多源遥感高分辨率时空连续地表实际蒸散发算法研究

The spatial-temporal information on the land surface actual evapotranspiration (ET) with high resolution is essential to the global change study and field water resource management application, while remote sensing science provide the most powerful method to monitor the spatial and temporal variation of ET. However, some challenges still exist in the ET estimation based on remote sensing after decades development, e.g. the spatial-temporal gaps due to the cloud impact, the lack of ET data both with both high spatial and temporal resolution. Thus, current study is contributed to develop ET data fusion method to fuse ET with different resolution, obtained from multiple satellite remote sensing data, to generate ET data with both high spatial and temporal resolution. To consider the impact of precipitation and irrigation on water use and improve the accuracy of ET time series, the root zone soil moisture will be modeled based on the water balance method with the assimilation of microwave remote sensing retrieved surface soil moisture and optical-infrared remote sensing retrieved evaporation ratio by data assimilation strategy. The root zone soil moisture will be estimated at coarse spatial resolution and high temporal resolution, which is further applied to capture the variation of soil water stress on ET. The coarse spatial resolution ET time series generated in this way will be applied in data fusion to generate spatial-temporal continuous ET both with both high spatial and temporal resolution, which is able to catch the impact of precipitation and irrigation. To improve the accuracy of fused ET in the heterogeneous region, an ET downscaling method based on high resolution remote sensing data will also be proposed and applied. The proposed method will applied and evaluated in the Heihe river basin located in the arid and semi-arid land of Northwestern China. The proposed study is helpful to quantitatively estimate land surface evapotranspiration, which can meet the national needs for climate change adaption, drought monitoring, and water resource management.

获取精确的高分辨率蒸散发时空分布变化是全球变化和田间水资源管理等重大科学和应用需求中迫切需要解决的关键科技问题，遥感科学的发展为其提供了最有效的手段，然而目前遥感估算蒸散发时空格局信息存在时空连续性不足和不能兼顾高时间与高空间分辨率等问题。本研究拟融合不同时空分辨率遥感估算蒸散发，改进异质性地表蒸散发融合精度方案，并结合水量平衡模拟和数据同化提高蒸散发时间序列估算精度，发展时空连续高时空分辨率遥感蒸散发时间序列获取算法，从而提高遥感获取高分辨率蒸散发时空分布信息的能力和精度。该项目研究有助于提高地表水热通量定量遥感研究水平，满足国家气候变化应对、干旱监测、水资源管理等方面的需求。

获取精确的高分辨率蒸散发时空分布变化是全球变化和田间水资源管理等重大科学和应用需求中的关键问题，遥感科学的发展为其提供了有效手段，然而目前遥感估算蒸散发时空格局信息存在时空连续性不足和不能兼顾高时间与高空间分辨率等问题。本研究通过蒸散发遥感模型改进、高时空分辨率算法发展等，提高了蒸散发遥感估算能力和时空分辨率，满足国家气候变化应对、干旱监测、水资源管理等方面的需求。在蒸散发模型方面，以ETMonitor模型为基础，开展提高公里级分辨率蒸散发遥感算法研究，包括：利用多源遥感获取的地表植被变化和降水特征，考虑降雨截留过程中主控因素的变化，优化湿冠层蒸发速率等参数化方法，提高冠层降雨截留蒸发（蒸散发分量）估算精度；构建土壤水分胁迫指数，结合晴天地表温度-植被指数特征空间方法估算土壤水分对蒸散发的胁迫作用，建立适用于全天候的地表冠层阻抗估算方法，提高蒸散发（尤其是植被蒸腾）估算精度；引入水量平衡公式模拟土壤水分动态，通过数据同化减小模型误差累积，提高蒸散发估算精度。基于黑河流域通量观测数据验证表明，改进后算法精度得到了较大提高。在高时空分辨率蒸散发估算方面，发展了一套蒸散发时空数据融合方案，基于灵活的时空数据融合方法，采用潜在蒸发比作为融合变量，融合ETMonitor估算全天候时空连续逐日1公里分辨率蒸散发数据和基于Landsat估算的晴天30米蒸散发数据，实现异质性地表逐日30米蒸散发估算，同时能够消除高低分辨率数据融合产生的“块效应”。采用该方法获得了黑河中游及异质性更强的下游绿洲研究区30米分辨率蒸散发，与站点观测具有较好的一致性，取得了良好效果。此外，本项目评估了微波土壤水分遥感产品对蒸散发估算的影响等，发现国产风云卫星土壤水分产品能较好的应用于蒸散发遥感估算，研发了基于国产风云卫星数据的蒸散发算法，对推动国产卫星在水资源管理方面的应用具有积极作用。

内容获取失败，请点击重试