Retrieval Of Canopy Parameters And Estimation Of Aboveground Carbon Storage For Typical Types Of Subtropical Forest Using Remote Sensing Data

Subtropical forest as an important part of China's forests plays a significant role in carbon cycling of forest ecosystems.Studying on canopy biophysical parameters and carbon storage of subtropical forest has an important theoretical and practical value,which well evaluates the contributions of subtropical forest to global climate change.Traditionally,many researches in canopy biophysical parameters are based on statistical model,which is simple in application but is hard to represent the radiative transfer mechanism.Canopy biophysical parameters,leaf area index(LAI)and chlorophyll content,in subtropical broad-leaved forest,coniferous forest and bamboo forest were estimated by using the radiative transfer model based on combining with the Landsat 8 OLI remote sensing image and field data.Then,the carbon storage in the three forest types was also estimated.This paper aims to provide the reference data for explaining the process of carbon sequestration and potential carbon sequestration capacity in subtropical forest based on remote sensing data.The main study contents include as follows:1.Eight kinds of forest types in mountain area of northwestern Zhejiang province were classified based on analysis of spectral features from the Landsat 8 OLI remote sensing image.Taken bamboo forest as an example,bamboo forest area and its distribution in Zhejiang province was extracted2.Spectral vegetation indices were constructed using wavelet transform.The inversion model between spectral vegetation indices and photosynthetic rate was built to estimate photosynthetic rate of leaves of Moso bamboo on the basis of measured spectral reflectance and photosynthetic rate.3.A look-up table for estimating LAI and chlorophyll content was built using radiative transfer model to achieve remote estimation of forest canopy biophysical parameters of subtropical forest with high accuracy.The different cost functions and regularization algorithm were applied in radiative transfer model.4.The model based on remote sensing data was built to estimate carbon storage of subtropical forest in the northwestern Zhejiang province.There are several conclusions from this study as follows:1.The classification accuracy of typical forests was over 80.0%.The product accuracy of broadleaf forest,coniferous forest and bamboo forest are 75.6%,82.08%,75.6%,and 75.9%,respectively.The total area of bamboo forest in Zhejiang province from classification was 902.7 thousand ha,which is consistent with forest inventory data(890.0 thousand ha).2.The net photosynthetic rate of bamboo leaf was sensitive to visible and infrared bands based on correlation analysis between the net photosynthetic rate and spectral features,such as 560 nm,680nm,920 nm,1600nm,and 2400 nm.Multi-variable regression model was built to estimate net photosynthetic rate with the R2 of 0.774 and RMSE of 0.288.Inversion model of net photosynthetic rate was constructed based on multi-scale wavelet transformation with the R2 of 0.88.3.LAI estimate from the radiative transfer model has accuracy of over 75% compared to observations.The correlation coefficients between measured LAI and estimated LAI were 0.935,0.842,and 0.798 for broadleaf forest,coniferous forest and bamboo forest,respectively.Canopy chlorophyll content estimate of broadleaf forest had the highest accuracy with R of 0.929,followed by coniferous forest with R of 0.771 and bamboo forest with R of 0.693.The results showed that canopy biophysical parameters of typical forests in subtropical region can be accurately estimated by using the radiative transfer model.Uncertainty in estimates was reduced through using cost functions and regularization algorithm.4.The model for estimation carbon storage of typical forests in subtropical region was built based on combination of the latest Landsat 8 OLI data and measured carbon storage data.The correlation coefficients between measured and estimated carbon storage were 0.985,0.876,and 0.736 for broadleaf forest,coniferous forest and bamboo forest,respectively.The RMSE of broadleaf forest,coniferous forest and bamboo forest were 2.7 Mg C/ha,6.3 Mg C/ha,and 5.2 Mg C/ha,respectively.The model with high accuracy can be used to estimate carbon storage of typical forests in subtropical region in large scale.