NDVI–rainfall relationship in the Semliki watershed of the equatorial Nile

Abstract

The use of remotely sensed data to describe, study, monitor and provide greater understanding on watershed’s landscape dynamics have proven to be useful worldwide, especially where timely and reliable ground information are neither available nor accessible like in most of the developing world. This paper, thus presents the first documented relationship between NDVI and rainfall in humid central Africa, and serves as a precursor in the understanding of the interaction between catchment characteristics and water resources of the Semliki region. This study investigates spatial and temporal relationship between NDVI and satellite-derived rainfall. Normalized Difference Vegetation Index (NDVI) time series derived from the NOAA–AVHRR (National Oceanic and Atmospheric Administration–Advanced Very High Resolution Radiometer) satellite data and FEWS satellite-derived rainfall are analyzed. The Semliki watershed (23,621 km 2) of the equatorial Nile basin is the study area. Monthly NDVI time series of 21 Semliki’s subcatchments (S3–S23) over 7 years (2001–2007) were processed and extracted with Windisp 5.1 from 10-day NDVI maximum value composite images. At the monthly time step, only 12 subcatchments had weak positive correlation (0.23–0.42) at 5% significance level and the rest were not significantly correlated. Incorporating a 1-month lag effect increased the correlation coefficients between the two variables (0.28–0.68) with 17 subcatchments being significantly correlated at the designated level. The topography within the catchment was found to play a defining role for NDVI values. S15, S16, S18 and S20, the subcatchments covering the Mount Ruwenzori characterized by elevations up to 4862 m above sea level, consistently recorded the lowest values of NDVI over time and no significant correlation could be established between rainfall and NDVI.