Exploring Rice Yield Variability Under Climate Change Through NDVI Analysis

Abstract

This study presents a novel approach to predicting paddy yields in Brunei's Wasan Rice Scheme using projected normalized difference vegetation index (NDVI) values derived from climate projections under three time periods: near future (2020–2046), mid-future (2047–2073), and far future (2074–2100). Employing CMIP6 socioeconomic pathways (SSP245, SSP370, SSP585), random forest (RF) and multiple linear regression (MLR) models were utilised to link historical NDVI with meteorological factors such as rainfall and temperature. Results indicate that main-season yields are expected to decline or stabilize across scenarios, while off-season NDVI consistently increases, reflecting robust vegetation recovery. These findings emphasise the differential impacts of climate change across growing seasons, providing critical insights for agricultural planning and adaptation strategies. By integrating scenario-based NDVI projections
and predictive modeling, this study offers a comprehensive framework for understanding future crop dynamics under changing climatic conditions.