Urbanization Effect On The Changes Of Extreme Temperature In East Asia

Monitoring and detection of climate change are inseparable from the reliable long-term temperature data series. At present, urbanization effect has become one of the main source of systematic error in land temperature observation data. Regional studies show that the surface air temperature data series are significantly affected by the increased urban heat island effect, and obvious systematic error exists in the current datasets used for studies. The urbanization effect on the trends of extreme temperature series in regions of rapidly economic development in East Asia has become a key scientific problems in the study of the regional temperature change. The paper analyzes the urbanization effect on the trend of extreme temperature indices based on the daily maximum temperature (Tmax) and minimum temperature (Tmin) in East Asia.During1951-2011, Tmin, Tmax and mean temperature (Tmean) significantly increase, and the trends of annual mean Tmin, Tmax and Tmean are0.24℃/10yr,0.19℃/10yr and0.21℃/10yr respectively in East Asia. The trend of diurnal temperature range(DTR) significantly decreases in the region. The rising trend of Tmin is greater than the trend of Tmax, resulting in the decrease of DTR. Days of cold night(Tn10p) and cold days(Tx10p) significantly reduce, and the trends are-3.92d/10yr and-1.84d/10yr respectively. The days of warm night(Tn90p) and warm days (Tx90p) significantly increase, and the trends are5.66d/10yr and3.61d/10yr. The trends of all the four extreme temperatures indices pass the0.01level significance test.The urbanization effects on the trends of annual Tmin, Tmax and Tmean are large and significant in East Asia during the period analyzed, and they all pass the0.01level significance test, with the magnitudes being0.05℃/10yr,0.04℃/10yr and0.04℃/10yr respectively. The urbanization effect for annual mean DTR is-0.02℃/10yr, statistically significant at the0.01level significance. The urbanization effect for Tmin is greater than that of Tmax, and this leads to a large urbanization effect in the DTR series. The urbanization contributions of annual Tmin, Tmax, Tmean and DTR are20.83%,21.05%,19.05%and33.33%respectively in East Asia. The urbanization effect for annual Tn10p and Tn90p are1.74d/10yr and1.33d/10yr respectively, and all the effects pass the0.01level significance test. The urbanization contributions of annual TnlOp and Tn90p are44.39%and23.50%. The urbanization effects for annual Tx10p, Tx90p respectively are0.34d/10yr and0.13d/10yr, and the urbanization contribution for Tx10p is18.48%. For Huairou station, the urbanization effects on minimum and maximum temperatures are-0.004℃/10a and-0.035℃/10a respectively in original temperature data, however, the urbanization effects increase to0.388℃/10a and0.096℃/10a in adjusted temperature series. Therefore, the homogenizing adjustments to the minimum and maximum temperatures at Huairou station solve the inhomogeneity problem, but the urbanization effects on the maximum and minimum temperatures are recovered, particularly for the minimum temperature.At Beijing station, urbanization effect results in a significant reduction in annual frost days, cold nights and DTR, and a significant rising of annual warm nights and mean minimum temperature during1960-2008. Urbanization effects on the trends of the indices series of Beijing Station relative to daily minimum temperature are all significant at the0.01confidence level. The urbanization effects on trends of the indices series of mean minimum temperature and mean DTR are the largest in wintertime, and the urbanization effects on trends of the indices series of warm nights occur mainly in summertime.At Shijiahuang station, the urbanization make the distribution of annual Tmean, Tmin and Tmax during1962-2011shift to high temperature side. Urbanization effects on annual Tmin distribution is very obvious, but the effects on annual Tmax distribution is very small. The increase of high value parts of Tmin is larger than low value parts, and this makes the shape of annual Tmin PDF at the urban station more flat than the those of the rural station. The difference of trends of the extreme temperature indices based on daily Tmin including cold nights and warm nights between urban and rural stations caused by urbanization is larger than the difference of the trends of the extreme temperature indices based on daily Tmax including cold days and warm days.