Dominant Modes of Rainfall Variability over East Africa in Response to Multiscale Global Climate Analyses

Thumbnail Image
October, 2016
Journal Title
Journal ISSN
Volume Title
The main interest was to carry out an analysis of time evolution of global climate during the climatologically prominent phase of El Niño Southern Oscillation (ENSO) from 1950-2008 and their relationship to the dominant modes of March-May (MAM) seasonal rainfall variability to provide useful climate information needed by end users for incorporation into sustainable climate change developmental goals. The study utilized monthly data consisting of horizontal global winds at 200 hPa from the National Centers for Environmental Prediction-National Center for Atmospheric Research (NCEP-NCAR) reanalysis data, Climatic Research Unit (CRU) gridded terrestrial precipitation data and Extended Reconstructed Sea Surface Temperature (ERSST). The three major timescales of investigation in the annual cycle included monthly, bimonthly and seasonal. The study employed Empirical Orthogonal Function (EOF) analysis, lagged heterogeneous grid point correlation, composite analysis technique and multiple linear regression analysis. The EOF analysis was performed on the East African long rains (MAM) and the four leading modes were retained. Lagged heterogeneous grid point correlation between the time coefficients of the four leading modes and global SST were computed to evaluate, delineate and monitor the specific SST signals that were connected to the long rains. On the monthly timescale, the grid-point correlation showed that, EOF 1, 3 and 4 MAM precipitation modes responded differently to the Pacific ENSO, Atlantic and Indian Oceans. However, the second mode apparently was not well related to the global SST features. Meanwhile, EOF 3 showed an indirect relationship with the Pacific while an Atlantic Multidecadal Oscillation (AMO)-like feature captured in the North Atlantic was identified as directly linked to the mode. The composite iv analysis revealed that divergent circulations and the centers of action at 200hPa level varied on the monthly, bimonthly and seasonal timescales. The distinction of the circulation patterns were based on their strengths, locations, and spatial extents. Similar observations were made on the combined timescale. The multiple linear regression model outputs between the rainfall modes and the climate indices, revealed the R2 values ranging between 0.0-0.4, 0.01-0.3, 0.01-0.25 and 0.02-0.3 for monthly, bimonthly, seasonal and combined timescales respectively. On all the timescales, the highest R2 values were recorded in January, December and May for EOF 1, EOF 3 and EOF 4 respectively. This is suggestive that the East African long rains variability is greatly modulated by global climate features on a monthly timescale. Overall, the findings provide useful prediction information required for improving capacity to adaptive climate change impacts.
A Thesis Submitted to the Department of Physics, Kwame Nkrumah University of Science and Technology in Partial Fulfillment of the Requirements for the Degree of Master of Philosophy in Meteorology and Climate Change,