ANALYSIS OF DATA ON PRECIPITABLE WATER VAPOUR OVER NIGERIA USING SATELLITE AND RADIOSONDE OBSERVATIONS

Abstract

Precipitable Water Vapour (PWV), which is a measure of the amount of water vapour in a vertical column, shows spatio-temporal variability in the atmosphere over any location. This research work analyzed the spatial distribution of daily precipitable water vapour (PWV) and temperature, globally and over Nigeria, using daily data covering the years between 2004 and 2011. These were obtained from the archives of the European Union of Meteorological Satellites’ (EUMETSAT’s) Satellite Application Facility on Climate Monitoring (CM-SAF), Offenbach, Germany. The satellite data were those of specific humidity and temperature received at six pressure levels (i.e. 1000, 850, 700, 500, 300 and 200hpa), reclassified into the lower (1000-850hpa), mid-(850-500hpa) and upper (500-200hpa) levels of the troposphere (i.e. UPW, MPW and LPW respectively). Global observations show high values at the tropical region (up to 5g/cm2), with a decreasing trend towards the temperate and polar regions. Over Nigeria, PWV decreases inland from the coastal zone towards the sub-Sahelian zone, while temperature increases from the coastal zone to the sub-Sahelian zone. These patterns are found to be generally applicable in all layers of the troposphere. Autoregressive Integrated Moving Average (ARIMA) technique was used to develop a model on Total Precipitable Water Vapour (TPWV) over Abuja using satellite data (2004-2009). The model in backshift notation is of the form 365 (1 0.51589 )(1 ) tt B B z a , with seasonal ARIMA (1,0,0)(0,1,0)[365]. The model when tested with TPWV data for 2010 and 2011, gave a coefficient of determination (R2) = 0.601, Root Mean Square Error (RMSE) = 0.2869 and Maximum Absolute Error (MAE) = 0.1908 and p-value=0.000. All these values show that the model is significant enough to describe TPWV variations at Abuja. Furthermore, for the purpose of comparison, satellite data obtained in 2007 over Abuja were also correlated with radiosonde data obtained from the African Monsoon Multidisciplinary Analysis (AMMA) Project in 2007 over the same station. The regression parameters gave R2 = 0.188, RMSE = 0.003, p-value = 0.000, t-test = 0.5357 and bias = -0.06 for the upper level PWV; R2 = 0.185, RMSE = 0.016, p-value = 0.000, t-test = -0.7443 and bias = 0.19 for the mid-level PWV; R2 = 0.007, RMSE = 0.014, p-value = 0.583, t-test = -0.6653 and bias = 0.32 for the low level PWV, and R2 = 0.234, RMSE = -0.6799, p-value = 0.000, t-test = -0.4864 and bias = 0.44 for the TPWV. The bias correction between satellite and radiosonde data over Abuja gave 0.003, 0.0164, 0.014 and 0.031 for the upper, mid-, lower and total components respectively. These values indicate that for the mid- and upper layers, satellite and radiosonde observations correlates better than the lower level of the troposphere. These comparisons generally show the low dry bias of the RS-92 at the upper layer and the overestimation of humidity retrieval algorithms of the satellite observation