基于现场光谱的黄河口湿地芦苇生物量估算模型研究

本文利用2014年春季黄河口湿地芦苇的现场光谱和生物量数据,以植被指数、光谱一阶导数及其导出量为特征参数,运用多种单变量回归模型,构建了基于现场光谱的芦苇生物量估算模型。结果表明:(1)总体上,特征参数与生物量的相关性由大到小排序为:植被指数 光谱一阶导数 光谱一阶导数导出量;(2)春季芦苇光谱在715~755 nm波段范围处的一阶导数与生物量显著相关。当构成植被指数的红光波段在722~751 nm范围内,近红外波段为765 nm或768 nm时,植被指数与生物量相关性最好;(3)S型模型和三次模型较其他单变量的估算模型结果好,所有估算模型中,修改型土壤调节植被指数MSAVI(modified soil adjusted vegetation index)的S型模型的估算结果最好,R2、MRE和RMSE分别为0.817、11.80%和0.085 kg/m2,且估算值总体上与实测值相当。

Research on biomass estimation model of reed in Yellow River Estuary wetland based on the in situ spectral data

Utilizing the in situ spectral data and biomass data of reed, collected in the spring of 2004 in Yellow River Estuary Wetland, and regarding vegetation index, first derivative spectrum and its derived parameters as characteristic parameters, this article builds biomass estimation model of reed based on the in situ spectral data through a variety of univariate regression models.The research shows that:(1) Generally, the correlation between characteristic parameters and biomass is:vegetation index first derivative spectrum first derivative spectrum derived parameters.(2)There is significant correlation between biomass and first derivative spectrum in spring when the spectrum fluctuates from 715 nm to 755 nm.Among the band combinations, when the red band changes from 722 nm to 751 nm and the corresponding infrared red is 765 nm or 768 nm, the correlation between vegetation index and biomass is the best.(3) Compared with other estimation models, S-type model and cubic estimation model are more effective.Among all the estimation models, MSAVI's S-type model is the most accurate while R2, MRE and RMSE are 0.817, 11.80% and 0.085 kg/m2 respectively.In addition, the estimated value is generally equal to the measured value.