An explainable machine learning framework to analyze the influence of environmental drivers on vegetation condition: Applications in agricultural drought assessment

Articles in Press

Document Type : Research Paper

Authors

Department of Civil Engineering, Sharif University of Technology, Tehran, I. R. Iran

Abstract

Effective crop management depends on a clear understanding of the interactions among environmental variables and their combined effects on plant health. This study aims to evaluate the influence of environmental drivers—such as precipitation, air temperature, solar radiation, and aerosol optical depth (AOD)—on vegetation condition by integrating analysis of variance (ANOVA) within a multiple linear regression (MLR) framework and machine learning approach, SHAP (Shapley Additive exPlanations) on XGBoost model, using satellite and reanalysis data. Monthly mean values of multi-source remote sensing data were analyzed for the Shushtar sub-basin from 2001 to 2024, including ERA5-Land meteorological variables, AOD from Terra and Aqua satellites, and Landsat-derived NDVI used to compute the Vegetation Condition Index (VCI). The final results show that XGBoost outperformed MLR, achieving a higher accuracy (R² = 0.86, RMSE = 9.32) compared to MLR (R² = 0.29, RMSE = 19.06). The results of the MLR based on fitted coefficients indicate that volumetric soil water and precipitation exert the strongest linear influence on VCI, while wind speed shows non-significant effect. In addition, the ANOVA results identify AOD (F = 20.11, P < 0.01) and surface pressure (F = 18.19, P < 0.01) as the most statistically significant environmental drivers of VCI. Furthermore, SHAP-based result reveals that solar radiation (mean |SHAP| ≈ 3.0) and air temperature (mean |SHAP| ≈ 1.6) contribute most significantly to the XGBoost model’s predictions. Overall, this framework enables interpretable quantification of complex and nonlinear relationships on vegetation condition, supporting data-driven agricultural monitoring and management.

Graphical Abstract

An explainable machine learning framework to analyze the influence of environmental drivers on vegetation condition: Applications in agricultural drought assessment

Keywords

References

Aboelghar, M., Arafat, S., Saleh, A., Naeem, S., Shirbeny, M., & Belal, A. (2010). Retrieving leaf area index from SPOT4 satellite data. The Egyptian Journal of Remote Sensing and Space Science, 13(2), 121-127. https://doi.org/10.1016/j.ejrs.2010.06.001
Anyamba, A., Tucker, C. J., & Eastman, J. R. (2001). NDVI anomaly patterns over Africa during the 1997/98 ENSO warm event. International Journal of Remote Sensing, 22(10), 1847-1860. https://doi.org/10.1080/01431160010029156
Bajgiran, P. R., Darvishsefat, A. A., Khalili, A., & Makhdoum, M. F. (2008). Using AVHRR-based vegetation indices for drought monitoring in the Northwest of Iran. Journal of Arid Environments, 72(6), 1086-1096. https://doi.org/10.1016/j.jaridenv.2007.12.004
Bichri, H., Chergui, A., & Hain, M. (2024). Investigating the impact of train/test split ratio on the performance of pre-trained models with custom datasets. International Journal of Advanced Computer Science and Applications (IJACSA), 15(2), 331-339. https://doi.org/10.14569/IJACSA.2024.0150235
Bojanowski, J. S., Sikora, S., Musiał, J. P., Woźniak, E., Dąbrowska-Zielińska, K., Ślesiński, P., Milewski, T., & Łączyński, A. (2022). Integration of Sentinel-3 and MODIS vegetation indices with ERA-5 agro-meteorological indicators for operational crop yield forecasting. Remote Sensing, 14(5), 1238. https://doi.org/10.3390/rs14051238
Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system, presented at the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA. New York, NY: Association for Computing Machinery. https://doi.org/10.1145/2939672.2939785
Cheng, M., Zhong, L., Ma, Y., Wang, X., Li, P., Wang, Z., & Qi, Y. (2023). A new drought monitoring index on the Tibetan Plateau based on multisource data and machine learning methods. Remote Sensing, 15(2), 512. https://doi.org/10.3390/rs15020512
Daniali, M., & Karimi, N. (2019). Spatiotemporal analysis of dust patterns over Mesopotamia and their impact on Khuzestan province, Iran. Natural Hazards, 97(2), 259-281. https://doi.org/10.1007/s11069-019-03641-0
Domenikiotis, C., Spiliotopoulos, M., Tsiros, E., & Dalezios, N. R. (2004). Early cotton yield assessment by the use of the NOAA/AVHRR derived Vegetation Condition Index (VCI) in Greece. International Journal of Remote Sensing, 25(14), 2807-2819. https://doi.org/10.1080/01431160310001632729
Dutta, D., Kundu, A., & Patel, N. R. (2013). Predicting agricultural drought in eastern Rajasthan of India using NDVI and standardized precipitation index. Geocarto International, 28(3), 192-209. https://doi.org/10.1080/10106049.2012.679975
Hao, Z., & Singh, V. P. (2015). Drought characterization from a multivariate perspective: A review. Journal of Hydrology, 527, 668-678. https://doi.org/10.1016/j.jhydrol.2015.05.031
Ji, L., & Peters, A. J. (2003). Assessing vegetation response to drought in the northern Great Plains using vegetation and drought indices. Remote sensing of Environment, 87(1), 85-98. https://doi.org/10.1016/S0034-4257(03)00174-3
Klisch, A., & Atzberger, C. (2016). Operational drought monitoring in Kenya using MODIS NDVI time series. Remote Sensing, 8(4), 267. https://doi.org/10.3390/rs8040267
Kogan, F. N. (1995). Application of vegetation index and brightness temperature for drought detection. Advances in Space Research, 15(11), 91-100. https://doi.org/10.1016/0273-1177(95)00079-T
Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions, presented at the 31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA. Red Hook, NY: Curran Associates, Inc. https://doi.org/10.48550/arXiv.1705.07874
Mondal, S., Jeganathan, C., Sinha, N. K., Rajan, H., Roy, T., & Kumar, P. (2014). Extracting seasonal cropping patterns using multi-temporal vegetation indices from IRS LISS-III data in Muzaffarpur District of Bihar, India. The Egyptian Journal of Remote Sensing and Space Science, 17(2), 123-134. https://doi.org/10.1016/j.ejrs.2014.09.002
Omia, E., Bae, H., Park, E., Kim, M. S., Baek, I., Kabenge, I., & Cho, B.-K. (2023). Remote sensing in field crop monitoring: A comprehensive review of sensor systems, data analyses and recent advances. Remote Sensing, 15(2), 354. https://doi.org/10.3390/rs15020354
Patel, N. R., Parida, B. R., Venus, V., Saha, S. K., & Dadhwal, V. K. (2012). Analysis of agricultural drought using vegetation temperature condition index (VTCI) from Terra/MODIS satellite data. Environmental Monitoring and Assessment, 184, 7153-7163. https://doi.org/10.1007/s10661-011-2487-7
Quiring, S. M., & Ganesh, S. (2010). Evaluating the utility of the Vegetation Condition Index (VCI) for monitoring meteorological drought in Texas. Agricultural and Forest Meteorology, 150(3), 330-339. https://doi.org/10.1016/j.agrformet.2009.11.015
Segarra, J., Buchaillot, M. L., Araus, J. L., & Kefauver, S. C. (2020). Remote sensing for precision agriculture: Sentinel-2 improved features and applications. Agronomy, 10(5), 641. https://doi.org/10.3390/agronomy10050641
Seiler, R. A., Kogan, F., & Wei, G. (2000). Monitoring weather impact and crop yield from NOAA AVHRR data in Argentina. Advances in Space Research, 26(7), 1177-1185. https://doi.org/10.1016/S0273-1177(99)01144-8
Wang, J., Price, K. P., & Rich, P. M. (2001). Spatial patterns of NDVI in response to precipitation and temperature in the central Great Plains. International Journal of Remote Sensing, 22(18), 3827-3844. https://doi.org/10.1080/01431160010007033
Iran Agricultural Research

Articles in Press, Accepted Manuscript
Available Online from 18 July 2026

Files

Share

How to cite

Statistics