Using the random forest model and validated MODIS with the field spectrometer measurement promote the accuracy of estimating aboveground biomass and coverage of alpine grasslands on the Qinghai-Tibetan Plateau

作者：Xiaoxia Gao, Shikui Dong, Shuai Li, Yudan Xu, Shiliang Liu, Haidi Zhao, Jane C. Yeomans, Yu Li, Hao Shen, Shengnan Wu, Yangliu Zhi · 发表于：Ecological Indicators · 年份：2020 · DOI：10.1016/j.ecolind.2020.106114 · 被引用次数：75 · 研究领域：Remote Sensing in Agriculture、Species Distribution and Climate Change、Ecology and Vegetation Dynamics Studies

Maintaining the aboveground biomass (AGB) and canopy cover (CC) of grassland is key for sustainable grassland management and rational conservation planning, especially for fragile regions of Qinghai-Tibetan Plateau (QTP). However, the estimation of AGB and CC with land use changes is often challenging due to poor availability of high-quality images of remote sensing or the large-scale field survey data. In this study, we used the transect survey method combining with quadrat measurements with 1200 plant samplings and near surface hyperspectral measurements at 180 sites to validate the MODIS data. An integrative approach combining the plant sampling, near surface hyperspectral measurements, remote sensing data interpretation, random forest modelling and Hurst phenomenon prediction was developed to accurately and efficiently estimate the CC and AGB of the alpine grasslands on the QTP from 2000 to 2017 and their future trends. The results showed: 1) estimation of AGB and CC in different land use types by integrating random forest with multiple factors (NDVI, latitude, longitude and altitude, grassland types) was more accurate; 2) the inter-annual changes of AGB and CC of the alpine grasslands presented significant fluctuations, with an increasing trend in the QTP; 3) the areas with improved vegetation conditions of AGB (66.55%) and CC (64.58%) were much larger than those with degraded vegetation conditions of AGB (33.45%) and CC (35.42%) in past 20 years. 4) there are 60.66% of ...