Community Monitoring And Carbon Storage Estimation Of Mangrove Wetland In Zhangjiang Estuary

Mangrove is one of the highest productivity and carbon storage ecosystems in the world.It plays an important role in maintaining ecological balance,wetland diversity and regional socio-economic sustainable development.Remote sensing technology is widely used in mangrove resource survey because of its advantages of large-scale,multi-scale monitoring and short data acquisition period.At present,scholars at home and abroad mostly focus on using foreign high spatial resolution and hyperspectral remote sensing satellite image data,while the application of domestic satellite remote sensing data is relatively small.The Zhangjiang Estuary Mangrove National Nature Reserve in Fujian Province was selected as the research area.Combining with HJ-1A and Landsat remote sensing image data,and using remote sensing technology,combined with object-oriented classification,supervised classification,unsupervised classification,SVM classification and BP neural network model,the mangrove communities in the study area were classified,and the carbon storage and spatial distribution of mangrove vegetation in the study area were estimated.The main conclusions are as follows:(1)The HJ-1A hyperspectral data and the full-color band of Landsat8 OLI remote sensing image data are fused,and the spectral,shape and texture characteristics of the fused image are extracted according to the field survey data.The information of mangrove vegetation is extracted from fused remote sensing image by object-oriented classification method.The overall classification accuracy is 99.22%,and the Kappa coefficient is 0.9559.The results show that the fused hyperspectral satellite remote sensing data can effectively distinguish mangroves from non-mangroves,and the results can be used for further classification of mangrove plant communities.(2)According to Google remote sensing images and field survey data,the training samples of mangrove plant communities were selected.Mangrove communities were divided into three communities:Avicennia marina,Kandelia candel and Aegiceras corniculatum by using four classification methods:maximum likelihood classification,nearest neighbor classification,SVM classification and ISODATA classification.The results show that the overall accuracy of SVM is the highlest,the overall accuracy of classification is 79.26%,and the Kappa coefficient is0.6681.Followed by the KNN classification method,the overall accuracy of classification is 70.97%,Kappa coefficient is 0.6060.The overall accuracy of maximum likelihood classification is 64.94%,and the Kappa coefficient is 0.4222.The results of ISODATA classification are the worst,with an overall accuracy of 47.31%and a Kappa coefficient of 0.1227.(3)The biomass of each part of mangrove was calculated by using the allometric growth equation of mangrove,and the carbon storage of mangrove vegetation was obtained according to the relationship between biomass and carbon storage.Based on the characteristics of mangrove vegetation,a remote sensing inversion model of mangrove vegetation carbon storage was established by using BP neural network model.The results show that the model R~2 of Avicennia marina is 0.53,the model R~2of Kandelia candel is 0.69 and the model R~2 of Aegiceras corniculatum is0.79.Based on the inversion results,combined with the classification results of mangrove plant communities and remote sensing technology,the spatial distribution of carbon reserves of mangrove vegetation was obtained:the carbon reserves of mangrove vegetation ranged from 0.23kg to 2.58 kg,the minimum was 0.23 kg in Aegiceras corniculatum community in the northern part of the reserve,and the highest was 2.58kg in the Avicennia marina community in the central and southern part of the reserve.Overall,the carbon storage of vegetation in Avicennia marina community was the highest,followed by Aegiceras corniculatum,and the lowest in Kandelia candel.The purpose of this study is to provide technical reference for remote sensing inversion of mangrove plant communities and carbon reserves,and to provide a scientific reference for the protection of mangrove wetland ecosystems,and to promote the application of domestic hyperspectral satellite remote sensing data in wetland monitoring and protection.