Satellite derived rainfall over tropical Africa : its pattern variability and calibration
Loading...
Date
2004
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The major spatial and temporal structure of rainfall and zonal wind at seasonal and intra-seasonal time-scales over tropical Africa are identified by principal component analysis. Pentad CMAP rainfall and NCEP 700hPa zonal wind for the period 1980-2000 were used. Wavelet analysis is used to filter the data sets for cycles between 20-70 days for the intra-seasonal oscillations.
The first mode of seasonal rainfall that explains 36% of the total variance is dominated by the annual cycle associated with the ITCZ and is loaded over equatorial Africa. The dominant mode for the zonal wind is loaded over eastern Africa with an extended axis in the Indian Ocean (25.8%). A strong relationship exists between the zonal wind and rainfall modes at the seasonal time scale.
The dominant mode for the intra-seasonal rainfall is loaded over northern Congo (29%). The second and third modes are over northeast Angola and East Africa respectively. The intra-seasonal oscillations are phase locked to the seasonal cycle and reveal cycles of 30-50 days typical of Madden Julian oscillation.
The 700hPa zonal wind intra-seasonal oscillation reveals a dominant mode over the equatorial east Atlantic Ocean. The second and third modes are over the west Indian Ocean and southeast Atlantic. A strong relationship between zonal wind over the equatorial Atlantic and rainfall over both northern Congo and east Africa is observed. Cross correlation reveals that the zonal wind leads rainfall by a few pentads.
Hovmoller plots for OLR and velocity potential reveal that the intra-seasonal oscillations over tropical Africa are dominated by eastward propagation. However, standing and westward propagation are also observed. Propagation is weak over east African coast due to strong meridional flow of the Indian Monsoon.
Results from the satellite rainfall calibration and validation are examined. While satellite radiance values generally underestimate rainfall over Uganda, they naturally provide better spatial distribution than the gauges. Following calibration algorithm development, validation tests indicate that satellite rainfall explains 81% of rainfall received in the year 2000. These results should be interpreted with caution since rain gauges give point measurements while satellites provide pixel values.
Description
Thesis submitted to the Faculty of Science for the degree of Master of Science in the Department of Geography and Environmental Studies at the University of Zululand, 2004.
Keywords
Rain and rainfall--Africa, Precipitation (Meteorology)--Africa