AI and GIS Based Climate Change Impact Analysis on Crop Production

This study analysed the relationship between climate variability, vegetation conditions, and soybean production in Maharashtra during 2015–2025 using an integrated AI and GIS-based approach. After standardizing and matching the datasets by district and year, 88 district-year observations from eight common districts were obtained. The analysis shows that soybean production increased over the study period, with the highest production and yield recorded in 2022, followed by some decline by 2025.   The climate analysis indicates considerable variation in Kharif rainfall across the study period, while annual average temperature showed a gradual increasing pattern. The correlation analysis found a weak positive relationship between rainfall and soybean yield (r ≈ 0.329), whereas the relationship between annual average temperature and yield was very weak (r ≈ 0.048). These results indicate that rainfall and temperature alone do not explain soybean yield variation and that other agricultural and environmental factors also contribute to production outcomes.   The NDVI analysis showed a strong positive association between seasonal vegetation conditions and soybean yield. The August–September NDVI showed the strongest relationship (r ≈ 0.940) among the analysed seasonal periods. However, these correlations are based on 11 annual observations and should be interpreted as associations rather than evidence of causation.   Machine-learning models were developed using Kharif rainfall, annual average temperature, August–September NDVI, and cultivated area as predictors. Using a chronological training period of 2015–2022 and testing period of 2023–2025, Linear Regression, Random Forest, and XGBoost were evaluated using MAE, RMSE, and R². The obtained results show that the models captured part of the variation in soybean yield, while substantial unexplained variation remained. This indicates the importance of incorporating additional factors such as soil characteristics, irrigation, agricultural practices, pests, and other climatic variables in future modelling.   The GIS component provides a spatial perspective by linking district-level climate, NDVI, soybean yield, and model-predicted yield information with geographical boundaries. The combined AI-GIS framework therefore provides a systematic way to examine both temporal changes and spatial differences in soybean production across the selected districts of Maharashtra.   Overall, the study demonstrates that combining climate data, NDVI, machine learning, and GIS can provide a useful framework for analysing soybean production variability. The findings should be interpreted within the limitations of the dataset, particularly the relatively small integrated sample of 88 observations and the limited number of common districts. Future studies can improve the analysis by incorporating longer and more detailed datasets, soil and irrigation information, additional remote-sensing indicators, and more extensive district-level GIS analysis

[1]    P. G. Athare, Dharam R. Singh, Nalini Ranjan Kumar, Girish Kumar Jha, P. Venkatesh. “A spatial assessment of agricultural vulnerability to climate change using multidimensional data in Maharashtra state of India.” 2024.   [2]       Sandeep Gupta. “AI-BASED SPATIO-TEMPORAL ANALYSIS FOR PREDICTING CLIMATE-RESILIENT CROP YIELDS IN INDIAN AGRICULTURAL SYSTEMS.” 2025. [3]    Shilpa Vatkar, Sujata Kulkarni. “Optimizing Crop Yield Predictions through Satellite Data Fusion and Machine Learning.” 2025.   [4]    Ketaki Ghawali, Abhishek Garg. “Comparative Analysis of Statistical, Machine Learning, and Transformer-based Deep Learning Models for Climate-driven Crop Yield Prediction in Maharashtra.” 2026.   [5]    P. P. Patil, J. Jadhav, S. Bagade, S. Apturkar, B. Gavit. “Climate Variability and Crop Land Degradation Assessment Using a Remote Sensing and GIS in the Nashik Division of Maharashtra, India.” 2026.   [6]     D. Golhar, S. Bagade, P.B. Pawar, S. S. Shinde, M. Neelima. “SEASONAL VARIABILITY AND SPATIO-TEMPORAL TRENDS OF NDVI FOR VEGETATION MONITORING IN SANGLI DISTRICT OF WESTERN MAHARASHTRA INDIA.” 2025.   [7]   Shilpa Vatkar, Sujata Kulkarni. “Monitoring Agricultural Stress Using Multi-Sensor Satellite Data and Machine Learning Techniques.” 2025. [8]     Vivek Bhosale, Tejas Nirmal, Aayush Bhakare, Prof. Manoj Kharde. “Impact Prediction of Climate Change on Crop Yield and It’s Solutions.” 2025.   [9]   Kiran Moraye, Aruna Pavate, Suyog Nikam, S. Thakkar. “Crop Yield Prediction Using Random Forest Algorithm for Major Cities in Maharashtra State.” 2021. [10]  Saurabh Kelkar, A. Kulkarni, K. Koteswara Rao. “Impact of Climate Variability and Change on Crop Production in Maharashtra, India.” 2020.