The primary function of the TRMM Ground Validation (GV) Program is to creat
e GV rainfall products that provide basic validation of satellite-derived p
recipitation measurements for select primary sites. Since the successful 19
97 launch of the TRMM satellite, GV rainfall estimates have demonstrated sy
stematic improvements directly related to improved radar and rain gauge dat
a, modified science techniques, and software revisions. Improved rainfall e
stimates have resulted in higher quality GV rainfall products and subsequen
tly, much improved evaluation products for the satellite-based precipitatio
n estimates from TRMM.
Early improvements in TRMM GV rainfall products involve replacing a default
radar reflectivity (Z(e)) - rain rate (R) relationship with convective and
stratiform relationships independently derived using bulk-adjusted, qualit
y-controlled rain gauge data. Upon the development of an automated gauge qu
ality control procedure, poorly correlated gauge-radar data in the bulk-adj
ustment process are ignored and Z(e)-R relationships are again refined. The
se Z(e)-R relationships, applied to base scan reflectivity data to produce
GV rain maps, are further modified as improvements to intermediate radar an
d gauge products result in more realistic and internally consistent rainfal
l statistics. Additionally, improvements to radar and gauge data from the p
rimary GV sites further increase the reliability of rainfall products. Rain
fall accumulation statistics for each primary site are presented, demonstra
ting the evolving improvement in radar rainfall estimation for each success
ive generation of products. Current generations of GV rainfall products, ba
sed on more robust data and science techniques, have removed known biases i
n radar rainfall estimates. Increased confidence in GV rainfall products re
sults in more meaningful comparisons with satellite-derived precipitation e
stimates from TRMM. (C) 2000 Elsevier Science Ltd. All rights reserved.