Satellite-based Retrieval Of Canopy Parameters Of Moso Bamboo Forest With PROSAIL Radiative Transfer Model

Forest canopy plays an important role in the spatial distribution of the entire forestecosystem material and energy transfer, solar radiation transmission, maintainingenvironmental factors, and physiological parameters. Moso bamboo forest (Phyllostachysheterocycla var. pubescens), distributed in subtropical region of China (such as Zhejiang,Anhui, Jiangxi, Fujian, and other provinces), is a special kind of forest types with acluster-like canopy structure, whose canopy texture is significantly different from otherforest types. Researches on traditional forest canopy parameters were mostly based onstatistical model, which focused on the accuracy of the model and neglected its absorptionof solar radiation, directional reflectance, transmittance, and the transfer of its radiation inbamboo canopy. Therefore, the research from the point of view of the mechanistic modelto quantitatively inverse Moso bamboo forest canopy parameters.Using satellite remote sensing data, this paper is based on the ground under thesupport of plot survey data, combined with PROSAIL to carry on the study on Mosobamboo canopy parameter quantitative inversion of Leaf area index(LAI) and Canopychlorophyll content(CCC), which for the interpretation from remote sensing mechanism ofbamboo carbon sequestration foundation. It mainly contains the following aspects:1. Data acquisition and processing. It includes the remote sensing data acquisitionand preprocessing, sample data (such as spectrum, canopy parameters) as well as otheracquisition and processing.2. Study on Moso bamboo forest area remote sensing information extraction. On thebasis of remote sensing image preprocessing, the maximum likelihood method is applied toextract Moso bamboo forest remote sensing information.3. Simulation of canopy reflectance.leaf reflectance combined with measured plotdate transform into canopy reflectance.4. PROSAIL model input parameters settings. It was combined with measured dataand reference value through the sensitivity analysis to determine input parameters changerange and simulate canopy reflectance.5. Lookup-table (LUT) Setup. Obtain various circumstances of canopy reflectancethrough the model input parameters of different values and set up the lookup-table(LUT) ofLAI-canopy reflectance-Leaf chlorophyll content (LCC).6. LAI and CCC inversion. Matching search reflectance image pixels with LUT toget corresponding leaf chlorophyll content (LCC) and canopy leaf area index (LAI)，and scale the leaf chlorophyll content LCC up to the canopy. Final inversion results werevalidated using measured data.Through the study, this paper gets the following conclusion:1. Maximum likelihood classification method was used to extract Bamboo Forestremote sensing information. Results showed the classification accuracy of89.06%, thetotal Kappa coefficient of0.8587, bamboo forest user accuracy of93.33%, and productionaccuracy of83.33%;2. Bamboo leaf reflectance inversed by PROSPECT model have a better match to andmeasured leaf reflectance, with the correlation coefficient R2of0.9716and root meansquare error (RMSE) of0.027.3. PROSAIL model input parameters of the sensitivity from high to low in turn isLAI>LCC>N>ALA>Cw>Cm. LAI and LCC are determined as the two main sensitivefactors applied to Moso Bamboo forest LAI-canopy reflectance lookup-table construction.Because the Landsat TM images of3,4, and5wave bands are more sensitive to LAI, thesethree wave bands are selected to participate in the LAI inversion;4. Due to the good consistency by PROSAIL bamboo LAI remote sensing inversionresults and measured LAI, the correlation of them (R2) is0.895. The root mean squareerror (RMSE) and relative RMSE are respectively less than0.58and12.98%, but alsothere is a issue that the inversion of LAI mean value is higher than the actual value;5. Chlorophyll content effects on the canopy reflectivity has related to leaf areaindex. The greater leaf area index is, the greater the chlorophyll has influence on thecanopy reflectivity. The greatest influence on canopy chlorophyll band are respectively725nm (LAI=2), and730nm (LAI=4and LAI=6) to form the two peak in the555nm and725nm/735nm standard deviation. The conclusion has significant influence on the furtherexploring the effect of chlorophyll on canopy reflectance using hyper-spectral remotesensing data to estimate the canopy chlorophyll content;6. The canopy chlorophyll content was scaled from leaf chlorophyll content and hasgood consistency with the measured canopy reflectance with the correlation coefficient of0.612and the root mean square error of13.98%.