Modelling Ultraviolet Irradiance in South Africa

Abstract

South Africa (SA) is a country with a variety of climatic regions and topological diversity. The southernmost town in SA, Cape Agulhas (34 S, 19 E), is on about the same latitude as Perth and Sydney in Australia. UV indices greater than 10 are common in SA. This high level of UV radiation potentially causes many health problems resulting in high rates of skin cancer, eye disorders etc. A method is presented for inferring a level of UV irradiance from imprecise measurements. The method uses nine measured or estimated variables to infer the UV index. It employs a system of five artificial neural networks to convert the information contained in measured/estimated data into the UV index. The results obtained are of considerable statistical significance. It should be mentioned that the other statistical techniques used, such as linear and/or non-linear regression, did not produce satisfactory results.