闽江口沼泽植被地上鲜生物量与植株密度高光谱遥感估算

湿地植被地上生物量与湿地生态系统生产力、碳循环和养分循环等密切相关，是当前研究中的主要关注点之一。利用ASDFieldSpec2500 便携式地物波谱仪测定闽江口鳝鱼滩芦苇（Phragmites australis）和短叶茳芏（Cyperus malaccensis Lam. var. brevifolius Bocklr.）冠层反射光谱，同时测定其地上鲜生物量和植株密度，分析地上鲜生物量与冠层反射光谱及一阶微分光谱的相关关系，确定敏感波段，进而改进植被指数；利用回归分析法，基于各个植被指数，构建地上鲜生物量和植株密度的估算模型。结果表明，芦苇和短叶茳芏地上鲜生物量与冠层反射光谱在蓝光、红光及近红外波段相关性较高；与一阶微分光谱在蓝边、红边相关性较好；芦苇地上鲜生物量与冠层光谱和一阶微分光谱的相关系数均优于短叶茳芏。芦苇和短叶茳芏地上鲜生物量估算模型R2分别在0.408 5~0.765 和0.101 9~0.315 3 之间，估算精度相对于其他参数较高的参数主要有BNDVI、NDCI和MGBNDVI，最佳估算参数均为BNDVI；植株密度估算模型R2分别在0.093 0~0.718 和0.138 9~0.233 7 之间，估算精度相对于其他参数较高的参数主要有GBNDVI、BNDVI、MGBNDVI、NDCI 及SR， 最佳估算参数分别为GBNDVI和NDCI。一定程度上来说，利用高光谱数据估算河口芦苇和短叶茳芏地上鲜生物量和植株密度是可行的。

Hyperspectral Estimating Models of Aboveground Fresh Biomass and Density in Minjiang River Estuary Marsh Vegetation

Aboveground biomass of wetland vegetation is closely related with wetland ecosystem productivity, carbon cycle and nutrient cycle, and so on, is one of the main concerns in the present study. The canopy hyperspectral reflectance data of two tidal marshes dominated respectively by Phragmites australis and Cyperus malaccensis Lam. var. brevifolius Bocklr. was determined by ASDFieldSpec 2500 in the Shanyutan wetland in the Minjiang River estuary in October 2011 and October 2012, and the aboveground fresh biomass and plant density of the two marshes were collected simultaneously. The correlation between canopy reflectance, first derivative spectra and aboveground fresh biomass of P. australis and C. malaccensis was analyzed. The sensitive band was confirmed so as to improve vegetation index. Furthermore, using the regression analysis method, based on various vegetation indexes, the aboveground fresh biomass and plant density estimation model was constructed. The results showed that the aboveground fresh biomass was more closely correlated with the two marshes canopy reflectance in blue, red and near infrared band, and was relatively well correlated with first derivative spectra at blue edge and red edge. No matter canopy reflectance or first derivative spectra, the correlation of P. australis is better than C. malaccensis. Compared with other parameters, BNDVI, NDCI and MGBNDVI also have higher estimation accuracy of aboveground fresh biomass for both species, and BNDVI was the best estimated parameter of aboveground fresh biomass for both species and the estimation accuracy was higher for P. australis (R²=0.4085－0.765), while for C. malaccensis R² ranged from 0.1019 to 0.3153. Furthermore, GBNDVI, BNDVI, MGBNDVI, NDCI and SR also have higher estimation accuracy in plant density for both species, and there is a higher plant density estimation accuracy for P. australis as well (R²=0.0930－0.718), while for C. malaccensis it ranges from 0.1389 to 0.2337, and the best estimate parameter are GBNDVI and NDCI respectively. To some extent, it is feasible to use remote sensing data to estimate aboveground fresh biomass and plant density based on the models built in this work.