Remote-sensing Estimate And Spatiotemporal Analysis On The Carbon Storage Of Subtropical Pinus Massoniana Forest In Changting County, China

The global climate change has led to an increasing concern on the dynamics of the carbon storage of the forest ecosystem.Based on multi-temporal remote sensing images,in situ measured hyperspectral data and field-measured growth index data of Pinus massoniana forest,this study has explored remote sensing techniques for estimating the carbon storage of Pinus massoniana forest,and revealed the spatiotemporal dynamics of the carbon storages of Pinus massoniana forest in Changting County,Fujian Province,southeastern China,during a 24-year period from 1988 to 2012.By comparative analysis of the image classification methods including maximum likelihood,BP neural network,support vector machine and stratified classification,it is found that,the stratified classification method is the best method for the information extraction of Pinus massoniana forest.Based on this method,Pinus massoniana forest extraction models for Landsat TM,SPOT,ALOS and RapidEye were established,respectively.It is found that the area of Pinus massoniana forest in the county grew from 104982.66 hm2 in 1988 to 168928.68 hm2 in 2012,with a net increase of 63946.02 hm2.On the basis of in situ measured hyperspectral data and field-measured growth index data of Pinus massoniana forest,the relationship between the carbon storage and the spectral reflectance characteristics of Pinus massoniana was investigated.It is found that 677 nm and 1468 nm are the most sensitive wavelengths.As a result,the best hyperspectral model for estimating the carbon storage of Pinus massoniana forest was developed by using a ratio of R1468/R677.Based on field-measured growth index data of Pinus massoniana forest and remote sensing images of TM,ALOS and RapidEye,the best multispectral models for estimating the carbon storage of Pinus massoniana forest were constructed.In addition,the study of the impact of the spatial resolution and spectral resolution of the data indicates that improving image spatial resolution and spectral resolution can effectively improve the estimation accuracy of the developed models.Furthermore,to make use of the 2010 estimation model to predict the carbon storage of Pinus massoniana forest in other images,one important step is to calibrate the model because the time gap between the images could result in the change in the spectral response of Pinus massoniana.The application of the calibrated models in each image has revealed a gradual increase in the carbon storage and carbon density of Pinus massoniana forest in Changting County during the study period.The carbon storage grew from 1.69×106 t in 1988 to 7.42×106 t in 2012,accompanied with an increase in carbon density from 16.12 t/hm2 to 43.92 t/hm2 during the period.In addition,spatial overlay analysis showed that the average carbon storage of Pinus massoniana forest increased with the elevation and gradient increase.The carbon storage of Pinus massoniana forest was mainly distributed in the area with an elevation between 400m and 600m,preferably with a gradient between 5° and 25°.This study also reveals that the area,carbon storage and carbon density of Pinus massoniana forest in 18 towns of Changting have increased to some extent,of which Gucheng,Zhuotian,Hetian and Sidu towns have the most obvious growth.