Examining the NDVI-Rainfall Relationship Under High Enso Event Influence Using Geographically Weighted Regression in Peninsular Malaysia

The NDVI and rainfall connections are complex, spatially non-stationary, non-linear, and scale-dependent. To address this, a local modelling approach, Geographically Weighted Regression (GWR) was employed, to accommodate for spatial variability. By employing Climate Prediction Center morphing method (CMORPH) satellite dataset, this study investigates the Normalized Difference Vegetation Index (NDVI)-rainfall relationship in Peninsula Malaysia (PM), focusing on periods influenced by El Niño-Southern Oscillation (ENSO). The outcomes of Moran's I spatial autocorrelation confirm the presence of rainfall spatial variability which endorses the application of the GWR model. The findings suggest that there is a contrasting spatial relationship between NDVI and rainfall during the El Niño events, but positive relationship during the La Niña events. The southern region generally exhibited a positive relationship, while the northern and northeastern regions showed negative or weak associations. These results indicate that for a more precise, localized representation of a consistent relationship (either increasing or decreasing), it is essential to identify specific regions under the influence of El Niño.