generates voltage and current. More specifically, water molecules in the atmosphere interact with the hydrophilic parts of the sorbent. This interaction causes ions to dissociate and move from the surface inward. This movement of charges creates voltage and current—thus generating electricity. Figure 1 illustrates this process, demonstrating how moisture uptake and evaporation drive charge movement within a hygroscopic material.

Likewise, Transpiration Electrokinetic Power Generation (TEPG), employs passive water gradients to generate electricity through nano/microscale interactions. However, the key difference between MSEH and TEPG lie on its source of water uptake. Whereas MSEH absorbs atmospheric moisture, TEPG implements a capillary action to absorb water through a porous, hydrophilic sorbent—just like in plant stems in transpiration. As water evaporates, a continuous water flow remains as the sorbent pulls more water upward, and electrical double layer forms at the carbon/water interface, gathering protons on the wetside. Then, water flows through charged nanochannels and drags along ions, creating potential difference and electric field as a result. Through placing electrodes on two ends of the channel, the process produces usable electricity. Figure (b) illustrates the principle of TEPG, indicating how capillary-driven water flow and ion transport across nanostructures lead to voltage generation in carbon-coated cotton fabric. 

As electricity and power generation for both MSEH and TEPG processes are highly dependent on the sorbent’s ability to attract more water, sorbent selection and optimization possess vital influence on the efficiency of these two processes. Notably, Yun et al. (2019) implements a cellulose membrane for its highly hydrophilic porous abilities, while Kaur et al. (2021) utilizes microporous alumina ceramic. Out of the ten most highly cited articles on MSEH, seven researches indicated uses of hygroscopic cellulose structures, while three researches indicated uses of applied hydrogels, such as the nylon and hydrogel-based composite fiber in Chen et al. (2023). For all sorbents in the top ten highly cited articles in Moisture-Sorption-Based Energy Harvesting, hydrophilicity and porosity stood out as the two main determinants of energy generation performance. Further, an innovative study has been conducted in Bae et al. (2022), as they combine MXene (Ti3C2Tx) with cellulose, applying MXene’s metallic-like electrical conductivity for enhanced charge transport, while enhancing sorbent hydrophilicity, as surface functional groups, including -OH, -O, and -F, attach to MXene. The following review conducts a detailed analysis on notable trends in the MSEH and TEPG renewable energy and researches the effects of different sorbents on power generation, analyzing sorbent selection in the top ten highly cited articles in the field.

II. Market Growth and Applications of Nanogenerators

Despite its “infancy” status in 2015, the nanogenerat market has since grown exponentially. As of 2022, the nanogenerator market is already valued at USD 39.5 Billion, and it is forecasted to reach USD 67.44 Billion by 2030, growing at a CAGR of 8.2% (Virtue Market Research, 2025). Others report the nanogenerator market to grow at a CAGR of 18.8% by 2033 (Verified Market Reports, 2025). More specifically, one forecast estimates the triboelectric nanogenerators (TENGs) market, a section of the nanogenerator market, growth from USD 1.47 Billion in 2023 to about USD 13.56 Billion by 2033, a staggering ~24.88% CAGR over that period (Spherical Insights, 2024). This rapid acceleration reflects not only early-stage promise but also a clear shift toward mainstream adoption across sectors. With applications expanding across healthcare, agriculture, smart infrastructure, and consumer electronics, nanogenerators are gaining traction as key power generation sources of self-powered devices (Market Growth Reports, 2025).

Despite its mere certification in the lab environment, this technology holds vast potential as a low cost, low maintenance energy harvesting method, especially as sustainability and energy efficiency become global priorities. One of MSEH and TEPG’s notable applications include power generation for various sensors and detectors. In example, MSEH can be applied to power air quality monitors, while TEPG may be implemented to soil or plant monitoring systems in precision agriculture. Moreover, MSEH can be applied in wearable electronics to power low-energy health trackers or e-textiles and in smart packaging to maximize freshness of fruit/vegetable deliveries. On the other hand, TEPG can be embed in walls or roofs to harvest energy from water flow and temperature gradients. These technologies can further combine with other popular renewable energy technologies in the market, achieving higher sustainability. For example, a layer of transparent MSEH sorbent may be applied as a final layer on top of photovoltaics, enabling battery-free solar cells to be powered all through night. Although there are various factors still to be considered, both technologies hold significant potential in the future renewable energy generation.

Moreover, nanogeneration plays a vital role in diversifying energy sources. Suburban America suffers from blackout after blackout—Texas alone saw 264 major outages since 2000—costing over USD 15,000 to 90,000 in less than 30 minutes (Statista, 2017). However, nanogenerators may help resolve this issue. Since most nanogenerators, including triboelectric, piezoelectric, thermoelectric, moisture-induced, and transpiration-electrokinetic, do not require a major applied external force to generate electricity, they can minimize the losses from power outages by consistently circulating electricity through the system, until major electricity sources begin to supply as normal. 

III. Research Trends in Nanogenerators

In Figure 3, six keywords have been researched, including triboelectric (TENG), piezoelectric (PENG), thermoelectric (TEG), moisture-induced (MING), transpiration-driven (TEPG), and nanogenerators, using the Web of Science database accessed through the University of Washington online library. The primary y-axis exhibits the number of peer-reviewed articles for each nanogenerator subcategory, while the secondary y-axis indicates the number of peer-reviewed articles related to nanogenerators as a whole. Among the various nanogenerators explored over the past decade, 90% of them involves mechanical energy conversion—TENG (67.5%) and PENG (28.4%)—while other technologies rely on thermal or chemical energy conversions—TEG (2.2%), MING (1.6%), and TEPG (0.3%).

As shown in Figure 3, the number of peer-reviewed articles on nanogenerators has been rapidly increasing over the past decade, with an approximate increase of 549.65% between 2015 and 2024—a near sixfold increase in under a decade. This growth far exceeds increasing rates in broader “energy harvesting” publications, which also grew substantially—from 3,858, in 2015 to 9,226 in 2024. Most notable increases happened in mechanical energy subcategories, including the triboelectric nanogenerators, more than tenfolding from 218 in 2015 to 2,788 in 2024, and piezoelectric nanogeneratores, tripling from 202 in 2015 to 694 in 2024. The sharp increase in the number of publications not only indicate rapidly growing academic curiosity and attention on nanogenerators but also validates the technical feasibility of nanogenerators. 

Although not as significant as triboelectric and piezoelectric nanogenerators, more niche branches, such as the moisture-induced nanogenerators and transpiration-driven electrokinetic power generators have attracted attention to increasing number of researchers, as shown from the data above: moisture-induced increased from 3 in 2015 to 71 in 2024, while transpiration-driven increased from 1 in 2015 to 15 in 2024. Nevertheless, since water harnessing-induced energy generation is in the early stage as demonstrated by the figures above, it follows the early growth trajectories of TENG and PENG, as most research on these subfields have been conducted in the recent five years. This highlights a potential rocketing interest in the following couple of years. In fact, upon further data analysis, as shown in Figure 4, all keywords, even including “energy harvesting,” recorded the maximum R^2 value when data points were put into quartic equations. This quartic regression implies a steep increase over the past decade and a continuously increasing trajectory in near future. 

Table 1. Correlation coefficient values of the number of publications related to nanogenerators as a function

Since Moisture-Sorption-based Energy Harvesting (MSEH) and TEPG operate in low-light, low-thermal, and off-grid humid environments, its versatility projects a rapid growth in the following decade. For example, MSEH devices can be integrated into smart textiles, where nanostructured hygroscopic layers embedded in clothing harvest moisture and generate electricity to power low-energy sensors or health-monitoring wearables, eliminating the need for external batteries. Although designing and applying the new technology may take some time, this opens possibilities for energy-autonomous wearables in remote or off-grid settings. Moreover, with MSEH and TEPG’s recent successes to power small electronics in laboratory settings, scientists increasingly blur the boundaries between the nanogenerators’ subcategories, experimenting hybrid systems to maximize the system’s energy efficiency. For example, MXene, initially studied in TEPG for its conductive and hydrophilic properties, are now being explored in pizoelectric systems for its high mechanical flexibility and strength as flexible 2D structures, ideal for flexible sensors and energy harvesters,  surface functionalization, as the functional groups, such as hydroxl, fluoride, and oxide, alter symmetrical structures to induce piezoelectricity, and high conductivity (Bae et al., 2022). This cross-disciplinary integration reinforces the high potential in the advances of currently less-spotlighted subcategories of nanogenerators. 

IV. TEPG Key Sorbents and Performance Comparison

This section introduces key sorbents in the most cited literature within the TEPG community in the last decade and compares its specific performance, analyzing which composite materials recorded the highest efficiency.

One of most notable cellulose-based TEPG research, Yun et al. (2019), introduced a thin planar form cotton fabric (90 x 30 x 0.12 mm) coated with carbon black as the capillary medium. This carbon black coating provides a conductive pathway to channel the charge, also known as the “psuedostreaming” current. A single 90 x 30 mm panel produced an open-circuit voltage of 0.53V and short-circuit current 3.91 uA, leading to 2.1 uA power generation. The energy output per volume was reported as 1.14mWh/cm^3, indicating a specific energy capacity of 4.1 J per cm^3. This implies that a 2.2V LED light requiring ~20 mA could be lighted with several TEPG panels in series/parallel circuits, highlighting its vast potential as a passive energy generation source. The application of carbon black, a conductive and high-surface-area material, while the hydrophilic cotton provided continuous flow, resulting in modest energy generation at a low cost. 

Moreover, a study by Bae et al. (2020) further enhances the set up in Yun et al., applying calcium chloride (CaCl2) on same carbon-coated cotton fabric, to create a closed-loop water cycle that mimics transpiration. After a few drops of water are initially added to begin the capillary flow, as water evaporates, the CaCl2 absorbs the vapor and releases it to re-wet the fabric. The device remained in similar dimensions to Yun et al. (2019) but with an additional layer of hygroscopic CaCl2. This research’s most significant achievement lies on prolonged period of TEPG operation, as what would normally have continued for merely an hour continued energy generation for several days. Results indicated that the implementation of CaCl2 enhanced the output, as the voltage rose to 0.74 V and current to 22.5 uA from approximately 0.5 V and 4uAwithout. This increased the power generation to 16-17uW, extending its period to light an LED from an hour or two to a week.

Another innovative study has been conducted by Lv et al. (2020), as the research presents a simple, low-cost generator using commercial air-laid paper with conductive carbon ink. Here, the paper acts as the sorbent, providing porous network for water transport, while the carbon ink forms a conductive film/electrode on the paper, simulating transpiration with commercially available products. Having one end of the paper strip slightly dipped in lithium chloride (LiCl) solution and the other end exposed, a continuous capillary flow has been formed to generate electricity. The strip is very thin and flexible (6 cm x 1 cm), and a single trip produced a continuous open-circuit voltage of 0.35 V and 33.9 uA under ambient conditions, corresponding to a raw 10^-5 W power generation. However, when dipped in de-ionized water, the performance recorded 0.17 V and 7.2uA. This further indicates that higher electrolytic solutions leads to better performance, due to enhanced electrical conductivity. The simple set up provided great stability, as the system had a consistent output over 120 hours, powering small electronics, including an LED, a pocket calculator, a digital watch, directly from TEPG. Despite cheap materials, the research widened the scope of materials implemented in TEPG, nearing power generation in advanced lab settings. 

At last, other than natural cellulose-based fabric, Kaur et al. (2021) employs a porous ceramic (α-alumina) block as the water transport medium. The alumina had ~35% porosity with 50-200 interconnected pores. Without any additional conductive coating, the bare ceramic generates electricity once wet, due to its surface charge—alumina pores carry a negative charge in water. In this study, a 3.0 × 3.0 × 0.3 cm^3 rigid alumnia block was used, substantially thicker than the fabrics used in other studies. To simulate plant transpiration process, one end of the block is placed in water (~1.5 cm deep), while the other end is exposed to air, establishing a capillary-driven water transport pathway that enables continuous upward water flow and evaporation. As the alumina block was a rigid ceramic, it demonstrated lack of flexibility. However, the device was durable enough to function over a year in ambient conditions, compensating for the inflexibility. The system recorded an open-circuit voltage up to ~0.27 V stable output in standard ambient conditions, while it recorded a low short-circuit current ~1.2uA. The maximum instantaneous power was calculated as only 0.324 uW, but this output could be sustained over months with stable set up. Overall, the microporous alumina proved effective due to Alumina’s surface chemistry: Alumina’s negative OH- creates a strong EDL, and as water moves H+ cations drag, accumulating charge. The alumina ceramic demonstrated clear trade-offs: while it remained stable over a long period, avoiding issues of material delaminatoin that carbon coatings faced, it also recorded a particularly low short circuit current, as the conduction is only via ionic streaming, not electronic, leading to high internal resistance. 

Table 2. Performance Comparison of TEPG Studies Published in Recent Literature. Specific power (mW/cm³) was calculated based on maximum power using Voc and Isc. Exponent [1] indicates testing in deionized (DI) water, while [2] indicates testing in ionized water.

Table 2 demonstrates the 8 most highly-impacted publications on TEPG, including information on each research’s performance (open-circuit voltage & short-circuit current), sorbent composition, form factor, and specific power. The table reveals natural cellulose-based sorbent to be the predominant sorbent model in TEPG, often enhanced with advanced materials that exhibit high hygroscopicity. In fact, the most highly cited and best-performing articles—Bae et al. (2020), Yun et al. (2019), Bae et al. (2022)—all used cotton combined with another material as their sorbent, highlighting the potential of enhanced cotton in TEPG.  Additionally, the exponent numbers in the performance section indicate whether the experiment was conducted in deionized water [1] or ionized water [2], for more accurate comparisons.
The specific power was calculated through the following two equations with available open-circuit voltage and short-circuit current data. The calculated specific power is the maximum specific power that could be achieved in each experimental setting. 

Some researches, including Yu et al. (2024) and Su et al. (2023), investigated performances under both deionized water and ionized water environments. As shown in Table 2,  both studies yielded better performance in ionized (salt) water. Moreover, although Bae et al. (2020) replicates the experimental settings in Yun et al. (2019), it yielded greater specific power, as Bae et al. (2020) conducts the experiment in ionized water while Yun et al. (2019) in deionized water. When TEPG is conducted in ionized water environments, the presence of ions enhances charge carrier density and electrokinetic flow, increasing both voltage and current outputs. Though, values from deionized (pure) water hold greater significance, as most situations where TEPG and MSEH will be applied comprise ambient moisture or deionized water. However, in Bae et al. (2020), it was noted that without CaCl2 or other salts, the device stopped working within ~1 hour due to lack of moisture retention and ion supply. This still implies that there is a long journey that awaits before TEPG is applied commercially.

In conclusion, the comparative analysis of recent TEPG studies highlights the significant impact of sorbent composition, form factor, and water environments on device performance. Cotton-based systems, particularly those enhanced with conductive additives like Mxene and polyaniline (Bae et al., 2022), exhibit the highest specific power outputs—reaching up to 658.93 uW/cm^3. In contrast, ceramic and simpler textile setups, such as those using microporous alumina or uncoated cotton, display notably lower performance. Although it current performance do not yield enough to power daily electronics, considering its continuously rising performance over the past years and its ‘supernova’ status suggest a promising future. 

V. Conclusion

By examining the current research trends in moisture-sorption-based and transpiration-driven electrokinetic energy harvesting, it is evident that they hold significant potential as rapidly growing subfields within nanogenerator research. Both the number of publications and the performance of reported devices have shown a steep upward trajectory in recent years. Although current outputs in TEPG remain in the microampere range—insufficient for commercial electronics—continued progress highlights its potential viability, especially when the system reaches the milliampere range. Moreover, as hybrid technologies, where MSEH and TEPG systems complement more promising energy generation technologies, develop, they will play a vital role in decentralized, location-optimized power solutions for sensors, wearables, and other low-energy devices. 

IV. Acknowledgement

I would like to thank Professor Dichiara for his guidance throughout this study. His consistent support and insights have been pivotal to completing this research.

REFERENCES

Xu, J., Wang, P., Bai, Z., Cheng, H., Wang, R., Qu, L., & Li, T. (2024). Sustainable moisture energy. Nature Reviews Materials. https://doi.org/10.1038/s41578-023-00643-0

Bae, J., Tae Gwang Yun, Bong Lim Suh, Kim, J., & Kim, I.-D. (2020). Self-operating transpiration-driven electrokinetic power generator with an artificial hydrological cycle. Energy & Environmental Science, 13(2), 527–534. https://doi.org/10.1039/c9ee02616a

‌ Yun, T. G., Bae, J., Rothschild, A., & Kim, I.-D. (2019). Transpiration Driven Electrokinetic Power Generator. ACS Nano, 13(11), 12703–12709. https://doi.org/10.1021/acsnano.9b04375

Bae, J., Soo Kim, M., Oh, T., Lim Suh, B., Gwang Yun, T., Lee, S., Hur, K., Gogotsi, Y., Min Koo, C., & Kim, I.-D. (2022). Towards Watt-scale hydroelectric energy harvesting by Ti 3 C 2 T x -based transpiration-driven electrokinetic power generators. Energy & Environmental Science, 15(1), 123–135. https://doi.org/10.1039/D1EE00859E

‌ Kaur, M., Ishii, S., Nozaki, R., & Nagao, T. (2021). Hydropower generation by transpiration from microporous alumina. Scientific Reports, 11(1), 10954. https://doi.org/10.1038/s41598-021-90374-5

‌Yu, Z., Mao, J., Li, Q., Hu, Y., Tan, Z., Xue, F., Zhang, Y., Zhu, H., Wang, C., & He, H. (2024). A Transpiration-Driven Electrokinetic Power Generator with a Salt Pathway for Extended Service Life in Saltwater. Langmuir, 40(10), 5183–5194. https://doi.org/10.1021/acs.langmuir.3c03390

‌ Su, H., Azadeh Nilghaz, Liu, D., Dai, L., Tang, B., Wang, Z., Razal, J. M., Tian, J., & Li, J. (2023). Self-operating seawater-driven electricity nanogenerator for continuous energy generation and storage. Chemical Engineering Journal Advances, 14, 100498–100498. https://doi.org/10.1016/j.ceja.2023.100498

‌ Lv, Y., Gong, F., Li, H., Zhou, Q., Wu, X., Wang, W., & Xiao, R. (2020). A flexible electrokinetic power generator derived from paper and ink for wearable electronics. Applied Energy, 279, 115764. https://doi.org/10.1016/j.apenergy.2020.115764

‌ Luo, G., Xie, J., Liu, J., Luo, Y., Li, M., Li, Z., Yang, P., Zhao, L., Wang, K., Maeda, R., & Jiang, Z. (2023). Highly Stretchable, Knittable, Wearable Fiberform Hydrovoltaic Generators Driven by Water Transpiration for Portable Self‐Power Supply and Self‐Powered Strain Sensor. Small, 20(12). https://doi.org/10.1002/smll.202306318

‌ Triboelectric Nanogenerator (TENG) Market Size, Forecasts 2033. (2023). Spherical Insights. https://www.sphericalinsights.com/reports/triboelectric-nanogenerator-teng-market

‌ Nanogenerator Market Size, Share & Growth By 2033. (2025). Marketgrowthreports.com. https://www.marketgrowthreports.com/market-reports/nanogenerator-market-113380

‌ Research,https://virtuemarketresearch.com, V. M. (n.d.). Nanogenerator Market | Size, Share, Growth | 2023 – 2030. Virtue Market Research. https://virtuemarketresearch.com/report/nanogenerator-market

Verified Market Reports. (2025, February 17). Nanogenerator Market in United States 2025 | Size, Share and Future Outlook. Verified Market Reports. https://www.verifiedmarketreports.com/product/nanogenerator-market/

‌ Number of power outages by key U.S. state 2017. (n.d.). Statista. https://www.statista.com/statistics/1078354/electricity-blackouts-by-state/

‌

About the author

Jinwook (James) Chang

Jinwook is an aspiring environmental innovator bridging environmental science, energy, and policy to drive sustainable solutions from the lab to the real world. Part scientist, part historian, James draws on lessons from the past to shape innovations that power a cleaner tomorrow. (Website: https://www.jamestechgarage.com/)

The post Novel Nano generation Technologies for Harvesting Electricity from Water-Induced Processes appeared first on Exploratio Journal.

Abstract

With more industrial developments, there is an increased demand for clean energy. One source that can provide clean energy is hydrogen-based fuels. In the paper, the term “hydrogen power” means all energy generation methods based on hydrogen, such as hydrogen combustion or hydrogen fuel cells. However, generating hydrogen at an industrial scale requires scaling up hydrogen generation processes such as electrolysis. This depends on selecting suitable catalysts to expedite the process. This paper provides a review of existing theories that help identify potential catalysts for this process. Specifically, d-band theory and spinel theory predict the activity descriptors for Oxygen Evolution Reactions and Hydrogen Evolution Reactions, respectively.

Keywords: water electrolysis, hydrogen evolution reaction, oxygen evolution reaction, catalyst, d-band theory, spinel theory

The current world and social structure depend on generating electricity. There are various approaches to generating electricity, with the main options being using fossil fuels (combustion), renewable energy, and nuclear energy. In most fossil fuel generators, byproducts, mainly in the form of greenhouse gases (GHGs), will be produced due to the combustion reaction, and GHGs are capable of warming up the Earth’s climate and polluting the atmosphere (Markandya & Wilkinson, 2007, p. 979). According to data collected by a research team that published the findings in the journal “Earth System and Scientific Data,” a rigorous academic journal with very transparent processes, which are then compiled by Climate Watch, an organization under the World Resource Institute, the annual GHG emissions from the entire world, measured in billion tons of CO2 equivalent, from 1850 to 2016, it increased from 1.4373 billion tons to 46.50 billion tons or approximately a 3135 % increase in emissions. Most of these emissions come from energy demands (World Resources Institute, 2022). According to the same source, about 33% of the world’s emissions in 2021 came from electricity and heating, the largest sector of global GHG emissions (World Resources Institute, 2022). The data shows that energy production is a considerable portion of the global GHG emissions. Thus, a clear and most impactful solution to climate change will be finding a clean or low-carbon energy source, as it will directly address 30% of global emissions.

There are many ways to produce clean and non-polluting energy, such as solar energy, which is generated directly from sunlight. However, these methods are not perfect. Most renewable energy sources, particularly solar energy, are intermittent or unstable, requiring additional infrastructure to account for the problem (Mathew, 2022, p. 5). This, combined with the lack of a large and powerful energy storage system, leads to grids with renewable sources having to depend on fossil fuels, creating additional GHG emissions (Mathew, 2022, p. 5). Additionally, these renewable energy sources consume many resources, particularly land. For instance, a 1000 MW fossil fuel power plant requires 1-4 km2 of land for the entire facility, while renewables require a lot more land, with solar requiring 20-50 km2, wind requiring 50-150 km2 , and biomass requiring 4000-6000 km2 (Rashad & Hammad, 2000, p. 213). These factors combined make most of the current renewable energy systems unable to generate electricity as effectively as methods like fossil fuel. They could potentially release additional GHGs from the extra land use and infrastructure. However, not all renewable energy sources have that problem, and using hydrogen power can prevent these problems.

Hydrogen has many advantages as an element in itself. It is a highly energy-dense element (in terms of mass), making it comparable with other standard energy production methods, such as fossil fuels like petroleum or coal, as shown in Figure 1 (U.S. Department of Energy, n.d.-b). Hydrogen power has an effective energy system that is proven by fossil fuels. The underlying principle of hydrogen power is the same as that of fossil fuels, converting hydrogen combustion’s thermal energy to steam’s kinetic energy by boiling water, finally pushing a turbine with that kinetic energy, and generating electricity. This process has been proven by decades of application and is widely used today. Approximately 42% of all electricity generation in the United States uses steam turbines (U.S. Energy Information Administration, 2023). Hydrogen also comes with the added benefit of not producing any pollutants when burned, with its byproduct being only water, which is the product of the hydrogen combustion reaction. Not only that, but hydrogen can also be used in fuel cells, which is another way to produce power efficiently, with its efficiency ranging from 40% to 60% (U.S. Department of Energy Energy Efficiency & Renewable Energy, 2010). Hydrogen is also an essential industrial element, as shown in Figure 2, commonly used in industries like agriculture, where it can synthesize ammonia, a key component in all modern fertilizers (WHA International Inc, 2023; World Nuclear Association, 2024).

Background and Literature Review

Despite the benefits of a hydrogen-based energy system, getting hydrogen clean is a challenge. There are various ways to produce hydrogen; however, most are produced using fossil fuels, which release GHGs. In fact, approximately 95% of the world’s hydrogen production is based on fossil fuels and releases GHGs (Rosenow, 2022). In the case of hydrogen power, generating power with hydrogen that has a considerable amount of carbon footprint attached to it during its production process will make the purpose of hydrogen power obsolete, and thus, using renewable or clean hydrogen is essential to ensuring the benefits of hydrogen power can be released at full potential. To better classify different types of hydrogens, color codes are assigned to them, with different colors representing different carbon footprint levels, as shown in Table 1 (National Grid, 2025). Based on the classification of hydrogen, for hydrogen power to be completely carbon-free, using green hydrogen is the best approach. Electrolysis is essential to creating green hydrogen, as detailed in Table 1. It works by splitting water molecules, which consist of two hydrogen atoms and one oxygen atom. Thus, running a specific voltage through the water molecules will break the chemical bond between the atoms and release the atoms themselves. This method is carbon neutral and does not release additional GHGs, assuming the electricity used for the electrolysis is carbon neutral.

In an electrolysis reaction, the electric current passed through serves as the activation energy of the reaction to dissociate water molecules into hydrogen and oxygen. This happens because when an electrical current is passed through the anode, cathode, and the water itself, the water molecules undergo oxidation at the anode, producing oxygen gas and releasing electrons. At the same time, the hydrogen from the oxidation is also reduced at the cathode, where hydrogen ions are being reduced by gaining an electron from the oxidation at the anode, finally creating both hydrogen and oxygen gas at the ends. This process, however, is very energy- intensive as it needs to overcome a strong energy barrier presented by the OH bonds in water. These bonds in water have an average bond energy of 461.5 kJ/mol, an accepted value, which is quite strong (Song & Le, 2013). As a result, for the reaction to occur, more energy has to be passed through, making the process less efficient and more challenging to complete, decreasing the possibility for it to be used in large-scale industrial processes such as generating hydrogen in large enough quantities to supply power plants without a method to decrease the amount of energy used.

As a result, catalysts are being used to lower the amount of energy needed in this process, as a catalyst can lower the activation energy of reactions while not consuming itself during the reaction (U.S. Department of Energy, n.d.-a). This can be used to boost the amount of hydrogen acquired from electricity, improving the efficiency of the electrolysis reaction. There are various types of catalysts with pros and cons, as well as having properties that are more inclined to support either the oxidation or the reduction reaction. In the industry, the key to successfully creating and commercializing the system for broad public use is to find a suitable catalyst that balances various qualities. This can be done by reliably identifying the activity descriptor, which represents a quantifiable indicator for a catalyst’s capability to catalyze a specific reaction, in this case, the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER). Thus, this paper focuses on reliably identifying the activity descriptors for the hydrogen and oxygen evolution in the water electrolysis reaction.

HER and OER

As established before, the HER reduces hydrogen ions, and the OER oxidizes water molecules to release oxygen. Generally speaking, the OER is more energy-intensive and slower than the HER because it has a more complex reaction mechanism. Specifically, there are four electron transfer processes during the OER, as shown in Figure 3. This multi-step process requires breaking the strong OH bond in water to generate oxygen. The four-electron transfer process also means that forming multiple intermediates is challenging (J. Li, 2022). This, in turn, creates kinetic barriers, making it much slower and requiring a higher overpotential (extra energy) to make the reaction happen. On the other hand, the HER only requires 2 electron transfers, meaning it is essentially a more straightforward process that is also easier to achieve comparatively (Dubouis & Grimaud, 2019).

Additionally, during the OER, various intermediates containing oxygen form on the catalyst and get absorbed onto its surface. The formation of these intermediates is a critical step in the process, as this is the only way catalysts can facilitate the breaking and formation of oxygen molecules, which are the intended product. However, this process is challenging to balance as too much binding force will slow down the overall kinetics of the reaction, and the reaction could become “stuck” (J. Li, 2022). This is not a problem for HER because hydrogen atoms and their intermediates are much smaller, easier to release from the catalyst, and involve fewer steps to convert to their molecular form of H2 (Dubouis & Grimaud, 2019).

d-band Theory and Catalyst Performance Prediction for OER

The d-band theory is a fundamental concept used to explain and predict the performance of a transition metal catalyst in a reaction, where it applies specifically to transition metals because of the filled d-orbitals (Bhattacharjee et al., 2016). The theory mainly revolves around the d-band center, which is the average energy level of electrons in the d orbital relative to the Fermi level and is the highest possible energy for electrons at absolute zero, which serves a crucial role in catalytic activity (Bhattacharjee et al., 2016).

The theory’s predictions depend on the relative position of the d-band center (Bhattacharjee et al., 2016). Regarding OER specifically, the theory can predict how well a catalyst binds with oxygen-containing intermediates such as OH, O2, and OOH. When the d- band center is closer to the Fermi level, the interaction between the catalyst and the reaction intermediates generally increases. Conversely, a larger proximity will weaken these interactions (Bhattacharjee et al., 2016). This does not mean that aiming for the highest d-band center (i.e., closest proximity) will be the best. As explained above, a too-strong interaction will hinder oxygen molecules’ release (desorption) and will have a lower efficiency overall. Using the same logic, if the d-band center is too low, the intermediates will not bind strongly, leading to a higher activation energy requirement for the reaction (Bhattacharjee et al., 2016). As a result, catalysts with an optimal d-band center are defined as those that can strike a good balance between adsorption and desorption, allowing for an efficient intermediate process without energy losses, meaning higher overall efficiency. This is because the optimal balance of the d-band center can minimize overpotentials, which will enhance catalytic performance by wasting less energy, leading to a higher energy efficiency, thus allowing more energy to be applied to the reaction itself, increasing the reaction rate (Bhattacharjee et al., 2016). Additionally, over-potential has side effects like releasing heat and increasing material/catalyst stress, which can lead to a shortened lifetime and be suboptimal for future industrial operations. On the other hand, a lower overpotential will help maintain the stability of the catalyst, reducing wear and tear and, as a result, prolonging its lifespan, ultimately leading to the development of cheap and practical catalysts (Dubouis & Grimaud, 2019).

Application of d-band Theory in Catalyst Manufacturing and Design

The d-band theory can be applied in catalyst designs and manufacturing, with it serving as the guiding principle for tailoring catalysts by adjusting the electronic structure via alloying or doping various materials (Chen & Zhang, 2022). The theory has been validated through Density Functional Theory (DFT) calculations, a standard tool for validating and studying catalysts, supporting the effectiveness of the d-band theory (Nørskov et al., 2011). As a result, the d-band Theory can be used as a foundational principle for designing more advanced catalysts that work best for water electrolysis because the d-band theory can enable researchers to have something tangible that they can change for different results (Chen & Zhang, 2022).

Limitations of the d-band Theory for Catalyst Design

Nevertheless, the d-band theory has downsides as it does not apply universally to all catalyst materials (Bhattacharjee et al., 2016). As different catalysts have different electronic structures and surface morphologies, these factors lead to different catalysts requiring a different descriptor that considers and optimizes these factors (B. Wang & Zhang, 2022). For instance, the d-band theory works because it is based on the d orbital electrons, which can play a significant role in reaction intermediates (Bhattacharjee et al., 2016). However, the theory cannot accurately predict the catalytic behavior for materials like metal oxides, on-transition metals, and complex systems, such as perovskites and platinum, because the materials’ structures are too complex to be oversimplified by the d-band behavior in d-band theory (Gorzkowski & Lewera, 2015; B. Wang & Zhang, 2022).

Brief Explanation of HER Mechanism

Moving on to the hydrogen side of the reaction. The HER is a crucial aspect of water electrolysis and the source of green hydrogen. However, despite being generally more straightforward regarding reaction mechanism, HER does not have a definitive or universal theoretical model to predict catalyst performance, unlike OER, which has the d-band theory (Zheng et al., 2018). This makes identifying the optimal catalyst for the reaction completely different and requires understanding and analyzing unique activity descriptors that are not universally applicable to HER. Thus, an overview of HER catalyst design theories is presented below.

Cation Distribution and Spinel Theory for HER Catalyst Design

For HER, a very promising approach in catalyst design is using Spinels. Spinels are a type of crystalline material with the general formula of AB2O4, where “A” and “B” represent different metal cations, and “O” represents oxygen (Elkholy et al., 2017). Spinels generally have a cubic crystal structure characterized by two potential types of sites where the cation can be situated: the Tetrahedral or A site and the Octahedral or B site (Elkholy et al., 2017). These materials are known for their robustness, high thermal stability, and electrical conductivity, making them ideal for industrial applications after an optimal catalyst based on Spinels is successfully developed (Elkholy et al., 2017). One example of spinel is CoFe2O4. In this case, the Co2+ ion occupies the tetrahedral (A) sites while the Fe3+ ion occupies the octahedral (B) sites, together with the four oxygen atoms forming the framework of the molecule (Gomaa et al., 2024).

As there are two sites where cations can reside, a balance needs to be reached between these cations, denoted by δ. The balance significantly affects the catalytic activity for HER, and research has demonstrated that catalysts with an optimal cation distribution can substantially improve catalytic performance. This is because there is a better electron transfer process for the reaction and a more optimized binding strength of the hydrogen intermediates similar to hydrogen (Gomaa et al., 2024). For instance, the spinel of CoFe2O4 is an optimal catalyst for HER. CoFe2O4 has a cation distribution of δ of 0.33, and further research shows that CoFe2O4 exhibits low overpotentials, as low as 66 mV, which is advantageous as low levels of overpotential generally translate to a higher reaction efficiency (Gomaa et al., 2024; Niu et al., 2020). The arrangement of cations in the sites will influence the electronic structure of the spinel, similar to the d-band theory but with much more complicated mechanics (Gomaa et al., 2024). This change in electronic structure will optimize the interaction with hydrogen intermediates, reaching the right balance of binding strength (Exner, 2022).

Hydrogen Adsorption and Desorption Energy for HER Catalyst Analysis and Design

In addition to cation distribution, hydrogen and hydroxyl ions (OH-) adsorption energy has a crucial role in HER, which is especially important in alkaline media with a higher concentration of hydroxyl ions. In a study conducted by Baghban and colleagues, they used DFT to calculate the adsorption energies and achieved a 96.7% accuracy on predicting the behavior of actual catalysts (2021). An ideal catalyst for HER will exhibit a Gibbs free energy close to zero for hydrogen adsorption, meaning it will need less and less energy for the reaction to happen, or, in other words, a lower activation energy given that the catalyst can also efficiently adsorb and dissociate water to provide hydrogen for the reaction (Hu et al., 2016).

Outlook for HER Catalyst Descriptor

Due to the unique nature of the HER, it is essential to consider the current outlook for catalyst design. There are promising developments for HER catalyst descriptor analysis, but a universal and consistently working descriptor theory for HER still does not exist (Dubouis & Grimaud, 2019). Unlike in the case of OER, HER cannot use d-band theory because of complications, and other theories suffer from the same problem, causing the lack of a consistently working universal descriptor theory for HER. The HER involves diverse electronic properties observed in materials available for HER, making it very challenging to establish a single definitive set of rules or descriptors that can apply universally (Du et al., 2025). Admittedly, the d-band theory cannot predict catalysts under all circumstances as previously established, but it is still a valuable OER catalyst analysis approach, which is “better” than the current HER situation. As such, future HER research should focus on developing a more comprehensive theory, allowing the community to progress towards a more comprehensive theory while enabling other potential research areas.

Conclusion

As established previously, energy is critical for society, so developing a clean energy source is also essential. However, current energy generation options have significant limitations, such as pollution or scalability. Specifically, despite being cheap and efficient, fossil fuels are very polluting, while on the other hand, despite being clean, solar power and other renewables are less efficient, intermittent, and consume large amounts of resources to create a working system (Rashad & Hammad, 2000). An alternative to all these methods exists: using hydrogen as a clean fuel source. Hydrogen is an excellent alternative to fossil fuel, as it has a high energy density and low emissions (Hossain Bhuiyan & Siddique, 2025). In addition to that, hydrogen also matters in other fields, as it is an essential industrial resource. The side product generated by green hydrogen production, oxygen, also has an essential industrial application, making green hydrogen production even more tempting (Eckl et al., 2025; U.S. Energy Information Agency, 2024).

Despite these benefits, hydrogen production is mainly achieved using fossil fuel (steam reforming), where 62% of all hydrogen production relies on natural gas (steam reforming), and around 99% of all hydrogen production requires fossil fuel and leads to carbon emissions (International Energy Agency, 2024). Therefore, when hydrogen is used in a hydrogen-based powerplant or a hydrogen fuel cell, it will likely have a carbon footprint comparable to that of normal fossil fuel. As such, developing a completely carbon-neutral method to produce hydrogen, specifically electrolysis, is essential. Nevertheless, electrolysis has problems because it is inefficient and not easily scalable, especially for industrial operations. To solve this issue, catalysts that can meet the requirements of an industrial system can be used to make the reaction less energy-consuming, hence allowing us to achieve efficient large-scale water electrolysis. This requires a theory that can reliably identify the respective activity descriptors for both the HER and OER. As discussed throughout the paper, the catalyst for the OER can be predicted using the d-band theory by optimizing the adsorption and desorption of oxygen-containing intermediates, whereas HER performance depends on more material-specific approaches and is generally harder to define. However, methods to identify the optimal HER catalysts exist, including understanding cation distribution in spinel systems that use catalysts of a specific format, such as CoFe2O4. These strategies demonstrate that by understanding electronic and atomic structures, specifically the d-band center in transition metal-based catalysts and spinel cation balance, the performance of catalysts for water electrolysis can be effectively and quantitatively predicted. In conclusion, optimizing catalyst selection and advancing in d-band and spinel theory or other potential theories are necessary for the bigger goal of large-scale clean hydrogen production. Therefore, more focus, funding, and research should be directed towards understanding and developing these catalysts to enable their industrial use, from energy production to industrial uses, not only laboratory applications.

This paper provides a review of the current available methods in identifying the activity descriptors for both the HER and OER and does not aim to find new methods or theories. To solve the problem of catalyst design, more experimental trials and data on catalyst designs need to be done, which will enable the potential for further understanding of catalysis or even potentially finding the catalyst that can be applied in the industry. Also, this review does not include all aspects of the research, as the paper only discussed the theories related to water electrolysis and their associated catalysts, which themselves can still benefit from catalysis research developments in other fields. Specifically, catalysis research has been done in fields other than water electrolysis catalysis, and we anticipate that future work could incorporate findings from those fields into the field of water electrolysis. Doing so can compile a more comprehensive and effective review, providing more value to the field.

References

Baghban, A., Habibzadeh, S., & Zokaee Ashtiani, F. (2021). On the evaluation of hydrogen evolution reaction performance of metal-nitrogen-doped carbon electrocatalysts using machine learning technique. Scientific Reports, 11(1), 21911. https://doi.org/10.1038/s41598-021-00031-0

Bhattacharjee, S., Waghmare, U. V., & Lee, S.-C. (2016). An improved d-band model of the catalytic activity of magnetic transition metal surfaces. Scientific Reports, 6(1), 35916. https://doi.org/10.1038/srep35916

Chen, Z., & Zhang, P. (2022). Electronic structure of single-atom alloys and its impact on the catalytic activities. ACS Omega, 7(2), 1585–1594. https://doi.org/10.1021/acsomega.1c06067

Du, J., Yan, Y., Li, X., Chen, J., Guo, C., Chen, Y., & Wang, H. (2025). A mechanism-guided descriptor for the hydrogen evolution reaction in 2D ordered double transition-metal carbide MXenes. Chemical Science (Royal Society of Chemistry: 2010), 16(21), 9424– 9435. https://doi.org/10.1039/d4sc08725a

Dubouis, N., & Grimaud, A. (2019). The hydrogen evolution reaction: from material to interfacial descriptors. Chemical Science, 10(40), 9165–9181. https://doi.org/10.1039/c9sc03831k

Eckl, F., Moita, A., Castro, R., & Neto, R. C. (2025). Valorization of the by-product oxygen from green hydrogen production: A review. Applied Energy, 378(124817), 124817. https://doi.org/10.1016/j.apenergy.2024.124817

Elkholy, A. E., El-Taib Heakal, F., & Allam, N. K. (2017). Nanostructured spinel manganese cobalt ferrite for high-performance supercapacitors. RSC Advances, 7(82), 51888–51895. https://doi.org/10.1039/c7ra11020k

Exner, K. S. (2022). On the optimum binding energy for the hydrogen evolution reaction: How do experiments contribute? Electrochemical Science Advances, 2(4). https://doi.org/10.1002/elsa.202100101

Gomaa, A. K., Zonkol, M. G., Khedr, G. E., & Allam, N. K. (2024). Cation distribution: a descriptor for hydrogen evolution electrocatalysis on transition-metal spinels. EES Catalysis, 2(6), 1293–1305. https://doi.org/10.1039/d4ey00121d

Gorzkowski, M. T., & Lewera, A. (2015). Probing the limits of d-band center theory: Electronic and electrocatalytic properties of pd-shell–pt-core nanoparticles. The Journal of Physical Chemistry. C, Nanomaterials and Interfaces, 119(32), 18389–18395. https://doi.org/10.1021/acs.jpcc.5b05302

Hossain Bhuiyan, M. M., & Siddique, Z. (2025). Hydrogen as an alternative fuel: A comprehensive review of challenges and opportunities in production, storage, and transportation. International Journal of Hydrogen Energy, 102, 1026–1044. https://doi.org/10.1016/j.ijhydene.2025.01.033

Hu, G., Tang, Q., & Jiang, D.-E. (2016). CoP for hydrogen evolution: implications from hydrogen adsorption. Physical Chemistry Chemical Physics: PCCP, 18(34), 23864–23871. https://doi.org/10.1039/c6cp04011j

International Energy Agency. (2024). Global Hydrogen Review 2024. https://www.iea.org/reports/global-hydrogen-review-2024 0

Li, J. (2022). Oxygen Evolution Reaction in Energy Conversion and Storage: Design Strategies Under and Beyond the Energy Scaling Relationship. Nano-Micro Letters, 14(1), 112. https://doi.org/10.1007/s40820-022-00857-x

Markandya, A., & Wilkinson, P. (2007). Electricity generation and health. Lancet, 370(9591), 979–990. https://doi.org/10.1016/s0140-6736(07)61253-7 Mathew, M. D. (2022). Nuclear energy: A pathway towards mitigation of global warming. Progress in Nuclear Energy, 143(104080), 104080. https://doi.org/10.1016/j.pnucene.2021.104080

National Grid. (2025). The hydrogen colour spectrum. Nationalgrid.com. https://www.nationalgrid.com/stories/energy-explained/hydrogen-colour-spectrum

Niu, S., Li, S., Du, Y., Han, X., & Xu, P. (2020). How to reliably report the overpotential of an electrocatalyst. ACS Energy Letters, 5(4), 1083–1087. https://doi.org/10.1021/acsenergylett.0c00321

Nørskov, J. K., Abild-Pedersen, F., Studt, F., & Bligaard, T. (2011). Density functional theory in surface chemistry and catalysis. Proceedings of the National Academy of Sciences of the United States of America, 108(3), 937–943. https://doi.org/10.1073/pnas.1006652108

Rashad, S. M., & Hammad, F. H. (2000). Nuclear power and the environment: comparative assessment of environmental and health impacts of electricity-generating systems. Applied Energy, 65(1–4), 211–229. https://doi.org/10.1016/s0306-2619(99)00069-0

Rosenow, J. (2022). Is heating homes with hydrogen all but a pipe dream? An evidence review. Joule, 6(10), 2225–2228. https://doi.org/10.1016/j.joule.2022.08.015

Song, K., & Le, D. (2013, October 2). Bond Energies. Chemistry LibreTexts; LibreTexts Chemistry. https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_ Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Chemical_Bonding/ Fundamentals_of_Chemical_Bonding/Bond_Energies

U.S. Department of Energy. (n.d.-a). DOE explains…Catalysts. U.S. Department of Energy. Retrieved January 25, 2025, from https://www.energy.gov/science/doe-explainscatalysts

U.S. Department of Energy. (n.d.-b). Hydrogen Storage. Energy.gov. Retrieved January 25, 2025, from https://www.energy.gov/eere/fuelcells/hydrogen-storage

U.S. Department of Energy. (2010). Hydrogen and Fuel Cell Technologies Program: Fuel Cells. https://www1.eere.energy.gov/hydrogenandfuelcells/pdfs/doe_h2_fuelcell_factsheet.pdf

U.S. Energy Information Administration. (2023, October 31). How electricity is generated. U.S. Energy Information Administration. https://www.eia.gov/energyexplained/electricity/how- electricity-is-generated.php

U.S. Energy Information Agency. (2024, June 21). Use of Hydrogen. U.S. Energy Information Agency. https://www.eia.gov/energyexplained/hydrogen/use-of-hydrogen.php

Wang, B., & Zhang, F. (2022). Main descriptors to correlate structures with the performances of electrocatalysts. Angewandte Chemie (International Ed. in English), 61(4), e202111026. https://doi.org/10.1002/anie.202111026

WHA International Inc. (2023, September 21). Top industrial uses of hydrogen, and the need for industrial hydrogen safety. WHA International, Inc. https://wha- international.com/hydrogen-in-industry/

World Nuclear Association. (2024, May 17). Hydrogen Production and Uses. World Nuclear Association. https://world-nuclear.org/information-library/energy-and-the- environment/hydrogen-production-and-uses

World Resources Institute. (2022). Climate Watch Historical Country Greenhouse Gas Emissions Data. Climate Watch. https://www.climatewatchdata.org/ghg-emissions

Yan, Z., Liu, H., Hao, Z., Yu, M., Chen, X., & Chen, J. (2020). Electrodeposition of (hydro)oxides for an oxygen evolution electrode. Chemical Science (Royal Society of Chemistry: 2010), 11(39), 10614–10625. https://doi.org/10.1039/d0sc01532f

Zheng, Y., Jiao, Y., Vasileff, A., & Qiao, S.-Z. (2018). The hydrogen evolution reaction in alkaline solution: From theory, single crystal models, to practical electrocatalysts. Angewandte Chemie (International Ed. in English), 57(26), 7568–7579.

About the author

Jiajun Li

Jiajun is currently a 12th grade student at St. Andrew’s College. He is interested in physics, specifically nuclear physics, as well as environmental science, specifically the energy aspect. Jiajun is currently investigating how a clean and renewable energy source can solve most of the environmental crises that we are currently facing and how to develop future energy sources, such as advanced fission reactors and nuclear fusion reactors, which could greatly benefit society.

Jiajun is the leader and founder of his school’s physics club and a vital member of the environmental council, which has made significant progress on helping the environment within his school, including reducing food waste by over 20%. At his previous school, three other students and Jiajun succeeded in installing a solar energy system, and he is also planning the installation of a larger solar power system to power his current school.

The post Engineering Catalysts for Water Electrolysis: A Review of Activity Descriptors for Hydrogen and Oxygen Evolution Reaction appeared first on Exploratio Journal.

Abstract

The rapid development in the technologies of Unmanned Aerial Vehicles has ushered in a new era in precision agriculture, revolutionizing farm practices across the world. Equipped with multispectral, hyperspectral, and thermal imaging systems, UA Vs offer unparalleled capability for crop health monitoring, assessment of soil conditions, optimization in the use of fertilizers and pesticides, and improved water management. The current study investigates the huge potential for UA V technology to answer critical agricultural challenges, such as over-fertilization, excess pesticide use, and water shortages in Turkey, which have historically prevented sustainability and further efficiency. UA Vs can help farmers improve resource use significantly, lessen environmental impact, and yield better predictions with real-time data and precise intervention. This research reviews international case studies from areas like the Mediterranean and carries out regional questionnaires among Turkish farmers to analyze their UA V adoption barriers, such as high initial costs, inability to access technical expertise, and limited awareness of UA V capability. Large-scale and government-subsidized small-scale implementation of UA Vs within Turkish agriculture could help develop an overall modernization of agricultural techniques, increase profitability, and contribute to national objectives related to sustainability. Through such UA V-driven innovations, resource-use inefficiencies can be resolved alongside other environmental challenges, while long-term agricultural productivity and resilience are ensured for Turkey.

Keywords: UA V technology, remote sensing, precision agriculture, Turkey, sustainability, crop management, resource optimization

Introduction

Agriculture has traditionally been based on labor-intensive methods whereby farmers would observe field conditions and crop health by manual means. Traditional approaches to farming have usually led to several inefficiencies in resource usage and thus resulted in poor crop yields with greater labor costs. However, precision agriculture (PA) brought new technologies such as GPS, sensors, and data analytics, which revolutionized farming. Precision agriculture helped apply the most critical inputs of water, fertilizers, and pesticides with unprecedented precision. The sector has thus been able to receive a remarkable boost in productivity and sustainability [1]. This has lessened the burden, in terms of time and effort, placed on farm management while also encouraging better resource utilization [25; 18].

Generally, the rate of adoption of PA technologies has been increasing across the world. Examples of this increase have been seen in countries as; United States, Canada, and Australia, where digital agriculture solutions have completely revolutionized farming. In contrast, in Turkey, the rate of adoption can be considered quite low, although interest in the use of such technologies to raise agricultural efficiency is rapidly growing globally [25; 2]. Several reports indicate that drone technology seems to be the most promising approach in combating agricultural challenges in Turkey for modernization and better productivity [9].

UA V technology provides a great opportunity for farmers with active real-time monitoring abilities to enable fast assessments over large areas of agricultural fields without necessarily having an excessive amount of manual intervention. These will help develop an effective potential for reducing operational costs, hence making good precision in decision-making for the benefit of crop management. Studies have shown that UA V-based systems can reduce pesticide and fertilizer use and increase yield prediction accuracy for better sustainability in farming [26]. The new use of UA Vs in remote sensing, especially with multispectral imaging, is adding depth to farmers’ conclusions on their field conditions. This technology is considered vital for balancing the increasing global demand for food with the pressing need for environmental sustainability [5, 16].

Multi-spectral imaging enabled with UA V-supported remote sensing holds much transformational potential for agriculture. With the ability to take images in multiple wavelengths, including those in the Near-Infrared (NIR) part of the spectrum, UA Vs can work out plant health, soil conditions, and nutrient levels with remarkable accuracy. This imagery makes use of reflected light from the crop canopy in calculating such indices as NDVI (Normalized Difference Vegetation Index) for the computation of nitrogen and pesticide needs [27]. It has been reported in countries like China and Brazil that UA Vs may decrease pesticide consumption by up to 37% and raise crop productivity [29, 30, 28]. The fact that UA Vs are cost-effective and can deploy much quicker than ground-deployed methods bolsters a solid foundation on which UA Vs should be employed in Turkey, since agricultural efficiency improves through such technologies [26].

Despite its significant potential, UA V technology remains underutilized in Turkish agriculture, presenting a substantial opportunity for improvement. This research examines how deploying UA Vs to detect excessive pesticide spraying and over-fertilization can enhance resource management, increase crop yields, and reduce operational costs across Turkey’s agricultural sector.

Our study investigates three primary benefits of UA V technology adoption: improved resource management, enhanced yield monitoring, and greater economic sustainability. We hypothesize that increased implementation of UA V-based precision agriculture technologies will significantly improve crop yields while enhancing cost efficiency and time management through optimized resource utilization.

The paper also addresses current economic and technological challenges limiting UA V adoption in Turkey and proposes solutions based on successful implementation strategies from other countries. Our framework includes a critical review of UA V technology, its agricultural applications, analysis of relevant case studies, and recommendations for improving UA V adoption rates across Turkey.

Advanced UA V Technologies in Precision Agriculture

Unmanned Aerial Vehicles (UA Vs), commonly known as drones, have emerged as a pivotal technology in modern farming, transforming traditional practices through precision agriculture. By capturing high-resolution imagery and collecting environmental data across fields, UA Vs enable farmers to make evidence-based decisions regarding crop health, irrigation needs, fertilizer application, and pest management. This technology addresses the dual challenge of meeting increasing global food demand while promoting environmental sustainability [18]. UA Vs offer particular advantages over conventional monitoring methods due to their ability to rapidly survey large areas and navigate challenging terrains, making them adaptable to diverse agricultural landscapes. The two primary UA V designs used in agriculture—rotocopters and fixed-wing models—serve different purposes: rotocopters provide detailed observation capabilities through their hovering functionality and maneuverability, while fixed-wing UA Vs, with their extended flight times and greater coverage range, are ideal for monitoring extensive agricultural operations [5, 16]. These technological attributes position UA Vs as essential tools for modernizing agricultural practices, especially in regions like Turkey, where terrain diversity and varying farm sizes create unique monitoring challenges.

Multispectral and Hyperspectral Imaging

UA Vs can be equipped with multispectral and hyperspectral sensors that capture data beyond the visible spectrum, providing valuable insights into plant health and vegetation indices [19]. UA Vs with multispectral cameras have transformed traditional crop monitoring by capturing data across several key electromagnetic bands beyond visible light [8]. These sensors typically collect information in four to ten distinct spectral regions, including visible light (red, green, blue) and near-infrared (NIR) wavelengths. This multispectral data enables farmers to calculate critical vegetation indices such as the Normalized Difference Vegetation Index (NDVI), which measures plant vigor by comparing near-infrared light reflection (high in healthy vegetation) against red light absorption (linked to chlorophyll activity). Similarly, the Green Chlorophyll Index (GCI) provides specific insights into chlorophyll concentration, while other indices like the Enhanced Vegetation Index (EVI) offer improved sensitivity in high-biomass regions [5]. These quantifiable measurements allow farmers to identify spatial variations in crop development, detect stress factors before visual symptoms appear, and implement targeted interventions with precision. Multispectral imaging has proven particularly valuable for irrigation management, fertilizer application optimization, and early pest detection, enabling zone-specific treatments that reduce input costs while maximizing yield potential.

Hyperspectral imaging represents a significant technological advancement beyond multispectral capabilities, capturing hundreds of narrow, contiguous spectral bands across the electromagnetic spectrum. While multispectral sensors might collect data in 4-10 bands with bandwidths of 40-150 nm, hyperspectral systems typically record 100-300 bands with much narrower bandwidths of 5-10 nm [19]. This extraordinary spectral resolution creates detailed spectral signatures unique to specific crop conditions, enabling identification of subtlephysiological changes imperceptible to other sensing methods. The technology can detect variations in biochemical properties, including nitrogen, phosphorus, and potassium concentrations within plant tissues, differentiate between various disease pathogens before visible symptoms manifest, and identify moisture stress with remarkable precision [5]. Hyperspectral data processing often employs advanced machine learning algorithms to interpret the complex spectral information, creating highly detailed prescription maps for variable-rate applications. Despite higher equipment costs and data processing requirements compared to multispectral systems, hyperspectral imaging provides unparalleled insights for research applications and high-value crop production, where early detection of stress factors and diseases can prevent significant economic losses.

Thermal Imaging

Thermal imaging sensors integrated into UA Vs can capture data related to the canopy temperature of plants, providing critical information about their water status and stomatal conductance [8]. These sensors detect infrared radiation that can supply the variable of temperature to farmers over their fields. Variations in canopy temperature might indicate water stress, inadequate irrigation coverage, or even disease outbreaks. As summarized by Husain et al. (2023), “thermal imaging helps in identifying water-deficient zones and optimizes irrigation schedules, thus conserving water resources and enhancing crop resilience.”

Olson and Anderson (2021) also highlight the significance of thermal imaging in monitoring crop water stress, noting that UA Vs equipped with thermal sensors offer the Crop Water Stress Index (CWSI), which is a real indication of the actual status of plant hydration. This index provides substantial decision support for irrigation management by farmers, thus averting under- or over-watering. Farmers can thereby achieve optimization of water use, which may become crucial in arid areas where water scarcity is prevalent.

Image Stitching and Data Processing

One of the principal challenges in UA V-based agricultural monitoring is effectively managing the substantial volume of overlapping imagery captured during flight operations. As Olson and Anderson (2021) explain, “Image stitching is a key component for delivering high-resolution maps over large areas and enables high-resolution crop variability assessment and site-specific input applications.” This process involves algorithmically combining hundreds of individual high-resolution images into a single geometrically corrected orthomosaic map [12, 5]. The resulting composite provides a comprehensive, spatially accurate representation of the entire field that preserves the detailed resolution necessary for identifying subtle variations in crop performance. These orthomosaics reveal specific regions experiencing stress due to nutrient deficiencies, pest infestations, soil compaction, or water limitations that might otherwise remain undetected through traditional scouting methods. The geometric correction applied during processing ensures that measurements taken from these maps maintain spatial accuracy, enabling precise localization of problem areas for targeted intervention.

The transformation of raw UA V imagery into actionable agricultural intelligence requires sophisticated data processing beyond basic image stitching. Specialized software analyzes orthomosaics to extract critical information, including vegetation indices, plant height estimates, and canopy cover measurements [5]. As Husain et al. (2023) note, recent advancements in artificial intelligence and machine learning have dramatically improved the efficiency of this process, enabling “real-time analysis and faster decision-making” that enhances farm management responsiveness. These technological improvements, coupled with the emergence of cloud-based computing services, have significantly streamlined data processing workflows, reducing farmers’ dependence on expensive on-site hardware and specialized technical expertise [5]. The processed data ultimately generates practical outputs such as fertilizer prescription maps for variable rate application equipment and targeted pest management recommendations [5]. This complete workflow—from image capture to actionable recommendations—creates a digital decision support system that empowers farmers to implement precision agriculture practices with greater confidence and efficiency, maximizing both economic and environmental benefits through optimized resource utilization.

Yield and Nutrient Assessment

UA V-based remote sensing has revolutionized agricultural monitoring by offering efficient and non-destructive methods for yield and nutrient assessment. These aerial systems, equipped with advanced sensors, can capture detailed multispectral and hyperspectral imagery across large agricultural areas in minimal time. When this imagery is processed and analyzed, it provides crucial data enabling researchers and farmers to accurately estimate key crop parameters, including leaf nitrogen content, leaf area index, and plant biomass, all of which serve as valuable indicators for potential yield prediction [12]. For example, specialized studies have successfully demonstrated how UA V-based 3D photogrammetry techniques can measure crop height with remarkable precision, establishing strong correlations between these measurements and final corn crop yields [12]. This application is further emphasized by Olsson and Anderson (2021), who confirm that drones excel at “estimating leaf nitrogen content, which is very important for the optimization of fertilizer application and maximization of crop yields.” By enabling real-time monitoring of plant nutrient status through detailed spectral analysis, UA V technology facilitates precision fertilization strategies that optimize nutrient application rates and timing, thereby promoting healthy crop development while significantly reducing the environmental risks associated with over-fertilization and nutrient runoff into water systems [5, 16].

Furthermore, UA Vs have demonstrated exceptional capability in comprehensive nutrient management and yield forecasting through their sophisticated sensor systems. These platforms can simultaneously collect and integrate multiple data types, generating detailed crop health profiles that go beyond basic assessment. Husain et al.’s research highlights how UA V systems effectively employ specialized vegetation indices such as the Soil-Adjusted Vegetation Index (SA VI) to evaluate not only crop conditions but also underlying soil health and nutrient bioavailability patterns across fields. This advanced spatial analysis enables the creation of high-resolution nutrient management zone maps that allow farmers to implement variable-rate application technologies with unprecedented precision. By adjusting fertilization strategies according to real-time field data [5, 16], farmers can target inputs specifically where and when they’re needed, substantially reducing unnecessary application in already nutrient-sufficient areas. This precision approach delivers significant economic benefits through input cost reduction while simultaneously advancing environmental sustainability goals by minimizing nutrient leaching, runoff, and greenhouse gas emissions associated with excess fertilizer application, creating a win-win scenario for both agricultural productivity and ecological protection.

Precision Pesticide Application and Weed Detection

UA Vs are increasingly employed for precision pesticide application and weed detection, offering an environmentally friendlier alternative to conventional methods that often result in chemical overapplication. Equipped with precision spray systems, UA Vs can apply pesticides in a site-specific manner, targeting only affected areas rather than entire fields [11, 1]. According to Husain et al., drone technology enables targeted pesticide application that can reduce chemical use by as much as 30% while simultaneously lowering operational costs. This precision approach significantly minimizes the risk of pesticide drift and potential negative impacts on surrounding ecosystems. Weed detection using UA V imagery often employs deep learning-based methods to accurately identify and map weed infestations, creating detailed maps that enable variable rate chemical weeding and lead to significant savings in crop protection products[6, 8].

Furthermore, UA Vs enhance weed management through sophisticated detection capabilities using high-resolution cameras and spectral data analysis. These systems can effectively differentiate between crops and weeds, enabling precise herbicide application only where needed. “UA Vs fitted with high-tech imaging are able to detect infestations of weeds before they get out of hand, thus enabling the farmer to take control of them while still in their tender stages”[5]. This early intervention capability is crucial for maintaining optimal crop health and preventing yield losses from weed competition [8]. The integration of advanced imaging technologies with precision application systems creates a comprehensive approach to pest and weed management that simultaneously improves agricultural efficiency, reduces environmental impact, and enhances economic outcomes for farmers [11, 1].

Advanced UA V technologies in multispectral and hyperspectral imaging, integrated with thermal imaging, advance the new era in precision agriculture. Enabled by such technologies, farming can be optimized with increased resource use, yields, and management of more sustainable agricultural methods. The analysis of existing literature, when put together, creates a scenario whereby UA V not only changes the traditional ways of monitoring agriculture but also form the future innovations for sustainable farming. Equipped with thorough and timely information on crop health, water stress, nutrient levels, and management of pests, UA Vs put farmers in a better position to make informed decisions for increasing yields while lessening environmental impacts.

Agricultural challenges caused by climate change, resource scarcity, and increasing food demand continue to worsen. With these further challenges to farming, the role of UA Vs in precision agriculture will continue to play a vital part. Their application can further enhance the efficiency of agricultural operations beyond what was previously mentioned, in turn contributing to or supporting food security and sustainability goals. In each source, research and continued investments into the technology of UA Vs are highly emphasized as imperative for truly exploiting their potential in changing global agriculture.

Support Section I (Analysis and Literature Review)

Research conducted in various parts of the world has demonstrated that UA V technologies are of vital importance for optimizing resource use in agriculture; thus, they are envisioned as a promising solution for improving crop management in Turkey. However, despite their potential for further development and wider adaptation, the use of UA V technology for remote sensing of excessive pesticide spraying, over-fertilization, and yield monitoring in Turkish agriculture remains minimal. This situation creates immense opportunities to optimize resource use and improve crop yields, thereby enhancing cost and time management. Therefore, this section will review the existing literature to support the hypothesis that UA Vs can effectively address resource management inefficiencies in the agricultural sector. With UA Vs capable of monitoring pesticide and fertilizer use as well as crop health, the potential for driving sustainability, reducing input costs, and enhancing productivity becomes significant. These capabilities could substantially transform the landscape of Turkish agriculture, offering new pathways toward more efficient and environmentally sound farming practices..

How UA V Technology Has Impacted Agricultural Practices Globally

Unmanned Aerial Vehicle (UA V) technology has revolutionized agricultural practices worldwide by providing producers with enhanced assessment capabilities and facilitating more efficient decision-making processes. The global adoption of precision agriculture technologies, with UA Vs playing a pivotal role, continues to accelerate, with industry projections estimating a market value of $43.4 billion by 2025[4]. This remarkable growth trajectory underscores the agricultural sector’s increasing recognition of UA V technology’s transformative potential.

UA Vs offer farmers sophisticated options for assessing critical factors affecting agricultural systems. Color sensors mounted on UA Vs can accurately estimate leaf color, plant height, lodging, canopy cover, stand count, and even flower and fruit quantities[5]. Research has demonstrated UA V-based height estimation reaching an R-squared value of 0.51 with dry grain yield when measured at the beginning of August, illustrating its significant potential for yield prediction and comprehensive crop monitoring[12]. Beyond visual data, spectral sensors can estimate indirect leaf nitrogen content, yield potential, leaf area index (LAI), leaf chlorophyll content, and plant biomass. Additionally, thermal sensors capture vital data for estimating canopy temperature, stomatal conductance, water use efficiency, and plant water potentials[5]. Notably, Green Normalized Difference Vegetation Index (GNDVI) data acquired by UA Vs during anthesis and full crop development has produced accurate biomass estimates, with some studies reporting higher accuracy yield estimates with GNDVI as early as 5 weeks into a crop’s lifecycle [5].

The fusion of multiple data types significantly enhances assessment capabilities, with combinations of color, spectral, and thermal data products improving reliability and accuracy. Three-dimensional characterizations such as crop height or volume, derived from LiDAR or high-resolution color cameras and combined with conventional 2D data, can increase the accuracy of yield and quality estimates[5]. Further, the research has shown that combining plant height measurements with NDVI values has improved biomass estimation accuracy[5]. Additionally, radiometric correction of UA V imagery can significantly reduce grey value variation in overlapping images, from 14–18% to 6–8% under varying illumination conditions, further enhancing data quality[19].

The rich data gathered by UA Vs drives strategic decisions that enhance input efficiency, potentially maximizing return on investment while simultaneously reducing environmental impact. Studies conducted on Australian grain farms have estimated an increase ranging from $4/ha to $23/ha, with an average of $13/ha (8%) attributable to precision agriculture technology, including benefits from variable rate fertilizer application and reduced overlap in input distribution[21]. In another compelling case study from an Italian cereal farm, researchers documented a 53% reduction in pesticide usage during the post-adoption phase, largely attributable to precision sprayers[6]. This same farm experienced a 28% decrease in fuel consumption due to the implementation of newer, more efficient machinery equipped with advanced satellite guidance systems[6]. It is evident that UA V technology helps ensure precise pesticide application, utilizing fewer chemicals and minimizing environmental contamination[5, 16].

Precision agriculture technologies incorporating UA V data have demonstrated impressive efficiency improvements across multiple categories. Recording and mapping technologies (RMT) can lead to fertilizer savings ranging from 1.6–82% in peer-reviewed papers and 5–70% in EU projects. Variable rate technologies (VRT) have shown fertilizer savings ranging from 5–59% in peer-reviewed literature and 22–30% in EU projects. Guidance and Controlled Traffic Farming (CTF) technologies have reported fertilizer savings ranging from 1–26%, while Farm Management Information Systems (FMIS) have demonstrated fertilizer savings ranging from 14.7–46% in peer-reviewed papers and 5–70% in EU projects[14].

The applications of UA Vs in agriculture span numerous domains, including creating detailed fertilizer prescription maps, conducting hail crop damage assessments, enabling early pest detection, and generating crop yield and quality predictions[5]. UA V-captured multispectral imagery is used to assess plant health and nutrient status, enabling the creation of variable rate fertilizer prescription maps [3]. Methods like Ordinary Kriging (OK) and Inverse Distance Weighting (IDW) are employed for spatial prediction of soil properties for these maps, with modern approaches integrating data from digital elevation models and multispectral satellite images to enhance prediction accuracy [16].

The integration of UA V-based imagery with machine learning algorithms offers enhanced yield assessment accuracies while reducing reliance on labor-intensive ground-based surveys. Machine learning applied to UA V imagery has achieved classification accuracies ranging from 67 to 95% for identifying diseases like yellow rust, grapevine bacterial disease, myrtle rust, and citrus greening[5]. UA Vs equipped with multispectral cameras effectively assess plant health using Normalized Difference Vegetation Index (NDVI) and other specialized indices, detecting stress factors not visible to the naked eye[5, 16]. One study demonstrated that vegetation fraction mapping in wheat crops using a low-cost camera mounted on a UA V achieved impressive accuracies of 87.73–91.99% when operating at a height of 30 meters[8].

For crop yield and quality predictions, Partial Least Squares Regression (PLS-R) is widely used to analyze spectral data obtained by UA Vs, with PLS-R models for maize grain yield showing coefficient of determination (R²) values up to 0.86 in validation[15]. Hyperspectral data from UA Vs have been employed for wheat growth monitoring and yield estimation using methods like Multiple Linear Regression (MLR), Simple Linear Regression (SLR), and Partial Least Squares Regression (PLSR)[8]. The deployment of hyperspectral cameras on UA Vs shows particular promise due to their ability to capture broader spectral ranges than simpler multispectral cameras.

A major contribution of UA V technology has been its ability to gather phenotypic data quickly, efficiently, and non-destructively, significantly advancing crop improvement through genetics and plant breeding programs[5]. For instance, NDVI data generated from UA V imagery successfully identified quantitative trait loci (QTL) associated with drought adaptive traits in durum wheat, accounting for 89.6% of the phenotypic variance[5]. Biplot analysis, which can utilize data collected by UA Vs, has proven to be an effective tool for selecting superior genotypes and increasing efficiency in selection [3]. The rapid and comprehensive data collection capabilities offered by UA Vs greatly expedite field data acquisition for developing germplasm with desired agronomic traits.

In conclusion, UA V technology has transformed global agricultural practices by enhancing precision, efficiency, and sustainability. Research indicates that Farm Management Information Systems supported by UA V data show potential for yield increases ranging from 5-14%, fertilizer savings of 14.7-46%, and water savings of 10-50% [21]. As the technology continues to evolve and become more accessible, its adoption across different agricultural contexts promises further improvements in productivity and sustainability, helping address the growing global demand for agricultural products.

How Does Remote Sensing Save Resources Compared to Traditional Methods?

One key advantage lies in crop health monitoring. Traditional methods often involve visual inspection and manual collection of ground samples from random locations. This can be difficult, time-consuming, and may not provide a comprehensive overview, especially on large farms [13, 8]. Drones, however, can perform daily monitoring of crops to detect potential threats such as diseases, pests, and slow growth rates more effectively. They utilize aerial remote sensing, capturing images of different wavelengths to measure various parameters [1]. Drones can be equipped with different sensors, such as color, spectral (including near-infrared – NIR), and thermal cameras, allowing for the assessment of plant health indicators that might be invisible to the naked eye. This enables earlier detection of problems and facilitates quicker corrective actions, potentially preventing crop spoilage [8, 3]. The ability of drones to fly at low altitudes allows them to obtain high-resolution images with greater precision compared to satellites, which can be affected by cloud obstruction and have lower spatial resolution. While manned aircraft can cover large areas, they are often more cost-prohibitive than drones [2, 13].

In pesticide and fertilizer spraying, drone-mounted sprayers offer enhanced coverage ability and increased chemical effectiveness compared to conventional methods, which can be less effective in controlling pests and diseases [11]. Traditional spraying can also result in a higher cost of application. Drones can carry pesticide tanks (up to 40 liters in some cases) and follow pre-mapped routes to spray crops according to specific requirements; this capability allows for spot spraying in infected areas, reducing the overall amount of pesticides used and minimizing environmental impact [1, 8]. Drones can effectively cover fields with difficult access for tractors and aircraft. Furthermore, they can provide a more uniform and precise application of pesticides, operating at a standard delivery height and speed. Similarly, drones can be used for the application of fertilizers, making the process easier and potentially more efficient than traditional on-ground tools like tractors [13, 3]. Variable rate application of fertilizers and pesticides, guided by drone-collected data, directly translates to savings in resources [4].

Drones contribute significantly to resource management. The high spatial and temporal resolution data provided by drones enables site-specific management. For instance, drones with thermal sensors can detect temperature variations indicating water stress, allowing for efficient utilization of water through targeted irrigation and preventing water wastage [1, 13]. By identifying areas of nutrient deficiency or pest infestation, drones facilitate the precise application of fertilizers and pesticides, leading to reduced usage of these inputs. Studies suggest potential herbicide savings of 20% to 50% through precision application [14]. The data collected by drones, such as soil moisture information, can also be used to automate irrigation scheduling systems, further reducing water usage [3].

The efficiency in data acquisition by drones is a major benefit. They can cover large areas in short time periods and offer freedom from positioning and timing limitations compared to other remote sensing platforms [1]. The ability to collect data frequently allows for the detection of changes over time, which is crucial for effective agricultural management [8]. This data can then be analyzed using Artificial Intelligence (AI) and deep learning models for tasks like crop yield prediction, disease detection, weed mapping, and nitrogen status estimation, providing valuable insights for better management decisions [17]. This instantaneous data analysis capability allows for quicker responses to emerging issues [13, 1].

Several sources highlight the potential for cost and time savings with the use of drones. Farmers have reported that spraying and fertilization activities using drones save time and reduce costs due to less labor being required and the low cost of battery charging compared to fuel costs for traditional machinery [[13, 4]. Drone usage can significantly reduce labor costs associated with manual crop monitoring and pesticide application. For example, some studies indicate that drones save up to 20 times the labor of farmers [13].

Drones offer greater accessibility and coverage, especially in complex or difficult-to-reach areas. They can navigate terrains that are challenging for tractors and manned aircraft, ensuring that all parts of a field can be monitored and treated effectively [1]. Fixed-wing drones are particularly suited for large-scale surveys and monitoring vast agricultural areas, while multi-copters are often preferred for crop health monitoring and targeted interventions due to their maneuverability and hovering capabilities [16, 5].

Furthermore, drones play a crucial role in high-throughput phenotyping, allowing researchers and plant breeders to gather phenotypic data quickly, efficiently, and nondestructively. This includes estimating leaf color, plant height, canopy cover, stand count, flower count, and fruit count using various sensors [20]. Spectral sensors can estimate leaf nitrogen content, yield, leaf area index, and plant biomass, while thermal sensors can provide data on canopy temperature and plant water status [5].

In conclusion, drones offer numerous advantages over traditional agricultural methods by providing more efficient and precise crop monitoring, targeted resource application, improved data acquisition and analysis, potential cost and time savings, enhanced accessibility, and valuable tools for research and development in precision agriculture [18]. The increasing technological advancements and decreasing costs of UA Vs are making them an increasingly important tool for modern agriculture [8, 11].

Cost and Efficiency Comparisons

The increasing accessibility of UA V technology has emerged as a viable option for agricultural applications [8]. By 2019, complete agricultural-grade UA V systems—which included a UA V platform like the DJI Matrice 100, a sensor such as the MicaSense Red Edge MX, image processing software (e.g., Pix4D Fields), and necessary hardware—became available for under $30,000 [5]. This represents a significantly lower initial investment compared to traditional aerial imaging systems such as piloted aircraft, making UA Vs a more financially feasible solution for farmers [1]. Additionally, the proliferation of cloud-based computing services has reduced the need for substantial upfront investments in on-site image processing infrastructure, enabling farmers to efficiently process UA V-captured imagery without dedicated hardware and software expenses [5].

In terms of efficiency, UA Vs can cover significant areas in relatively short periods, leading to considerable time and labor savings compared to traditional ground-based monitoring methods [11]. Rotocopter UA Vs can cover 1–8 hectares per battery charge, while fixed-wing UA Vs can cover 10–40 hectares, depending on the altitude [5]. This ability to rapidly assess large fields allows for frequent monitoring and the acquisition of high-resolution data, which is difficult to achieve with manual ground inspections [1]. Farmers often find it challenging to regularly inspect every part of extensive land holdings, a task that drones can perform effectively through routine aerial surveillance, providing aerial views of their harvest and information on water systems, soil variations, pests, and fungal infestations [1]. Unlike satellite images, which can be obstructed by clouds, UA Vs offer greater reliability, along with freedom from positioning and timing limitations associated with satellite and manned aircraft-based remote sensing [11].

The adoption of UA V technology offers substantial cost-saving benefits for farmers, primarily through the enablement of Variable Rate Technologies (VRT). UA V-collected data facilitates the targeted application of inputs, resulting in significant reductions in resource use [13, 18]. Studies have indicated the potential for fertilizer usage reductions of up to 80% and pesticide use reductions of up to 80% [14]. This precise application, sometimes guided by GPS coordinates in drone systems, not only lowers input costs but also minimizes environmental impact by reducing the overall amount of chemicals used. For example, savings in herbicides have been reported in the range of 20% to 50% through precision application [5, 16, 14].

Moreover, the integration of UA Vs within broader robotic systems and smart machines has shown potential for significant decreases in labor costs (up to 97%) and fuel consumption (up to 50%) for specific agricultural tasks like spraying and fertilization [14]. Some reports suggest that drones can save up to 20 times the labor of farmers [13]. The automation of procedures through digital agricultural technologies (DATs), including UA Vs, streamlines operations and diminishes the discomfort associated with protective clothing worn during manual spraying [14].

Beyond cost reductions, improved yield and better resource management contribute to overall economic benefits for farmers. VRT, enabled by UA V data, has been linked to yield increases of up to 62% [14]. UA Vs equipped with sensors like thermal cameras can also enable efficient utilization of water through targeted irrigation, preventing water wastage [4]. Projects utilizing IoT and satellite data have demonstrated fertilizer cost savings ranging from 5% to 70%, highlighting the potential of data-driven approaches [14]. The comprehensive data provided by UA Vs, when effectively analyzed using techniques like vegetation indices and machine learning, translates into an economic benefit that enhances farmer profitability. In fact, the perceived economic benefit is a significant driver for farmers considering the adoption of drones [4].

In conclusion, the decreasing cost of agricultural-grade UA V systems, coupled with their efficiency in data collection and the accessibility of cloud-based processing, makes them a powerful tool for implementing precision agriculture practices like VRT. These practices directly lead to substantial cost savings for farmers through reduced input usage, potential decreases in labor and fuel, and improved yields, ultimately contributing to enhanced economic sustainability [1, 5].

Research Challenges

While the global success of UA V technologies in agriculture is evident, applying these innovations to Turkey’s unique agricultural landscape brings a number of challenges that must be addressed. One major issue pertains to limited data on UA V technology adoption and its impacts within Turkey. Because UA Vs are relatively recently introduced to Turkish agriculture, there is a noticeable deficiency in terms of localized research and empirical studies that should be pursued regarding the practical benefits and limitations of UA Vs under Turkish farming conditions.

This scarcity of data itself acts as an adoption barrier because many Turkish farmers and policymakers receive relatively few insights regarding how UA Vs can directly affect yield improvement, cost reduction, and resource management within their specific agricultural sectors. More importantly, scaling up the knowledge globally to the local level requires attention to specific problems that farmers in Turkey face with heterogeneous soil types and climate conditions, and crop varieties grown across regions. Other than this data limitation, two big challenges are specialized training and technical capability for flying UA Vs and analyzing generated data. While the technology of UA Vs in and of itself holds tremendous promise, certain relatively complex aspects involved in flying drones and making analyses of high-resolution multispectral images place demands for a level of technical expertise that may not be readily available in Turkish countryside settings. In this regard, the relatively limited availability of experienced UA V operators and data analysts within farming communities in Turkey is likely to dampen the rate of diffusion of this technology.

Besides these, very high upfront costs of UA Vs and continued maintenance and training requirements create financial barriers against their adoption by small and medium-sized farms that do not have the required capital investment in such technology. Overcoming these challenges will involve strategic efforts, including the establishment of local UA V service providers, government incentives to reduce adoption costs, and initiatives to build technical capacity within the workforce in rural areas.

Additional Barriers and Considerations

Other considerations and barriers to UA V implementation involve the cost of training and technical expertise. Mediterranean studies indicate that the ability to process NDVI and other UA V data requires a high degree of technical knowledge on the part of UA V operators and data analysts alike- a resource that may be limited in rural Turkish areas [6].

A developed network of local UA V service providers would make precision agriculture more accessible to Turkish farmers because the individual farms would not need to make investments in expensive equipment and training.

Overall, international experiences underline the transformational role that UA V technology can play in agriculture: from reducing input costs to yield prediction and supporting sustainable farming. A suitably adapted application can henceforth provide the means for UA V-driven precision agriculture to contribute toward attaining environmental sustainability, productivity, and economic resilience in the farming sector of Turkey.

Support Section II (Surveys and Turkey)

Agriculture in Turkey, representing nearly 23% of the country’s total workforce, faces significant challenges related to resource efficiency, environmental sustainability, and adaptation to climate variability [10]. With an average farm size of only 5.9 hectares—considerably smaller than EU (17.4 ha) and US (18.0 ha) averages—Turkish agriculture operates under different structural constraints than more consolidated agricultural systems elsewhere. The Cukurova region, accounting for 5% of Turkey’s total cultivation area with 1,091,000 hectares of cropping land, exemplifies both the potential and challenges of Turkish agriculture [10].

Turkish farming is experiencing transformation pressures from multiple directions, including climate change effects, freshwater supply shortages, and economic pressures from rising input costs like fertilizers [9]. The current agricultural production system faces particular strain as inflation (44.38%) and high interest rates (42.50%) create financial barriers for farm investments [23, 24]. Further, the average interest rate of 46.7% of Turkey from the 1970s to the last decade indicates the long-lasting economic trifle farmers face [22]. These economic conditions make cost-effective technological solutions particularly relevant for Turkish farmers.

Precision farming technologies offer promising solutions specifically tailored to Turkey’s agricultural challenges. Studies demonstrate that variable-rate fertilizer application could achieve fertilizer savings ranging from 4% to 37%, depending on soil variability—a significant consideration given that fertilizer costs currently represent approximately $643.33 per hectare for wheat production [9]. These savings would be particularly impactful in regions with high soil nutrient variability, though implementation potential varies significantly between geographic regions like Central Anatolia and the more diverse Cukurova region.

The economic viability of precision farming in Turkey depends heavily on farm size, with annual equipment costs ranging from $13-131 per hectare (for 500-50 ha farms, respectively). This suggests that while larger operations may readily adopt these technologies, structural adaptations may be necessary for Turkey’s predominantly small-scale farming sector to benefit from precision agriculture advancements [9].

Farmer Readiness and Perceived Benefits

Survey-based research across multiple Turkish provinces provides valuable insights into farmers’ readiness to adopt drone technology and their perceptions of its benefits. In the Southeastern Anatolia Region, a survey conducted between September 2022 and January 2023 with 249 producers in Mardin and Şanlıurfa revealed that 27.7% of respondents had used unmanned aerial vehicles for agricultural purposes in the previous production season. The study found a statistically significant correlation between drone use and education levels (p=0.000), with higher education corresponding to greater adoption rates. The surveyed farms averaged 26.4 hectares, substantially exceeding both the national average of 5.9 hectares and the regional average of 10.4 hectares, with farmers averaging 19.5 years of agricultural experience [4].

Regarding economic benefits, farmers generally recognized the potential advantages of drone technology, with a mean score of 3.497±1.07 on a 5-point Likert scale. Specifically, they acknowledged that drones could increase crop productivity (mean 3.534), reduce work costs (mean 3.426), increase profit (mean 3.438), and be cost-effective (mean 3.590). The perceived usefulness of drones scored even higher at 3.561±1.10, with farmers believing that drones would make their jobs easier (mean 3.598), provide better results than other systems (mean 3.534), enable faster work (mean 3.554), improve work performance (mean 3.554), and allow for work diversification (mean 3.566) [4].

Despite recognizing these benefits, farmers expressed moderate concerns about the technical aspects of drone operation. Statements regarding ease of use received somewhat lower scores, with “I think drones are easy-to-use tools” averaging 3.317 and “It will be easy for me to learn to use a drone” averaging 3.385. Confidence in their ability to use drones was similarly moderate, with statements like “I think using a drone is complicated and difficult” (mean 2.767), “It is difficult for me to use a drone” (mean 2.671), and “I think I am not a farmer who is good at working with digital tools like drones” (mean 2.590) indicating some hesitation. These lower confidence scores likely reflect the prevalence of rental services, where farmers might not directly operate the drones themselves [4].

The research identified two significant factors positively influencing drone adoption intention: perceived usefulness and perceived economic benefit. Interestingly, perceived ease of use and trust attitude did not significantly impact adoption intentions. Farmers were primarily motivated by potential reductions in workload, enhanced productivity, and lower operational expenses rather than by how easy the technology might be to use [4, 7].

High initial investment emerged as the primary barrier to adoption, as evidenced by a survey of 384 farmers in Aydın province. Only 3.12% of these farmers were initially willing to purchase drone technology, while a much larger percentage (32.6%) expressed willingness to rent Autonomous Unmanned Aerial Vehicle (AUA V) technology at an average price of TRY 287.54 per hectare. The significance of this barrier was further confirmed by the 67.2% of farmers who strongly agreed that the high initial purchase cost was a major obstacle to adoption. Notably, when government subsidies were mentioned, potential purchase interest more than tripled to 10.67%; as Parmaksiz and Cinar (2023) note, “Farmers preferred cooperatives as providers of UA V rental services”, highlighting the importance of policy interventions in technology adoption. The study also found that young, technology-oriented farmers with higher agricultural incomes showed greater interest in drone adoption [7].

A complementary survey in the Çukurova region (covering Hatay, Osmaniye, Adana, and Mersin provinces) assessed awareness of precision agriculture (PA) technologies among agricultural stakeholders. While 90.2% of participants (including agricultural engineers, farm equipment dealers, and farmers) reported following new agricultural trends, 51.8% had not heard the term ‘Precision Agriculture’ before. However, after receiving brief training, 97.6% believed PA technologies would benefit Turkish agriculture, 88.4% wanted more detailed training, and 89.7% expressed interest in using these technologies. This dramatic shift in interest after minimal education suggests that knowledge gaps represent a significant but potentially addressable barrier to adoption [10, 3].

Based on these findings, several promising pathways emerge for increasing drone technology adoption in Turkish agriculture. First, given the significant difference between purchase and rental willingness, cooperative-based rental services could significantly accelerate adoption. Second, the tripling of purchase interest when subsidies were mentioned highlights the critical role of government support in overcoming financial barriers. Third, the correlation between education and adoption, along with high interest in training after basic exposure, suggests that educational initiatives could substantially boost adoption rates. Finally, developing favorable borrowing channels specifically designed for agricultural technology investments could help address the initial cost barrier that currently limits widespread adoption [7, 10]

These findings collectively demonstrate that while Turkish farmers recognize the potential efficiency and economic advantages of drone technology, successful widespread adoption will require addressing financial barriers and knowledge gaps through coordinated efforts by government agencies, agricultural cooperatives, and educational institutions. The relatively high willingness to rent suggests a practical intermediate step toward technology integration in Turkish agriculture, particularly if facilitated by appropriate organizational structures and policy frameworks [10, 3].

How UA Vs Address Turkey’s Agricultural Challenges

The agricultural sector in Turkey has long been plagued by resource-wasting and environmentally destructive practices, which have led to issues such as overuse of water, soil degradation, and excessive chemical inputs. Traditional farming techniques, while deeply rooted in the culture, have often failed to adapt to the growing need for sustainability and efficiency. However, the advent of UA V (Unmanned Aerial Vehicle) precision agriculture presents a significant opportunity to address these challenges. UA Vs offer the potential to revolutionize Turkey’s agricultural practices by providing real-time data and insights that can lead to smarter decision-making, increased efficiency, and reduced environmental impact.

Improved Efficiency in Fertilizers and Pesticides

Turkey’s agricultural sector faces challenges related to efficient resource management and increasing productivity. Precision agriculture technologies (PATs) can be considered a tool for farm management that allows the agricultural entrepreneur to optimize inputs and reduce costs [6]. UA V technology can address these challenges by enabling precision application of fertilizers and pesticides, based on site-specific information collected by multispectral and hyperspectral sensors [7]. UA Vs can carry out a wide variety of agricultural operations in support of precision agriculture, including fertilizer application. This targeted approach can help reduce the overuse of agricultural inputs, which is crucial as Turkey has begun to become more foreign-dependent in agriculture. By optimizing the quantity and timing of fertilizer and pesticide applications according to the needs of different areas within a field, UA Vs can contribute to cost reduction and environmental sustainability [12, 7]. For instance, Recording and Mapping Technologies (RMT), which include UA Vs with sensors, can lead to significant fertilizer savings. One study achieved an 80% reduction in fertilizer doses applied in a vineyard through the use of aerial imagery and ground detection. Variable Rate Technologies (VRT), often implemented using UA Vs, have demonstrated the potential for fertilizer savings of up to 59.6%. In one case, a 31% reduction in potassium-based fertilizer and a 59% reduction in phosphate consumption were achieved [14]. Furthermore, regarding pesticides, UA Vs help ensure very precise pesticide spraying, which uses fewer chemicals as the zones are specific in number, minimizing environmental contamination. VRT has shown potential for pesticide savings ranging between 8% and 80%. Sensor-based variable-rate application (VRA) achieved an 8% fungicide savings in winter wheat, and another study reported a 51.9% reduction in spray volume for air-assisted spraying based on real-time disease spot identification. Autonomous robotic systems integrated into modern agricultural practices have also led to substantial Plant Protection Product (PPP) savings, with one system achieving 66% herbicide savings [14, 5, 16]. These precise application methods not only reduce input costs but also contribute to protecting the environment and human health through the controlled use of fertilizers and pesticides [7]. In Turkey, Tekin et.al.’s study examining variable-rate fertilizer application in wheat production suggested that applying fertilizer considering the variations in soil nutrients could be economically justified, with the degree of variability being a key factor [9].

Water Conservation and Irrigation Efficiency

Water scarcity is a significant concern in many agricultural regions, including parts of Turkey. UA Vs equipped with thermal sensors can play a vital role in improving irrigation efficiency and promoting water conservation [7]. By identifying areas experiencing water stress through canopy temperature measurements, farmers can implement variable rate irrigation (VRI), applying water only where and when needed. This precision in irrigation can lead to significant water savings, as demonstrated in international case studies where VRT irrigation systems have achieved water savings of 20% to 50% in vineyards and pear orchards [8, 14]. Recording and Mapping Technologies (RMT) contribute to the necessity of water conservation in agriculture. One study implementing an automatic irrigation scheduling system in an olive orchard, using soil moisture sensor data, resulted in a 24% reduction in water usage. Innovative irrigation scheduling software using model-predicted crop water stress achieved water savings between 16% and 35% [14]. Empirical data from the Precision Crop Management initiative, utilizing IoT sensors, satellite imagery, and drone technology, successfully reduced irrigation costs by 10% in wheat cultivation. Guidance and Controlled Traffic Farming (CTF) technologies have also been linked to water savings, with one study observing water reductions of 13.0% and 9.0% for different crops through precision irrigation in CTF systems. Furthermore, CTF was linked to a 65% increase in rainfall-use efficiency, leading to reduced runoff and water conservation. Variable Rate Technologies (VRT) have also demonstrated potential in significantly reducing water consumption, with computer simulations showing variable water savings up to 26% with optimized specific zone control in center-pivot irrigation. Soil moisture sensor-based irrigation scheduling achieved water savings ranging from 7.5% to 19%, and VRI systems reduced irrigation water use by 8–20% for soybeans and 25% for corn. The HydroSense project in Greece, using VRI in cotton fields, showed 5 to 34% savings in water consumption. These examples highlight the potential for UA V-based technologies to address water scarcity challenges in agriculture, including in regions like Turkey [14].

Environmental Impact on Turkish Agriculture

The detrimental ecological footprint of traditional agricultural practices in Turkey poses significant environmental challenges that demand urgent attention. Unconscious watering, spraying and fertilization methods widely employed across Turkish agricultural regions severely damage ecosystems, contaminate groundwater, and destroy beneficial insects and soil microorganisms [3]. Traditional pesticide application techniques enable these harmful chemicals to permeate the environment, posing risks to human health, with farmers often directly exposed to toxins during manual spraying operations. The risk of pesticide residues and their negative impact on ecosystems represents one of Turkey’s most pressing pollution problems, mirroring similar concerns worldwide [7].

Conventional fertilization approaches in Turkey typically treat agricultural land as homogeneous units, applying standardized amounts of fertilizer without accounting for spatial variability, resulting in substantial environmental degradation. This uncontrolled use of fertilizers and pesticides introduces heavy metals into soil systems, triggering contamination that affects all living organisms and ultimately compromises human health through the food chain [16]. Additionally, unconscious stubble burning practices severely impact groundwater quality and destroy beneficial soil organisms [3]. Traditional agricultural development’s singular focus on increased productivity has historically pushed natural systems beyond sustainable limits, while practices such as deep tillage and vegetation burning alter the living conditions of soil microorganisms, causing deterioration of their environment and decreasing soil biota diversity [7]. These conventional methods have contributed to alarming rates of water resource depletion, soil contamination, and excessive salinization of irrigated lands, leading to significant losses of arable territory [3].

The strategic deployment of drone technology in Turkey’s agricultural landscape offers a promising pathway to reduce environmental harm through precision farming methods. Research indicates that drones can transform conventional agricultural practices by enabling farmers to apply resources with unprecedented accuracy, addressing a fundamental environmental challenge in modern farming [5, 16]. This technological intervention represents a significant advancement in global efforts to balance agricultural productivity with environmental conservation.

Drones equipped with spraying capabilities enable site-specific delivery of pesticides and fertilizers, which substantially reduces overall chemical usage. This precision minimizes residue risks and their negative environmental impacts, addressing a globally recognized concern [8, 9]. Variable rate application (VRT) technology, facilitated by drones, proves essential for environmental sustainability by enabling precise application of water, fertilizers, and pesticides, thereby improving resource efficiency [14].

Research on Digital Agricultural Technologies (DATs) shows that VRT can achieve up to 80% reduction in pesticide use and 60% decrease in fertilizer usage. Additionally, initiatives like the Smart Orchard Spray Application within the IOF2020 project demonstrated significant environmental benefits: 22% to 39% reduction in greenhouse gas emissions through optimized pesticide application, along with 48% reduction in spray drift. Similarly, an autonomous system achieved 66% herbicide savings through precise weed detection and targeted spraying [14].

Another major environmental advantage comes from improved crop health monitoring capabilities. Drones effectively monitor crop health, enabling early detection of diseases, nutrient deficiencies, and water stress [1]. This early detection allows for targeted interventions rather than broad, indiscriminate treatments, potentially reducing the overall need for extensive chemical use. For example, vegetation indices like ExG and VEG, derived from low-cost camera UA V imagery, have achieved high accuracy (83.74–91.99%) in vegetation fraction mapping, demonstrating their potential for precise weed management [8, 5]. Multispectral and hyperspectral sensors on drones can identify changes in crop nutrient status, enabling precision fertilization and preventing over-fertilization and nutrient runoff [5, 16].

Agricultural drones (AUA Vs) are widely recognized as a technology that significantly contributes to sustainable agriculture by reducing input use. This aligns with the global need for policy changes aimed at reducing environmental problems in agriculture. The optimization of input consumption—applying only the required quantities of fertilizer and pesticide—represents a key aspect of this contribution [7, 3].

By enabling more precise farming methods and offering opportunities to optimize inputs and reduce costs, drone technology supports environmentally conscious production. The integration of UA V data with GIS and predictive modeling can further optimize resource usage and enhance crop resilience to stress [8, 4]. A case study of an Italian cereal farm that adopted precision agriculture technologies demonstrated a 53% reduction in pesticides after adoption [6]. These advancements collectively indicate a strong potential for drones to play a crucial role in fostering more sustainable agricultural practices in Turkey.

Farmer Concerns

Turkish farmers face significant economic hurdles when considering new agricultural technologies, particularly agricultural drones (AUA Vs). The high initial investment cost, ranging from USD 15,000 to 25,000, represents a major barrier to adoption, with 67.2% of surveyed farmers citing this as prohibitive. This substantial upfront cost translates to extremely low purchase intention, with only 3.12% of farmers indicating they would consider buying such technology [7]. Beyond the purchase price, farmers worry about ongoing expenses including subscription fees, training costs, maintenance requirements, and eventual replacement costs. These cumulative “adjustment costs” significantly impact the perceived economic viability of new technologies [18]. Without clear evidence of sufficient return on investment and profitability, farmers remain reluctant to make such substantial financial commitments [4]. Farm size also plays a crucial role in technology adoption decisions, with approximately 10.3% of farmers rejecting Precision Agriculture technologies, specifically citing high investment requirements and small field size as significant barriers [10].

Farmers express considerable hesitation regarding the ease of use and practicality of new technologies like drones. Despite acknowledging potential time and cost savings, many remain uncertain about operational aspects such as spraying effectiveness, coverage capabilities, and their own ability to operate drones even after training. Factor analysis reveals that some farmers feel uneasy using new technological tools, consider themselves too old to learn, find the learning process troublesome, and view new technology as an additional burden [7]. The technological complexity of solutions like drones and aircraft/satellite systems has limited their adoption even in more technologically advanced markets like the United States [18]. Interestingly, while approximately half of the surveyed Turkish farmers believed traditional spraying methods were inefficient (47.6%) and potentially harmful to their health (53.4%), they still expressed uncertainty about alternatives [7].

Legal procedures and regulations governing advanced agricultural technologies create additional adoption barriers. Complex or restrictive legal arrangements can impede the implementation of potentially beneficial technologies. In certain regions of Turkey, particularly near the Syrian border, security issues and the presence of signal jammers introduce additional restrictions on drone usage. Farmers must obtain necessary permits regarding flight conditions and adhere to legislation from the General Directorate of Civil Aviation [4, 7]. Data ownership, privacy, and security have emerged as significant concerns in digital agriculture, with ongoing debates centering on farmers’ concerns about ownership of data generated from their fields and potential compensation, alongside data privacy and security considerations [18].

Government intervention could significantly impact adoption rates, with recommendations including incorporating drones into existing state machinery and equipment support programs, providing financial assistance such as interest-free loans for AUA V purchases and extending current agricultural equipment subsidies to cover AUA V technology [7]. For example, in Malaysia, research suggests the government should offer different incentives specifically for rural farmers using advanced technologies like drone cameras [13]. Agricultural cooperatives are identified by Turkish farmers as organizations best positioned to deliver rental services for expensive technologies like drones, potentially accelerating adoption among small and medium-sized operations by leveraging collective farmer funds [7].

TARNET, a technology company affiliated with Agricultural Credit Cooperatives, could play a vital role in facilitating access to AUA V technology by managing subsidy programs or offering affordable rental services [7]. Agricultural extension services can facilitate drone adoption by providing information and guidance, while technology providers increasingly offer cloud-based farm management information systems [18]. Research institutions and industry representatives provide essential evidence about farmers’ technology adoption behaviors, informing policy decisions. The social aspects of technology diffusion—including communication, exhibitions, fairs, and field days—play crucial roles in increasing awareness of new technologies. Developing efficient rental markets represents a viable approach to overcoming high initial purchase costs, especially since most farmers who used drones did so by
renting them [10, 7].

Data Collection Challenges

One of the considerable limitations in this research was the very limited nature of previous studies and case analyses on the adoption of UA Vs in the agricultural sector of Turkey. In contrast to countries where various aspects of UA V technologies have been studied and piloted, there is limited, if any, prior research undertaken within Turkey to fully understand the potential for the technology within a local context. The limited number of case studies and field data restricted the depth and breadth of this research; it was required to rely heavily on international examples or small-scale projects. In this respect, this shortcoming in current localized research reflects a serious requirement for a more extensive research investigation related to the potential of integrating UA Vs within Turkish agriculture.

Moreover, the very process of gaining relevant data was particularly difficult due to the fact that there are no records or organized systems tracking the use of UA Vs within Turkey. This lack of structured data systems made it extremely challenging to assess the current state of UA V implementation or measure its impact across different farming regions. Future research would greatly benefit from the establishment of comprehensive data collection frameworks, such as national registries or cooperative databases, to track UA V utilization and outcomes. Such initiatives would ultimately allow researchers to conduct more powerful analyses in the development of a clear understanding of UA Vs’ potential benefits and challenges within Turkey’s unique agricultural landscape.

Implications for Turkey’s Agricultural Sector

Regardless of the very promising potential, several obstacles stand in the way of this technology being widely used in Turkey’s agricultural sector. A very important factor is that the farmers are not aware, and there is a lack of technical knowledge, especially in the less developed regions. Farmers are skeptical about such technologies being viable; often, they think that it would be too expensive or complicated to use. This is a major knowledge gap, as it prevents farmers from fully grasping how UA Vs could improve productivity by reducing costs. Such problems need awareness campaigns, hands-on training, and demonstration projects that will outline the practical benefits of the technology.

Another major barrier is, of course, the economic one, especially for small-scale farmers who cannot afford the initial costs of the UA Vs or the related services. In this regard, Turkey can introduce government-backed financial incentives, such as subsidies, soft loans, or tax exemptions, to make UA Vs more accessible. Cooperative-led initiatives could also play an important role, whereby farmers are allowed to pool resources and rent UA Vs rather than purchase them outright, reducing the economic burden on individual farmers.

Furthermore, the absence of centralized data systems and extensive research on UA V usage in Turkey represents a barrier to scaling. Without reliable data regarding UA V performance and trends in adoption, it is challenging to develop effective policies or tailor solutions to farmers’ needs. Building national databases and encouraging partnerships between universities, cooperatives, and government bodies could help close this gap and make informed, data-driven decisions to further promote UA V use.

Overcoming these challenges will help Turkey unlock the transformational potential of UA V agriculture. Such technologies can lead to sustainable development through the efficient use of resources, reduction in environmental impact, and enhanced agricultural productivity. With proper support mechanisms, UA Vs could act as an agent of change for modernizing the agricultural sector of Turkey by helping farmers adapt to global challenges with long-term economic and environmental sustainability.

Conclusion

Unmanned Aerial Vehicles (UA Vs) hold a revolutionary capability for the modernization and sustainability of Turkish agriculture. This study has explored the benefits of UA V-based remote sensing technologies in addressing some of the most pressing challenges in Turkey’s agricultural sector, such as over-fertilization, excessive pesticide use, and inefficient resource management. Drawing upon global and local case studies, the findings confirm the critical role UA Vs can play in achieving significant advancements in productivity, cost efficiency, and environmental sustainability.

Utilization of airborne UA V technologies in agriculture has yielded impressive worldwide results because it allows access in real time to high-resolution, spatially detailed data that can be used for precision farming. For example, multispectral and hyperspectral imaging are able to help unlock information about plant health, soil and nutrient availability, and empower Turkish farmers to apply fertilizer and pesticide at the end of their fields. Not only does it decrease the amount of chemical waste, but also this targeted app reduces the environmental pollution, which should be seriously taken into account in areas such as the Konya Plain, where intensive irrigation and fertilizing lead to soil degradation. Additionally, UA Vs equipped with thermal imaging sensors optimize water use by identifying areas of water stress, which is essential for arid regions in Turkey facing water scarcity.

Examples from elsewhere, including the Mediterranean and Australia, showcase the economic and operational benefits of UA V technology. Research has demonstrated pesticide use reductions up to 30% and substantial cost savings via better fertilizer application. For Turkish farmers, these results do so place UA Vs at the forefront of a potential cost reduction by containing or even increasing crop yield. Additionally, pilot projects at local levels in Turkey, encompassing examples of UA V-aided pesticide spraying in paddy fields, have also demonstrated the capacity of the technology to save time, reduce labour costs, and support sustainable farming.

Despite these benefits, several barriers impede the widespread adoption of UA Vs in Turkey. High upfront costs, a lack of technical expertise, and limited awareness about the capabilities of UA Vs are among the most significant challenges. Surveys show that many farmers, especially farmers running small and medium-sized fields, consider UA Vs to be too expensive or too complicated. Moreover, the absence of centralized data systems and empirical studies on UA V adoption within Turkey creates additional obstacles, as farmers and policymakers lack the localized insights necessary to fully understand the technology’s impact.

In order to address these limitations, some strategic efforts should be taken in order to encourage the use of UA V technologies. Specifically, government-sponsored measures like subsidies, interest-free loans, and tax breaks can help alleviate the economic pressure on farmers. Cooperative-based rental services for unmanned air vehicles represent an alternative solution that can be used by farmers to get access to the latest technology with minimal capital outlay. Furthermore, awareness campaigns and hands-on skills programs are crucial in addressing the knowledge gap, especially in rural areas where technical skills are lacking. Such efforts should be supported with the establishment of national databases and research protocols to record UA V applications and results that would provide data-driven decisions regarding the integration of UA Vs in agriculture in Turkey.

In the future term, the application of UA V technology has the promise to turn Turkish agriculture into a model of sustainability and productivity. Through providing precision agriculture capability, UA Vs can alleviate losses of resources, environmental impairment, and yield enhancement, all of which are highly relevant to addressing climate change and the demand for more food in the world. On the other hand, the popularization of UA Vs is compatible with Turkey’s sustainability goals, potentially leading to sustainable long-term economic resilience and environmental protection.

In conclusion, while the road to widespread UA V adoption in Turkey’s agricultural sector is fraught with challenges, the opportunities far outweigh the obstacles. Under the optimal policy environment, along with technological advancement and farmer training, UA Vs may provide disruptive changes to Turkish agriculture, an increase in productivity, sustainability, and competition in the global market. The success stories from other countries and the initial results from local pilot projects provide a clear roadmap for how Turkey can harness the full potential of UA V technology to modernize its agricultural practices and ensure food security for future generations.

References

[1] Hafeez, A., Husain, M. A., Singh, S., Chauhan, A., Khan, M. T., Kumar, N.,
Chauhan, A., & Soni, S. (2023). Implementation of drone technology for farm
monitoring & Pesticide spraying: A review. Information Processing in
Agriculture, 10(2), 192-203. https://doi.org/10.1016/j.inpa.2022.02.002

[2] Meivel, S., Maguteeswaran, R., Gandhiraj, N., & Govindarajan, S. (2016).
Quadcopter UA V Based Fertilizer and Pesticide Spraying System. Journal of
Engineering Sciences, 1(1), 8-12.

[3] Gürbüz, S. G., & Belli ̇ türk, K. (2021). Determination of the Nutritional
Status of Agricultural Lands in Tekirdag Province Ergene District. In A.
Çelik, K. Bellitürk, & M. F. Baran (Authors),
AGRICULTURAL-RESEARCHES-RESOURCEBOOK (pp. 45-80). IKSAD
PUBLISHING HOUSE.

[4] Acıbuca, Veysi. (2024). The possibility of using unmanned aerial vehicles in
agricultural activities in Turkey. Egyptian Journal of Agronomy, 0(0),
0–0. https://doi.org/10.21608/agro.2024.260593.1406

[5] Olson, D., & Anderson, J. (2021). Review on unmanned aerial vehicles, remote
sensors, imagery processing, and their applications in agriculture.
Agronomy Journal, 113(2), 971–992. https://doi.org/10.1002/agj2.20595

[6] Finco, A., Bucci, G., Belletti, M., & Bentivoglio, D. (2021). The Economic
Results of Investing in Precision Agriculture in Durum Wheat Production: A
Case Study in Central Italy. Agronomy, 11(8), 1520. https://doi.org/10.3390/
agronomy11081520

[7] Parmaksiz, O., & Cinar, G. (2023). Technology Acceptance among Farmers: Examples
of Agricultural Unmanned Aerial Vehicles. Agronomy, 13(8), 2077.
https://doi.org/10.3390/agronomy13082077

[8] Aslan, M. F., Durdu, A., Sabanci, K., Ropelewska, E., & Gültekin, S. S. (2022).
A Comprehensive Survey of the Recent Studies with UA V for Precision
Agriculture in Open Fields and Greenhouses. Applied Sciences, 12(3), 1047.
https://doi.org/10.3390/app12031047

[9] Tekin, A. B. (2010). Variable rate fertilizer application in Turkish wheat
agriculture: Economic assessment. African Journal of Agricultural Research,5, 647-652.

[10] Keskin, M., & Sekerli, Y . E. (2016). Awareness and adoption of precision
Agriculture in the Cukurova region of Turkey. Agronomy Research, 14.

[11] Borikar, G. P., Gharat, C., & Deshmukh, S. R. (2022). Application of Drone
Systems for Spraying Pesticides in Advanced Agriculture: A Review. IOP
Conference Series: Materials Science and Engineering, 1259(1), 012015.
https://doi.org/10.1088/1757-899x/1259/1/012015

[12] Furukawa, F., Maruyama, K., Saito, Y . K., & Kaneko, M. (2019). Corn Height
Estimation Using UA V for Yield Prediction and Crop Monitoring. In Unmanned
Aerial Vehicle: Applications in Agriculture and Environment (pp. 51–69).
Springer International Publishing. https://doi.org/10.1007/
978-3-030-27157-2_5

[13] Suwandej, N., Meethongjan, K., Loewen, J., & Vaiyavuth, R. (2022). The
Efficiency of Using Drones to Reduce Farming Costs and Yields. Journal of
Positive School Psychology, 6(5).

[14] Papadopoulos, G., Arduini, S., Uyar, H., Psiroukis, V ., Kasimati, A., & Fountas,
S. (2024). Economic and environmental benefits of digital agricultural
technologies in crop production: A review. Smart Agricultural Technology,
8, 100441. https://doi.org/10.1016/j.atech.2024.100441

[15] Herrmann, I., Bdolach, E., Montekyo, Y ., Rachmilevitch, S., Townsend, P. A., &
Karnieli, A. (2019). Assessment of maize yield and phenology by
drone-mounted superspectral camera. Precision Agriculture, 21(1), 51–76.
https://doi.org/10.1007/s11119-019-09659-5

[16] Radočaj, D., Jurišić, M., & Gašparović, M. (2022). The Role of Remote
Sensing Data and Methods in a Modern Approach to Fertilization in Precision
Agriculture. Remote Sensing, 14(3), 778. https://doi.org/10.3390/rs14030778

[17] Chlingaryan, A., Sukkarieh, S., & Whelan, B. (2018). Machine learning approaches
for crop yield prediction and nitrogen status estimation in precision
agriculture: A review. Computers and Electronics in Agriculture, 151,
61–69. https://doi.org/10.1016/j.compag.2018.05.012

[18] U.S. Department of Agriculture. (2023, February). Precision Agriculture in the Digital Era: Recent Adoption on U.S. Farms (J. McFadden, E. Njuki, & T.
Griffin, Authors; Economic Information Bulletin 248). Economic Research
Service.

[19] Adão, T., Hruška, J., Pádua, L., Bessa, J., Peres, E., Morais, R., & Sousa,
J. (2017). Hyperspectral Imaging: A Review on UA V-Based Sensors, Data
Processing and Applications for Agriculture and Forestry. Remote Sensing,
9(11), 1110. https://doi.org/10.3390/rs9111110

[20] Jang, G., Kim, J., Yu, J.-K., Kim, H.-J., Kim, Y ., Kim, D.-W., Kim, K.-H., Lee,
C. W., & Chung, Y . S. (2020). Review: Cost-Effective Unmanned Aerial
Vehicle (UA V) Platform for Field Plant Breeding Application. Remote
Sensing, 12(6), 998. https://doi.org/10.3390/rs12060998

[21] CSIRO. (2007, March). The economic benefits of precision agriculture: case
studies from Australian grain farms (M. Robertson, P. Carberry, & L.
Brennan, Authors). Grains Research and Development Corporation.

[22] Turna, Y ., & Özcan, A. (2021). The relationship between foreign exchange rate,
interest rate and inflation in Turkey: ARDL approach. Journal of Ekonomi,
5, 19-25.

[23] TCMB Faiz Oranları (%) Gecelik (O/N) [TCMB Interest Rates (%) Overnight (O/N)].
(n.d.). Türkiye Cumhuriyeti Merkez Bankası. Retrieved December 23, 2024,
from https://www.tcmb.gov.tr/wps/wcm/connect/tr/tcmb+tr/main+menu/
temel+faaliyetler/para+politikasi/merkez+bankasi+faiz+oranlari/faiz-oranlari

[24] Fiyat Endeksi (Tüketici Fiyatları) [Price Index (Consumer Prices)]. (n.d.).
Türkiye Cumhuriyeti Merkez Bankası. Retrieved December 23, 2024, from
https://www.tcmb.gov.tr/wps/wcm/connect/TR/TCMB+TR/Main+Menu/Istatistikler/
Enflasyon+Verileri/Tuketici+Fiyatlari

[25] United States Government Accountability Office. (2024, January). Precision
Agriculture Benefits and Challenges for Technology Adoption and Use. United
States Government Accountability Office.

[26] Çukur, F., & Çukur, T. (2024). Bazı Avrupa Birliği ülkelerinin tarımsal
yapılarının TOPSIS yöntemi ile incelenmesi [Examination of agricultural
structures of some European Union countries using TOPSIS method]. Ege
Üniversitesi Ziraat Fakültesi Dergisi, 61(3), 357–366. https://doi.org/10.20289/zfdergi.1462784

[27] Tsouros, D. C., Bibi, S., & Sarigiannidis, P. G. (2019). A Review on UA V-Based
Applications for Precision Agriculture. Information, 10(11), 349.
https://doi.org/10.3390/info10110349

[28] Nahiyoon, S. A., Ren, Z., Wei, P., Li, X., Li, X., Xu, J., Yan, X., & Yuan, H.
(2024). Recent Development Trends in Plant Protection UA Vs: A Journey from
Conventional Practices to Cutting-Edge Technologies—A Comprehensive
Review. Drones, 8(9), 457. https://doi.org/10.3390/drones8090457

[29] Tang, Y ., Hou, C. J., Luo, S. M., Lin, J. T., Yang, Z., & Huang, W. F. (2018).
Effects of operation height and tree shape on droplet deposition in citrus
trees using an unmanned aerial vehicle. Computers and Electronics in
Agriculture, 148, 1–7. https://doi.org/10.1016/j.compag.2018.02.026

[30] Deng, Q., Zhang, Y ., Lin, Z., Gao, X., & Weng, Z. (2024). The Impact of Digital
Technology Application on Agricultural Low-Carbon Transformation—A Case
Study of the Pesticide Reduction Effect of Plant Protection Unmanned Aerial
Vehicles (UA Vs). Sustainability, 16(24), 10920. https://doi.org/10.3390/
su162410920

About the author

Alp Yörük

Alp is currently a junior at Robert College in Istanbul, Turkey. He aspires to become an aeronautical engineer and has a deep passion for engineering, drones, and their applications in environmental sustainability. From the past to the present, Alp has never stopped creating; he’s participated in robotics competitions and project development contests that allowed him to bring his ideas to life.

Alp is especially interested in how drone technology can support sustainable agriculture and help unravel global challenges. One day, he hopes to make a meaningful impact on himself, his community, and the world.

The post Remote Sensing with UAVs: Addressing Over-Fertilization and Pesticide Management for Improved Agricultural Efficiency in Turkey appeared first on Exploratio Journal.

Abstract

This paper assesses the introduction of tomatoes to China and the environmental and economic factors that influenced their adaptation. Based on historical and agricultural data, the study identifies two primary routes through which tomatoes entered China: the Heilongjiang route in the north and the Guangdong route in the south. The evolutionary progression of the two tomato varieties is influenced by regional differences in climate types and soil fertility. In northern China, lower temperatures and significant day-night temperature variation promoted carbohydrate accumulation, resulting in the cultivation of Pink Fruit Tomatoes (PFT) with a higher extent of sweetness. In contrast, the warmer southern climate favored the synthesis of red pigments, leading to the development of Red Fruit Tomatoes (RFT), which contain fewer carbohydrates. The paper also highlights how the dissemination of tomatoes influenced local dietary preferences and led to agricultural and economic growth. These findings underscore the importance of ecological factors in crop adaptation and demonstrate the broader cultural and economic impacts of tomato cultivation in China.

Keywords

Solanum lycopersicum, tomato, China, food science, environment, climate, potassium content, cuisine history.

Introduction

Can you imagine a world where people consume food made of bland chemical compounds? Can you imagine the world that we live in without the existence of crops? Crops are indispensable for the development of human civilizations. Many crops were introduced to China from exotic regions ^[1]. One of China’s most significant exotic crops is tomato (Solanum lycopersicum). 

The tomato is believed to have originated in the Andes Mountain region. Specifically, the ancestor of wild-type tomatoes (S. pimpinellifolium) was thought to have first originated from Peru. In contrast, the ancestor of the cultivated tomatoes (Lycopersicum esculentum var. cerasiforme) was considered to have first originated in Mexico. There are three subgroups of tomatoes: small (S. pimpinellifolium), moderated (S. lycopersicum var cerasiforme), and large (S. lycopersicum). They differ in economic values, are suitable for domestication, and have a wide cultivation scope ^[2]. 

History records and sequencing analyses suggest that tomatoes were introduced to China in the Ming Dynasty, around the 17th century. In addition, historical recordings indicate that tomatoes were reintroduced during the early period of the Republic of China. The general communication routes by which tomatoes were introduced to China can be classified into two relatively late-diverging pathways. 

In the first route, tomatoes were introduced to the Philippines and later cultivated there before being introduced to Guangdong Province and its vicinity. In the second route, Russians constructed the Chinese Eastern Railway and introduced tomatoes to Heilongjiang. The differences in latitude between these two regions led to changes in tomatoes as they adapted to their respective environments. Additionally, the genetic mutations arose due to the differing dietary habits of residents in various areas. 

Specifically, S. lycopersicum has two subspecies in Heilongjiang and Guangdong. The Heilongjiang variant exhibits a lower level of expression of red pigments, which correlates with a reduced synthesis level of flavonoids in its peel. This characteristic makes it easier to peel the tomatoes, aligning with the dietary habits of residents in Heilongjiang. In contrast, the variant found in Guangdong has a normal synthesis level of flavonoids, resulting in a slight reddish peel. This adaptation benefits residents in southern areas where reliable sources of flavonoid intake are limited. 

The characteristics that mainly affect the taste and the appearance of tomatoes are controlled by the fifth chromosome. There are several quantitative trait loci (QTLs) that control the hardness, total soluble solids, and other kinds of characteristics of tomato fruits that are related to the choice of species of tomatoes.

The variation in tomato species not only influenced cultivation practices but also led to significant changes in local economies and other aspects of society. The impacts of introducing tomato species can be explained from two different perspectives. This paper aimed to explore the reasons and genetic mutations that allowed tomatoes to thrive in China, the factors influencing their introduction, and the resulting economic and social impacts of these mutations in different areas.

Section 1 

Introduction of Tomato to China

Species, characteristics, introduction

There are three major subgroups of tomatoes currently in the world. Among them, S. lycopersicum has the highest economic value for its flavor and nutrition ^[2], unlike the other two subgroups, S. pimpinellifolium and Lycopersicum esculentum var. cerasiforme. Specifically, the ancestor of wild-type tomatoes (S. pimpinellifolium) was considered to first originate from Peru, which has relatively small-sized fruit with a diameter of about 1 cm and mass of 1-2 g ^[3]. The ancestor of the cultivated tomatoes (Lycopersicum esculentum var. cerasiforme) was considered to have first originated in Mexico. This species of tomato has a diameter of about 3-8 cm and a mass of 10-30 g ^[3]. The most widely spread tomato species is S. lycopersicum, which originates from S. pimpinellifolium. It has more nutritious substances and flavor storage than S. pimpinellifolium, and it shows a higher potential of becoming an economic crop because it gradually approaches the standard. Because of these characteristics, Mexicans began to domesticate tomatoes relatively early. After Columbus discovered the New World, the tomato spread to a more extensive range than the rest of the world. In the Mid-16^th century, the tomato was introduced to Europe; in the 17^th century, the tomato was introduced to Filipino; North America and Japan were shown that the tomato was introduced in the 18^th century; the tomato stabilized its role for being a kind of global vegetable (economic crop) after 19^th century followed by Miller first defined tomato botanically in 1768 ^[4]. The history of the tomato can be further explored in terms of how it was introduced in China, since it has been proven that it caused impacts. Two major regions where tomato was introduced in China were Guangdong, a southern province in China, and Heilongjiang, a northern province in China ^[5,6]. The species has the most abundant storage of nutritious substances and varied flavors, making it the species with relatively higher economic value among the three existing subgroups of tomatoes.

The route to Guangdong and its vicinity was introduced in the Ming Dynasty during the 17^th century. Additional evidence proving that Guangdong was a region where the tomato was introduced is listed. First, in 1625, “Dian (Yunnan, an interior province close to Guangdong) Annals” mentioned that the tomato was introduced from Yue (Guangdong), meaning that in the Ming Dynasty, the tomato was introduced in Yunnan, following the introduction to Guangdong, indicating an earlier period that the tomato was introduced in Guangdong. Additionally, “Qian (Guizhou, an interior province close to Guangdong) Plants” mentioned the name of the tomato as “June Tomato” ^[7,8]. It is noteworthy that the writer of “Qian Plants” had taken office in Guangdong ^[5], proving that the origin of the introduction of tomatoes to China was in Guangdong.

Meanwhile, Heilongjiang and its vicinity also introduced tomatoes in the early period of the Republic of China ^[5]. In 1915, 1930, and 1932, the “Hulan County Annals” ^[9] and the “Heilongjiang Annals Collection” ^[8] gave descriptions of tomatoes introduced there with similar descriptions of S. lycopersicum and mentioned that the tomato was introduced from Russia, indicating the Russian origin of the tomato in the northern route. Also, Russia constructed the Chinese Eastern Railway, which propelled the prosperity of Harbin ^[10]. During this period, Russia also brought seeds as food to northern China.

Section 2

Environmental factors that affect the introduction of tomato

Several environmental factors can affect the introduction of tomatoes because different subspecies of tomatoes require different environments to survive. This section of the essay will focus on three environmental factors: conditions of soil, conditions of climate, and the ability of tomatoes to adapt to the local environment. Because of the alteration of the environment, the tomato made some changes, including the differentiation of two subspecies, which are commonly called “Pink Fruit Tomato (PFT)” growing mainly in Heilongjiang, and “Red Fruit Tomato (RFT)” growing mainly in Guangdong.

Condition of the soil

The types of soil in Guangdong and Heilongjiang both allow it to be permeable to air and water ^{[11, 12]}.

Figure 1 describes the distribution of potassium in the soil in China. The content of potassium is abundant in the plain in northern China, with total potassium greater than 25 g kg^-1. This was formed by the accumulation of humus, since Heilongjiang has a warm spring and summer season and a cold, long winter season, making vegetation thrive in spring and summer and wither in winter. However, the formation of snow in the cold, long winter season makes it impossible to decompose the wasted plant tissues because the low temperature inhibits the activity of the microbes responsible for decomposing the tissues. With the repetition of this same process for millions of years, Heilongjiang finally formed a thick layer of black soil with an abundance of potassium and other kinds of nutrients.

Figure 2 describes the relationship between the existence of addition of potassium on the growth of tomatoes and the final yield of tomatoes. It was shown that with the addition of potassium chloride, the yield of tomato is 8012 kg hm^-2, which is higher than the group without the addition of potassium chloride with a yield of 6020 kg hm^-2. These data suggest the ability of potassium to improve the yield of tomatoes; considering the data in Figure 1, it is suitable for the tomato to grow in northern China, especially the Northeastern Plain. 

According to Figure 1, the soil condition in Guangdong does not show as excellent as in Heilongjiang because of the lower content of potassium. However, the strong adaptability of tomatoes to the local environment can compensate for this loss. If the basic requirement, such as soil with deep soil layers and good drainage, is achieved, the tomato can adapt to the environment without too much dependence on potassium. To be specific, the tomato requires high soil aeration, and when the oxygen content in the soil drops to 2% ^[15,16], the plants will wither and die. Therefore, it is not suitable to cultivate tomatoes in low-lying and poorly structured soils that are prone to flooding. Sandy loam has good permeability and a rapid rise in soil temperature, which can promote early maturity in low-temperature seasons. Clay loam or organic-rich clay with good drainage has strong fertilizer and water retention abilities, which can promote vigorous plant growth and increase yield.

Tomatoes are suitable for slightly acidic soils, with a pH of 6-7 ^[16]. However, the soil in Guangdong has relatively strong acidity, and it stores a high content of aluminum, which inhibits the growth of tomatoes. Tomato can withstand and survive is also caused of the resistance gene that can be activated by the high-acidity signal from upstream. Guangdong’s unique climate conditions allow tomatoes to adapt to nutrient deficiency, though harsh soil conditions include heavy metals and aluminum, unlike Heilongjiang’s high nutrient content. Granted, the high nutrient content in Heilongjiang allows the synthesis of more nutrients, such as soluble solids, but some colored substances including flavonoids and carotenoids may have a limiting synthesis because while the huge temperature differences between days and nights would favor the synthesis of sugar, the excessively low temperature might impede the synthesis of colored substances. Therefore, PFT would acquire a pink appearance and more content of sugar content, which is more compatible with the cuisine methods used by the Northern residents. Conversely, Southern residents prefer to cook food by stewing and boiling for a long time, and tomato fruits with excessive hardness enhanced by the absence of reaction between flavonoids and pectin in tomato fruits can withstand the condition, so the RFT becomes popular. 

Condition of climate

In addition to the advantage of soil, Heilongjiang takes advantage of its climate conditions to be a great place for tomatoes to grow ^[11]. Heilongjiang is in the temperate monsoon climate region with relatively high rainfall and heat in summer and low temperature and relatively low rainfall in winter. The abundant sunlight and a huge gap in the temperature between days and nights lead to metabolic activity and storage of organic compounds in tomatoes ^[18]. 

Although the nutrients in the soil of Guangdong are limited, it still provides sufficient heat, rainfall, and sunlight because of its position in subtropical and tropical monsoon climate regions.

According to Table 2 and Table 3, Guangdong presents a high amount of precipitation ^[12,17]; another research about the impacts of relative humidity on the synthesis of nutrients of tomato fruits reveals that with the increase in relative humidity in the environment where the tomato grows, the activity of synthesis of nutrient increases, meaning that higher relative humidity will lead to a higher synthesis of nutrients, which sustains the survival of tomatoes, including flavonoids. This is caused by the variance that the 603 bp deletion in the promoter region of the SIMYB12 gene inhibits its expression, resulting in the inability to accumulate flavonoids in mature tomato peel ^[2].

On the other hand, low temperatures can activate the calcium ion signal in plant cells. This regulates the transcriptional activation of downstream cold signals and response factors. Calcium channel protein (CNGC) is believed to be involved in temperature sensing and response. One of the main known pathways for plants to sense low-temperature signals is through the ICE1-CBF-COR transcriptional cascade signaling. The ICE1 (inducer of CBF expression 1) gene encodes a bHLH transcription factor. After sensing the low-temperature signal, ICE1 directly binds to the promoter region of the CBF (C-repeat binding factors) gene to activate the expression of the CBF gene. This gene exists in tomatoes, which means that it can withstand the environment with low temperatures and survive ^[19]. The tomato, hence, will ignore the negative impacts caused by low temperatures, proving that Heilongjiang is an appropriate region to receive the introduction. 

Mini Conclusion

In conclusion, the introduction of tomatoes to China can be attributed to the advantageous climate and soil conditions in Guangdong and Heilongjiang, while the adaptability of the tomato itself is also a possible reason that it can survive in regions that are extremely far away from its “homeland” — the Andes Mountains Region. 

Section 3: Impacts of the introduction of tomatoes

The mentioned advantages and conditions of China propelled the introduction of tomatoes to China, and the introduction of tomatoes also made a difference in several aspects. The main impacts can be attributed to economic impacts and dietary impacts.

Economic impacts

The vegetable economy was affected by the introduction of tomatoes. Since 1961, the yield of tomatoes has experienced an unprecedented rate of growth. According to Figure 3, the yield of tomatoes increased from about 5,000,000 tons to about 35,000,000 tons at the end of 2005. In addition, according to Figure 4, by 2004, the export value of tomatoes in China increased to about 10,000,000 dollars. Meanwhile, the rate of production of tomatoes in China was raised to 1.35%, ranked as the 12^th worldwide in that year. These data show the economic impacts brought by tomatoes and the gradual improvement in economic participation of China in the world.

Dietary impacts

PFT’s soft peel makes it easy for humans to scramble it into juices, which is affinitive with the dietary habits in northern China. The low temperature made food with high energy necessary, so stir-fried food with oil became popular in northern China. “Tomato with scrambled eggs” and similar kinds of meals that use tomato and oil as fried food became popular in the early period of the Republic of China in northern regions ^[5]. The unique nutrients in tomatoes and their unique flavors make it possible for the combination of this exotic vegetable and local methods of cooking. Consequently, the range of ingredients in Chinese meals was extended, leading to impacts on regional dietary changes. Nowadays, tomatoes are promoted by businesses to encourage the public to add them to their daily diets.

RFT has a harder peel compared to PFT because of the accumulation of flavonoids. This is relevant to the subtle improvement of food in southern regions. The characteristics of the peel are relevant to the boiling method in Southern China. It is noteworthy that tomatoes are used as flavors instead of main “roles” in meals in southern regions, like the “Tea with Ganzimi,” which uses tomato as a kind of flavor with sugar to boil with tea. 

In conclusion, the introduction of tomatoes made a difference in the structure and the attitude of Chinese residents toward exotic food.

Conclusion 

The paper mainly discusses the introduction of tomatoes to China and the conditions that facilitated their successful adaptation. The tomato, originating from North and South America, was spread to other continents with the advancement of navigation technology. The two main routes for the introduction of tomatoes to China were the Heilongjiang and Guangdong routes. 

The soil and climate conditions in different regions created distinct environments for the tomatoes to adapt to, eventually leading to the development of different subspecies. The species cultivated by residents in the north evolved as PFT, while other species, such as RFT, developed in Guangdong. RFT exhibits a red appearance due to a favorable climate for synthesizing colored substances, including flavonoids, though it contains a lower level of sugar. In contrast, PFT shows a faint red or pink appearance because of the excessively low temperature. However, the significant temperature difference between day and night in northern regions favors sugar synthesis. Additionally, the nutrient-rich soil in the northern area contributes to the synthesis of essential nutrients ^[13-16].

The introduction and spread of the tomato have triggered economic development and preferences of the residents towards the food. In conclusion, the introduction of tomatoes to China was mainly divided into two routes, each with its unique advantages and characteristics. The impacts of the introduction are divided into several aspects, highlighting the tomato’s potential as an economic crop and its influence on residents’ diets. 

References 

[1] Yao, Y., Huang, J., Yan, Y., & Chen, H. (2005). The agricultural sciences and technologies introduced by the Journal of Beizhi Agricultural Science from the West and their significance in sci-tech. JOURNAL OF AGRICULTURAL UNIVERSITY OF HEBEI, 28(3).

[2] Lin, T., Zhu, G., Zhang J., Xu, X., Yu Q., Zheng, Z., … & Xue, Y. (2017). Genomic analysis reveals the history of tomato breeding. Hereditas, 36(12), 1275-1276.

[3] Rick, C. M., & Holle, M. (1990). Andean Lycopersicon esculentum var. cerasiforme: genetic variation and its evolutionary significance. Economic Botany, 44(Suppl 3), 69-78.

[4] Xiao, Y., An, Z., Huang Y., Li, S., & Zhang B. (2017). Preliminary exploration of the history of tomato development and dissemination. CHINA VEGATABLES, (12), 77-80.

[5] Liu, Y. (2007). RESEARCH ON THE SPREAD OF TOMATO AND ITS INFLUENCE IN CHINA (Doctoral dissertation). Nanjing: Nanjing Agricultural University.

[6] Zhang, P. (2006). Explanation and exploration of the names of Chinese vegetables.

[7] Liu, W. (1991). Dian Annals. Yunnan Education Press.

[8] Zhang, Y., & Xiang, M. (2016). The Names of Tomatoes in Chinese Dialects. Modern Linguistics, 4, 56.

[9] Liao, F., & Ke, Y. Annals of Hulan County during the Republic of China. Local Gazetteer of Alechu Kago during the Guangxu reign.

[10] Ji, Y. (2024). Art as a catalyst: a revitalization strategy for the cultural landscape heritage of Middle Eastern railways. International Academic Forum on Cultural and Artistic Innovation, 3(9), 39-42.

[11] Li, L., Peng, X., Qian, R., Wang, J., Du, H., & Gao, L. (2024). Spatial Variation Characteristics and Influencing Factors of Black Soil Quality in Typical Water-Eroded Sloping Cropland. Journals of Soil and Water Conservation.

[12] Leyva, R., Constán-Aguilar, C., Blasco, B., Sánchez-Rodríguez, E., Romero, L., Soriano, T., & Ruíz, J. M. (2014). Effects of climatic control on tomato yield and nutritional quality in Mediterranean screenhouse. Journal of the science of food and agriculture, 94(1), 63–70. 

[13] Liu, F., Zhang, G. L., Song, X., Li, D., Zhao, Y., Yang, J., … & Yang, F. (2020). High-resolution and three-dimensional mapping of soil texture of China. Geoderma, 361, 114061.

[14] Jin-Xin, W. A. N. G., Qing-jun, L. I., & Yan, Z. H. A. N. G. (2021). The influence of potassium fertilizer on the growth, yield, and the quality of different tomato species. Soil and Fertilizer Sciences in China, (3), 96-101.

[15] Zhang, C., Li, X., Yan, H., Ullah, I., Zuo, Z., Li, L., & Yu, J. (2020). Effects of irrigation quantity and biochar on soil physical properties, growth characteristics, yield, and quality of greenhouse tomato. Agricultural Water Management, 241, 106263.

[16] Tian, M. (2020). Do you know the requirements for growth of the tomato? Xinlang News. https://k.sina.cn/article_7233330430_1af23dcfe00100o61j.html

[17] Liang, J., Liu, Z., Tian, Y., Shi, H., Fei, Y., Qi, J., & Mo, L. (2023). Research on health risk assessment of heavy metals in soil based on multi-factor source apportionment: A case study in Guangdong Province, China. Science of the Total Environment, 858, 159991.

[18] Singh, D. P., Rai, N., Farag, M. A., Maurya, S., Yerasu, S. R., Bisen, M. S., … & Behera, T. K. (2024). Metabolic diversity, biosynthetic pathways, and metabolite biomarkers analyzed via untargeted metabolomics and the antioxidant potential reveal high-temperature tolerance in a tomato hybrid. Plant Stress, 11, 100420.

[19] Shi, Y., Liu, H., Ke, J., Ma, Q., & Wang, S. (2024). Research advances in cyclic nucleotide-gated channels in plants. Chinese Bulletin of Botany, 0-0.

About the author

Tengyu Cui

Tengyu is a student from the A-Level program, Grade 11, Beijing National Day School, China. He currently resides in Beijing, China, and he is 17 years old. He has a passion for exploring food science and food history. With a background of a high school student in Beijing, Tengyu has explored disciplines relevant to biology and food science. Having an interest in extracurricular research, Tengyu has done studies about postharvest fruit preservation and composed a review essay about the pros and cons of food preservation in solving a global challenge. 

Having entered an international competition called iGEM, Tengyu has collaborated with students from other high schools in China to manage to make an aromatherapy product from tea residue by using E. coli, as an expression carrier to catalyze the β-carotene in tea residue to β-ionone. Frequently sanguine to explore and review knowledge, Tengyu continues to do research on the relationships among people, the environment, and the food.

The post The Development of the Tomato (Solanum lycopersicum): Introduction Routes and Factors and Social Impacts in China appeared first on Exploratio Journal.

Abstract

The purpose of this research is to compare the two popular marine microplastic sampling methods: Manta net method and the Neuston net method. Specifically, the precision in the data that each method collected is compared, and a conclusion is deduced. This will be done using various viewpoints including a histogram, a world map microplastic density diagram, and a box-and-whisker plot consisting of count of density measurements vs. density measurements, pure density measurements, and time vs. density measurements respectively. The comparison of precision is done for each viewpoints, and the method that has a higher precision in the data it collected is concluded in the conclusion section.

1. Introduction

Since the worldwide commercialization of plastic products, microplastic particles have been mixed into our drinking water. These microplastic particles are small plastic particles which their diameters are less or equal to 5 millimetres. These microplastic particles can possibly be hazardous to human body when consumed, although there are some debates about the specific effects, including lipid metabolism, induce oxidative stress, and include neurotoxic responses. In order to avoid these harmful effects, there is a need to filter out the microplastic particles from our drinking water. 

However, as microplastic particles are distributed around the globe differently by region, there is a need to use different filtering plans depending on the distribution. And in order to formulate specific plans by region, there is a need to apprehend the specific densities in marine environments by region. Since it is realistically impossible to observe microplastic particles in every ocean, samples have to be used to determine the approximate densities of microplastic particles in the oceans. 

Two useful methods for sampling of microplastic densities are using a Manta net and using a Neuston net. These two sampling methods have different properties, and the selection of which method is going to be used should be determined by evaluating these two methods critically, and concluding which method has more reliability in the measurements. This paper will compare the data collected by using these two sampling methods, and determine which sampling method has higher precision in their measurements.

1.1 Background Research

1.1.1 Manta Net Method

The Manta Net’s name comes from Manta Rays, a sea animal that feeds from small sea organisms at the ocean’s surface. As this name indicates, the Manta Net’s original function was to collect small sea organisms from the surface of the ocean. But as the Manta Net is capable of collecting micro-sized objects, it was started to be used as a microplastic density sampling tool from oceans. 

The specifics of the properties of the Manta Net’s structure are the following:

• The opening of the net has varying dimensions, the width varying from 30 cm to 120 cm, and the height varying from 10 cm to 75 cm. The most common values for these dimensions are 60 cm for width and 15 cm for height according to various researches. 
• Following the dimensions of the opening, the net’s length varies from 200 cm to 300 cm
• The mesh size the of the manta net vary from 300 μm to 350 μm, and the most common mesh size is 330 μm. 

1.1.2 Neuston Net Method

The Neuston net is named after the species “Neustons,” aquatic organisms that stays mostly on the surface of water that originates from planktons. As can be known for the origin of the name, Neuston nets are used for surface sampling of water surface organisms such as zooplankton, but it also used to sample marine microplastic on the surface of water used for research. 

The specifics of the properties of the Neuston nets vary significantly by its function, usage, and design, but the most common properties are the following:

• The opening of the net has varying dimensions, the opening area varying from 0.5 to 1 square meters 
• Following the dimensions of the opening, the net’s length varies from 3 to 8 meters.
• The mesh size the of the neuston nets are usually 333 μm and 335 μm, 

1.1.3 Properties of microplastics in the ocean

A crucial piece information that is essential for the filteration of microplastic is their properties. Microplastic particles are plastic pieces that consists of dimensions shorter than 5 mm, which are part of beads, fragments, pellets, film, foam, and fiber. These microplastic particles are made from different types of polymer chains, the most abundant and dominant type of polymers being polyethylene and polypropylene. 

Polyethylene microplastic particles come from plastic bottles, water tanks, and bags. A reason for these applications is because of its hydrophobic property. As the molecular structure of the polyethylene is non-polar, water, or other liquid with polar structures, is repelled and as a result is water-proof. And as polyethylene has relatively low density, it tends to float on the surface of water, which can be sampled using Manta net or Neuston net.

Polypropylene microplastic particles come from food packaging, automobiles, and electronics. Polypropylene is water-proof, as its water absorption rate is 0.01% after 24 hours in water. Also, polypropylene has a very low density, which makes it capable of being sampled by Manta net or Neuston net.

1.2 Data and Method

1.2.1 Data Base used for Data Collection

The data base used in this research is from “National Centers for Environmental Information” (NCEI), which is based on various sampling methods of microplastic particles in the ocean surface. The sampling methods that are going to be examined are Manta net method and the Neuston net method. 

The data base consists of the date the data was collected, the Latitude of the collection position, the longitude of the collection position, the ocean that the data was collected, region of collection (including subregions), measurement of the density of the particles (pieces/m³), density class range, concentration class, sampling method, references, organization of collection, and accession numbers. The database can be accessed through an article published by the NCEI which has a title “Marine Microplastics,” which will be cited below in the Work Cited section (section 4).

1.2.2 Software used for Data Analysis

In this research, the Tableau Public 2024.2 software was used to analyze and organize the data. Using this software, the pieces of data in the data base was organized into different diagrams in order to explore the different aspects of the effectiveness of the Manta net and the Neuston net method in the sampling of microplastic in marine environments.

2. Research Question

Between the Manta net sampling method and Neuston net sampling method, which method has more precision in sampling microplastic particles in marine environments?

3. Data Collection and Analysis

3.1 Measurement vs. Count of Measurement

The histogram shows the count of measurements for each measurement values according to the two sampling methods, Manta Net method and Neuston Net method represented as blue and orange respectfully. The range of values for the count of measurements are set to be larger or equal to 1, zeros in the y axis will create many holes in the histogram, which will make the data be difficult to compare. Also, for the difference to be shown clearly shown in the model, the upper maximum value for the measurement values has been set to 0.1, since the values of the count of measurement for the two models above the measurement values of 0.1 were consistently 1, which is insignificant for the comparison of the data to conclude which sampling method has higher precision than another. 

This histogram has several implications, first showing the significant difference in count of measurements by sampling method. The highest count of measurement for all measurements is 6 for Manta Net method, and 536 to the third significant figures for the Neuston Net method. The big difference in the number of values gives a portion of the indication that the Neuston Net method has a higher precision than the Neuston Net method, as the higher number of the count of measurement may show the measurements not varying in a large amount throughout repeated trials. However, the distribution of the measurements, specifically the range of the count of measurements has to be analyzed in order to conclude that the Manta Net method has a higher precision than the Neuston Net method.

The range of the count of measurement for the two marine microplastic sampling methods is a strong indication of their precision, as a larger range indicates that the values measured were more consistent throughout the measurement process, especially for the measurement value that the maximum value for the count of measurements is associated with. The range values can be deduced by the following:

As the count of measurement values for the histogram is set to have values of larger or equal to 1, the lowest count of measurement that is shown in the histogram is 1. The Manta Net method’s range of the count of measurement values can be calculated by subtracting the minimum value from the maximum, which gives 6-1=5. Therefore, the Manta Net method’s range of the count of measurement is 5. The Neuston Net method’s range of the count of measurement values can be calculated by subtracting the minimum value from the maximum, which gives 536-1=535. Therefore, the Neuston Net method’s range of the count of measurement is 535. In terms of the whole distribution of values, the range of 5 for the Manta Net method and the range of 535 for the Neuston Net method is a significant difference. And as the Neuston Net method has a larger range than the Manta Net method, it can be deduced that the Neuston Net method had consistent measurements, specifically for the measurement value of 0.00216 which the 536 count of measurements is associated with, and therefore it can be pre-concluded that the Neuston Net method has higher precision than the Manta Net method in terms of sampling marine microplastic densities.

However, there are limitations in this model. First, the region that these measurements were taken are not considered in the histogram. As precision is defined as the consistency of the measurements of microplastic densities in water within a certain region, the region that the values were taken from has to be analyzed in order to accurately conclude which sampling method has higher precision. For the conclusion made above using the histogram, the region that the measurement value of 0.00216 was measured has to be considered, and if they are taken from different regions, the conclusion will vary according to the portion of values that were taken from the different regions. Secondly, the amount of time that each sampling method used to collect the measurement values are not considered. Longer time taken for the collection of the data may result in larger counts of measurements, which can contaminate the data. The time taken for the measurements should also be taken into account in order to make an accurate conclusion for the precision of the two sampling methods.

3.2 World Map Microplastic Density Diagram Analysis

The world microplastic density diagrams show the density measurements for each measurement values according to the two sampling methods, Manta Net method and Neuston Net method represented as blue and orange respectfully. The range of values for the measurements are set to be larger or equal to 0.0007 pieces/m³, as the density is measured in this case so the values can be measured lower than 1, but has to be higher than 0. The lowest value found in the set was 0.0007 pieces/m³, which is applied to the filter in the software. Also, for the difference to be shown clearly shown in the model, the upper maximum value for the measurement values has been set to 1 pieces/m³, since the values above 1 pieces/m³ are not many in count but much higher in value, which makes the majority of the values not significantly seen as visuals as this diagram’s features higher density values as larger area of the color in respect to the sampling methods. The data base as a whole has most of the measurements done in the Pacific and Atlantic Ocean, and the other oceans’ measurements has insignificant difference between the two sampling methods. Therefore, the other regions will be neglected from the discussion for the reason of concision and significance. 

As this world map diagram shows the microplastic density in terms of regions, it can be used to evaluate the consistency of data collection in specific regions. First, to examine the Pacific Ocean region, there is not a very notable difference in the orange and blue regions shown in Figure 2 and Figure 3 respectively. However, there is a difference in area of the colored region. Figure 2 shows a variety of size of area of the colored region, consisting of many smaller areas and relatively fewer larger areas. Conversely, Figure 3 shows relatively higher consistency in the size of blue colored areas, mostly larger colored areas, also in a more compacted region in the Pacific Ocean. Therefore, it can be pre-concluded that in the Pacific Ocean region, the Manta net method has a higher precision than the Neuston net Method, as it has less random errors in the measurement in the same region.

The Atlantic Ocean region has a very significance in the measurement regions between the Manta net and Neuston net sampling methods. In figure 2, the Neuston net method shows a relatively consistent measurement in a compacted area near North America and Central America, while almost no measurement was done with the Manta net method in Figure 3. In this region, the Neuston net method shows high precision as the size of the colored areas are relatively consistent, and the region of measurement is compacted. In the European region of the Atlantic Ocean, there are relatively more measurements done using Manta net method than the Neuston net Method. In this region, the measurements made using the Manta net method has a variety of size of the colored region, having more small colored areas compared to the Neuston net method’s sampling near the North and Central America. Also, the region the measurements were made is relatively larger, which also impacts the amount of data collected by decreasing the consistency of the measured regions. Therefore, in the Atlantic Ocean region, it can be pre-concluded that the Neuston net method has higher precision than the Manta net method in marine microplastic sampling.

However, there are limitations in this model. First, the total amount of measurements done for each method was not taken account. The number of trials is crucial to deducing the precision of a sampling method, as difference in the number of trials can impact the data set to have higher or lower precision than the other method. Therefore, the conclusions made were under the assumption that the number of trials of the sampling using Manta net and the Neuston net does not have a large difference, which will allow a conclusion to be deduced from this diagram. Also, the duration of the data collection has not been considered in this diagram. If the duration of the measurement is very long, the measurement of the microplastic density for both sampling method is capable of varying as time passes, which will impact the precision for both methods as the values themselves change during the trials. Therefore, the conclusion is under the assumption that the values changed not in a large amount between trials.

3.3 Box-and-Whisker Diagram by Time vs. Measurement

The box-and-whisker diagrams show the density measurements for each measurement values according to the two sampling methods, Manta Net method and Neuston, both according to the year that the data was collected. The range of values for the count of measurements are set to be larger or equal to 0.0007 pieces/m³, as the density is measured in this case so the values can be measured lower than 1, but has to be higher than 0. The lowest value found in the set was 0.0007 pieces/m³, which is applied to the filter in the software. Also, for the time of measurement, the years that both methods commonly had been used for the convenience of comparison between the data of two sampling methods. 

To analyze the box-and-whisker plot to compare the precision of the two sampling methods, the number of outliers is the factor. As precision is defined as the consistency of measuring the values between trials of the same condition, in this case, the year that the measurement was done, the higher number of outliers shows that the method has a low precision. In order to calculate the number of outliers, the upper and lower whiskers of the box-and-whisker diagrams have to be calculated, since the values that are higher or lower respectively are outliers of the data. The calculation for the minimum value, or the lower whisker, is done as: 

Minimum Value = Q1 – 1.5(IQR)

where Q1 is the 25^th percentile of the data and IQR is the interquartile range of the data. For the upper The calculation for the maximum value, or the lower whisker, is done as:  

Minimum Value = Q3 + 1.5(IQR)

where Q3 is the 75^th percentile of the data. The values for these maximum and minimums are calculated for each year for figure 4 and 5, which is shown in the table below:

The count of the number of outliers for each year for the two sampling methods is shown in the table below:

As can be seen in Table 3, the box-and-whisker diagram for the Manta net method (Figure 4) has much less years that has more outliers than the Nueston net method, which shows that the precision for each year is dependent on the consistency of the values measured in each trials in the same condition, which means that the outliers reflect an inconsistency in the measurements. Therefore, in the box-and-whisker diagram aspect in respect to the time the measurement was done, it can be pre-concluded that the Manta net method has a higher precision than the Neuston net method.

4. Conclusion

By examining the three diagrams, it could be known that Manta net method and Neuston net method has higher precision when viewed in certain aspects but lower in another; specifically, the histogram viewpoint shows that the Neuston net method has higher precision than the Manta net method, the world map view shows varying results for different regions of ocean that the measurement took place, and the box-and-whisker diagram shows that the Manta net method has higher precision than the Neuston net method. Therefore, it can be concluded that the two method has similar precision in general. However, as the world map showed more significant results in the Atlantic Ocean region where Neuston net method is shown to be more precise, it can also be concluded that the Neuston net sampling method has higher precision than the Manta net method in sampling marine microplastic. 

Works Cited

Author links open overlay panelGabriel Erni-Cassola a, et al. “Distribution of Plastic Polymer Types in the Marine Environment; a Meta-Analysis.” Journal of Hazardous Materials, Elsevier, 21 Feb. 2019, www.sciencedirect.com/science/article/pii/S0304389419301979. 

 “Marine Microplastics.” National Centers for Environmental Information (NCEI), 18 Nov. 2024, www.ncei.noaa.gov/products/microplastics. 

“Neuston.” Neuston – an Overview | ScienceDirect Topics, www.sciencedirect.com/topics/earth-and-planetary-sciences/neuston. Accessed 26 Nov. 2024. 

Nyadjro, Ebenezer S., et al. “The NOAA NCEI Marine Microplastics Database.” Nature News, Nature Publishing Group, 20 Oct. 2023, www.nature.com/articles/s41597-023-02632-y. 

Pasquier, Gabriel, et al. “Manta Net: The Golden Method for Sampling Surface Water Microplastics in Aquatic Environments.” Frontiers, Frontiers, 21 Feb. 2022, www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2022.811112/full. 

Sharma, Kabita. “Polyethylene: Structure, Properties, Types, Uses.” Science Info, 3 Apr. 2024, scienceinfo.com/polyethylene-structure-properties-types-uses/. 

Useon. “Polypropylene (PP) Plastic: Types, Uses and Processing.” USEON, 18 Nov. 2024, www.useon.com/polypropylene/. 

US Department of Commerce, National Oceanic and Atmospheric Administration. “NOAA National Ocean Service Education: Coastal Pollution Tutorial.” NOAA’s National Ocean Service, 22 Oct. 2019, oceanservice.noaa.gov/education/tutorial-coastal/marine-debris/md04.html. 

About the author

Namwoo Cho

Namwoo is a grade 12 student at the Shanghai American School. He is a future engineering student, and is interested in physics and chemistry.

The post Between the Manta Net Sampling Method and Neuston Net Sampling Method, Which Has More Precision in Sampling Microplastic Particles in Marine Environments? appeared first on Exploratio Journal.

1. INTRODUCTION

1.1. Background on Airbnb

Airbnb is a company that was formed in 2008 which provided a service allowing people to rent their extra room or their entire house for temporary lodgers, typically tourists. It developed fast and became a threat to conventional forms of hospitality as it offered an accommodation alternative to hotels. In subsequent years, Airbnb went international and today it has spread to over 220 countries, hosting millions of listings (Ekeroma, 2023). However, as the company grew in popularity, so did its unintended effects on local housing markets hence creating what has become to be referred to as the “Airbnb effect.”

The “Airbnb effect” is the impact that the short-term rentals bring about in terms of housing accessibility and prices. Critics say that companies, like Airbnb, lead to a decline in the number of apartments for mid to long-term rent, which then pushes up the price and can fuel a housing crisis (Hoffman & Heisler, 2020). This impact has turned into a debate in various cities in the world where the cost of rent is a problem. Besides the economic impacts, Airbnb can change cultures and social structures in neighborhoods including introducing new populations of people who are often temporary and can replace the local residents and change their cultures, beliefs, and practices.

1.2. Purpose of the Study

This study aims to explore the extent of Airbnb’s effect on affordable housing and neighborhood dynamics in two major U.S. cities: Seattle and New York. This report will assess two facets of how Airbnb affects urban environments and neighborhoods. This research explores the ‘Airbnb effect’ on housing affordability and the actions local authorities are taking to reduce its impact on the housing rental market.  Thus, it will assess how Airbnb influences the social and cultural context of these cities’ neighborhoods.

1.3. Research Questions

1. What efforts have local governments made to combat the “Airbnb effect” and increase the availability of affordable housing?
2. How does the Airbnb phenomenon affect the cultural and social living environment of neighborhoods?

1.4. Research Focus

This research focuses particularly on Seattle and New York, which have experienced major growth in Airbnb listings and have taken different actions to deal with the challenges short-term rentals present. the selection of these cities was because of their disparate characteristics. With a booming tech sector, Seattle, a city that is relatively smaller, has experienced fast urbanization and an uptick in tourism, positioning it as an important case study of how Airbnb affects a growing urban environment. The large, densely populated, and longstanding tourism and housing issues in New York position it as a unique case study revealing how Airbnb can exacerbate current challenges in established urban locales.

1.5 Methodology

To perform this study, the analysis of both the qualitative and quantitative information was used. The method used in the research was the consideration of scholarly resources, city housing reports, and case studies pertaining to the effects of Airbnb on housing markets, culture, and social relations. Some of the sources used were journals, both peer-reviewed and government publications, and market research papers on the Seattle and New York housing economies. Further, the newspapers’ articles and other policies that focus on public policy within the cities were also considered important in ascertaining the regulations adopted by City authorities and the efficacy of such measures. The comparison of the housing affordability issue, neighborhood change and social effect in both cities provided a fair and diversified understanding of the Airbnb impact based on quantitative data collected from different sources.

2. LITERATURE REVIEW

2.1. The “Airbnb effect” on Housing Markets

The so called “Airbnb effect” has generated a number of theories with aims at describing its effects on housing markets and cities. The term “Airbnb effect” emerged around 2010 as a shorthand for the complex and multifaceted impacts of short-term rental platforms like Airbnb on housing markets, neighborhoods, and cities. However, it is difficult to pinpoint a single author or specific date for the term’s origin. These theories range from the negative impacts of displacement and housing commodification to more positive arguments around urban revitalization and economic growth through tourism. Among the theories, there is the displacement theory which posits that the conversion of long-term rental units to short-term letting results in the displacement of long-term occupants. When property owners want to get more returns from short-term rentals, they take units off the long-term rental market leaving many residents with no option other than to move out, often to sub-standard units (Jiao & Bai, 2020). There is evidence that Airbnb listing growth is directly linked to raised housing prices mainly in cities that attract a lot of tourists (Ekeroma, 2023).

Another important theory is the theory of commodification of housing, which says that platforms like Airbnb are making housing a commercial asset instead of a basic need. From this perspective, Airbnb’s impact is detrimental and exploits the residential spaces for the sake of monetary gains rather than for the main societal purpose of providing shelter. This shift results into a situation whereby housing becomes a market commodity for generating income rather than being a necessity for accommodation (Marsella et al., 2024). This theory also reflects the general culture of commodification of housing where the property is mainly seen in terms of its financial value rather than a place to live thus deepening housing inequalities.

An opposing theory is the urban revitalization theory which postulates that Airbnb and similar platforms can transform deserted neighborhoods for the better by bringing in investment and tourists. The proponents contend that the temporary or short-term letting of homes generates economic growth and results in the revitalization of what might be dormant or forgotten neighborhoods or towns (Mody et al., 2021). Nevertheless, this theory is usually criticized as failing to provide for the social consequences of gentrification such as provisions of accommodation to the poor (Jiao & Bai, 2020). 

In combination, these theories capture the complex nature of the “Airbnb effect” and how it offers both benefits as well as negative impacts on urban neighborhoods. It is therefore important for policy-makers to fully appreciate these different views in order to begin to strike the right balance in regulating short-term rentals for all the ill effects that they pose, while at the same time fully harnessing their benefits in support of local economies.

2.2. Seattle’s Housing Market

Airbnb has disrupted housing in Seattle for several reasons, with a big effect on some specific neighborhoods that attract tourists. The housing market has further tightened and rented space prices have gone up, while availability to these services has gone down in the city (Negi & Tripathi, 2023). Because of the growing number of property owners who change their units into short-term rentals, there is a further reduction in the housing stock that is available for ordinary residents in the city – “a problem that the city already faces,” while “exacerbating what is already a severe housing shortage.” This trend tends to be more significant in towns such as Seattle within which the hi-tech sector has led to an unprecedented increase in housing costs; forcing many residents to live in substandard housing facilities (Richards et al., 2020). This pressure has been compounded by the presence of Airbnb, which continues to drive up rental prices and make it nearly impossible for those in low and middle wage jobs to find accommodation.

Apart from increasing rents, Airbnb listings in Seattle have led to the gentrification of several neighborhoods.  Areas like Capitol Hill and the Central District have seen an influx of short-term rental properties, which has driven up property values and displaced long-term, lower-income residents. As a result, these once-diverse communities are experiencing demographic shifts, with rising rents and home prices, making it harder for existing residents to remain in the area. Neighborhoods that used to be mainly cheap and diverse have become more gentrified with top end demographics, thus forcing out many people (Sarkar et al., 2020). The transition from a long-term rental system to the short-term arrangement has changed the population in the areas in a negative way, specifically losing the cultural and social structures of a neighborhood. What used to be neighborhood shops that primarily served the residents have since been supplanted by tourist-oriented shops which alter the nature of these areas. Seattle’s experience shows how Airbnb’s multifaceted social and economic impacts on urban housing systems present significant difficulties for cities as they attempt to promote economic development without negatively affecting housing affordability and social cohesion.

2.3. New York’s Housing Market

New York City’s housing market has also not been spared the big blow by Airbnb, most especially in areas that are already congested in terms of housing. For a long time, this city has been impacted by a high population density, and huge demand for houses, making it prone to short-term rental impacts (von Briel & Dolnicar 2021). There have been many cases in which Airbnb has been accused of contributing to the hike in rents and shortage in the supply of affordable accommodation as landlords shift from low-cost, long-term rental housing to high-yield short-term rental properties. With these changes, the already existing housing affordability problem in the city has further worsened, forcing residents to struggle to find decent homes to rent or even buy (Laskin, 2020). There is high competition in the housing market and Airbnb is part of the problem as it causes displacement of residents and a change of demographics in previously affordable neighborhoods.

2.4 Governmental Efforts to Combat the “Airbnb effect”

Due to the increasing importance of the affordable housing problem, Seattle has succeeded in implementing new strategies to control the use of short-term rentals. In December 2017, the city put in place new regulations that took effect in January 2019. In most regions of the city, these rules set the limit on a host’s units to two properties. The design of this cap limits how many short-term rental units commercial operators can own, thus preserving the amount of affordable housing for long-term residents. The intention of these regulations is to restrict the shift from long term accommodation towards short term rental to resolve the housing supply problem for residents (Marsella et al., 2024). In conjunction with these policies, Seattle intends to grow the production of affordable housing by offering developers incentives to include affordable units in their developments. Still, the performance of these actions is uncertain, and some critics say the approaches are not adequateto resolve the problem. They propose that the policies do not completely meet the economic pressures behind rising housing prices and create loopholes that still promote short-term rentals, thereby hindering their potential to solve the affordable housing crisis.

One of the pioneering cities in taking serious action to regulate short-term rental services, New York City, has put into effect some of the strictest laws across the United States. In 2010, the amendment to the New York State Multiple Dwelling Law banned rentals lasting less than 30 days in almost all apartment buildings, unless the permanent tenant is also on the premises during the rental. The goal behind this law was to diminish unlawful hotel operations through services like Airbnb. As of 2016, further rules came into effect with Local Law 146, which necessitated that Airbnb, along with other platforms, supply the city with listing information. In 2018, Local Law 18 was passed, enforcing stricter penalties for illegal short-term rentals and requiring more transparency from hosts. These laws collectively prevent individuals from listing entire apartments for extended periods, and hosts must reside in the unit for any rental under 30 days (Jiao & Bai, 2020). The city has also sued some hosts who have infringed on these laws, resulting to hefty fines being charged on the culprits. These steps are taken in line with increased efforts to save affordable housing and mitigate long-time inhabitants from eviction. Such regulations have had some impact in the reduction of Airbnb listings, although they have not been without their problems; for instance, property owners have resisted these laws and fight legal battles against the enforcement of such laws.

2.5. Impact on Cultural and Social Living Environment

It is clearly proven that having short term rentals in residential areas has a ripple effect on the way a community is formed. Specifically, short-term rentals disrupt the social fabric by replacing long-term residents with transient occupants, which undermines community stability and cohesion. Also, short term occupancy displaces long-term occupiers, thus undermining the social structure of the community, and erasing the shared community identity (Hoffman & Heisler, 2020). At times, tourism may lead to the development of negative effects, such as gentrification, where locals are forced out and the cultures of communities are changed. This occurrence is especially prevalent in places where hosts offer many Airbnb spaces for rent, and the locals’ organic community nature is eroded by tourism.

Some neighborhoods in Seattle have been profoundly affected by Airbnb, while others have undergone more moderate change. For instance, areas such as Capitol Hill and Ballard have been found to be effectively experiencing a form of gentrification as more and more of them are becoming subsumed into the world of short-term lets, thereby raising questions on the disintegration of social bonds. Old timers, residents who have lived in the neighborhood for most of their lives, complained of feeling alienated from their own communities by the ever-flowing fury of tourists (Mody et al., 2021). Moreover, it arose that some of the previously local-oriented stores are changing their target, turning into tourist-oriented ones, and thus changing the culture of the district.

In New York the effects of Airbnb on neighborhood characteristics are even more apparent. On the outskirts of New York’s boroughs, such as Williamsburg and the Lower East Side, which are renowned for their ethnical diversity and nightlife, the rise of short-term rental has hugely been realized. This outcome has caused a dramatic decrease in the production of housing units that are reasonably affordable, along with an increase in tensions between residents and tourists (Nieuwland & Van Melik, 2020). A number of longtime residents in these areas have identified the influx of tourist shops and services along with the vanishing of traditional business staples as proof that gentrification is currently underway. The social bonds of these neighborhoods have worn thin due to resident displacement caused by costs that have become unaffordable.

2.6. Comparative Analysis: Seattle vs. New York

Similarities in the “Airbnb effect”

Seattle and New York recognize similar problems with high rent and shortages in longer term accommodation because of the impact of Airbnb. In each city, there are grumbles about the consequences of short-term rentals, which frequently displace established residents and redefine some areas into high-end tourist hotspots (Jiao & Bai, 2020). The results of this include housing costs, availability, and neighborhood connections; all of which have led to official interventions. In both municipalities, there is local community unease about the diminishing neighborhood character and its effects on social relationships. In spite of these activities, the effectiveness of regulatory measures continue to be a matter of debate, with sustained conflicts between the rise of short-term rentals and the affordability of housing.

Differences in Governmental Response

The two municipalities have striven to create measures to control the impact of short-term rentals, but they have done so in disparate manners. In addition to tough regulations regarding the number of properties a host can register, Seattle encourages affordable housing (Marsella et al., 2024). New York, alternatively, is more proactive and has  banned most short-term rental platforms and sanctioned only owner-occupied homes for short-term rentals, legitimizing the rules through legal power and potential consequences. These differences can therefore be attributed to the specificities of each city’s coping strategies, as well as the political and economic realities it presents. Critics who argue that these cities have not gone far enough often suggest more stringent measures, such as capping the number of short-term rental units, imposing higher taxes or fees on Airbnb hosts, or increasing enforcement mechanisms to ensure compliance with existing laws (Ekeroma, 2023). They believe these steps could help mitigate the impact on local housing markets and preserve affordable housing for long-term residents.

Comparative Impact on Neighborhoods

Although the cultural and social significance of Airbnb is the same in both Seattle and New York, many aspects of the phenomenon are city-specific. In Seattle, the changes appear to be a lot more focused and can be felt at the neighborhood level, with some areas experiencing dramatic shifts in community demographics (Jiao & Bai, 2020). In contrast, New York’s neighborhoods face more widespread disruption, with entire communities being reshaped by the influx of short-term rentals. This effect is manifested to a greater degree in New York since there are significantly more Airbnb listings concentrated in this area and their adverse effects on the availability of housing and the overall cohesiveness of communities are even more apparent.

2.7. Discussion

Synthesis of Findings

Observations made in Seattle and New York demonstrate that both markets have been changing due to the presence of Airbnb: rental prices are going up, and options for affordable housing are shrinking. Seattle hosts are using Airbnb to generate income while increasing housing costs and making it harder for residents to find affordable housing. Because the city’s market size is relatively small, the consequences of short-term rentals have affected this area more. While this is an issue in many cities all over the world, New York’s housing stock has been further stressed, and Airbnb has acted as a vehicle for gentrification which has led to the evictions of longer-term residents. These outcomes suggest that the “Airbnb effect” is present in various markets, and is not unique to small towns or cities.

At the same time, the cities have significant differences in the scale and nature of Airbnb’s impacts. New York has a larger population, and the population density is higher; therefore, the impact is felt even more as the effects are more drastic in the neighborhood and the cohesiveness of the community. Still, these changes have happened in Seattle too, albeit somewhat smaller in scale. The responses from the local governments of both cities while differing in the extent show that managing the sharing economy is not easy. Nonetheless, despite the attempts to regulate short-term rentals, their regulation remains weak and relative, and both cities continue to share concerns about Airbnb’s economic advantages and detrimental effects on community stability and affordable housing.

Implications for Policy and Practice

The implications of these findings for policymakers are clear: It is supposed to underline the necessity of developing more profound and proactive models to tackle the issue connected with Airbnb. Currently, the regulation performed in Seattle and New York has failed to address negative effects on affordable housing and community stability. The law makers should ascertain the possibility of enforcing more strict measures; not just restricting the essence of short-term rental but also increasing the consequences of violation of the set laws. Further, providing higher incentives to property owners to focus on long-term rental accommodations rather than short-term ones can also relieve the housing market pressures.

However, it is also crucial for the cities to take a more active role in tracking the impact of Airbnb and other similar services. This may include monitoring the capacities of the housing market and the socio-economic effects of short-term rentals so that changes can be made to the existing policies if necessary. The findings from the Seattle and New York cases are therefore noteworthy as they illustrate how different cities can deal with similar issues. It is imperative to achieve equilibrium in the sharing economy that protects rights of residents but encourages innovation towards making housing more affordable in the future.

Future Research Directions

Further studies should examine the consequences of Airbnb regulations in the long run concerning housing markets and neighborhoods. Also, it would be useful to compare the current state of affairs with other cities that have different legal frameworks so that the results can then show the efficacy of different methods. Additional research may also seek to analyze other areas of life in cities affected by short-term rental operations, including mobility, economy, and fairness.

Limitations

This study focused on Seattle and New York, two large, economically prosperous cities. The findings may not be directly applicable to smaller cities or those with different economic conditions. Additionally, the study focused primarily on housing affordability and neighborhood dynamics, and further research may be needed to explore other potential impacts of Airbnb.

3. CONCLUSION

3.1. Recap of Key Points

In this analysis, we explore the effects of Airbnb on affordable housing and community change in Seattle and New York. The conclusions presented in the paper show that, over time, Airbnb has led to the increase in rent and the decrease in housing stock and community cohesion in both cities. Governmental actions were discussed in the study along with the differences of success when it comes to facing these challenges.

3.2. Final Thoughts

Airbnb has both potential benefits and drawbacks for cities. While it can offer economic opportunities, it can also contribute to housing challenges. Cities need to find a balance between regulation and innovation to ensure that Airbnb benefits residents while mitigating its negative impacts. Comparing the progress of Seattle and New York can give insights into other cities that experience similar problems with the sharing economy and the importance of enhanced and rational regulation.

REFERENCES

Ekeroma, J. E. (2023). The Airbnb Phenomenon: A Qualitative Analysis of Its Consequences on Urban Housing Markets.

Hoffman, L. M., & Heisler, B. S. (2020). Airbnb, short-term rentals and the future of housing. Routledge.

Jiao, J., & Bai, S. (2020). An empirical analysis of Airbnb listings in forty American cities. Cities, 99, 102618.

Laskin, K. (2020). Is Airbnb Polluting the Big Apple? The Impact of Regulating the Short-Term Rental Service in New York City. Journal of Civil Rights and Economic Development, 33(3), 5.

Marsella, A., Wagner, G. A., Melo, V., Anastasi, S., & Stephenson, F. (2024). The Political Actors Behind Airbnb Bans: Evidence from New York City. Available at SSRN 4873434.

Mody, M., Suess, C., & Dogru, T. (2021). Does Airbnb impact non-hosting Residents’ quality of life? Comparing media discourse with empirical evidence. Tourism Management Perspectives, 39, 100853.

Negi, G., & Tripathi, S. (2023). Airbnb phenomenon: a review of literature and future research directions. Journal of Hospitality and Tourism Insights, 6(5), 1909-1925.

Nieuwland, S., & Van Melik, R. (2020). Regulating Airbnb: how cities deal with perceived negative externalities of short-term rentals. Current issues in tourism, 23(7), 811-825.

Richards, S., Brown, L., & Dilettuso, A. (2020). The Airbnb phenomenon: The resident’s perspective. International Journal of Tourism Cities, 6(1), 8-26.

Sarkar, A., Koohikamali, M., & Pick, J. B. (2020). Spatial and socioeconomic analysis of host participation in the sharing economy: Airbnb in New York City. Information Technology & People, 33(3), 983-1009.

Von Briel, D., & Dolnicar, S. (2021). The evolution of Airbnb regulations. Airbnb before, during and after COVID-19. University of Queensland. https://doi. org/10.6084/m9. figshare, 14195972.

About the author

Jeremy Li

Jeremy is a 12th-grade student interested in economics who is now exploring the complex relationship between short-term rental platforms like Airbnb and the effects on local housing markets and communities. Through his most recent research, Jeremy seeks to elucidate the “Airbnb effect,” so affecting social dynamics and housing affordability in metropolitan settings. His work focuses on the many experiences of cities like Seattle and New York, where he examines the challenges brought by rising rental prices and displaced long-time residents. Developing effective policies that balance the interests of tourists and local people depends on an appreciation of these dynamics.

The post The Impact of Airbnb on Affordable Housing and Neighborhood Dynamics: A Comparative Analysis of Seattle and New York appeared first on Exploratio Journal.

INTRODUCTION

Let us first understand what is circular economy. The circular economy is a system in which materials are never wasted and nature is replenished. A circular economy keeps items and materials in circulation through activities such as maintenance, reuse, refurbishing, remanufacturing, recycling, and composting. The circular economy addresses climate change and other global issues such as biodiversity loss, waste, and pollution by separating economic activity from the consumption of scarce resources. (Macarthur et al., 2022) Humans’ perspective of resource management and economic growth has drastically evolved as a result of the circular economy. The linkage between circular economy and sustainability needs to be well understood. The circular economy is a way to achieve SDGs. Sustainability is a larger umbrella which has a circular economy as its handle. As per research, it is shown that CE is the tool that needs to be used to achieve sustainable development, especially in terms of the environment(M. Walker ., 2018) which we are going to focus on in this research. With the start of the Industrial Revolution, the take-make-dispose paradigm of resource consumption has come to represent the linear economy. However, with the introduction of mass manufacturing, consumption, and trade globalisation, it quickly began to drastically decline, giving rise to the Circular economy (CE). The term “circular economy” (CE) was initially used by Pearce and Turner (1990) in their work “Economics of Natural Resources and Environment.” They claimed that the current economic system should be changed to a circular economy because producing activities are constantly generating waste and pollution, endangering the environment so the system should be transformed into a circular one by considering waste as a source for more resources. (Andrade et al., 2022))

There are still several gaps and challenges impeding the digital technology-aided circular business model. There are still a lot of problems related to information that act as barriers to the integration of technology and the circular economy. Among these include inadequate information, high transaction and search expenses and a lack of knowledge. Data integration issues are also overlooked while considering the role of digital solutions in the circular economy. This is an essential issue as it is not always clear how to utilize data to promote the shift from a linear economy to a more circular one. Data integration is important for businesses as it combines the information sources with the business goals of the stakeholders involved.

Expanding on the issues mentioned above, the lack of information and data about various products and supply chains can cause problems as without relevant data it is difficult to manage the resources properly. The existing knowledge gap can also prevent businesses from reducing their waste generation. The costs required to obtain and utilize relevant data can be very high, especially for smaller companies and start-ups. This not only includes the money associated but also the time and energy required to evaluate the data. The integration of data is also an essential issue as it is not always clear how to utilize data to promote the shift from a linear economy to a more circular one. Various stakeholders use different formats, standards, tools and platforms for data integration causing difficulties in decision making.

AIM:

The topic of research the author proposes is a study to investigate the impact of data science in better adoption and implementation of circular economy in manufacturing. The study is going to look into the different variables and their extent of impact on the adoption of circular economy in manufacturing.

Research question –

What is the influence of data science on circular economy in manufacturing?

Objectives:

1. To gain insights and explore the scope of the circular economy.
2. To identify and investigate how data science impacts the circular economy.
3. To establish a relationship between data science and circular economy in manufacturing
4. To observe the influence of data science in circular economy in manufacturing

2. LITERATURE REVIEW

2.1. Circular economy

The circular economy is a model which is framed to keep existing resources in use and increase their lifespan for as long as possible. Unlike traditional linear economy which follows a “take-make-dispose” model where products are disposed of after being used, circular economy aims at keeping products and materials in use and extracting maximum value from them by eliminating the waste produced by them. Furthermore, the circular economy revolves around product sustainability design development, reuse, remanufacture, and recycling principles. The process of circular economy starts by reducing unnecessary consumption and product design development. Secondly, efforts are made to increase the product lifecycle and reduce the need for new raw materials by refurbishment, remanufacturing, and repurposing. In the end, materials which can be recycled are taken out from products whose lifecycle has ended, and the emphasis on new virgin materials required as raw material is reduced. Product sustainability design refers to the creation of new products in such a way that their negative effects on the environment can be minimized by following various steps such as material selection, and energy efficiency. Remanufacturing is the process of restoring old and used products to new-like conditions so that the need for new materials and extra energy can be reduced.

2.2 Development of circular economy over a period

(1990s): Attempts were being made to frame the concept of “Circular economy” by one of the well-known economists, D. Pierce who was later joined by K. Turner for the same. Other terms like “cradle-to-cradle” and industrial ecology were also put forward, which were identical and awareness regarding the change of approach to using materials started developing. “Eco-industrial parks were developed in North and South America, Southeast Asia, Europe and southern Africa during the 1990s”. Talking about the example of Copenhagen, the waste of coal in one of the power plants started being used as raw materials in fertilizers.

(2000 to 2010): A need to change production and consumption methods along with conserving natural resources was felt by some of the world’s greatest political leaders. Germany introduced its “Closed Substance Cycle Management Act” and Japan also added basic laws to approach a sustainable economy and recycle-based society in 2000. In addition to this, China became the world’s first country to adopt legislation on circular economy in 2002 and these changes were made in its 12th Five-Year Plan.

(2010 to present): The main reason for the shift from a linear model to a circular model of the economy is to ensure the sustainable development of the world in the context of the fourth industrial revolution. Moreover, it has been studied that this paradigm shift from and abandonment of “single-use” resources has the potential to generate economic benefits worth $4.5 trillion by 2030. Adding on, 44% of the Fortune Global top-100 companies have already adopted the mechanisms of circular economy. (Walter et al ., 2018) One of the major factors which led to consumers rethink their approach towards sustainability and widen the use of digital technology in the circular economy was COVID-19, significantly increasing consideration of the impact of their purchase on ESG (Environment, Social, and Corporate Governance). “The transition to a circular economy model in companies in these areas of economic activity will help reduce raw material costs, expand markets, improve brand reputation and dialogue with customers, increase their loyalty, create a competitive business model and much more.” According to a survey by e-commerce delivery platform Soundcloud, 38% of the respondents showed their willingness to pay an extra amount for the delivery of their goods in environmentally friendly vehicles.

2.3 Data Science and Technology

Now coming to digital technologies, IoT in business may help companies access real-time remote tracking of product location, status and consumption. IoT and other digital technologies help companies to gain knowledge about consumer behaviour and gain great opportunities thereby, changing the nature of the relationships between manufacturers and customers from negotiation to communication and increasing the scope of collaboration & partnership. As the products become smarter, manufacturers may upgrade their digital components such as firmware. Product sustainable upgradability will improve the transition to a Circular Economy by contrasting product obsolescence and its resulting material waste.

In written works, Big Data is usually defined through 4’s Volume, Variety, Velocity and Veracity. They are distinguished by i) huge volumes of data generated continuously, ii) different and unstructured forms, such as imaging, texting, etc. iii) high data generating frequency and iv) a good quality and a clear, validated application. As Big Data is massive and changes too quickly, it cannot be analysed using typical software or database procedures. In research, big data is usually seen as a legitimate strategy to facilitate better decision-making, when paired with the right analytical methods. To be more precise, analytics uses software and data mining techniques to create decision support systems and business intelligence to identify patterns in data and make decisions. Analyzing data may help in gaining the right implications and in terms of business it helps in gaining insights, for comprehensive decision-making. Thus, the combination of Big Data and analytics which leads in providing valuable insights leading to the right sustainable decision making may help management move closer to Circular Economy.

The adoption of a circular economy has given an advantage to various stakeholders across the businesses thus, promoting brand image, sustainability, economic and national growth. Shifting to circular business practices provides opportunities for cost optimization, and resource efficiency, and helps businesses escape price volatility. Ruling bodies and government benefit directly because of job creation and economic resilience features of the circular economy. Moreover, it allows them to reduce pollution levels and achieve their green initiatives and national sustainability goals much faster and more effectively as compared to linear manufacturing models. Consumers get the advantage of access to more durable, sustainable, and affordable products. Through processes such as remanufacturing and refurbishment, consumers can purchase the same products at much lower prices. Furthermore, consumers can track the supply chains of the products they are using because of transparent supply chain practices in a circular economy and make informed decisions regarding which products to purchase. For organizations, reduction in resource consumption, waste generation, and disposable costs leads to cost efficiency and cost optimization. Moreover, when processes like refurbishment and remanufacturing are implemented, the manufacturing costs of the business are also saved as the firm does not need to manufacture from scratch. Adding on, this also gives huge innovation opportunities to businesses as they can develop new products, and services by re-thinking about what can be further worked upon by using materials from other products. In addition to this, circular economy allows businesses for market differentiation as businesses which follow the model of circular economy, transparent supply chains, and fair sourcing of the materials can attract a loyal base of customers.

Although there are many advantages of circular economy as mentioned above, there are some challenges and limitations too that cannot be ignored. The standardization and interoperability of the models and formats of the data hinder the process of exchanging data seamlessly within multiple organizations. Moreover, organizations resist sharing their data due to extra efforts needed to transform the data and adopt other data modelling practices. This further leads to one bigger challenge, which is huge investments needed by organizations trying to adapt to the circular economy because of the upgradation and software cost needed for technology. Though AI (Artificial Intelligence), IOT (Internet of Things), and blockchain have emerged as promising technologies which help in the circular economy, the fact cannot be ignored that data privacy has also raised as a huge constraint because of huge concerns over privacy and security of data while using these technologies. Furthermore, strong intent of collaboration is the need of the hour due to the complex supply chain processes which include multiple stakeholders across the globe. Thus, product takeback, remanufacturing, and recycling which are the core of the circular economy become intensively complex and require transparency and collaboration.

Regulatory frameworks and policies also sometimes prove as a huge limitation in the implementation of digital technology in circular business models. The primary reasons can be complex regulatory requirements such as water and waste management or standardization of products, which can require huge resources to fulfil which can be troublesome for small and medium-sized enterprises (SMEs). Adding on, regulatory policies may lag to cope with fast-paced upcoming technologies like Artificial Intelligence and blockchain ultimately, creating a huge gap between the pace of both. Furthermore, many nations have outdated legal policies and regulations regarding the circular economy, which further creates an uncertain legal environment thus, preventing firms from investing and upgrading to circular economy models. There is a strong need for collaboration which enables circular-models organizations to assist governing bodies in framing and altering these policies. The digital skills gap from a global perspective is also a significant challenge as it reduces the effectiveness of technology-infused circular economy models. Human resources with expertise in various fields such as data analytics, supply chain optimization, data programming, and cybersecurity are most needed. But unfortunately, there is a shortage of these professionals in industries which are less-technology oriented such as manufacturing and agriculture. A potential solution can be collaboration between academics, training institutes and industry to bridge this gap.

2.4.Theories

Across the literature, the author has come across various theories which would help in establishing relationships between different variables within the theoretical framework. These theories are found to be aligned with the study being conducted.

2.4.1.Theory of decoupling

Disentangling “economic goods” from “environmental bads” has been suggested by the UN and the OECD as a way to achieve sustainability. The type of decoupling required for ecological sustainability is indicated by scientific consensus findings on resource consumption and environmental repercussions (such as greenhouse gas emissions): global, absolute, fast enough, and long enough. This objective provides a basis for classifying the various forms of decoupling according to their applicability. (UNEP,2011) The word refers to a situation in which economic progress occurs without a corresponding increase in the consumption of material resources or a detrimental influence on the environment, in contrast to most prior experiences (IRP, 2017, 23). The idea is particularly important for formulating policies that, in the first place, consider economic growth to be desirable or necessary, and, in the second place, acknowledge the unsustainable nature of the existing levels of material consumption and environmental harm brought on by the economy. Economic growth must be divorced from rising material consumption and environmental effects if it is to continue (OECD, 2001; UNEP, 2011). When one wishes to promote economic growth while acknowledging previous or present unsustainable development brought on by the economy, they can utilise the prospect of decoupling as support (Jackson and Victor, 2019; Hickel and Kallis, 2019)

2.4.2. Resource efficiency theory

Economic and environmental policy heavily weighs the importance of resource efficiency and the circular economy. The circular economy and resource efficiency are normative notions that advocate maximising prosperity and well-being while minimising losses and promoting enhanced material circulation while respecting the boundaries of the environment. Long-term advantages to society and the economy come from resource efficiency, which preserves the environment and lessens reliance on the natural resources that power the economy. In addition to assimilating air pollutants and wastes, ecosystems give society access to resources, food, clean water and air, as well as aesthetic and recreational enjoyment. Ecosystem collapse, deterioration, and depletion can be prevented with resource efficiency.

2.4.3. PSS theory

A product-service system is the outcome of an innovation strategy that moves the company’s emphasis from creating and marketing physical products to creating a system of products and services that work together to meet particular customer needs. A PSS is referred to as a Sustainable Product-Service System when it helps to reorient the present unsustainable trends in production and consumption practices. By using this strategy, a business offers extra services to ensure the product’s longevity and usefulness (also known as product life extension) before it is sold to a customer or consumer. Over a predetermined length of time, maintenance, repair, upgrading, and replacement services would all be included in a normal service contract. Following the conclusion of the contract, the PSS provider may take back the product, deciding about its possible sale or disposal.

2.4.4. Closed loop development theory

Closed-loop manufacturing systems work to achieve sustainability by concurrently advancing environmental and economic objectives. We point out that businesses can prevent a lot of harmful environmental effects, like waste, energy use, transportation, and packaging, by implementing closed-loop production methods. We present the idea of constructing closed-loop production systems using the framework of sustainable supply chain networks. The closing of process linkages across the businesses in a supply chain is made possible by the implementation of SSCN, which speeds up the shift from a flow economy to a circular economy. Circular economies are more sustainable, which eventually boosts both the economy and the environment. Designing is more than just picking and mixing conventional production parameters to achieve high output and cost-effectiveness. It is now more crucial than ever to take into account all forms of waste generated during the production process and to take the appropriate steps to prevent, minimise, reuse, or recycle waste that already exists.

2.4.5. Collaborative consumption theory

The collective usage of an item or service by a group is known as collaborative consumption. In contrast to collaborative consumption, which allows several people to have access to a good and share its cost, regular consumption involves a single person paying the entire cost of an item and having exclusive use of it. Flea markets, swap meets, garage sales and second-hand shops are examples of the types of events that have historically given rise to collaborative consumption, which is defined as the practice where consumers both obtain and provide valuable resources or services through direct interaction with peers or through intermediary platforms (Felson & Spaeth, 1978; Belk, 2014). But in recent years, it has attracted more attention and momentum, especially with the introduction of digital technologies like Web 2.0, mobile technology, and social media.

2.4.6. Social sustainability theory

The word “social sustainability,” which has multiple definitions, generally refers to the social aspects of sustainability. Social sustainability theory highlights the need for integrated approaches to sustainable development by acknowledging the interdependence and connectivity of social, economic, and environmental systems. In a normative sense, it refers to the social goals of sustainability initiatives.

2.4.7. Cognitive dissonance theory

It is essential to understand the behaviour of consumers to design strategies regarding the adoption of the circular economy. Several existing theories explain the psychology behind consumer behaviour towards digitalization and circular economy. The first theory is the cognitive dissonance theory which was proposed in 1975 by psychologist Leon Festinger. It provides us insights into consumer behaviour after they have evaluated and used the product or the service and studies their behaviour when they face a discrepancy between the pre-service and post-service perception of the product (Park et al., 2015, Venkatesh and Goyal, 2010). These contradictory cognitions trigger the expectations and psychological state of the user which leads to negative emotions. This can harm the reputation of the brand which can demotivate the individual to adopt the new technology. Previous research shows that consumers experience three emotions namely anger (Harmon-Jones, 2004, Harmon-Jones et al., 2017), guilt (Gosling et al., 2006, Turel, 2016) and regret (Roese and Summerville, 2005, Gilovich et al., 1995b) while experiencing the performance shortcoming.

2.4.8. Unified theory of consumer acceptance

The second theory is the Unified Theory of Consumer Acceptance and Use of Technology which aims to give an understanding of the variables impacting the user behaviour across different contexts and has laid out two key determinants for studying behavioural intention. It is a sub-part of the TAM model. The theory focuses on two main key determinants which are perceived ease of use and perceived usefulness.

2.4.9. ICT theory

The Integration of Information and Communication Technology theory was the result of several theories and past developments. The TAM model and the diffusion of innovations theory have been greatly influential in understanding the behaviour towards ICTs. ICT theory studies the impact and influence of technological developments on communication processes, information, individuals and groups. The theory helps to navigate the complex relationship between the behaviour of individuals and technological tools.

2.4.10. Digital twin theory

Digital twin by definition is the virtual representation of an object. The theory of digital theory revolves around the concept of Industry 4.0 and the transformation of technologies. (IEEE,2020) NASA’s Apollo program had a huge role in the development of this theory as they used digital models for simulations to ensure the reliability of the Apollo missions. It further developed after the emergence of Industry 4.0 where several replicas were created for the systems which helped to improve the overall efficiency of data. (Philips et al., 2020) In recent times the theory has been used due to its operational efficiency and reduced cost. Similarly, this theory can be used in the production of products that hold sustainability. This utilization will lead to better urban planning and use of resources.

2.4.11.Social network theory

Social Network theory examines the social interactions between people, communities and organisations. It looks into various aspects like communication, friendship or resource exchanges. Three main lines of research led to the development of these theories namely sociometric analysis, interpersonal relations and anthropology relations. This framework did not develop till the 1960s and advancements took place with the help of old research studies. In terms of digitalization, it highlights the importance of partnerships and collaborations which can lead to the adoption of a more circular model. The usage of digital tools can also lead to the sharing of ideas and co-creation among people creating a sense of belongingness with one another.

3. RESEARCH METHODS

3.1. Research Instruments

In this inquiry project, research will be conducted through a survey, to be distributed among students aged 20 and above, sustainability-related venture employees and ESG dept employees of various companies. The intention is to try and collect results from a broad cross-section of people and gain valuable insight into variations of perceptions. The survey will comprise a mixture of quantitative and qualitative questions. The quantitative questions will utilise a Likert scale to apply nominal values to responses gauging agreeableness, allowing for a clearer analysis of the data. Each question which requires the respondents to answer on a Likert scale about their perception (i.e., non-definitive questions, such as age) will include a midpoint. This research project will be conducted as a short-term, ‘action research’ (Gray, 2020), this methodology is most suitable as the results are intended to offer action points and promote organisational change if potential improvements are found. While a long-term study could offer an analysis of any change implications, the available timeframe of this project does not allow for this. However, using a short-term, cross-sectional study enables research insight from a broad range of participants.

The results of the survey will first use the analytical tools built into MS Forms, additional analysis will be undertaken by various means, looking to identify correlations between responses. Significant findings will be represented graphically for ease of interpretation in the final report, with discussion drawing attention to them.

3.2. Data Collection

The data collection schedule includes two weeks starting from mid-August, the window has been shortened due to time constraints. To collect data from consumers for this research, questionnaires were distributed online through social media platforms like LinkedIn, X (formally Twitter) etc. and interviews with 5 professionals from sustainability-related industries were conducted. The final questionnaire consisted of eighteen questions in addition to the demographic information. There were 3 questions to collect demographic data like field of occupation, designation, and work experience. A 10-point Likert scale was used,1 being Very Strongly Disagree and 10 being Very Strongly Agree. For analysis, 52 responses were collected and improper responses and outliers were discarded. The demographic results showed that 49% of respondents were students and the rest 51% respondents were middle-level employees. The SmartPLS 4.0 software was used for analysis and structural equational modelling was followed. Structural equational modelling (SEM) is a statistical technique used to analyze between variables using hypothesis testing. SEM was selected as it tests multiple relationships while accounting for measurement error and it can handle latent variables.

3.3. Theoretical Framework

The author has created a theoretical framework using different variables to establish a relationship between them using theories to prove the hypotheses.

Variables:

Independent variable-

1. Collaboration & partnership

2. Optimum waste management

3. Product lifecycle

4. Sustainable design thinking

5. Consumer acceptance

6. Regulatory framework

Sub variables-

1. Silo effect

2. Cloud sharing

3. Shelf life tracking

4. Remanufacturing

5. Automation

6.IOTs (internet of things)

7. Infrastructure for innovation

8. Updated cloud data

9. Greenwashing

10. Price Sensitivity

11. Artificial Intelligence

12. Cloud computing

Dependent variable – Adoption of circular economy in manufacturing

Mediating variable – Data science

Moderating variable- Organizational culture

Outcome variable-

1. Economic benefit

2. Enhanced product lifecycle

3. Extended use of waste material

4. Sustainable living standards

5. Sustainable economic innovation

Hypothesises-

H1:Collaboration and partnership help data science for the adoption of circular economy in manufacturing

H2:Consumer Acceptance helps data science for the adoption of circular economy in manufacturing

H3:Data Science in Adoption of circular economy in manufacturing positively impacts economic benefit

H4: Data Science in Adoption of circular economy in manufacturing positively impacts enhanced product lifecycle

H5: Data Science in the Adoption of circular economy in manufacturing positively impacts the extended use of waste material

H6: Data Science in Adoption of circular economy in manufacturing positively impacts sustainable economic innovation

H7: Data Science in Adoption of circular economy in manufacturing positively impacts sustainable living standards

H8: Optimum waste management helps data science for the adoption of a circular economy in manufacturing

H9:Product lifecycle helps data science for the adoption of circular economy in manufacturing

H10: Regulatory framework helps data science for the adoption of circular economy in manufacturing

H11: Sustainable design thinking helps data science for the adoption of circular economy in manufacturing

4. Research Design and Methods

4.1. Project Design

The inquiry will be conducted primarily by a questionnaire distributed to target respondents. This will be distributed, utilising various social media means. The aim of this is to allow for a diverse range of respondents. The decision to distribute the survey should remove noncoverage errors, though there is a risk of nonresponse errors occurring (Sills & Song, 2002). Further to allowing this first level of privacy where possible an option for randomised responses will be utilised to combat evasive answer bias (Warner, 1965).

Distributing the survey online presents a risk of non-response, with electronic surveys shown to record lower response rates than physical means of distribution (Nayak & Narayan, 2019). (Dilman, et al. 2009) showed that the highest response rates were to incentivised surveys, it was decided that offering a monetary incentive to complete the survey was not suitable for this inquiry project.

Much of the survey will provide quantitative data to analyse, through the use of Likert scale questions. Using a Likert scale allows the conversion of ordinal data to interval data, either directly or by use of composite scoring (sums and averages) (Willits, Theodore, & Luloff, 2016), this allows quick calculation of average responses and simple plotting of correlated data sets to graphs. Care will be taken to consider missing information when calculating averages. When analysing Likert scale responses mean values will be used to show central tendencies, but these may produce an excess of neutral results, if this is the case standard deviations can be used to express the variability of results (Boone & Boone, 2012). Other than that P values, Cronbach alpha values and AVE will be used for reliability and validity Testing along with hypothesis testing. The structure of interviews defines the qualitative analysis. Semi-structured interviews have been used and for analysis, the outcomes of each interview have been divided into common, contradicting and different views and later these are used along with the quantitative results to prove the hypotheses.

Multiple response methods are being used to produce a more accurate picture of people’s beliefs and feelings, offering holistic and demonstrable action points. Having to analyse and synthesise the findings of both quantitative and qualitative data will take longer, however, the benefit of a more complete insight far outweighs the additional time and effort (Shah & Corley, 2006).

4.2. Timescale and Resources

It is intended that this research inquiry be completed in 10 weeks, with the final report available at the end of August 2022.

The literature review was conducted in the first week, with findings used to inform the survey structure and lines of question. The aim is for all approvals and data collection schedules to be completed by the fourth week to maximise the available time for analysis. The final week of the inquiry project will be composing a full and final report document, with review and scrutiny of results. This review and confirmation of result validity will be the final stage of the research to verify the accurate reporting of the findings

It is not expected that any additional resources will be required, other than the time of respondents to complete the survey, this is expected to not be any more than 10 minutes per respondent. Initial response numbers were considered insufficient for analysis, which is why there was a need for semi-structured interviews to aid in the quality of results through qualitative analysis. Analysis has been completed as a sole endeavour, providing a singular perspective on the results of the survey and reducing the chance of compromising the anonymity of responses.

5. Data Analysis

The first four questions in the survey were the only mandatory responses in the survey, as they and the participant information sheet formed the consent for respondents, as such all completed surveys returned yes responses. The demographic shows 49% of respondents as students and 51% as middle-level employees.

Next questions were composed using a Likert scale, all with responses 1-10, the wording associated with the high and low scores was tailored to suit each question, maintaining like phrases where possible. The common approach of a ten-point scale was maintained for continuity and fair comparison and to ensure the tangible validity of results.

Regarding all the survey questions we set the positive benchmark as 8 out of 10 and the lowest average achieved was 8.54 out of 10 which strongly proved the likeliness of the proposed hypothesis. With an overall average of 9.05 out of 10, there was no problem in proving the likeliness of the hypothesises. The hypothesis whose likeliness is proved in the weakest was consumer acceptance helps data science for the adoption of circular economy in manufacturing. It had 17.30% of responses did not support the claim or hypothesis. This was the highest percentage for a non-supporting population in the whole analysis but as we can see the values are still acceptable as the likeliness score percentage is above 80 I.e.82%.

Now to strengthen the claims and hypotheses, the author used SmartPLS4.0 to establish relationships and get deeper analytic statistics.

The reliability was assessed by using Cronbach’s alpha value as shown below. The respective values of Cronbach alpha for all the variables were much higher than 0.7 which led to considerable acceptance of all the values. For determining model reliability, Cronbach’s alpha was used where all the values were greater than 0.7 respectively. (Cronbach, 1951).

Next, the validity was tested using AVE, where the acceptable value was set to AVE > 0.5 and all the values were found to be acceptable as shown below, thus proving the validity of the data.

Now to strongly prove all the hypothesises, the P value was taken into consideration. The hypothesis was tested through a P-test (Armstrong & Overton, 1977) where respondents with missing responses were considered non-respondents. (Kam & Meyer, 2015)It was found that all the P values were below 0.05 and hence all were strongly supported and proved all the hypotheses true.

5.1.Findings

H1:Collaboration and partnership help data science for the adoption of circular economy in manufacturing was highly supported ( p <0.05)

Transition to circular economies is often troublesome due to differences in objectives. The majority of the time, stakeholders face many conflicts due to each other’s different priorities and objectives. Accordingly, cooperation is regarded as a crucial key for unlocking everything that circular economies have to offer a business. Transition mostly stops due to overlying costs of implementation that stakeholders do not want to face They fail to see the broad picture or the betterment of the planet. Partnerships come into play here. Also, public government partnerships can contribute to the collection of waste and recycling rates. Overall, valuable partnerships and collaborations with the government and private institutions with the help of data science can lead to easier adoption of circular economy in manufacturing.

H2:Consumer Acceptance helps data science for the adoption of circular economy in manufacturing was highly supported ( p <0.05) 

It has been highlighted that low consumer acceptance is a major boon in case one wants to scale one’s business models (CBMs) towards circular economies. Digitalisation urges innovative ways to exchange and share products and certain services. Technologies such as big data and IoT keep inventory checks and aid in utilising underused products and assets. Furthermore, organisations leverage these technologies such as digital platforms, and augmented reality for collecting advanced product imagery to address obstacles with security, convenience, trust and hygiene. They increase consumer acceptance with increased transparency and remove quality and authenticity concerns. Thus, waste reduction and up-cycling of by-products are made possible by an increment in consumer understanding of circular economy principles. Consumers can be made more aware through trust-building, addressing potential risks, and spreading awareness.

H3:Data Science in the Adoption of circular economy in manufacturing positively impacts economic benefit was highly supported ( p <0.05)

Manufacturers can save operational costs by reducing waste, optimising resource utilisation, and identifying inefficiencies by analysing massive datasets. Predictive maintenance and lifecycle analysis are further made possible by data science, which increases product longevity and lowers the requirement for raw materials. In addition to reducing expenses, this also creates new opportunities for profit through remanufacturing, recycling, and cutting-edge business concepts like product-as-a-service. All things considered, applying data science to circular economy strategies increases revenue while advancing sustainability.

H4: Data Science in the Adoption of circular economy in manufacturing positively impacts enhanced product lifecycle was highly supported ( p <0.05)

Manufacturers may create more robust and repairable products by using advanced analytics to obtain insights into product usage, wear, and failure trends. With the use of this data-driven method, a product’s lifecycle can be precisely monitored, allowing for timely maintenance and upgrades that increase the product’s lifespan. Additionally, data science ensures that resources are continuously reused, enabling the circular economy, by optimising material recycling and repurposing at the end of a product’s lifecycle.

H5: Data Science in the Adoption of circular economy in manufacturing positively impacts extended use of waste material was highly supported ( p <0.05)

Data science finds ways to recover and reuse materials that would otherwise be thrown away by examining waste streams and material flows. This lowers the requirement for virgin resources by enabling producers to turn trash into useful inputs for new goods. Additionally, by ensuring that materials are effectively sorted, processed, and reintegrated into production cycles, predictive analytics aids in the optimisation of the recycling process. This expanded utilisation of waste not only promotes sustainability but also lowers expenses and improves manufacturing resource efficiency.

H6: Data Science in the Adoption of circular economy in manufacturing positively impacts sustainable economic innovation was highly supported ( p <0.05)

We can develop novel approaches to creating goods that are intrinsically sustainable—designed for disassembly, repair, and reuse—by examining patterns of material consumption, waste production, and product lifecycles. Additionally, data science opens the door for creative economic models that reduce waste and repurpose it as a resource, like closed-loop systems and circular supply chains. With this strategy, sustainability becomes a source of economic value and changes the way we see consumption and production. We can test and scale these advances with data science’s rigorous analysis and iterative process, guaranteeing that environmental stewardship and economic growth are mutually reinforcing rather than conflicting.

H7: Data Science in the Adoption of circular economy in manufacturing positively impacts sustainable living standards was highly supported ( p <0.05)

Manufacturers may use data-driven insights to optimise resource utilisation, reduce waste, and build products that are more durable, efficient, and environmentally friendly. This transition not only minimises the environmental footprint but also makes sustainable products more accessible and inexpensive, hence increasing consumers’ quality of life. Furthermore, data science enables the creation of innovative business models that prioritise sustainability, such as product leasing or take-back programs, which help to integrate circular economy ideas into everyday life. As a result, integrating data science into manufacturing contributes to a more sustainable future in which economic growth and higher living standards coexist with environmental care.

H8: Optimum waste management helps data science for adoption of circular economy in manufacturing was highly supported ( p <0.05)

The implementation of the circular economy in manufacturing can be supported by data science to a greater extent through optimal waste management. Data science can track and analyse waste streams, spot trends, and locate opportunities for resource recovery and recycling by managing trash effectively. In line with the circular economy’s tenets, producers can minimise waste, recycle materials, and develop closed-loop systems with the aid of this data-driven strategy. In the end, efficient waste management offers the information and understanding needed to create more environmentally friendly production processes that also maximise resource efficiency.

H9:Product lifecycle helps data science for the adoption of circular economy in manufacturing was highly supported ( p <0.05)

The product life cycle emphasises sustainable practices and optimization of resources. It directly influences the adoption of a circular economy. The alteration from traditional linear models to circular economy practices comprises reusing, repairing, upgrading and recycling goods throughout their life cycle to mitigate waste generation, consumption of resources and impact on the environment. Additionally, including circularity indicators by using IOTs and cloud computing throughout the design, manufacturing, distribution and use phases of products is important for improving and measuring circular economy performance in sectors like electrical and electronic manufacturing with data science elements.

H10: Regulatory framework helps data science for adoption of circular economy in manufacturing was highly supported ( p <0.05)

Regulatory Framework is an important factor in the implementation of the circular economy. The adoption of a circular economy is driven by Government policies and initiatives in various industries. The European Union has also adopted regulations and guidelines, such as the Circular Economy Package, to encourage circularity in various industries and provide a legal framework that facilitates the shift towards sustainability. Furthermore, the implementation of guidelines such as the Regulation of Taxonomy for sustainable operations aims to establish a standard language for investors, providing incentives for environmentally beneficial initiatives and encouraging businesses to implement circular economy principles. These regulatory measures would not only help in guiding companies but also provide incentives to drive towards a more sustainable future and with cloud structures and AI, it will be easy to track the implementation of regulatory frameworks.

H11: Sustainable design thinking helps data science for adoption of circular economy in manufacturing was highly supported ( p <0.05)

Sustainable design thinking integrates circular economy principles with data science to transform manufacturing processes. By using data-driven insights, manufacturers can identify inefficiencies, optimize resource use, and minimize waste. Sustainable design thinking guides these efforts, ensuring that solutions not only improve operational efficiency but also align with circular economy goals, such as reducing waste, promoting material reuse, and extending product life cycles. This synergy between data science and sustainable design creates a more resilient, eco-friendly manufacturing system

6. Conclusions

6.1. Introduction

In this concluding chapter the findings will be summarised according to the research objectives, followed by a short section on the contribution to knowledge. Then the limitations of the study will be discussed. Finally, recommendations with implications for theory, further research and practice are presented.

6.2. Summary of findings

This study aimed to achieve the following:

1. To gain insights and explore the scope of the circular economy.
2. To identify and investigate how data science impacts the circular economy.
3. To establish a relationship between data science and circular economy in manufacturing
4. To observe the influence of data science in circular economy in manufacturing
5. To gain insights and explore the scope of the circular economy.

We concluded how important and relevant circular economy is as by shifting from a traditional linear model of production and consumption to a circular approach, businesses and societies can reduce waste, optimize resource use, and create new economic opportunities. This model emphasizes the importance of designing for longevity, repairability, and recyclability, ultimately contributing to environmental conservation and economic resilience. As we continue to investigate and implement circular economy principles, it becomes increasingly evident that such strategies are essential for addressing the pressing challenges of resource depletion and climate change.

1. To identify and investigate how data science impacts the circular economy

In conclusion, identifying and investigating how data science impacts the circular economy revealed its pivotal role in enhancing resource efficiency and sustainability. Data science enables more precise tracking and analysis of resource flows, product lifecycles, and waste management processes, leading to optimized circular practices. By leveraging advanced analytics, machine learning, and predictive modelling, businesses can improve decision-making, forecast demand, and identify opportunities for recycling and reuse. This data-driven approach not only supports the transition to a circular economy but also drives innovation and operational efficiency

1. To establish a relationship between data science and circular economy in manufacturing

In conclusion, we got to know how collaboration & partnership, Optimum waste management, product lifecycle, sustainable design thinking consumer acceptance and regulatory framework play a pivotal role in helping data science in the adoption of a circular economy in manufacturing.

1. To observe the influence of data science in circular economy in manufacturing

In conclusion, we got to know that data science in circular economy in manufacturing leads to economic benefit, enhanced product lifecycle, extended use of waste material, sustainable living standards and sustainable economic innovation

6.3. Limitations

As with any study, this research suffers from limitations. The most important ones are mentioned ahead

Firstly, the existing literature is sparse. Manufacturing-specific literature was hard to find. Available literature lacked narrowing down the topic of the circular economy into the manufacturing sector and then linking it up with data science. Additionally, the small sample size might have biased the results due to the fact, that the data collection through simple random sampling failed to achieve the desired sample size of 384 respondents (Biggam, 2017, 171 and Saunders et al., 2019, 302) during the data collection period. Instead, only 52 cases could be collected and analyzed. Hence, the sample might not reflect the total population. A small sample size also increases the error margin resulting in less precise conclusions on the population.

6.4. Recommendation

The findings and limitations of this study have been used to develop recommendations with implications for theory, further research and practice.

6.4.1. Implications for theory and further research

1. Increase the sample size in further research

A larger sample size of 384 respondents is recommended for further research. This ensures the representativeness of the evidence gathered regarding the population as a confidence level of 95% is reached, subsequently reducing the error margin to 5% (Biggam, 2017, 171 Saunders et al., 2019, 302).

1. Extend the data collection period in further research

In the current study, the data collection was limited to two weeks, partly in the summer holiday period. An extension to a period of one month would reduce the non-response rate due to the absence or busy schedule of the respondents.

REFERENCES

Bressanelli, G., Saccani, N., & Perona, M. (2022, January 6). Investigating Business Potential and Users’ Acceptance of Circular Economy: A Survey and an Evaluation Model. Sustainability. https://doi.org/10.3390/su14020609

Camacho-Otero, J., Boks, C., & Pettersen, I. N. (2018, August 4). Consumption in the Circular Economy: A Literature Review. Sustainability. https://doi.org/10.3390/su10082758

Kulviwat, S., Bruner, G. C., Kumar, A., Nasco, S. A., & Clark, T. (2007, November 7). Toward a unified theory of consumer acceptance technology. Psychology & Marketing. https://doi.org/10.1002/mar.20196

Marikyan, D., Papagiannidis, S., & Alamanos, E. (2020, July 25). Cognitive Dissonance in Technology Adoption: A Study of Smart Home Users. Information Systems Frontiers. https://doi.org/10.1007/s10796-020-10042-3

Wen, H., Lee, C., & Song, Z. (2021, May 20). Digitalization and environment: how does ICT affect enterprise environmental performance? Environmental Science and Pollution Research. https://doi.org/10.1007/s11356-021-14474-5

The Digital Twin Concept in Industry – A Review and Systematization. (2020, September 1). IEEE Conference Publication | IEEE Xplore. https://ieeexplore.ieee.org/abstract/document/9212089

Fleischmann, M., Bloemhof-Ruwaard, J. M., & Dekker, R. (2000). Logistics Impact on Inventory Management. International Journal of Physical Distribution & Logistics Management, 30(3/4), 282-305.

Tesla. (2022). Tesla, Inc. 2021 Impact Report.

IKEA. (2022). IKEA Sustainability Report FY21.

Womack, J. P., Jones, D. T., & Roos, D. (1990). The Machine That Changed the World: The Story of Lean Production. Simon and Schuster.

Harry, Schroeder, R. (2000). Six Sigma: The Breakthrough Management Strategy Revolutionizing the World’s Top Corporations. Currency.

Signify. (2021). Sustainability Annual Report 2020.

Picot, A., Reichwald, R., & Wigand, R. T. (2008). Information, Organization and Management: Expanding Markets and Corporate Boundaries. Springer Science & Business Media.

Top of Form

European Commission. (2020). A new Circular Economy Action Plan for a cleaner and more competitive Europe. Retrieved from https://ec.europa.eu/environment/circular-economy/pdf/new_circular_economy_action_plan.pdf

European Parliament. (2019). EU legislation is in place for the circular economy. Retrieved from https://www.europarl.europa.eu/RegData/etudes/IDAN/2019/631060/EPRS_IDA(2019)631060_EN.pdf

Philips. (n.d.). Digital Twin Technology for Healthcare. Retrieved from https://www.philips.com/a-w/research/connected-care-and-health-solutions.html

Walter Fraanje a b, a, b, AbstractCollaborative consumption indicates the emergence and rapid spread of a new set of consumption practices. Originally praised as an antidote to an unsustainable market economy, Belk, R., Cheng, M., … MyWheels, n. d. (2018). What future for collaborative consumption? A practice theoretical account. Retrieved from https://www.sciencedirect.com/science/article/abs/pii/S0959652618329305

Data Silos, why they’re a problem, & how to fix it. (n.d.). Retrieved from https://www.talend.com/resources/what-are-data-silos/

T, K. (2023a). What does a memorandum of understanding accomplish? Retrieved from https://vakilsearch.com/blog/what-does-memorandum-of-understanding-accomplish/

Vadén, T., Lähde, V., Majava, A., Järvensivu, P., Toivanen, T., Hakala, E., & Eronen, J. T. (2020, October 1). Decoupling for ecological sustainability: A categorisation and review of research literature. Environmental Science & Policy. https://doi.org/10.1016/j.envsci.2020.06.016

Figure 1. Relative and absolute decoupling (modified from UNEP, 2011). (n.d.). ResearchGate. https://www.researchgate.net/figure/Relative-and-absolute-decoupling-modified-from-UNEP-2011_fig1_331234240

Resource efficiency and the circular economy: concepts, economic benefits, barriers, and policies – UCL Discovery. (n.d.). https://discovery.ucl.ac.uk/id/eprint/10054117/

Van Ewijk, S. (2018, January 1). Resource efficiency and the circular economy: Concepts, economic benefits, barriers, and policies. ResearchGate. https://www.researchgate.net/publication/327868697_Resource_efficiency_and_the_circular_economy_Concepts_economic_benefits_barriers_and_policies

Winkler, H. (2011, January 1). Closed-loop production systems—A sustainable supply chain approach. CIRP Journal of Manufacturing Science and Technology. https://doi.org/10.1016/j.cirpj.2011.05.001

Chen, J. (2023, January 30). Collaborative Consumption: What it is, How it Works. Investopedia. https://www.investopedia.com/terms/c/collaborative-consumption.asp

National Environment Agency. (2020). Singapore Circular Economy Roadmap. Retrieved from https://www.nea.gov.sg/programmes-grants/circular-economy/singapore-circular-economy-roadmap

Wasserman, S., & Faust, K. (1994). Social Network Analysis: Methods and Applications. Cambridge University Press.

Camacho-Otero, J., Tunn, V., Chamberlin, L., & Boks, C. (2020, December 15). Consumers in the circular economy. Edward Elgar Publishing eBooks. https://doi.org/10.4337/9781788972727.00014

Partridge, E. (2014, January 1). Social Sustainability. Springer eBooks. https://doi.org/10.1007/978-94-007-0753-5_2790

Milios, L. (2021, January 21). Towards a Circular Economy Taxation Framework: Expectations and Challenges of Implementation. Circular Economy and Sustainability, 1(2), 477–498. https://doi.org/10.1007/s43615-020-00002-z

Kirchherr, J., Yang, N. H. N., Schulze-Spüntrup, F., Heerink, M. J., & Hartley, K. (2023, July 1). Conceptualizing the Circular Economy (Revisited): An Analysis of 221 Definitions. Resources, Conservation and Recycling. https://doi.org/10.1016/j.resconrec.2023.107001

Morseletto, P. (2020, February 1). Targets for a circular economy. Resources, Conservation and Recycling. https://doi.org/10.1016/j.resconrec.2019.104553

Kirchherr, J., Reike, D., & Hekkert, M. P. (2017, December 1). Conceptualizing the circular economy: An analysis of 114 definitions. Resources, Conservation and Recycling. https://doi.org/10.1016/j.resconrec.2017.09.005

A Model for the Transition to the Circular Economy: The “R” Framework. Symphony, doi: 10.4468/2022.1.08fornasari.neri

A., Dwivedi., Sanjoy, K., Paul. (2022). A framework for digital supply chains in the era of circular economy: Implications on environmental sustainability. Business Strategy and The Environment, doi: 10.1002/bse.2953

Jinghua, Liu., Muhammad, Umer, Quddoos., Muhammad, Hanif, Akhtar., Muhammad, Sajid, Amin., Muhammad, Junaid, Tariq., Arij, Lamar. (2022). Digital technologies and circular economy in supply chain management: in the era of COVID-19 pandemic. Operations Management Research, doi: 10.1007/s12063-021-00227-7

Pietro, De, Giovanni. (2023). The Impact of Digital Technologies and Sustainable Practices on Circular Supply Chain Management. Logistics doi: 10.3390/logistics7010001

Angelina, Pavlović., Snežana, Nestić., Goran, Bošković. (2021). Circular economy management in business organizations using digital technologies. doi: 10.5937/SJEM2101022P

Fiona, Charnley., Fabienne, Knecht., Helge, Muenkel., Diana, Pletosu., Victoria, A., Rickard., Chiara, Sambonet., Martina, Schneider., Chunli, Zhang. (2022). Can Digital Technologies Increase Consumer Acceptance of Circular Business Models? The Case of Second Hand Fashion. Sustainability, doi: 10.3390/su14084589

Petros, Spachos., Weiwei, Li., Mark, Chignell., Alberto, Leon-Garcia., Leon, Zucherman., Jie, Jiang. (2015). Acceptability and Quality of Experience in over-the-top video. doi: 10.1109/ICCW.2015.7247424

Boriana, Rukanova., Jolien, Ubacht., Ben, Turner., Yao-Hua, Tan., Jonathan, Schmid., E., Rietveld., Wout, Hofman. (2023). A Framework for Understanding Circular Economy Monitoring: Insights from the Automotive Industry. doi: 10.1145/3598469.3598530

Tommaso, Fornasari., Neri, Paolo. (2022).

Fiona, Charnley., Fabienne, Knecht., Helge, Muenkel., Diana, Pletosu., Victoria, A., Rickard., Chiara, Sambonet., Martina, Schneider., Chunli, Zhang. (2022). Can Digital Technologies Increase Consumer Acceptance of Circular Business Models? The Case of Second Hand Fashion. Sustainability, doi: 10.3390/su14084589

Miroslav, Jurkovič. (2022). How to Increase Consumer Acceptance of the Transition to Circular Economy. Conference proceedings, doi: 10.53465/edamba.2021.9788022549301.226-233

Anindita, Prabawati., Evi, Frimawaty., Joko, Tri, Haryanto. (2023). Strengthening Stakeholder Partnership in Plastics Waste Management Based on Circular Economy Paradigm. Sustainability, doi: 10.3390/su15054278

P., Giovani, Palafox-Alcantar., Dexter, V., L., Hunt., Christopher, D., F., Rogers. (2021). Current and future professional insights on cooperation towards circular economy adoption. Sustainability, doi: 10.3390/SU131810436

Sepani, Senaratne., KC, Abhishek., Srinath, Perera., Laura, M., M., C., E., Almeida. (2021). Promoting stakeholder collaboration in adopting circular economy principles for sustainable construction. doi: 10.31705/WCS.2021.41

Foivos, Psarommatis., Gökan, May. (2022). Achieving Global Sustainability Through Sustainable Product Life Cycle. doi: 10.1007/978-3-031-16407-1_46

Olga, Timofei. (2023). Circular business models for increasing product usage. doi: 10.53486/cike2022.08

Kam, C. C. S., & Meyer, J. P. (2015). How careless responding and acquiescence response bias can influence construct dimensionality: The case of job satisfaction.

Armstrong, J. S., & Overton, T. S. (1977). Estimating nonresponse bias in mail surveys. Journal of Marketing Research, 14(3), 396–402.

B., M., Hapuwatte.. (2023). Optimizing Product Life Cycle Systems for Manufacturing in a Circular Economy. doi: 10.1007/978-3-031-28839-5_47

Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. https://doi.org/10.1007/BF02310555.

Patwa, N., Sivarajah, U., Seetharaman, A., Sarkar, S., Maiti, K., & Hingorani, K. (2021, January 1). Towards a circular economy: An emerging economies context. Journal of Business Research. https://doi.org/10.1016/j.jbusres.2020.05.015

About the author

Ritam Bhandari

Ritam is a motivated individual who is really interested in the field of sustainability.

The post Role of Data Science in Adoption of Circular Economy in Manufacturing appeared first on Exploratio Journal.

Abstract

This study establishes the relationship between the water requirements of crops and reservoir management in one of the most fertile regions of the world—Kandi, India. Utilizing simulations built in a computing notebook, we analyze the water table dynamics for crops, specifically maize and black gram, alongside the depth of a canal feeding a reservoir over time. The findings highlight the temporal changes in water availability and the impact of drought on agricultural sustainability. Graphs generated through the analysis effectively illustrate these dynamics, providing insights into water management strategies that can enhance crop resilience and optimize resource utilization in environments prone to depleting water resources.

Introduction

Globally, there is evidence of the effects of climate change on agriculture; however, nations such as India are particularly vulnerable because of their large agricultural population, overuse of natural resources, and inadequate adaptation strategies. India’s warming trend over the last century has shown a 0.60°C increase. Food security may be impacted as a result of the anticipated effects, which are likely to exacerbate variations in many crops. In certain regions of India, the output of wheat and pulses has already been shown to be negatively impacted by rising temperatures, higher water stress, and fewer rainy days. Such dangers pose a serious risk to India’s agriculture-based economy, which is why irrigation and water conservation are essential.

Although maize is the main crop in many areas due to its resilience, it is susceptible to changes in water supply and temperature. Higher temperatures have been linked to poorer yields, especially during the crucial flowering and grain-filling stages. Similar difficulties affect black gram, an important pulse crop; these include rising drought conditions and rainfall variability, which can reduce productivity and jeopardise the food security of vulnerable populations.

The negative impacts of climate change go beyond lower yields. For example, higher temperatures and humidity aggravate the prevalence of pests and diseases that affect both black gram and maize. Furthermore, the long-term sustainability of these crops is seriously threatened by nutrient depletion and soil degradation brought on by unpredictable weather patterns. Considering how vital maize and black gram are to the local diet and how they sustain farmers’ livelihoods, creating resourceful irrigation and water management techniques is crucial to preserving agricultural resilience in the face of these difficulties.

Area of study

Kandi, situated in the Punjab and Jammu Kashmir regions of India, is one of the country’s most fertile areas, producing 573.19 lakh tonnes of maize and black gram in the 2022-2023 crop year. However, this region is facing a severe decline in water resources due to climate change and the over-extraction of groundwater.

Climate

The region’s climate ranges from semi-arid to sub-humid. The highest average temperature, reaching 41 °C, is recorded during the first two weeks of June, while the lowest average temperature, at 6 °C, occurs in January.The region receives average annual rainfall of 800-1500 mm with a very high coefficient of variation.

Soils

The soils in the region, primarily loamy sand to sandy loam, exhibit low to medium moisture retention and are highly erodible, with inherent fertility being very low and limited organic carbon content.

Crops and cropping systems

Maize is the dominant crop in the Kandi area, covering around 46,000 hectares, along with key kharif crops like pearl millet and black gram. The region also grows major rabi crops such as wheat and lentil, and supports diverse fruit production.

Problems in the agriculture

Erratic Rainfall: The region experiences annual rainfall of 800-1500 mm, with about 80% occurring during the kharif season (July-September). Rainfall is highly variable, often leading to droughts, especially during sowing. A significant chance (55-98%) of dry spells exceeding six days exists monthly, and pre-monsoon showers in June are unpredictable, delaying kharif sowing and impacting yields. Early monsoon withdrawal can also lead to severe droughts, complicating the establishment of subsequent rabi crops due to inadequate moisture.

Soil and Water Erosion: The area is characterized by hilly regions, seasonal streams (choes), and cultivated zones, with gullies and rills commonly present. The low organic carbon content (<0.04%) makes the soil highly dispersible and erodible, causing 30-40% of rainfall to result in runoff and water erosion.

Model Development and Methodology

The research aims to design an irrigation model for the study area, seen in Figure 1, that utilizes real-time data to optimize water application. By integrating soil moisture sensors and local terrain features with weather data, this approach seeks to revolutionize irrigation practices.

Employing a combination of remote sensing, data analysis, and modelling techniques, the project creates a model that assesses soil moisture, weather patterns, and crop water requirements to generate optimal irrigation schedules. The goal is to significantly reduce water consumption while enhancing crop yields.

Irrigation Model Software

The irrigation model was developed and executed using Python, utilizing libraries like NumPy for mathematical calculations and Matplotlib for visualizing water table dynamics. Additionally, ArcGIS was employed to map the terrain and analyse spatial data. This combination of tools allowed for accurate simulation of water dynamics and crop irrigation needs, helping to optimize water usage for maize and black gram crops.

Graphs and Results

Drought Period (Hour 50 to Hour 120): The water table graph for maize and black gram matches the highlighted yellow area seen in Figure 3, between hours 50 and 120, which plainly shows a protracted period without rainfall. Water stress for both crops is exacerbated at this time due to the decline in water table levels caused by the absence of rainfall.

Post-Drought Rainfall Spike: There is a discernible increase in rainfall after the drought, indicating an abrupt entry of water to the area. The low water table readings in the following graphs, however, suggest that this rainfall may not be enough to restore the ideal water levels for agricultural development given the earlier depletion of the water table.

Correlation with Crop Water Usage: The period of higher water consumption seen in both black gram and maize coincides with the absence of rainfall during the hours highlighted. This emphasises the necessity of efficient water management plans to lessen the effects of these dry spells and guarantee that crops continue to be resilient even in the face of infrequent rainfall.

Initial Water Availability: Over the course of the first two days, maize’s water table levels constantly exceeded those of black gram, suggesting that during its early growth phase, maize needs more water.

Drought Impact Period (Days 2-5): The yellow shaded area represents the dramatic decline in water table levels that both crops experienced between

Days 2 and 5. This indicates a time of heightened drought or water stress that is affecting the availability of water for both crops. Interestingly, maize maintained a little higher water table than black gram even throughout this dry period, highlighting its higher water requirement.

Post-Drought Water Levels: After day 5, there was a noticeable decline in the amount of water available, with the water table of maize falling below that of black gram. This suggests that the crop might have been more affected by the prolonged drought or that the water resources may have been rapidly depleted due to maize’s water requirements.

Implications for Irrigation Management: The general downward trend in both crops’ water table levels highlights the necessity of implementing focused irrigation techniques to promote crop growth, particularly in times of drought. By comprehending these dynamics, water allocation may be optimised to guarantee that both crops receive enough moisture for growth.

Initial Inflow Phase: The increase in reservoir level in the first two days of the simulation is indicative of an efficient canal inflow of water, which is essential for supplying maize and black gram with the irrigation they require in their early growth stages (see Figure 4).

Stability Period: Consistent soil moisture levels are ensured between days two and five due to the relative stability in depth, which is essential for improving crop health and yield potential.

Peak Depth Observations: The peak depth attained by day 7 indicates a sufficient supply of water, offering chances for well-timed irrigation scheduling that coincides with crucial stages of crop growth.

Recommendations for Adaptive Management: Constant monitoring of reservoir levels is necessary to make real-time adjustments to irrigation techniques, which maximises water allocation and reduces crop stress, particularly in advance of dry spells.

Agricultural Sustainability Consequences: Reservoir dynamics insights emphasize the significance of efficient management of water resources, which can boost crop resilience, maximize resource use, and increase food security in the face of climate instability.

A reservoir of water supports irrigation as designed during the simulation (see Figure 5). The irrigation network has a visible effect on average water table level as seen in Figure 6.

Proposed Solutions for Enhanced Water Management

Effect of Drought on Water Table Levels

During the simulated dry spell (days 2 to 5), the data shows a significant drop in the water tables for black gram and maize. Maize exhibits a slower recovery of the water table post-drought compared to black gram, which can be attributed to its greater water needs, while black gram displays greater adaptability. This decrease demonstrates the substantial effect that protracted dry weather has on crop water availability, ultimately impacting crop growth and productivity.

Solution: It is critical to use efficient irrigation scheduling and water-saving measures, such as drip irrigation, to preserve soil moisture during these dry spells. Introducing crop types resistant to drought may also improve resistance to water stress.

Reservoir and Canal Management

The water level graphs show variations, with a noticeable period of stability throughout the drought (seen by the yellow-shaded area). This implies that reservoirs are essential for acting as water scarcity buffers.

Solution: To guarantee sufficient water storage for use during the dry seasons during the rainy season, better reservoir management techniques are necessary. Stable water levels may be maintained and sharp fluctuations in availability can be avoided with regular monitoring of canal inflow rates.

Rainfall Dependency and Variability

The rainfall data highlights the region’s dependence on erratic rainfall patterns that have a direct impact on crop water supplies. It also reveals a sharp rise in rainfall after a protracted dry spell (between hours 50 and 120).

Solution: By installing rainwater harvesting devices and enhancing soil water retention techniques, farmers may lessen their reliance on erratic rainfall and give their crops a more reliable supply of water during dry seasons.

Efficient Water Resource Utilization

A comparative study shows that maize tends to use up water resources more quickly than black gram, pointing to the latter’s higher water requirement and possible sustainability problems during dry spells.

Solution: Crop rotation techniques that combine water-intensive crops like maize with drought-tolerant crops like black gram, can maximise water consumption while preserving soil moisture.

Overall, these preliminary findings suggest that the proposed model could decrease water usage, offering a practical solution for areas like Kandi facing water scarcity. By simulating various rainfall scenarios and their effects on water availability, this model provides critical insights for conserving water and promoting sustainable agricultural practices.

Conclusion

The relationship between crop water demand, reservoir management, and water availability in India’s Kandi region is clarified by the research model. It illustrates how drought times have a major impact on crop water intake, especially for black gram and maize, with maize needing noticeably more water, by simulating water table dynamics. Although efficient reservoir management has been key to mitigating the effects of drought, stable crop yields remain elusive due to the unpredictability of rainfall patterns. Because of the model’s adaptability, it can be used to simulate different crops and produce customised insights about their water requirements and drought resilience.This is achieved by incorporating crop-specific coefficients and parameters, such as root water uptake rates, soil porosity, evapotranspiration rates, and crop growth stages, which allow the model to generate tailored insights for various crops.

The results emphasize the critical need for sustainable water management solutions, such as enhanced irrigation practices, rainwater collection, and the development of drought-tolerant crops, in order to increase agricultural resilience in this semi-arid area. The research shows a feasible way to improve water usage efficiency and guarantee continued agricultural productivity in areas vulnerable to water constraint by utilising these adaptive tactics.

References

• Alagh, Y. K. (2013). The Future of Indian Agriculture. National Book Trust, India.
• Government of Punjab, Irrigation Department. (2015). Kandi Master Plan: A Comprehensive Flood Management Plan for the Kandi Area. Retrieved from https://irrigation.punjab.gov.in/kandi-master-plan
• Indian Council of Agricultural Research (ICAR). (2021). Efficient Water Management in Agriculture – Challenges and Opportunities. Retrieved from https://icar.org.in
• Indian Meteorological Department (IMD). (2020). Rainfall and Temperature Data of Punjab (2010-2020). Retrieved from https://www.imdpune.gov.in
• Singh, N. P., & Verma, A. (2019). Socioeconomic impacts of climate change on agriculture in India. International Journal of Agricultural Economics, 5(2), 99-112.
• Kumar, R., Sidhu, P. S., & Sharma, B. D. (2005). Mineralogy of soils of the Kandi area in the Siwalik hills of semi-arid tract of India. Agropedology, 15(2), 40-50.
• Pathak, H., Nayak, A. K., Jena, M., Singh, O. N., Samal, P., & Sharma, S. G. (2020). Socioeconomic Impacts of Climate Change on Agriculture in India. Agricultural Research Journal, ICAR.
• Rathore, A. L., & Dabas, J. P. S. (2018). Rainwater Harvesting: A Sustainable Water Resource Management Strategy. Water Research Institute.

About the author

Johanna Nikhil

Johanna is a high school student passionate about pursuing computer science, with a focus on integrating technology with sustainability. She enjoys exploring how IoT can address environmental issues and has taken on projects that support underprivileged communities, including writing a coding book, organizing crowdfunding campaigns, and implementing irrigation solutions for farmers. In her free time, Johanna enjoys acting, cooking, cycling, and exploring new places.

The post Exploring Water Table Dynamics for Sustainable Crop Production: A Case Study of Maize and Black Gram in the Kandi Region appeared first on Exploratio Journal.