基于DEM的天山山区气温和降水序列推算方法研究

Temperature and Precipitation Series in the Tianshan Mountains Calculated Based on DEM

为了计算天山山区区域气候要素平均值序列,提出了一个新的气候序列计算方案.对天山山区17个气象站和10个水文站的1961—2000年的年气温、降水资料进行了自然正交分解(EOF),并以DEM(Digital Elevation Model)的1 km×1 km网格数据为基础,结合多元回归等方法,分别建立了前3个特征向量与经度、纬度及海拔高度因子的插值模型,由此推算出天山山区(海拔≥1 500 m)年平均气温、降水序列.误差分析表明,27站实测的年平均气温序列与计算的区域平均序列值的相关系数为0.996,系统性偏差小,平均相对误差为5.5%;年降水量序列的实测值与计算值平均相对误差略大,为14.8%,相关系数为0.972.区域平均序列的计算值与测站简单的算术平均序列在量值上存在明显的差异,计算出的气温平均偏低4.3℃,降水量平均偏高43.2 mm.此方法为计算站点稀少、地形复杂的区域平均的要素时间序列方面提供了一个解决手段.

A new interpolating scheme is applied to calculate the time series of temperature and precipitation in the Tianshan Mountains. The data of 17 meteorological stations and 10 hydrological stations in 1961—2000 in the Tianshan Mountains are used. An interpolating model based on DEM （Digital Elevation Model, the spatial resolution is about 1 km× 1 km） are formed, respectively, by using empirical orthogonal function （EOF） and polynomial regression. Time series of annual mean temperature and annual precipitation on each grid in the area higher than 1 500 m a. s. l. are calculated by using the interpolating model. According to an error analysis, the correlation coefficient of the annual mean temperature between the average value observed from the 27 stations and the area average value calculated by each grid is 0. 996, and the systematic error is small, with an average relative error of 5.5%. For annual precipitation, the average relative error is 14.8%, which is higher than that of the temperature, the correlation coefficient of annual precipitation series is 0.972. There is different between the value calculated from the time series and the observed mean value from the 27 stations. As compared with the observed mean value, the calculated annual mean temperature is 4.3℃ lower and the calculated annual precipitation is 43.2 mm higher. This method is useable to obtain an area mean climate time series in mountain area with complicated terrain and less observation stations.