Development of high spatial resolution rainfall climatology for Ghana

Thumbnail Image
July 2015
Journal Title
Journal ISSN
Volume Title
Various sectors of the country’s economy (health, energy, agriculture, planning and many others) depend on climate, and as such availability of quality climate data becomes essential for climate impact studies in these sectors. In this study, rainfall climatology database has been developed for Ghana using GMet station datasets distributed over the four agro-ecological zones and spanning a 33-year period (1980 – 2012). Datagaps within the rainfall time-series were filled by Regularized Expectation Maximization (RegEM) and homogenization of the time-series was performed by Quantile Matching Adjustments (QMadj). The homogenized datasets were then gridded at a high-spatial resolution (0:25 o x 0:25 o ) using Minimum Surface Curvature (MSC) with tensioning parameter. Seasonal rainfall for the four agro-ecological zones have been derived based on the grids covering the entire country and this allowed a clear evidence of the migration of Inter-Tropical Discontinuity (ITD) from the South of the country to the North and back; thus, establishing a uni-modal rainfall regime over the Northern part of the country and a bi-modal rainfall regime over the Southern part of the country. Finally, Climatic Research Unit Time-Series 3.22 (CRU TS 3.22) monthly precipitation data was used to validate the gridded dataset, obtaining high Pearson’s correlation co-efficients (0.5 – 0.9), low relative mean difference (0 – 0.3) and low relative root mean square error values (0 – 8). At present, a country-wide rainfall climatology has iii been developed from GMet rainfall time-series which will serve as a precursor for further climate impact study, in the aforementioned sectors, across the country.
A thesis submitted to the Deparment of Physics, Kwame Nkrumah University of Science and Technology in partial fulfillment of the requirements for the degree of Master of Philosophy (Meteorology and Climate Science).