Adjustment And Evaluation Of Urbanization Bias In Monthly Mean Surface Air Temperature Dataset Over Mainland China

The reference stations network of surface air temperature in Mainland China was improved in this study, and a method for urban-bias adjustment for monthly mean surface air temperature of national stations observation network was also developed. Applying historical temperature dataset of updated reference stations network and the renewed urban-bias adjustment method, the urban-bias of monthly surface air temperature of519urban stations in the national stations observation network was evaluated and adjusted for time period1961-2010, and the adjustment method was examined by existing research of a few case urban stations. Then it was contrasted and analyzed that the means, variances and linear trends of519urban stations before and after urban-bias adjustment. At last, a monthly mean temperature dataset after urban-bias adjustment in the mainland of China was established.There were two factors to be considered in building reference stations of corresponding urban station, which one is the distance between urban and rural stations and the other is the consistency of their natural climate characteristics. The single temperature series of urban station was weighted by original temperature series of corresponding reference stations according to their correlation coefficients which was removed linear trend. And the linear trend difference between urban and reference series was regarded as adjusted value to verify the warming trend of urban stations.The evaluation results for the adjustment show that most urban stations witness a significant urbanization effect on the trends of annual mean surface air temperature especially in North China, Central China and coastal areas of South China. Among the stations with positive urbanization bias, about78%stations pass the significance test at the0.01confidence level. There are74stations with negative urbanization bias, which accounts for14%of total urban stations, and they are mainly distributed in the northwest arid region, southwest and south of northeast. Among the stations with negative urbanization bias, nearly57%stations do not pass the significance test at the0.01confidence level. By examining the urban bias adjustment results for a few case urban stations, we are able to show that the adjustment method developed and the adjustment results in this study are reasonable, and they can be applied in the studies of regional climate change.The contrast of the means, variances and linear trends of519urban stations before and after urban-bias adjustment shows that most of the stations change with increasing of means, reducing of variances and decreasing of linear trends after adjustment, which is according with the distribution characteristic of urbanization bias in Mainland China.It is established a monthly mean temperature dataset with the urban-bias adjusted in the mainland of China, which consists of589national stations. The trend of annual mean surface air temperature in Mainland China is0.235℃/10a during the period1961-2010. During the trends of four seasons, the biggest value appears in winter with an increasing rate of0.337℃/10a, the smallest value appears in summer with an increasing rate of0.159℃/10a. In the spring and autumn times, the linear trends are0.206/10a and0.227/10a respectively.