Performance assessment of satellite rainfall estimates in rain detection capabilities at different thresholds over Nigeria
Afeez Alabi Salami, R. M. Olanrewaju, J. F. Olorunfemi
Abstract
In Nigeria, where rainfall plays a pivotal role in agriculture and disaster management, assessing the accuracy of satellite rainfall estimates at various thresholds is imperative. This study assesses the areal averages of nine satellite precipitation estimates (SPEs) with 48 ground-based rain gauge observations in Nigeria. Employing categorical statistical metrics and compromise programming, the research assesses the performance of SPEs at seven rainfall thresholds. Results reveal that 82-94% of rainfall estimates align with ground truth, while 5-14% of rainy days are incorrectly detected by rain gauges. Notably, SPEs tend to overestimate low rain between 1mm and 5mm d−1 and underestimate low heavy rain rates (10mm ≤ rain < 20mm), impacting flood monitoring and disaster preparedness. GPCC, TAMSAT-TARCAT, and CHIRPS emerge as more reliable SPEs, particularly in highland regions. This research underscores the importance of choosing appropriate satellite precipitation estimates for enhanced disaster preparedness and resource management across different Nigerian regions.