Bejide Oluwafemi David*
This research assesses the accuracy of gridded precipitation datasets in Plateau, Jos, Nigeria, and examines their relationship with in situ ground data. Four datasets-PERSIANN-CD, SM2Rain ASCAT, TAMSAT and MERRA 2-are compared with ground data from Nigeria Meteorological Station (NIMET) and the Geography Department, University of Jos. The study area, characterized by a single-maxima rainfall pattern, experiences a pronounced dry season from November to January and a rainy season from March to November.
The research findings reveal that all datasets have the capability to detect precipitation events. However, the TAMSAT dataset demonstrates the closest relationship with in situ data and exhibits the lowest bias and root mean square error. Nonetheless, weak correlations between gridded and ground data suggest that factors such as measurement methods, spatial and temporal resolutions and elevation differences may influence results. Notably, SM2Rain ASCAT shows the highest probability of detection and critical success index.
Furthermore, the research identifies a negative relationship between the resolutions of gridded datasets, the critical success index and probability of detection. This implies that coarser gridded datasets tend to increase the probability of detection and critical success index.
In conclusion, the research highlights the importance of considering various factors when selecting precipitation datasets for specific applications and regions.
