Study On The Carbon Sequestration Capacity Estimation Of Typical Vegetation Based On Hyperspectral Data

With the acceleration of the urbanization process,the city’s carbon emissions have risen sharply,and thus bringing a series of environmental and ecological problems.Therefore,in order to control the city’s carbon level and build eco-city and garden city,the study of urban carbon sequestration is also increasingly urgent.Green plants transfer atmospheric carbon dioxide into its organic matter,release oxygen at the same time,and finally finish the process of carbon fixation through photosynthesis function.As for city environment,the forest plays a major role in reducing CO2,while the economic forest around the city and the crops in farmland,as green plants too,also play an important role in carbon sinks.With the continuous development of hyperspectral remote sensing and quantitative remote sensing technology,it is possible to estimate various physical and chemical parameters of plants in a wide range of area with remote sensing methods.The Gross primary productivity(GPP)of plants,as an important link in regional and global carbon budget,is the amount of organic carbon that is fixed by photosynthesis per unit time,which is the focus of remote sensing inversion research in a long time.And the GPP of plant leaves,which are the main organs of plant photosynthesis,reflects the strength of the carbon sequestration capacity of vegetation,and is also a good indicator of the ability of the carbon sinks assessment.Therefore,it is very important to estimate the GPP of typical vegetation leaves in urban areas by remote sensing ways quickly and accurately,and then evaluate their carbon sequestration capacity in both ecology and economy value.In this thesis,the GPP estimation of 5 different kinds of vegetation,oilseed rape,wheat,ginkgo,citrus and camphor trees was studied at the leaf level by using the measured hyperspectral data and photo synthetic data.The regression model of leaf GPP and VI×PARin was established and 8 different vegetation indices,SR,NDVI,EVI2,GNDVI,CIgreen,red edge NDVI,MTCI and CIred edge,were tested and analyzed in the leaf GPP inversion.And we analyzed the influence of the environmental factors such as phenology,water stress and external light conditions on the inversion model and why these influence happening.Finally,we tried to establish a comprehensive leaf GPP inversion model in the case of oilseed rape and wheat,ginkgo and camphor,and the combination of all these four different vegetation types.The main work and conclusions of this paper are as follows:(1)The establishment of inversion model:The leaf GPP of oilseed rape in the flowering period,pod period and whole period,ginkgo in the weak light and strong light and the camphor can be well estimated by the model established based on VI×PARin.The coefficient of determination R2 is all above 0.79.The leaf GPP of wheat can also be estimated by the model based on VI×PARin,but the accuracy of the model is lower than that of the previous three vegetation types.The relationship between leaf GPP of citrus and VI×PARin cannot be established in this study.In addition,except for the wheat under controlled water condition,all the regression curves of established GPP inversion models have similar shape and trend,mostly in the form of quadratic or linear function,and no obvious saturation phenomenon is found.(2)The influence of phenology and environmental condition on inversion model:there is a certain difference in leaf GPP inversion model of oilseed rape between the flowering period and the pod period,and the model curve slope of pod period is relatively more moderate,but leaf GPP in all periods can be expressed by one unified model.The accuracy of the model in all periods is lower than that of any of the single period,but the gap is not significant.The water control condition has obvious effect on the inversion model of wheat leaf GPP,and the water control condition will affect the light use efficiency(LUE)of wheat,and its Chlorophyll Efficiency(ChlE)will increase significantly,thus the inversion of leaf GPP of wheat should best to be grouped under different water conditions.The GPP inversion model of Ginkgo and citrus leaves were different in different light intensity grades(strong light or low light),and a certain degree of shading or higher scattering direct ratio will significantly improve the LUE of these two vegetation types,thus it is better to choose different leaf GPP inversion model of ginkgo and citrus under different light conditions.(3)The selection of vegetation indices.Under the condition of five different vegetation types and their different environmental conditions,the leaf GPP inversion model based on indices(red edge NDVI,MTCI and CIred edge),which are based on red edge band,is always the optimal or suboptimal model,and usually has a higher coefficient of determination R2 and a lower RMSE.And the accuracy of non-optimal model is still not that much lower than that of optimal model,showing a strong applicability and stability.(4)The influence of vegetation types.For the experimental data in this paper,the leaf GPP inversion model of ginkgo and citrus is more sensitive to the light intensity conditions than that of oilseed rape,and the ginkgo is more sensitive than citrus under different light intensity conditions;and the GPP of oilseed rape and wheat leaves can be estimated by a comprehensive model under similar environmental conditions.The GPP of ginkgo and camphor leaves can also be estimated by a comprehensive model.But the leaf GPP of all four Vegetation types(except for citrus)cannot be estimated by a comprehensive model.The three indices based on red edge band(red edge NDVI,MTCI and CIred edge)showed the lowest sensitivity and highest versatility compared with other vegetation indices when used to establish inversion model between rape and wheat,ginkgo and camphor and all four vegetation types,and were the most suitable indices for comprehensive estimation of leaf GPP for different vegetation.