通过水稻淹水实验模拟洪涝胁迫状态, 分析不同生育期、 不同洪涝胁迫强度下的水稻叶面积指数变化及其冠层高光谱响应规律, 建立洪涝胁迫下水稻叶面积指数(LAI)的估测模型。 结果表明, 分蘖期、 拔节期、 抽穗期水稻LAI均随淹水深度的增加而降低; 水稻冠层680~760 nm光谱的一阶微分具有“双峰”现象, 主峰位于724~737 nm, 701和718 nm有次峰出现, 并随淹水深度的增加出现“三峰”; 冠层红边位置在各个生育期出现“蓝移”, 红边幅值、 红边面积在分蘖期、 拔节期、 灌浆期均呈现“蓝移”, 在抽穗期出现“红移”; 生育前期水稻LAI与光谱特征参量存在显著相关, 生育后期相关性不显著; 最后以Dλ737/Dλ718光谱参量建立了水稻生育前期的LAI估测模型。 表明在洪涝胁迫下, 水稻LAI的变化幅度直接反映胁迫强度的大小, 估测模型LAI=3.138(Dλ737/Dλ718)-0.806可用于估测洪涝胁迫不同强度下的水稻叶面积指数。

In order to provide the foundational theoretical support for flood loss estimation of rice with RS, the change of leaf area index (LAI) and canopy spectral response during four developmental stages and three waterlogging depths were studied, and the LAI estimation model was established with spectra characteristics parameter using regression analysis method. The results show that LAI value decreases as water depth increases in tillering, jointing and heading stages, and LAI value under complete submergence decreased by 36.36% than CK in jointing stages. “Double-Peak” presented in the canopy first derivative spectra of 680~760 nm where the red edge parameters existed, and the main peak is located in the 724~737 nm with 701 and 718 nm exhibiting secondary peak. With water depth increasing, “Triple-Peak” emerges especially. The red edge position moves to long-wavelength direction in each developmental stage. Blue shift of red edge amplitude and red edge area was detected in tillering, jointing and filling stages, while red shift appeared in heading stage. The relationship between spectra characteristics parameters and LAI were investigated during 4 growth stages, results were not consistently significant at any wavelengths, and the leaf area indices were significantly correlative to the spectra parameters before heading stage, so the spectra parameters before heading stage can be used to estimate the leaf area indices, and a regression model based on parameter Dλ737/Dλ718 was recommended. Therefore the variation range of LAI for rice could response to the stress intensity directly, and the regression model LAI=3.138(Dλ737/Dλ718)-0.806 can precisely estimate the leaf area index under flooding and waterlogging stress.