Reconstruction Of Ultraviolet Radiation Using A Radiative Transfer Model

In recent years, the depletion trend of stratospheric ozone layer has been monitored either in the polar regions or the midlatitude of the northern hemisphere. This decrease of ozone amount would lead to a significant increase in the ultraviolet (UV) radiation reaching the earth’s surface. Solar UV radiation plays a major role on human’s health due to its strong biological, environmental and climatic effects. There is also a potential threat to other animals and plants and even the whole ecosystem. Thus, to make the research on UV radiation has gradually become important highlights in current society. However, the observations of UV radiation are almost started later compared to the corresponding global solar radiation (G). The large-scale, long-term continuous UV values are more scarce. Due to the urgent demand of prolonged UV radiation data, the reconstruction of UV values has been proposed and widely applied by both domestic and foreign scholars.A semi-empirical reconstruction method for UV radiation is selected in this paper by calculating the cloud modification factors. The G and UV radiation under clear-sky conditions is simulated by a radiative transfer model, SMARTS. A comparison between the results of reconstruction and measurement of UV radiation shows that, the reconstruction method offers high resolution and accuracy, and can be well applied to the reconstruction of historical U V radiation data. In addition, due to the diversity of ecosystems and climate characteristics in most regions of our country, there are9stations such as Beijing, Hailun, Fukang, Eerduosi, Taihu, Minya Konka, Qianyanzhou, Dinghu Mountain and Sanya bay were selected in our studies. The main results are following as below:(1) The variation characteristics of inputs for the model shows that, W, namely the the atmospheric precipitable water, which has a significant seasonal variations with the maximum annual values appeared in summer (mostly in July) while the minimum values occurred in winter (mostly in January). There are smaller values of W occurred in Fukang and Eerduosi compared with another stations, and the larger values observed in Dinghu Mountain and Taihu. The daily total column ozone (O3) at the high latitudes was significantly greater than the low latitudes, and a different seasonal variation was also appeared, the maximum O3at the high latitudes always appeared in spring and minimum occurred in autumn, but at the low latitudes are summer and winter, respectively. The aerosol optical depth (AOD) mostly required from the satellite remote sensing results, MODIS. The AOD values are smaller in Sanya bay and Minya Konka while larger values observed in city-sites such as Beijing and Taihu.(2) Values of AOD, O3, W and surface albedo were selected to give a sensitivity test, which results indicate that the sensitivity of UV radiation on AOD and surface albedo variation is superior to it on O3and W variation, UV radiation reaching the earth’s surface is mainly affected by the impact of aerosols.(3) The measurements of G and UV radiation for five or six years were analyzed in these stations. The daily G radiation is substantially maintained at0～30MJ·m-2d-1and UV radiation at0～1.5MJ·m-2·d-1,the ratio of UV radiation to G radiation (UV/G) are always within2.5%-6%. In addition, both of them have a consistent seasonal variation of maximum values appeared in summer while minimum values occurred in winter.(4) The G and UV radiation under cloudless-sky conditions were simulated by a radiative transfer model, SMARTS. By way of a comparison with measurements, the simulated G and UV radiation are higher and have smaller dispersion, this mainly due to the ignorance of cloud radiation effects. But compare to G, the difference between simulated and measured UV values are larger and more discrete.(5) The cloud and aerosol attenuation ratio of G and UV radiation are defined respectively in this paper using measured and simulated values. The maximum cloud attenuation ratio of G radiation occurred in Minya Konka with a average value of32.8%, and the remaining stations are between10%to20%. The cloud attenuation ratio of UV radiation in Minya Konka are also maximum with a average value of45.9%, Beijing and Hailun are followed with average value of38.9%and39.4%, the minimum value occurred in Taihu with11.9%and the remaining stations are between20%to30%. Different with cloud attenuation ratio, the aerosol attenuation ratio of UV radiation in all stations are larger compared with G radiation. The average aerosol attenuation ratio of G radiation has a decreasing order of Taihu (27%), Dinghu Mountain (23.1%), Qianyanzhou (18.4%), Beijing (18.2%), Fukang (15.2%), Eerduosi (12.7%), Sanya bay (10.9%), Hailun (8.1%) and Minya Konka (5.4%). Similarly, the average aerosol attenuation ratio of UV radiation is decreased as Taihu (41.4%), Dinghu Mountain (35.9%), Qianyanzhou (29.5%), Beijing (27.6%), Fukang (24.8%), Eerduosi (20.7%), Sanya bay (18.4%), Hailun (13.1%) and Minya Konka (9.2%).(6) A semi-empirical reconstruction method of UV radiation is established using the cloud modification factors. A comparison between the results of reconstruction and its corresponding measurement show that, all of the mean absolute bias error (mabe) and root mean square error (rmse) are less than20%, and even most stations are less than10%. There is no obviously overestimate or underestimate occurred in Beijing, mabe is7.70%and rmse is9.74%. The UV radiation reconstructions in Fukang have a slight underestimate, but only5.3%of mabe and8.26%of rmse are given. The mabe and rmse in Eerduosi is8.12%and9.62%while10.96%and12.72%in Taihu,7.96%and9.88%in Minya Konka,6.28%and7.86%in Qianyanzhou,9.48%and12.09%in Dinghu Mountain and4.67%and5.94%in Sanya bay, respectively. Hailun have a highest mabe and rmse compared to other stations, but only14.68%and17.51%, respectively.Thus, a high accuracy and applicability in the semi-empirical method were confirmed.