Estimation And Retrieval Of Carbon Content In East Dongting Wetland

Wetland ecosystem is one of the most important ecosystems on the surface of the Earthand the largest component of the terrestrial ecosystem carbon pools. Wetlands occupy only anarea of6%of the total global land, but stores20%of the total carbon stocks of the terrestrialecosystems. Wetlands play a crucial role in the global carbon cycle. Due to global climatechange and human activity induced interference, the carbon cycle of wetland ecosystems hasbeen greatly impacted, even functioning as carbon sink turned into carbon source. Therefore,this study investigated spatial distribution of wetland carbon content and developed wetlandecological system carbon estimation model and thus provided a great potentail for monitoringthe dynamic of wetland carbon content and predicting the impact of human induceddisturbance and climate change on the carbon cycle of wetland ecosystems.In this study, the East Dongting Wetland Nature Reserve was selected as the study area.Combining SPOT-5multi-spectral images with field measurements, we investigated theestimation methods of wetland vegetation biomass, soil organic carbon and DOC of water. Thecorrelation of image derived spectral variables with vegetation biomass, soil organic carbon,and DOC concentration of water, respectively, were first analyzed. The estimation models ofcorresponding carbon contents were then developed. It is expected that this study could providean effective means for monitoring the dynamics of wetland ecosystem carbon content. At thesame time, this study led to the spatial distribution maps of the wetland carbon content and thusprovided the suggestions and guidelines for monitoring the changes of wetlands at regionalscale and for wetland protection and restoration. The significant findings include:(1) On the basis of analyzing the correlation of vegetation biomass with remote sensingderived spectral variables, the vegetation biomass estimation models were built. Resultsshowed that the RVI, ME, NDVI and MSAVI driven multivariate linear regression model wasthe best, which led to the best wetland vegetation carbon content model. (2) By analyzing the correlation between soil organic carbon and its impact factors, wefound that the vegetation biomass had the largest correlation coefficient, then soil bulk densityand soil total nitrogen content. The soil organic carbon estimation models were developed bytaking into account the important impact factors as independent variables. The optimalprediction model was the multivariate linear regression model based on vegetation biomass andsoil total nitrogen content.(3) Among all the remote sensing derived spectral variables, R/G had the highestcorrelation with DOC of the water. The transformation of the measured DOC concentrationusing common logarithm, natural logarithm, square root and cube root, to some extent,increased the correlation. Logarithm transformations resuled in the greatest increase ofcorrelation. The model that accounts for the relationship of natural logarithm of DOC with R/G,ME and SWIR-NIR led to the best fit.(4) The carbon content of the East Dongting lake wetland was mapped using the aboveoptimal models. Results showed that the carbon content of the land areas varied from5368.38g/m~2to11784.63g/m~2. The carbon content in lush vegetation area was higher than inother regions. Water DOC concentrations ranged from0.5mg/L to3.02mg/L. The concentrationin the shore areas was higher than in the central area and in the areas disturbed by humanactivities was higher than in other regions. In the East Dongting lake wetland ecosystem, thesoil contained the largest part of the carbon. The soil carbon content density was about twentyto thirty times of the total aboveground carbon content. The soil carbon content thus was thelargest carbon storage carrier in the East Dongting Lake wetland ecosystem.