Mitigating Climate Change Impact Through Smart Agriculture—Soil Temperature Analysis and Forecasting

Climate change is the drastic alteration of average weather conditions over several decades, especially due to man-made activities. The rapid climate change has thus led to uncertainty in many environmental factors responsible for healthy crop production. Among them, soil temperature plays a vital role, impacting crop growth with even slight fluctuations. Machine learning models can be a useful tool for predicting agricultural output due to the rising unpredictability of soil temperature. This paper conducts a comparative analysis of multiple machine learning models, like random forest, support vector machines for regression, k-nearest neighbour, gradient boosting, and AdaBoost. Soil temperature and other meteorological factors taken from INCOMPASS land stations were observed between 2016 and 2018 by the NERC Environmental Information Data Centre. AdaBoost outperformed the rest, with 0.986 and 0.809 R2 scores for the two datasets used. This paper also provides insight into the impact of climate change on soil temperature and crop production. However, the availability of more data on the correlation between soil temperature and crop yield would lead to even smarter models with more direct applications in the field. This research demonstrates how machine learning models could assist farmers in reducing the hazards brought on by climate change.