植被生物量反映了生态系统获取能量的能力,分析其分布特征对了解生态系统结构和功能具有十分重要的意义。传统的反演植被地上生物量的方法往往由于样本的缺少,以及影响因子的不确定性而导致预估精度不高。本文选用ELM对105块实测样本的遥感因子(TM影像灰度值和植被因子等10个因子)进行训练,用余下34块样地进行验证,结果表明:ELM反演植被地上生物量,可以获得较高的精度,模型预测结果与实测结果的曲线拟合决定系数R2达0.89。此外,对2010—2015年玛纳斯河流域的植被地上生物量进行反演,认为流域内上游山区生物量大部分较为稳定,中游平原区生物量呈现增加趋势,下游荒漠区生物量则呈现退化趋势。
Aboveground biomass reflects the capability of ecosystems to obtain energy.Analysis on the spatial distribution pattern is of great significance for understanding the structure and function of ecosystems.The accuracy of inverting aboveground biomass with the conventional approach is low due to the lack of samples and the uncertainty of impact factors.In this study,Extreme Learning Machine (ELM) was used to train the remote sensing factors of 105 samples which included seven-band pixel values of TM image and vegetation factors,and the remaining 34 samples were used for verification.The results confirmed that ELM approach could invert vegetable aboveground biomass with a higher accuracy,and its determination coefficient of curve fitting reached 0.89.In addition,the inversion of vegetation aboveground biomass in the Manas River Basin from 2010 to 2015 found that the biomass was relatively stable in the upper area of the Manas River Basin,was in an increase trend in the middle plains and was in a deterioration trend in the downstream desert.
[1] 程晓莉,安树青,钦佩,等.鄂尔多斯草地退化过程中植被地上生物量空间分布的异质性[J].生态学报,2003,23(8):1526-1532.
[Cheng Xiaoli,An Shuqing,Qin Pei,et al.The heterogeneity in spatial distribution of the above-ground biomass in the degraded grasslands in Ordos[J].Acta Ecologica Sinica,2003,23(8):1526-1532.]
[2] 吴桂平,叶春,刘元波.鄱阳湖自然保护区湿地植被生物量空间分布规律[J].生态学报,2015,35(2):361-369.
[Wu Guiping,Ye Chun,Liu Yuanbo.Spatial distribution of wetland vegetation biomass in the Poyang Lake National Nature Reserve,China[J].Acta Ecologica Sinica,2015,35(2):361-369.]
[3] 王淑君,管东生.神经网络模型森林生物量遥感估测方法的研究[J].生态环境,2007,16(1):108-111.
[Wang Shujun,Guan Dongsheng.Remote sensing method of forest biomass estimation by artificial neural network models[J].Ecology and Environment,2007,16(1):108-111.]
[4] Moreau S,Toan T L.Biomass quantification of Andean wetland forages using ERS satellite SAR data for optimizing livestock management[J].International Journal of Remote Sensing,2003,84(4):477-492.
[5] Cao L,Coops N C,Innes J L,et al.Estimation of forest biomass dynamics in subtropical forests using multi-temporal airborne LiDAR data[J].Remote Sensing of Environment,2016,178(14):158-171.
[6] Sarabandi K,Polatin P.An iterative inversion algorithm with application to the polarimetric radar response of vegetation canopies[J].IEEE Transaction Geosciences Remote Sensing,1994,32(1):62-71.
[7] Ma C,Ouyang J H,Chen H L,et al.A novel kernel extreme learning machine algorithm based on self-adaptive artificial bee colony optimisation strategy[J].International Journal of Systems Science,2016,47(6):1342-1357.
[8] 王渊刚,罗格平,冯异星,等.近50 a玛纳斯河流域土地利用/覆被变化对碳储量的影响[J].自然资源学报,2013,28(6):994-1006.
[Wang Yuangang,Luo Geping,Feng Yixing,et al.Effects of land use/land cover change on carbon storage in Manas River Watershed over the past 50 years[J].Journal of Natural Resources,2013,28(6):994-1006.]
[9] 贺可,吴世新,周宏飞,等.玛纳斯河流域两种典型土地利用变化分析[J].干旱区研究,2018,35(4):954-962.
[He Ke,Wu Shixin,Zhou Hongfei,et al.Two typical land use modes in the Manas River Basin[J].Arid Zone Research,2018,35(4):954-962.]
[10] 李菊荣,王延华,唐湘玲,等.新疆玛纳斯河流域土地利用变化特征及影响因素研究[J].土壤通报,2018,49(1):61-68.
[Li Jurong,Wang Yanhua,Tang Xiangling,et al.,The spatial-temporal characteristics of land use change in the Manas River Basin and influencing factors[J].Chinese Journal of Soil Science,2018,49(1):61-68.]
[11] 梁二敏,张军民,杨卫红.新疆玛纳斯河流域绿洲景观生态脆弱性时空分异[J].干旱区研究,2017,34(4):950-957.
[Liang Ermin,Zhang Junmin,Yang Weihong.Spatiotemporal variation of landscape ecological vulnerability in oasis in the Manas River Basin,Xinjiang[J].Arid Zone Research,2017,34(4):950-957.]
[12] 冯异星,罗格平,周德成,等.近50 a土地利用变化对干旱区典型流域景观格局的影响——以新疆玛纳斯河流域为例[J].生态学报,2010,30(16):4295-4305.
[Feng Yixing,Luo Geping,Zhou Decheng,et al.Effects of land use change on landscape pattern of a typical arid watershed in the recent 50 years:A case study on Manas River Watershed in Xinjiang[J].Acta Ecologica Sinica,2010,30(16):4295-4305.]
[13] 姜亮亮,刘海隆,包安明,等.玛纳斯河流域景观格局演变特征与驱动机制分析[J].水土保持研究,2014,21(4):256-262.
[Jiang Liangliang,Liu Hailong,Bao Anming,et al.Analysis on landscape pattern change and driving mechanism in Manas River Watershed[J].Research of Soil and Water Conservation,2014,21(4):256-262.]
[14] 李慧,雷晓云,包安明,等.基于SWAT模型的山区日径流模拟在玛纳斯河流域的应用[J].干旱区研究,2010,27(5):686-690.
[Li Hui,Lei Xiaoyun,Bao Anming,et al.Application of simulation about montanic daily runoff volume in the Manas River Basin based on SWAT model[J].Arid Zone Research,2010,27(5):686-690.]
[15] Ding S F,Zhang Y N,Xu X Z,et al.A novel extreme learning machine based on hybrid kernel function[J].Journal of Computers,2013,8(8):2110-2117.
[16] Huang G B,Zhu Q Y,Siew C K.Extreme learning machine:Theory and applications[J].Neurocomputing,2006,70(1):489-501.
[17] 柳小桐.BP神经网络输入层数据归一化研究[J].机械工程与自动化,2010(3):122-123.
[Liu Xiaotong.Study on data normalization in BP neural network[J].Mechanical Engineering and Automation,2010(3):122-123.]
