基于Landsat OLI和ETM+的杭州城市绿地地上生物量估算比较研究  

Estimation of Above-ground Biomass of Urban Green Land in Hangzhou Based on Landsat OLI and ETM+ Data

作　　者：邱布布 徐丽华[1,2] 张茂震 张结存[3] 钱军朝 陆张维 

机构地区：[1]浙江农林大学环境与资源学院,浙江临安311300 [2]浙江省森林生态系统碳循环与固碳减排重点实验室,浙江临安311300 [3]宁波远见旅游研究有限公司,浙江宁波315000

出　　处：《西北林学院学报》2018年第1期225-232,共8页Journal of Northwest Forestry University

基　　金：国家自然科学基金(E080201);浙江省自然科学基金(LY15D010006)

摘　　要：城市绿地地上生物量的定量估算为城市碳循环研究提供重要依据。以杭州市辖区为研究区,选取Landsat 8OLI影像,提取影像的原始波段及植被指数信息,基于199个样地实测数据,建立研究区绿地地上生物量估算模型,反演研究区生物量,并与同时期Landsat 7ETM+遥感影像生物量估算结果进行对比分析。结果表明,基于OLI影像建立的逐步回归和随机森林模型R2分别为:0.643、0.514;城市绿地生物量主要集中在5~15、35~55t·hm^(-2),城区内绿地主要以带状、块状分布,其值偏低,中部公益林及城郊马尾松林生物量值较高;基于OLI影像估算的研究区生物量平均值为32.520t·hm^(-2),相应的ETM+估算结果为34.454t·hm^(-2),两者估算结果的差值基本上以0为中心,差值的平均值为-1.934t·hm^(-2),2种影像估算结果的Pearson系数为0.901,显著性<0.001;总体上,2景影像的生物量估算结果差异较小,可通过线性关系相互补充利用。Quantitative estimation of the biomass of urban green space system is of significance to study of urban carbon cycle. In this paper, Landsat 80LI was selected to extract the original band and vegetation in- dex information. Based on the measured data of 199 samples,a model for estimating the above ground bio- mass of the vegetations occurring in Hangzhou urban green land was established to invert of biomass in the study area. The results of the estimation was compared with those obtained from Landsat 7 ETM＋ in cor- responding perios. The coefficients of determination （R2） of step-wise regression and random forests based on Landsat80LI were 0. 643 and 0. 514,respectively. Urban green space biomass mainly focused in the 5- 15 t hm-2 ,35-55 t hm-2. In urban area,the green spaces were in strip and block distributions,with low biomass, the biomass of the public welfare forests in middle area pine forests in suburb exhibited was high. The mean above-ground biomass was 32. 520 t hm-2 by the estimation results based on OLI,while for the results of ETM＋ image was 34. 545 t hm-2. The differences between the two results was basically cen- tered around 0 as the center,and the average value of the difference was -1. 934 t hm-2. The Pearson correlation between the two images was 0. 901,significance was less than 0. 001. Generally,the biomass estimation re- sults of the two images had little difference, which could be supplemented by the linear relationship.

关 键 词：OLI ETM+ 生物量 逐步回归 随机森林 

分 类 号：S731.2[农业科学—林学]