Diurnal distribution of absolute humidity over West Africa using data retrieved from era-interim

Abstract

The variations of absolute humidity (ρ) have been studied using yearly averages of air and dew point temperature data taken at twenty five selected weather stations which were further grouped into four climatic regions (Coastal, Derived, Guinea and Sahel) over West Africa for the years 2005 to 2009 retrieved from ERA-Interim. It uses a December 2006 version of the European Centre for Medium- Range Weather Forecast (ECMWF) Reanalysis Integrated Forecast Model (IFS Cy31r2) which covers dates from 1 January 1979 to present, at spectral resolution of about 80km. It was observed that over all the stations and regions, ? shows strong diurnal variations.Here, ? decreases from 00:00hr in the midnight to 06:00hr in the morning. This is then followed by a sharp increase to 18:00hr in the evening at all the stations in both the dry and wet seasons. For all stations and regions, ? is minimum at 06:00hr and maximum at the midnight or evening period. Considerable seasonal variation was also observed as ? was highest in the wet season months but low in the dry season months.The variability of absolute humidity was highest in the dry season months and lowest in the wet season months at 00:00hr, 06:00hr and 18:00hr but reverse is the case at 12:00hr. Variability at the Coastal region is uniformly low throughout the years under consideration while variability at the Sahelian region is high. Using Multivariate Linear Regression (MLR) modeling technique, multiple linear regression models were developed for the stations and regions. The statistical indicators such as Coefficient of determination (R2), Mean Percentage Error (MPE), Mean Bias Error (MBE), and Root Mean Square Error (RMSE) at 95% confidence level for R2 and desirable lower values of MPE, MBE and RMSE were calculated to monitor the efficiency of the developed models. Results show that the models are adequate to predict ρ at most of the stations and regions.

Keywords: Diurnal Variation, Reanalysis, Forecast, Multivariate Linear Regression, Absolute Humidity