多时相极化SAR与作物生长模型耦合的区域水稻产量差当季估算方法研究

As the increase in rice yield through chemical fertilizer, pesticide and genetic improvement will be facing limitations in the next decades, closing the existing yield gaps between the attainable potential and actual yields is essential to ensure food security. In order to address the issues of rice yield gap research that scaling up to entire regions, the spatial and temporal inconsistency of potential and actual yield, and poor timeliness, this proposal aims to make full use of the advantages of SAR and crop growth model, regionally estimating rice yield gap in-season by integrating crop growth model with polarimetric SAR (POLSAR) data. Polarimetric backscattering behaviors of rice fields with different sowing methods，canopy density and panicle types will be measured and analyzed using an anechoic chamber. And a rice microwave scattering model will be modified to consider direct-seedling and transplanted sowing methods as well as erect/curved panicle types by proposing two-cyclinders and multi-sphere panicle scattering models. Based on the scattering model, rice canopy density, a necessary input of crop model to calculate yield potential, will be estimated using POLSAR data by considering different sowing methods; panicle traits, such as length, diameter, inclination angle, number of grains and weight etc., will be also retrieved to calculated actual yield. Consequently, in-season and regional yield gap will be acquired from the difference between potential and actural yields. Finally, a comprehensive interpretation of yield gap causes will be conducted to explore spatial and temporal variation in yield gaps, identify potential environmental constraints to increasing yields and explore potential ways to increase yields using existing varieties.

近年来，随着化肥农药投入增加、品种改良等带来的水稻增产将趋近极限水平，缩小产量差成为保障粮食安全的重要途径。本项目针对水稻产量差研究中区域扩展困难、潜在与实际产量时空不匹配以及时效性差等问题，协同利用极化SAR与作物模型，充分发挥极化SAR结构、介电敏感性、空间连续性和作物模型时间连续性的优势，建立区域水稻产量差当季估算方法。在纯净可控电磁环境下，研究水稻植株密度与不同穗型的极化散射机理，发展考虑撒播/插秧种植方式的双圆柱和多球水稻微波散射模型。在此基础上，研究构建考虑种植方式差异的极化SAR水稻植株密度反演方法，为作物模型潜在产量估算提供输入；发展基于双圆柱和多球稻穗模型的极化SAR水稻穗部性状反演方法，估算实际产量，最终实现区域水稻产量差当季估算。在此基础上，研究水稻产量差的时空分异规律，结合品种、田间管理、社会经济等因素，综合解析产量差形成原因，为充分挖掘水稻生产潜力提供科学依据。

近年来，随着化肥农药投入增加、品种改良等带来的水稻增产将趋近极限水平，缩小产量差成为保障粮食安全的重要途径。本项目针对水稻产量差研究中区域扩展困难、潜在与实际产量时空不匹配以及时效性差等问题，协同利用极化SAR与作物模型，充分发挥极化SAR结构、介电敏感性、空间连续性和作物模型时间连续性的优势，建立区域水稻产量差当季估算方法。.通过项目研究工作，获得了插秧/撒播不同种植方式，以及弯曲/直立不同穗型水稻田的多视向、多角度、多频段极化散射特征；构建了考虑种植方式的水稻植株密度反演方法，为区域潜在产量计算提供准确输入；构建了基于Feko电磁计算的水稻微波散射模型，考虑了插秧/撒播不同种植方式，以及弯曲、直立稻穗形态和微观拓扑结构特征，在此基础上，针对水稻实际产量反演，简化水稻微波散射模型，保留了对稻穗形态和微观拓扑结构的刻画，并结合改进的基于物理模型的极化分解方法，构建了极化参量与穗部性状之间的关系，利用遗传算法实现了水稻实际产量的估算。基于极化SAR反演的水稻物候、种植方式和植株密度，结合土壤、气象、品种空间插值数据，驱动DSSAT-(CRERS-Rice)模型，实现区域水稻潜在产量估算，并结合实际产量构建了区域水稻产量差当季估算方法，解决了水稻产量差研究中区域扩展困难，潜在产量与实际产量时空不匹配以及时效性差的问题，为综合评估区域水稻生产潜力、缩小产量差，稳步提高水稻产量提供科学依据。