Six satellite-based rainfall estimates (SRFE)—namely, Climate Prediction Center (CPC) morphing technique
(CMORPH), the Rainfall Estimation Algorithm, version 2 (RFE2.0), Tropical Rainfall Measuring Mission
(TRMM) 3B42, Goddard profiling algorithm, version 6 (GPROF 6.0), Precipitation Estimation from Remotely
Sensed Information using Artificial Neural Networks (PERSIANN), Global Satellite Mapping of Precipitation
moving vector with Kalman filter (GSMapMVK), and one reanalysis product [the interimECMWFRe-Analysis
(ERA-Interim)]—were validated against 205 rain gauge stations over four African river basins (Zambezi, Volta,
Juba–Shabelle, and Baro–Akobo). Validation focused on rainfall characteristics relevant to hydrological applications,
such as annual catchment totals, spatial distribution patterns, seasonality, number of rainy days per year,
and timing and volume of heavy rainfall events.Validation was done at three spatially aggregated levels: point-topixel,
subcatchment, and river basin for the period 2003–06. Performance of satellite-based rainfall estimation
(SRFE) was assessed using standard statistical methods and visual inspection. SRFE showed 1) accuracy in
reproducing precipitation on amonthly basis during the dry season, 2) an ability to replicate bimodal precipitation
patterns, 3) superior performance over the tropical wet and dry zone than over semiarid or mountainous regions,
4) increasing uncertainty in the estimation of higher-end percentiles of daily precipitation, 5) low accuracy in
detecting heavy rainfall events over semiarid areas, 6) general underestimation of heavy rainfall events, and
7) overestimation of number of rainy days in the tropics. In respect to SRFE performance, GPROF 6.0 and
GSMaP-MKVwere the least accurate, andRFE 2.0 andTRMM3B42were themost accurate. These results allow
discrimination between the available products and the reduction of potential errors caused by selecting a product
that is not suitable for particular morphoclimatic conditions. For hydrometeorological applications, results support
the use of a performance-based merged product that combines the strength of multiple SRFEs.
THIEMIG Vera;
ROJAS MUJICA Rodrigo Felipe;
ZAMBRANO Hector;
LEVIZZANI Vincenzo;
DE ROO Arie;
2013-01-14
AMER METEOROLOGICAL SOC
JRC66477
1525-755X,
http://journals.ametsoc.org/doi/abs/10.1175/JHM-D-12-032.1,
https://publications.jrc.ec.europa.eu/repository/handle/JRC66477,
10.1175/JHM-D-12-032.1,
