20+ Million Readerbase
Indexed In
  • Open J Gate
  • RefSeek
  • Hamdard University
  • OCLC- WorldCat
  • Publons
  • International Scientific Indexing
  • Euro Pub
  • Google Scholar
Share This Page
Recommended Webinars & Conferences

9th International Conference on GIS and Remote Sensing

Edinburgh, Scotland
Journal Flyer
Flyer image


Assessment of Spatio-temporal Relationship between Gauge and Satellite Rainfall Estimates over Awash River Basin, Ethiopia

Salih Edris, Daniel Teka, Amanuel Zenebe

Accurate information on rainfall data is essential for numerous operational and research fields. Conventionally, ground-based measurement is the main source of rainfall data. However, in developing countries, networks of ground-based measurements are very sparse or nonexistent. An alternative to this measurement could be satellitebased rainfall estimates (SREs). However, SREs need to be validated as their accuracy can be affected by topography and climate. This study seeks to investigate the spatiotemporal relationship between gauge and SREs over the Awash River basin. The Climate Hazards Group Infrared Precipitation (CHIRP), CHIRP combined with station observations (CHIRPS), and African Rainfall Climatology version 2 (ARC2) are evaluated at dekadal (10-day), monthly and annual time-scales for selected normal years against 37 ground-based measurements located at different elevations of the basin. A point-to-grid-based comparison is adopted, using continuous statistical validation tools. Temporal and spatial analysis indicates the basin exhibits tremendous spatial variability in the rainfall amount which varying from 190 mm in lowland to 1300 mm yr-1 in highland, with significant correlation. From the overall analysis at dekadal, monthly, and annual temporal scale, CHIRPS followed by CHIRP exhibited better performance in comparison to ARC2. ARC2 product is poorly performed SREs with underestimations of high rainfall rate. The agreement between the SREs and ground-based measurement improved with increase in time scale from dekadal (for instance CHIRPS has correlation > 0.77, Nash-Sutcliff efficient coefficient (Eff) > 0.59, root mean square error (RMS) < 22.1, and bias ≤ 1.1) to monthly (correlation > 0.89, Eff > 0.79, RMSE < 39.0 and bias ≤ 1.0), but the performance of these products decrease when aggregated into annual time scale ( correlation > 0.40, Eff > -0.56, RMSE < 161.10). In general, the SREs shows good agreement with ground-based measurements over the highland parts of the basin at dekadal and monthly time scale, however, at an annual time scale, all products show better performance over the lowland parts of the basin. This study demonstrates the reliability of satellite rainfall estimates in regions where the spatial network of the gauge is highly sparse and inaccessible.

Published Date: 2021-09-10;