Abstract
This article examined rainfall enhancement over Lake Victoria. Estimates of overlake rainfall were compared with rainfall in the surrounding lake catchment. Four satellite products were initially tested against estimates based on gauges or water balance models. These included TRMM 3B43, IMERG V06 Final Run (IMERG-F), CHIRPS2, and PERSIANN-CDR. There was agreement among the satellite products for catchment rainfall but a large disparity among them for overlake rainfall. IMERG-F was clearly an outlier, exceeding the estimate from TRMM 3B43 by 36%. The overestimation by IMERG-F was likely related to passive microwave assessments of strong convection, such as prevails over Lake Victoria. Overall, TRMM 3B43 showed the best agreement with the “ground truth” and was used in further analyses. Overlake rainfall was found to be enhanced compared to catchment rainfall in all months. During the March–May long rains the enhancement varied between 40% and 50%. During the October–December short rains the enhancement varied between 33% and 44%. Even during the two dry seasons the enhancement was at least 20% and over 50% in some months. While the magnitude of enhancement varied from month to month, the seasonal cycle was essentially the same for overlake and catchment rainfall, suggesting that the dominant influence on overlake rainfall is the large-scale environment. The association with mesoscale convective systems (MCSs) was also evaluated. The similarity of the spatial patterns of rainfall and MCS count each month suggested that these produced a major share of rainfall over the lake. Similarity in interannual variability further supported this conclusion.
Significance Statement
The article underscores the difficulty of IMERG-F and other passive microwave retrievals in assessing rainfall over inland water bodies.
© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).