Abstract:
Accurately forecasting rainfall trends is vital for the socio-economic development of a nation. Observed daily rainfall data from the Ghana Meteorological Agency (GMet) spanning 1980 – 2015 was deployed to study the spatio-temporal rainfall trend in the Odaw river basin (ORB) in Accra. Using Mann–Kendall (MK), Sen’s slope and set of homogeneity tests, the monthly, annual and decadal rainfall variability based on the indices – rainfall total, daily maximum and rainy/wet day count were examined. Principal Component Analysis (PCA) and wavelet (WT) analysis were applied on the monthly time series in order to verify the spatio-temporal variability. It was observed that a mean value ranging from 760 to 1200 mm, 77.1–94.4 mm and 56–90 day count were recorded for Annual Rainfall Total (ART), Annual Daily Maximum (ADM) and Annual Rainy Days (ARD) respectively in the ORB. Mann–Kendal test recorded an overall non-significant positive trend in the Odaw basin for ART and ARD with an average positive slope of 3.7 mm/year and 0.23 wet days/year and a significant positive trend with slope value of 0.55 mm/year for ADM in the basin. On the monthly scale, a significant upward trend was observed for the dry seasonal months; December, January and February for the Monthly Rainfall Total (MRT) and Monthly Daily Maximum (MDM) rainfall series during the period at all stations. Homogeneity for both annual and monthly rainfall time series were observed over the stations in the basin by the Pettitt, Alexandersson’s SNHT and Buishand’s tests. PCA revealed that the spatial variability of rainfall in ORB is very diverse with 55.9% of the variability located in the middle to highland part while 44.0% of the variation is occurring mostly in the low-lying area at the southeastern part of the ORB. The wavelet analysis also revealed a strong annual periodicity at all the selected stations in the basin. The outcome of this study provides valuable information to formulate adaptation measures through appropriate strategies for managing flood in the study area.