Developing Regression Models for Estimating Atmospheric Visibility over Ikeja, Nigeria

Abstract

In this study, the meteorological parameters of measured monthly average daily visibility, mean temperature, relative humidity, atmospheric pressure, rainfall and wind speed covering a period of 12-years were utilized to develop and evaluate one, two, three, four and five variable correlation models for the estimation of atmospheric visibility for Ikeja, Lagos State, South Western, Nigeria (Latitude 6.58°N, Longitude 3.33°E and altitude 40 m above sea level). The accuracy of the newly proposed models for all the variable correlations are statistically tested using correlation coefficient (R), coefficient of determination (R2), Mean Bias Error (MBE), Root Mean Square Error (RMSE), Mean Percentage Error (MPE) and t-test. Based on the overall results, the four variable visibility models relating the relative humidity, atmospheric pressure, rainfall and wind speed was recommended having the highest values of R (98.9%), R2 (97.9%) and the lowest values of RMSE (0.1821 km) and MPE (0.4025%). The one variable visibility model relating the rainfall was recommended based on the lowest values of MBE (0.0012 km) and t-test (0.0038). The results of this study revealed that despite the very high values of R and R2 for the five variable correlation model, it is not appropriate for visibility estimation as it has very high values of MBE, RMSE, MPE and t-test as compared to other estimated models with high overestimation of the measured visibility values. The newly recommended visibility models can be used for estimating daily and monthly visibility with higher accuracy for Ikeja and regions with similar climatic information.