Modelling Cropping Intensity of Kerala Using ARIMA with Exogenous Predictors

Abstract

In order to comprehend long-term patterns and the impact of climatic variables, this study used linear time-series modeling approaches to analyze Kerala's cropping intensity over a 58-year period (1965–66 to 2022–23). The dependent variable was cropping intensity, and the exogenous predictors of yearly rainfall, maximum temperature, and minimum temperature were included to evaluate their possible influence. To find the most appropriate and reliable forecasting model, a variety of model configurations were assessed using common model selection criteria, such as the Bayesian Information Criterion (BIC) and the Akaike Information Criterion (AIC). The ARIMA with Exogenous Variables (ARIMAX) and Auto-Regressive Integrated Moving Average (ARIMA) models were both used and contrasted. Because it incorporates external factors, the ARIMAX model is especially good at capturing how climate variables affect cropping intensity. The ARIMAX(0,1,1) model consistently performed better than the ARIMA(0,1,1) model across all assessment measures among the many model combinations studied. Significantly, temperature variables were found negatively influencing cropping intensity. Relative diagnostic tests were performed to verify the residuals behaved like white noise and to make sure the selected model was adequate. These tests included checks for autocorrelation and non-linearity. These results demonstrate how important climate variables are in determining cropping intensity over time, particularly temperature trends. The findings provide valuable empirical information for creating evidence-based legislation, climate-resilient agricultural practices, and sustainable resource management that are suited to Kerala's changing agroclimatic environment.