Effect of Topography in Satellite Rainfall Estimation Errors: Observational Evidence across Contrasting Elevation in the Blue Nile Basin
No Thumbnail Available
Date
2013-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
In this study, the effect of topography on spatial and temporal variability of rainfall on local
(25 km by 25 km grid) scale was analyzed across two contrasting elevation locations.
We deployed 10 in high elevation grid (average elevation 2097m) and 9 tipping rain
gauges in low elevation grid (average elevation= 695 m). Based on high quality data
from the dense rain gauges, we evaluated the spatio-temporal properties of rainfall and
evaluated the errors in widely used high resolution satellite rainfall estimates. Three
satellite rainfall estimates; TMPA 3B42RT, TMPA 3B42V7 and CMORPH were
evaluated using statistical and categorical verification measures. Our results showed that
there is significantly large spatio-temporal variability of rainfall at local scale. Diurnal cycle
also indicated that, low elevation receives more rainfall at early morning and the high
elevation gets more rainfall between the afternoon and mid-night. Compared to the high
elevation grid, low elevation grid receives less seasonal accumulation which is less
frequent but intense rainfall. High elevation area has more rain events (varying 91 to
105) than low elevation (range; 66 to 7 5) using 6hour minimum inter-event time. In high
elevation maximum event rain rate is 67mm/hour but low elevation is l 22mm/hour. Our
results also demonstrated clear difference between the two grid cells not only in terms of
rainfall magnitude and property but also on the accuracy of satellite rainfall estimates which
could be attributed to topography. 3B42RT, 3B42V7 and CMORPH overestimated total
3hour mean rainfall in low elevation (with bias 1.2, I.land 1.5 and ME 0.2, 0.1 and
0.5 respectively). All underestimated in high elevation by 0.9, 0.8 and 0.7, and ME
0.2, -0.2 and -0.5 for 3B42RT, 3B42V7 and CMORPH respectively. The categorical
statistics indicated that CMORPH performed better in its probability of detection (POD), but
it is influenced by non rainy clouds. All products highly affected by miss rainy events on high
elevation area and by more false alarms at low elevation area. From our results 3B42V7 is
best in the low elevation and 3B42RT performed better in the high elevation area.
Description
Keywords
Topography, Elevation, Rainfall variability, satellite rainfall estimate accuracy