ANALYSIS OF NIGERIAN HYDROMETEOROLOGICAL DATA

Abstract

Rainfall and runoff like most hydrologic events are governed by the laws of chance; hence, their predictions cannot be done in absolute terms. Since there is no universally accepted method for determining the likelihood of a certain magnitude of rainfall or runoff, common probabilistic models were used in this research to predict the magnitude and frequency of their occurrence. Missing records were determined by the mass curve analysis for rainfall and regression analysis for runoff involving runoff data at neighbouring site. Tests on time homogeneity, showed that the annual rainfall records at Port Harcourt, Enugu and Lokoja were stationary and random, the annual runoff records of River Niger at Onitsha, Lokoja and Baro were non-stationary, showing a decreasing trend of mean annual runoff. Various models were tested for suitability in predictions of annual rainfall at Port Harcourt, Enugu and Lokoja, also for annual runoff of River Niger at Onitsha, Lokoja and Baro. The mean annual rainfall was found to diminish from the coast to inland with values' of 2, 400, 1, 700, and 1,200mm for Port Harcourt, Enugu and Lokoja respectively. The mean annual runoff for River Niger at Onitsha, Lokoja and Baro were 117, 000, 169, 300, and 60, 525 Mm³ respectively. The application of the models showed that the lognormal distribution should be adopted for predictions of annual rainfall at Port Harcourt and Lokoja and annual runoff of River Niger at Onitsha and Lokoja, The normal distribution should be adopted for predictions of annual rainfall at Enugu and annual runoff of River Niger at Baro. However, for a return period of 5 years, the annual rainfall  appear to be best described by the Pearson type III and log Pearson type III distributions.