三江源区高寒草地地上生物量遥感反演模型研究

Modeling Aboveground Biomass of Alpine Grassland in the Three-River Headwaters Region Based on Remote Sensing Data

为了发展适用于三江源区高寒草地(包括高寒草甸和高寒草原)地上生物量(aboveground biomass,AGB)模拟的遥感反演模型,基于2006—2014年逐年7—8月三江源区高寒草地70个采样点地上生物量数据与同期MODIS-NDVI和MODIS-EVI数据,通过回归分析方法建立高寒草地地上生物量遥感反演模型,并利用长时间序列MODIS数据对2005—2014年三江源区高寒草地地上生物量的时空格局进行模拟分析.结果表明:基于EVI建立的乘幂模型对三江源区高寒草地地上生物量的拟合效果最好,其判定系数(R2)最大,达到0.654;均方根误差(RMSEP)最小,仅为27.86 g/m^2.根据三江源区70个采样点的地上生物量数据最终确立的估算模型为y=348.769x0.783(R2=0.655,P<0.001).估算模型模拟结果显示,2005—2014年三江源区高寒草地地上生物量空间特征基本一致,总体表现为从东南到西北逐渐减少的变化趋势,这与该区域的降水量、气温、海拔和植被类型等因素有关;2005—2014年三江源区高寒草地地上生物量平均值为169.25 t/a,最高值为2010年的178.36 t/a,最低值为2008年的162.80t/a,年际变化趋势表现为2005—2008年逐年下降、2008—2014年则在波动中逐年有所上升.研究显示,三江源区高寒草地地上生物量遥感反演模型及其确定的模型参数可对该区域草地地上生物量进行有效评估.

To estimate aboveground biomass（ AGB） of alpine grassland including alpine meadow and alpine steppe in the Three-River Headwaters Region from remote sensing data,a remote sensing-based model was developed through various regression analyses based on70 aboveground biomass observations during July and August from 2006 to 2014 and corresponding MODIS-NDVI and MODIS-EVI data.The spatial and temporal patterns of aboveground biomass from 2005 to 2014 were then estimated by using long time series MODIS data.Regression analyses of aboveground biomass and NDVI or EVI showed that the power model of EVI performed well to estimate alpine grassland aboveground biomass,with the highest determination coefficient of 0. 654,and the smallest root mean square error of 27. 86g/m^2. According to 70 AGB observations,the final established model was y = 348. 769x^0. 783,R^2= 0. 655, P 0. 001. The spatial patterns of aboveground biomass from 2005 to 2014 were similar,and reduced gradually from southeast to northwest. This trend was related to the rainfall, temperature, altitude and vegetation types in the region. The average aboveground biomass from 2005 to 2014 was 169. 25 t/a,with the highest value of178. 36 t/a in 2010,and the lowest value of 162. 80 t/a in 2008. The inter-annual variability of aboveground biomass had suffered significant changes,decreasing from 2005 to 2008 and increasing in fluctuation from 2008 to 2014. The results suggested that the developed model structure and parameters of aboveground biomass of alpine grassland in the Three-River Headwaters Region provide a reliable method for aboveground biomass research in this region.