Assessing predictability of post-monsoon crop residue fires in Northwestern India

Abstract

Over the past five decades, the Green Revolution in India has been a great success resulting in significantly increased crop yields and food grain productivity. Northwestern India, also known as the country’s breadbasket, alone produces two-thirds of the wheat and rice grains under the crop rotation system. Our previous study has shown that the post-monsoon rice crop production in the Punjab state of India has increased by 25%. The crop yields produce proportionate amounts of residue, a large part of which is subjected to burn in the open fields due to the near-absence of a wide-scale, affordable, and environmentally sustainable removal mechanism. A significant increase in crop productivity coincides with a 60% increase in post-harvest crop residue burning during 2002–2016. The study also demonstrated a robust relationship between satellite measurements of vegetation index—a proxy for crop amounts, and post-harvest fires—a precursor of air pollution events, for predicting seasonal agricultural burning. In this report, the efficacy of the proposed prediction model is assessed by comparing the forecasted seasonal fire activity against the actual detection of active fires for the post-monsoon burning seasons of 2017–2021. A simple linear regression model allows efficient prediction of seasonal fire activity within an error of up to 10%. In addition to forecasting seasonal fire activity, the linear regression model offers a practical tool to track and evaluate the effectiveness of the residue management system intended to reduce fire activities and resulting air pollution.