Forecasting of Cropping Intensity of Kerala in India: A Multivariate Approach

Abstract

A state like Kerala in Indian subcontinent, even though, blessed with the typical physiographical features contributing to its agricultural diversity, has an average cropping intensity of 132 percentages which shows the huge gap between its potential and actual production. A multivariate statistical approach known as Vector Auto Regressive (VAR) modeling was done in the present study to analyse the behavior of Cropping Intensity in reference to gross cropped area, net cropped area and annual rainfall of the state from 1965-66 to 2022-23 and to forecast the same. These forecasts are aimed at enhancing resource optimization, sustainability, or agricultural planning. VAR(1) model was found to be better fitting for prediction as well as forecasting based on validation tests like ADF Test, Granger Causality and decision criteria like Akaike Information Criterion (AIC), Hannan-Quinn Information Criterion (HQ), Schwarz’s Criterion (SC) and Akaike's Final Prediction Error criterion (FPE). The model suggested that Cropping Intensity is influenced by past Cropping Intensity, Gross Cropped Area, Net Cropped Area and Annual Rainfall.