基于MODIS LST产品估算青藏高原地区的日平均地表温度

Estimating Mean Daily Surface Temperature over the Tibetan Plateau Based on MODIS LST Products

地表温度是衡量地-气界面热状况的重要指标,在冻土制图、城市热岛效应等研究领域有重要应用.MODIS数据只提供了每日4次过境的瞬时温度值,而实际应用中往往需要日平均地表温度.对实测0cm温度数据进行分析后,发现其日变化曲线呈分段函数特性,白天按正弦曲线变化,夜间则呈线性下降.据此提出对白天和晚上的数据进行分段拟合的方案,并与已有的两种拟合方法作对比验证.地面验证工作包括两个方面:一是模拟的验证,按卫星过境时间从逐小时的实测0cm数据中挑出4个瞬时值,将上述拟合方法应用于这4个瞬时值,得到拟合的日平均温度,再同24h的算术平均值比较.结果表明:Sin-Linear法精度最高,平均误差在1K以内;其二是地面对MODIS的验证,将3种方法应用于MODIS数据,再与地面平均值比较,3种方法中Sin-Linear法误差最小,且具有最好的线性相关,R2超过0.9.

Land surface temperature（LST） is one of the key factors that indicate the thermal state of the land-atmosphere boundary.It has frequently been applied in many fields such as permafrost mapping and analyzing urban heat island effect.The MODIS LST products merely provide four instantaneous observations every day,whereas,the daily mean LST is often required in practice.In this study,a piecewise function is used to figure out the diurnal circle of ground temperature at the surface.In daytime,the surface temperature takes on the form of a sinusoidal function,while in nighttime it declines almost linearly.Furthermore,two existing methods are reviewed and one new approach（Sin-Linear） is proposed to transfer the instantaneous MODIS LST data to daily mean LST.In order to validate the new approach,four instantaneous temperature data are selected from the hourly observed 0 cm temperature according to the overpass time of the satellites.Then the three methods are applied to estimate the daily mean LST,which are compared with the arithmetic averaged data of the hourly temperature records.It is found that the Sin-Linear method proposed here has the highest precision,with annual mean error within 1 K.On the other hand,the three methods are applied to the MODIS LST data,and the results are compared with the ground daily averaged data（0 cm）.The Sin-Linear method indicates the best consistency with the ground observation,with R2 greater than 0.9.