VARIATION OF PRECIPITABLE WATER VAPOUR OVER NIGERIA

Abstract

Precipitable water vapour is the amount of water vapour in the atmosphere contained in a vertical column of unit cross-sectional area, extending between any specified two levels. The monthly surface data of relative humidity and air temperature taken over Nigeria between 1970 and 1994 were used to develop models for the twenty-one selected stations, four climatic regions and Nigeria while the models developed were validated using surface data of relative humidity and air temperature taken over Nigeria from 1995-2000. In addition, the efficiency of the models developed was assessed using statistical indicators. The stations were evenly distributed across the country with five from the Coastal and two from the Derived Savannah region, seven each from the Guinea Savannah and Sahelian regions. The monthly mean distribution of precipitable water vapour (PWV) shows double peaks in February and November with a dip in July/ August in all the stations at the Coastal and Derived Savannah region and a single peak, lasting from May to September/October in all the selected stations at the Guinea Savannah and Sahelian regions. Analyses have also shown that there is an existence of strong correlation between precipitable water vapour and the two meteorological parameters at the selected stations. Results have shown that all the multiple linear regression models were statistically significant at all the stations at 0.05 alpha level. The results of the accuracy test carried out using statistical parameters, namely: mean bias error (MBE), root mean squared error (RMSE), mean percentage error (MPE), coefficients of determination (R2) and t-statistics (t-test) also showed the performance and significance of the models developed for the twenty one selected stations, the four climatic regions and Nigeria, MBE values (0.026), RMSE values (0.064), MPE values (-0.542), R2 values (0.984) and t-test values (0.460) which fall on the desirable values. The scattergram of the linear relationship between actual and models ‘calculated values for all the stations and regions, showing positive slopes and high R2 values. , this implies that the variation of PWV over Nigeria is well monitored. The models developed for the twenty one selected stations, the four climatic regions and Nigeria gave good performances according to their MBE and RMSE which are small and positive