Estimation of daily ultraviolet radiation in Beijing using a semiempirical method

This study proposes a semiempirical method to reconstruct daily ultraviolet (UV) radiation from global solar (G) radiation measurements using a radiative transfer model. The attenuation ratio and cloud modification factors are calculated based on measured and simulated data under cloudless-sky conditions. A reconstruction method of UV radiation is established using cloud modification factors; based on comparisons among reconstructions and measurements, the reconstruction model is demonstrated to offer high resolution. The bias errors for daily measured and reconstructed UV radiation are maintained within ±20%, the mean absolute bias error (mabe) is 7.7% and the root mean square error (rmse) is 9.7%. Furthermore, the model performance and transferability were tested by comparison with a simple empirical model in Beijing, Eerduosi and Hailun. A comparison of the measured and estimated UV values for the two methods in the aforementioned three locations revealed that smaller mabe and rmse were observed in our method, with both of these values in the three locations being less than 14%. Thus, a better applicability and transferability has been confirmed. The results and analysis should contribute to improving the knowledge about actual UV climate characteristics.