A Comparison of Two Satellite Rainfall Estimates for GATE

View More View Less
  • 1 Joint Institute for Marine and Atmospheric Research (JIMAR), University of Hawaii, Honolulu 96822
© Get Permissions Rent on DeepDyve
Restricted access

Abstract

Rainfall estimates obtained for the GATE experiment by two satellite rainfall estimation techniques are compared for different time and space scales. The Kilonsky-Ramage technique uses polar-orbiting satellites for 1° resolution rainfall estimates over the tropics. The Griffith-Woodley technique uses geostationary satellite data to generate much higher resolution estimates. There is good correspondence between the A scale isohyetal patterns and rainfall volumes estimated by both techniques for periods ranging from 1 to 80 days. Largest discrepancies appear over Africa and nearby ocean areas, where the Griffith-Woodley estimates are higher. Correlation coefficients between the two estimates are higher when only oceanic points are compared, ranging up to 0.92 for the entire GATE period. When continental areas of the A scale are included, correlation coefficients are lower, reaching 0.66 for phase 2. The large discrepancies over Africa and nearby ocean areas are seen to be associated with a nocturnal peak of rainfall which is not detected by the daytime passes of the polar orbiting satellite used in the Kilonsky-Ramage calculations.

Comparison of results of both techniques with shipboard radar ground-truth over the GATE B scale show that both techniques produce comparable rainfall volume estimates at this scale, generally within 15% of the radar estimates. Overall, results indicate that the Kilonsky-Ramage technique can produce low cost and fairly reliable estimates of rainfall over the tropical oceans.

Abstract

Rainfall estimates obtained for the GATE experiment by two satellite rainfall estimation techniques are compared for different time and space scales. The Kilonsky-Ramage technique uses polar-orbiting satellites for 1° resolution rainfall estimates over the tropics. The Griffith-Woodley technique uses geostationary satellite data to generate much higher resolution estimates. There is good correspondence between the A scale isohyetal patterns and rainfall volumes estimated by both techniques for periods ranging from 1 to 80 days. Largest discrepancies appear over Africa and nearby ocean areas, where the Griffith-Woodley estimates are higher. Correlation coefficients between the two estimates are higher when only oceanic points are compared, ranging up to 0.92 for the entire GATE period. When continental areas of the A scale are included, correlation coefficients are lower, reaching 0.66 for phase 2. The large discrepancies over Africa and nearby ocean areas are seen to be associated with a nocturnal peak of rainfall which is not detected by the daytime passes of the polar orbiting satellite used in the Kilonsky-Ramage calculations.

Comparison of results of both techniques with shipboard radar ground-truth over the GATE B scale show that both techniques produce comparable rainfall volume estimates at this scale, generally within 15% of the radar estimates. Overall, results indicate that the Kilonsky-Ramage technique can produce low cost and fairly reliable estimates of rainfall over the tropical oceans.

Save