Distribution analysis and autoregressive modelling of ultraviolet radiation over Akure,Nigeria

摘要

Management of health risks associated with excessive exposure to ultraviolet radiation involves understanding its characteristics within any location. This work employed five-year archived data of UV index for analysis and autoregressive modelling of ultraviolet radiation over Akure (7.15°N, 5.12°E), Nigeria. In-situ measurements of UV index were made every day between January 2007 and December 2011 at 30 min interval using Davis 6162 vantage Pro2 weather station. Prevalence of high intensity UV index, which indicates human susceptibility to UV-related health risks was investigated. The statistical model that best describes UV distribution and its autoregressive characteristics were also determined for the location. Annual UV index was found to fit a Nakagami distribution and well modelled by a 3rd order polynomial equation to at least 95% accuracy. Non-linear autoregressive (NAR) artificial neural network (ANN) analysis also returned regression coefficient values of 0.95, 0.94 and 0.94 for each of training, validation and test parameters respectively.