loading page

Validation and intercomparison of satellite-based rainfall products over Africa using TAHMO in-situ rainfall observations
  • +2
  • Denis Macharia,
  • Katie Fankhauser,
  • John Selker,
  • Jason Neff,
  • Evan Thomas
Denis Macharia
Regional Centre for Mapping of Resources for Development

Corresponding Author:[email protected]

Author Profile
Katie Fankhauser
University of Colorado Boulder
Author Profile
John Selker
Oregon State University
Author Profile
Jason Neff
University of Colorado at Boulder
Author Profile
Evan Thomas
University of Colorado at Boulder
Author Profile

Abstract

Sparse rain gauge networks and declining observations in Africa limit climate research in the region. However, the proliferation of satellite-rainfall products (SRPs) and the growth of citizen-science-driven in-situ observations driven by cheaper data collection technologies have provided a pathway to overcoming the data scarcity problem. In this paper, we used rain gauge data from 596 stations operated by the Trans-African Hydro-Meteorological Observatory (TAHMO) across Africa to evaluate the performance of two widely-utilized satellite-based rainfall products: the Climate Hazards InfraRed Precipitation with Stations (CHIRPS) and the Tropical Applications of Meteorology using Satellite data (TAMSAT), and two under-validated and underutilized products: the satellite-only Global Satellite Mapping of Precipitation (GSMaP) and the gauge-corrected GSMaP version (GSMaP_Gauge). We also inter-compared the performance of the four products over Africa, East Africa, Southern Africa and West Africa at daily, pentadal, and monthly timescales. Our findings indicated that the GSMaP products had better performances at daily timescales whereas CHIRPS and TAMSAT matched or outperformed the GSMaP products at pentadal and monthly timescales. GSMaP_Gauge daily rainfall detection was almost 1.5 times the CHIRPS detection scores at the same temporal scale. The Pearson correlation coefficient increased with temporal aggregation but the volumetric errors increased for all products. Additionally, all the products overestimated (underestimated) low (high) intensity rainfall events. Our analysis adds to a growing number of validation studies in Africa and presents an opportunity for developers of satellite-rainfall products to integrate the new TAHMO observations in bias-correction algorithms to improve the accuracy of SRPs in the region.