天山中段近地表气温的遥感反演及时空分布特征

Remote sensing inversion and spatio-temporal distribution of near-surface air temperature in the middle Tianshan Mountains

为天山中段近地表气温研究提供可靠、空间分辨率较高的数据源,该文以气象站气温数据、再分析气温数据以及遥感数据为主要数据源,采用多元回归模型反演了近地表气温并进行精度验证。在此基础上,揭示天山中段近地表气温时空分布特征及其变化趋势。结果表明:①反演精度与实测值相近,且精度较好。②天山中段近地表气温呈四周高,中部低,平原区高,丘陵、山区低的空间分布格局;年际变化呈先升高后降低,年内呈周期性单峰变化的趋势。③2001-2016年天山中段近地表气温以不显著增加为主,占显著降低的区域甚少。研究结果证明用多元回归模型可以提高近地表气温的遥感反演精度,为今后近地表气温研究提供空间分辨率较高的数据及其获取方法。

In order to provide a reliable and spatially resolved data source for the near-surface air temperature study in the middle Tianshan Mountains,this paper used the monthly average temperature data of the weather station,the reanalysis temperature data and the remote sensing data as the main data sources,and used the multiple regression model to invert near-surface air temperature. Based on the near-surface air temperature and accuracy verification,the spatial and temporal distribution characteristics of the near-surface air temperature in the middle Tianshan Mountains and their changing trends were analyzed. The results showed that:①the inversion accuracy was close to the measured value,and the accuracy was better. ②The near-surface air temperature in the middle Tianshan Mountains was high in the surrounding area,low in the middle,high in the plain,low in the hills and mountains. The interannual growth was first increased and then decreased,and the cyclical single peak changed during the year. ③In the past 16 years,the near-surface air temperature in the middle Tianshan Mountains had not increased significantly,and the area with a significant decrease had been rare. It is feasible to use the multivariate regression model to invert the near-surface air temperature,and provide a spatially high-resolution data source for the near-surface air temperature study.