关键词: 湿地植被生物量, 遥感, BP神经网络模型, 影响因素, 盐城自然保护区

Abstract:

Taking the ETM + remote sensing image of the core area of Yancheng Nature Reserve and the field aboveground biomass measured in the same period as the data source,we built BP artificial neural network model and simulated biomass distribution of the study area.The conclusions of the study are as follows:The BP neural network model was used to retrieve the biomass of wet weight and dry weight,their accuracy arrived at 70% and 74% respectively.The total biomass of dry weight is 9.370×10⁷kg and the total biomass of wet weight is 4.996×10⁸kg.In the spatial variation,it mainly presents obvious difference from the land to the sea and slow change along the coast.The biomass of Spartina,reed and Suaeda range from high to low.The dry weight of Suaeda salsa’s biomass mainly concentrated in 1.5kg/m²,while its biomass of wet weight is concentrated in 0~6kg/m²; The dry weight of Spartina alterniflora’s biomass mainly concentrated in 1~2kg/m²,while its biomass of wet weight is over 8kg/m²; The dry weight of Reed’s biomass mainly concentrated in 0~2kg/m²,while its biomass of wet weight is concentrated in 2~6kg/m².The biomass is positively correlated with plant height and coverage,and the correlation between dry weight and the two was stronger.The biomass was positively correlated with ecological niche and soil environmental factors,especially with soil nutrients.

Key words: wetlands vegetation biomass, remote sensing, BP neural network model, impacts, Yancheng Natural Reserve