In the face of escalating climate hazards, a groundbreaking review published in *Environmental Research Communications* (which translates to *Environmental Research Communications*) offers a critical synthesis of how remote sensing (RS) and machine learning (ML) are transforming our ability to predict and mitigate the impacts of climate hazards on crop yields. Led by Salomon Obahoundje of the International Water Management Institute (IWMI) in Accra, Ghana, the study scrutinizes 177 studies on climate hazards and 197 on RS–ML applications in crop yield modeling, providing a roadmap for future advancements in climate-resilient agriculture.
The research reveals a significant concentration of studies in Asia, followed by Africa and the Americas, with agricultural drought emerging as the most frequently studied hazard. “The dominance of statistical approaches, such as the coefficient of variation, highlights the need for more sophisticated methods to analyze climate variability,” Obahoundje notes. The review identifies Random Forest (RF) as the most widely used ML algorithm for drought detection, followed by Support Vector Machines (SVM), Artificial Neural Networks (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Extreme Gradient Boosting (XGBoost). For drought impacts on crop productivity, RF again leads, followed by Least Absolute Shrinkage and Selection Operator (LASSO), while for climate variability impacts, RF, SVM, ANN, Long Short-Term Memory (LSTM), Multiple Linear Regression (MLR), and Convolutional Neural Network (CNN) are prominently used.
The study underscores the regional disparities in the adoption of advanced ML and deep learning (DL) techniques. Asia leads in integrating these advanced methods, while Africa, due to infrastructure and data limitations, predominantly employs simpler and more interpretable models. “This disparity highlights the urgent need for standardized protocols and real-time, microclimate-aware monitoring systems to improve model reliability and applicability in underrepresented, data-scarce regions such as sub-Saharan Africa,” Obahoundje emphasizes.
The review also points out the limitations of widely used RS products like MODIS, TRMM, CHIRPS, and ERA5, which, despite their accessibility, have restricted effectiveness due to limited spatial resolution. The research gaps identified include a limited investigation at the sub-national level, insufficient ground-truth validation, and inadequate monitoring of complex, compounding hazards like drought–flood–heatwave interactions. Moreover, the research remains skewed toward economically dominant crops (maize, cotton, and soybeans), neglecting marginal crops (cocoa, cashew, cassava, plantain, and coffee) critical to food-insecure regions.
The implications for the energy sector are profound. As climate hazards intensify, the demand for climate-resilient crops will rise, driving innovations in agricultural technologies and practices. The integration of advanced ML and DL techniques with RS products can enhance predictive capabilities, enabling farmers and policymakers to make informed decisions. This, in turn, can stabilize food supplies and reduce the economic impacts of climate-induced crop failures.
Obahoundje’s review recommends hybrid modeling frameworks that integrate process-based and data-driven methods, broader spatial and crop coverage, and standardized protocols. These advancements can strengthen climate-resilient agriculture and global food security, ultimately contributing to a more sustainable and secure energy sector.
As the world grapples with the escalating impacts of climate change, this research offers a beacon of hope, highlighting the transformative potential of ML and RS in safeguarding our food systems and energy security. The study not only sheds light on current practices but also paves the way for future developments, ensuring that agriculture remains resilient in the face of climate hazards.