基于神经网络的三江源区草地地上生物量估算

发布时间：2018-02-04 01:56

本文关键词： 草地地上生物量三江源 BP神经网络时空变化　出处：《环境科学研究》2017年01期 　论文类型：期刊论文

【摘要】：三江源区位于青藏高原腹地,作为长江、黄河、澜沧江三大河流的发源地,是我国重要的水源涵养和生态功能保护区.为了及时准确地获取该区域草地生物量信息,根据三江源区高寒草甸、高寒草原采样点的地上生物量实测值,结合遥感植被指数、海拔、气象观测数据(光合有效辐射、年均气温、年降水量)构建BP神经网络模型,估算2001—2010年三江源区的草地地上生物量,并对其进行分县统计和年际变化分析.结果表明:1通过多次反复的训练与测验得到的BP神经网络模型,对高寒草甸、高寒草原的地上生物量模拟值与实测值的R2分别为0.73、0.79,表明BP神经网络模型具有较好的模拟效果.22001—2010年三江源区草地地上生物量多年平均值为172.34 g/m~2,其中高寒草甸为214.81 g/m~2,高寒草原为130.07 g/m~2.3三江源区草地地上生物量的空间分布具有明显的空间异质性,呈从东南向西北递减的趋势.其中,位于东部的河南县草地地上生物量最高,为413.46g/m~2;而北部的曲麻莱最低,仅为69.04 g/m~2.42001—2010年三江源区草地地上生物量呈缓慢波动上升趋势,平均升幅为0.93g/(m~2·a).研究显示,利用站点地上生物量实测数据构建BP神经网络模型并对地上生物量进行模拟,对于分析区域尺度的草地地上生物量分布格局和变化趋势行之有效.
[Abstract]:As the birthplace of the Yangtze River, the Yellow River and the Lancang River, the source area of the three rivers, located in the hinterland of the Qinghai-Tibet Plateau, is an important water conservation and ecological protection area in China. Based on the measured results of aboveground biomass of alpine meadow and alpine steppe in the source of the three rivers, the BP neural network model was constructed by combining with remote sensing vegetation index, elevation and meteorological observation data (photosynthetic effective radiation, average annual temperature, annual precipitation). The aboveground biomass of grassland in the source area of the three rivers from 2001 to 2010 was estimated and analyzed by county statistics and interannual variation. The results showed that the BP neural network model obtained by repeated training and test was applied to alpine meadow. The simulated and measured values of aboveground biomass of alpine steppe were 0.73n0.79, respectively, which indicated that BP neural network model had better simulation effect. 2The annual average of aboveground biomass of grassland in the source region of the three rivers from 2001 to 2010 was 172.34 g / m ~ (-2), among which the alpine meadow was 214.81. The spatial distribution of aboveground biomass in the alpine steppe was 130.07 / g / m ~ (2.3) in the source region of the three rivers, and the spatial distribution of the aboveground biomass in the alpine steppe had obvious spatial heterogeneity. The aboveground biomass of grassland in Henan county in the east was the highest (413.46 g / m ~ (-2)), while Qumailai in the north was the lowest, which was only 69.04 g / m ~ (2. 4) from 2001 to 2010. The aboveground biomass of grassland in the source region of the three rivers increased slowly from 2001 to 2010. The average increase was 0.93 g / m ~ (2) 路a ~ (-1). The results showed that the BP neural network model based on the measured data of aboveground biomass was used to simulate the aboveground biomass, which was effective in analyzing the distribution pattern and variation trend of aboveground biomass on the regional scale.
【作者单位】：中国科学院地理科学与资源研究所生态系统网络观测与模拟重点实验室;中国科学院大学;中国科学院地球化学研究所;青海省生态环境遥感监测中心;中国环境科学研究院;
【基金】：中国工程院重点咨询项目(2014-XZ-31) 国家重点基础研究发展计划(973计划)项目(2014-YKY-003) 国家科技支撑计划项目(2015CB954102)
【分类号】：S812
【正文快照】： 三江源位于青海省南部,是长江、黄河和澜沧江的源头,同时也是重要的草地生态功能区,该区域的生态系统安全问题受到国内外研究人员的广泛关注[1-3].草地作为三江源区分布最广的生态系统类型,一直是该区域生态研究的重点对象.生物量作为全球碳循环的重要组成部分,在全球碳平衡中

【相似文献】