高时空分辨率微波遥感信息与农作物模型同化研究

Fully understanding the mechanism and process of crop growth and obtaining timely and accurate crop biophysical parameters in the main stages of crop growth are key factors to improving the accuracy of crop yield estimation. The study aims to explore the mechanism of microwave scattering, develop methods of quantitative retrieval of biophysical parameters of crop using remote sensing data, further understand the coupling mechanism of remote sensing data and land surface process model and improve the accuracy of crop yield predictions. .Most paddy rice crops grow in rainy and cloudy environment, for example, southern part of China. The regular acquisition of optical remote sensing data is hampered by frequent cloud cover. Radar can acquire remote sensing information with a high temporal resolution due to its all-weather capability. Crop growth models can be used to estimate crop growth state and yield. But the need of a large number of input parameters and nonoptimal and changing growing conditions hampers the application of such models at large spatial scale and limits the confidence in their output.In this study, with a focus on rice yield estimation in rainy and cloudy areas of Leizhou of Guangdong province in southern China, time series of high temporal and spatial resolution multi-polarization TerraSAR-X data will be collected over the whole rice growth season and the measurements of ground microwave scatterometer are to be made during the dates of acquisition of TerraSAR-X. The dataset will be used to study the X band microwave temporal backscattering characteristics and mechanism of rice fields to develop a new scattering model to retrieve rice LAI (leaf area index) in the whole rice growth season. WOFOST crop model will be employed to study the mechanism of the process of rice growth and simulate the dynamic changes of biomass and LAI of rice in study areas. The model will be initialized using ground measurements, including soil properties, weather data, genetic characteristics of the rice and rice management information. Based on the above results and analysis of several data assimilation methods, A new method is to be developed to assimilate rice LAI retrieved from TerraSAR-X data into WOFOST crop model with the objective to improve the accuracy of the rice yield estimation at catchment scale. The assimilation process is to re-initialize the crop model by revising the above selected input parameters to minimise temporal difference between the rice LAI simulated by the crop model and the rice LAI from TerraSAR-X. .The research will prompt the application of microwave remote sening, help better understand the significance of the assimilation of remote sensing data in crop model and provide an important support for the national agriculture and regional sustainable development.

对农作物生长机理和过程的充分了解以及在农作物生长期及时准确地获取农作物生物物理参数信息是提高农作物估产精度的关键问题。本研究以华南多云雨地区水稻估产为研究重点,选择广东省雷州半岛水稻产区作为典型样区,通过地表散射计测量和高时空分辨率多极化微波遥感TerraSAR-X数据研究水稻的时域X波段微波后向散射特征和机理，建立散射模型定量反演水稻整个生长期的叶面积指数；并利用农作物生长模型研究水稻生长过程机理和水稻生物量及叶面积指数的动态变化。在此基础上进一步研究微波遥感信息和农作物模型的同化方法,将微波遥感信息与农作物模型耦合，以获取高精度的水稻估产结果。本研究旨在开发高时空分辨率微波遥感数据定量反演水稻生物物理参数的方法，探索微波散射机理和遥感信息与陆表过程模型的耦合方法，促进高时空分辨率主动微波遥感在南方多云雨地区中的应用，提高作物估产精度，为国家农业和区域可持续发展提供重要的科学依据。

提高农作物估产精度对于中国农业政策的制定，粮食价格的宏观调控以及国际国内粮食贸易具有很重要的意义。主动微波遥感具有不受云雨影响，全天候、全天时监测的优点，逐渐成为南方水稻监测和估产重要数据源。但遥感信息往往反映的只是作物表面瞬间物理状况，还不能真正揭示作物生长发育和产量形成的内在机理、个体生长发育状况及其与环境气象条件的关系。一些农作物生长模拟模型在国内外已经被广泛用来预测农作物的生长状态和估算农作物的产量。但是随着农作物品种的变化和种植面积的扩大，尤其是在非理想种植条件下，这些模型应用的精度受到了很大影响。本研究以华南多云雨地区水稻估产为研究重点,选择广东省雷州半岛水稻产区作为典型样区,获取了研究区农作物的多个生长周期的生物物理参数，在此基础上调整现有农作物模型对研究区农作物生长过程进行模拟。通过分析高时空分辨率多极化微波散射机理和农作物的时域微波后向散射特征，进一步理解了微波极化散射过程，同时研发了研究区农作物分类方法和建立散射模型对农作物种植面积和农作物生长期的参数进行了估算。在此基础上进一步研究微波遥感信息和农作物模型的同化方法,将微波遥感信息与农作物模型耦合，以获取高精度的水稻估产结果。本研究结果有助于深入理解高时空分辨率微波散射机理，提高微波遥感定量反演农作物生物物理参数的精度，进一步探索微波散射机理和遥感信息与陆表过程模型的耦合方法。研究结果也可促进高时空分辨率主动微波遥感在南方多云雨地区中的应用，提高作物估产精度，为国家农业和区域可持续发展提供重要的科学依据。