基于Landsat TM数据的洪河湿地植被地上生物量遥感估算模型

Remote sensing estimation models based on Landsat TM data for vegetation biomass in Honghe Wetland

利用Landsat5 TM数据和同期的野外实测生物量数据分别建立了洪河湿地草甸湿地和沼泽湿地生物量估算模型,并结合土地覆被数据,利用模型估算草甸湿地和沼泽湿地的地上生物量。研究表明,沼泽、草甸湿地生物量与遥感信息相关性较高,而湿地植被总体样本与遥感信息相关性较低。以总体样本建模的精度远远低于以草甸湿地植被、沼泽湿地植被单独建模的精度,以不同植被类型分别建立生物量估算模型更适合湿地植被生物量的反演。草甸湿地植被的最优估算模型为以TM第七波段反射率为自变量的一元二次回归模型,沼泽湿地植被的最优估算模型为以TM第一波段反射率、归一化植被指数、湿度为自变量建立的多元线性模型。经统计分析得到洪河湿地草甸湿地的平均生物量为1084.92 g/m2,沼泽湿地的平均生物量为628.48 g/m2。

Biomass estimation models for the meadow and marsh wetlands in Honghe Wetland were established by using Landsat5 TM data and homochronous field investigation data.Combined with the land cover data,biomasses of the meadow and marsh wetlands were calculated by the models.The results show that:the correlation of the remote sensing information to the biomasses of meadow and marsh wetlands are higher than to the overall vegetation population sample.The estimation models for the meadow and marsh wetlands were established separately,and their accuracy are much higher than the model based on the overall vegetation population sample.The biomass estimation model based on different vegetation types was approved to be more suitable to biomass inversion of wetland.The best estimation model for the meadow wetland is a quadratic regression model with an independent variable of the reflectivity of TM7.The best estimation model for the marsh wetland is a multivariate linear model with the independent variables of TM1,NDVI,and humidity.The average biomasses of 1084.92 g/m2 and 628.48 g/m2 were obtained through statistical analysis for the meadow and marsh wetlands in Honghe Wetland,respectively.