遥感模型耦合数据同化方法的北方典型农田陆面过程模拟

Timely, accurate estimation of land surface heat flux is the key issue in solving the problem of low agriculture water resources usage in China. The study focuses on land surface data assimilation method coupled with remote sensing evapotranspiration model, taking Hetao irrigation area located in Inner Mongolia as research region. The remote sensing evapotranspiration model based on satellite data is first used to estimate latent heat flux at satellite over-passing time and is validated by land observation at pixel scale. Then the instantaneous latent heat flux is extrapolated to daily evapotranspiration, and long term continuous evapotranspiration at regional scale is simulated based on the remote sensing model and extrapolation method. Meanwhile, distributed land surface model coupled with crop growth model is calibrated to continuous simulate the regional fluxes, water balance and yield. In order to further improve the accuracy of long term simulation of land surface energy fluxes, the data assimilation algorithm based on ensemble Kalman filter technique is assembled to the land surface model. A sensitivity analysis at pixel scale is discussed in the first place in order to choose the appropriate modeling and observation errors as well as ensemble size. On that basis, the validated data assimilation method is assembled to the land surface model. The remote sensing models including evapotranspiration model and soil moisture model is used as observation value in the data assimilation system to update model states such as soil water contents and surface temperature, etc. to evaluate the improvement of data assimilation system in estimating land surface heat fluxes, which will lead to scientific support in drought forecast, irrigation management and comprehensive utilization of water resources.

我国农业水资源利用效率较低，解决这一问题的关键基础性研究内容在于及时准确地估算陆地表面水热通量时空分布。本项目以内蒙古河套灌区为研究对象，开展基于遥感模型的陆面过程数据同化方法研究。首先构建基于卫星遥感数据的遥感蒸散发模型模拟瞬时潜热通量，并在象元尺度利用地面观测值对模型进行验证。随后对模拟的瞬时潜热通量进行时间尺度扩展，在区域尺度进行蒸散发长系列连续模拟。同时，构建遥感数据驱动的分布式陆面过程模型并耦合作物生长模型，在灌区尺度对水分、能量、产量进行连续模拟。在此基础上，构建数据同化算法，在点尺度验证同化方法并进行敏感性分析，提出选择最优同化方案的依据。将验证后的数据同化方法集成到陆面过程模型中，结合遥感蒸散发模型和遥感土壤含水量模型对陆面过程模型的状态变量进行实时优化校正，分析数据同化方法对提高地表水热通量模拟精度的作用，为干旱预测、灌溉管理及水资源综合利用提供科学支撑。

在我国水资源总量紧缺、用水缺口加大的背景下，如何在保证粮食安全的前提下最大程度地减少农业用水量，提高农业灌溉效率是农业节水和水资源可持续利用所面临的关键问题。解决这一问题的关键基础性研究内容在于及时准确地估算陆地表面水热通量时空分布。本项目以我国西北干旱-半干旱地区内蒙古河套灌区为研究对象，开展基于遥感模型的陆面过程数据同化方法研究。首先在农田多尺度作物耗水量及土壤墒情观测的基础上，研究利用卫星遥感数据反演作物耗水及土壤墒情时空分布的方法，实现作物耗需水量时空格局的连续模拟；随后构建数据同化方法对遥感蒸散发模型模拟的潜热通量和土壤水模型模拟的土壤含水量进行同化，获取高精度、高时空分辨率的河套灌区水热耦合的实时模拟信息；在此基础上，分析研究区基于作物实际耗水量的灌溉水利用效率变化，分析复杂作物种植结构条件下农田耗水的时空格局及年内变化规律。项目研究成果可为干旱预测、灌溉管理及水资源综合利用提供科学支撑。