为快速准确获取区域内土壤盐分含量(Soil salt content,SSC)信息及空间分布特征,选取莱州湾南岸滨海平原地区为研究区,系统采集裸土期土壤样品95个并获取同时相Sentinel-2多光谱影像,进一步利用变量重要度评估技术选取SSC的敏感波段作...为快速准确获取区域内土壤盐分含量(Soil salt content,SSC)信息及空间分布特征,选取莱州湾南岸滨海平原地区为研究区,系统采集裸土期土壤样品95个并获取同时相Sentinel-2多光谱影像,进一步利用变量重要度评估技术选取SSC的敏感波段作为输入自变量,测定得到的SSC值为因变量,分别建立基于随机森林和空间关联随机森林算法的遥感估算模型,完成区域尺度上的SSC反演制图。结果表明:Sentinel-2影像近红外范围内波段(B11、B12、B2和B8a)对SSC响应较为敏感,波段11的重要度值最高。空间关联随机森林模型的估算效果指标R^(2)(决定系数)和RMSE(均方根误差)分别为0.86和0.38,相对比随机森林模型估算效果提升了16.22%和35.60%,模型的计算效率和稳健性也相应提高。SSC在区域上整体呈现较低含量水平(<2 g/kg),高值区分布在西北部和东部部分区域(>6 g/kg),主要受到海水入侵影响。此外,在微地貌和干旱蒸发的作用下形成的离散块状分布的中度土壤盐化区(2 g/kg≤SSC<4 g/kg)应引起重视。展开更多
大兴安岭林区土壤有机碳储量对于区域碳源汇变化及气候响应研究具有重要意义。本文基于土壤剖面数据,结合多种环境变量,采用随机森林(RF)模型,估计了大兴安岭卡马兰河流域0~30 cm深度的土壤有机碳储量(SOCS)及其空间分布特征。结果表明...大兴安岭林区土壤有机碳储量对于区域碳源汇变化及气候响应研究具有重要意义。本文基于土壤剖面数据,结合多种环境变量,采用随机森林(RF)模型,估计了大兴安岭卡马兰河流域0~30 cm深度的土壤有机碳储量(SOCS)及其空间分布特征。结果表明:RF模型(R2 = 0.75, RMSE = 6.81)对小尺度区域的模拟细节表现较好,研究区0~30 cm的SOCS主要分布在地势相对平坦的河流沿岸,与河流走向基本保持一致。研究结果得出了相对精确的卡马兰河流域表层土壤有机碳储量及空间分布特征,该成果能够丰富大兴安岭多年冻土区土壤有机碳储量的认识,并为生态过程相关模拟研究提供数据支持。Soil organic carbon stocks in the forested areas of Great Khingan are of great significance for the study of regional carbon source and sink changes and climate response. In this paper, based on soil profile data and combining multiple environmental variables, we estimated the soil organic carbon stock (SOCS) and its spatial distribution characteristics at 0~30 cm depth in the Kamalan River watershed of Great Khingan by using the random forest (RF) model. The results showed that the RF model (R2 = 0.75, RMSE = 6.81) performed well in simulation details for small-scale areas, and the SOCS at 0~30 cm in the study area was mainly distributed along the river with relatively flat topography, which was basically consistent with the river course. The results of the study yielded a relatively accurate characterisation of the surface soil organic carbon stock and spatial distribution in the Kamalan River basin, which can enrich the understanding of soil organic carbon stock in the perennial permafrost region of the Great Khingan and provide data support for simulation studies related to ecological processes.展开更多
文摘大兴安岭林区土壤有机碳储量对于区域碳源汇变化及气候响应研究具有重要意义。本文基于土壤剖面数据,结合多种环境变量,采用随机森林(RF)模型,估计了大兴安岭卡马兰河流域0~30 cm深度的土壤有机碳储量(SOCS)及其空间分布特征。结果表明:RF模型(R2 = 0.75, RMSE = 6.81)对小尺度区域的模拟细节表现较好,研究区0~30 cm的SOCS主要分布在地势相对平坦的河流沿岸,与河流走向基本保持一致。研究结果得出了相对精确的卡马兰河流域表层土壤有机碳储量及空间分布特征,该成果能够丰富大兴安岭多年冻土区土壤有机碳储量的认识,并为生态过程相关模拟研究提供数据支持。Soil organic carbon stocks in the forested areas of Great Khingan are of great significance for the study of regional carbon source and sink changes and climate response. In this paper, based on soil profile data and combining multiple environmental variables, we estimated the soil organic carbon stock (SOCS) and its spatial distribution characteristics at 0~30 cm depth in the Kamalan River watershed of Great Khingan by using the random forest (RF) model. The results showed that the RF model (R2 = 0.75, RMSE = 6.81) performed well in simulation details for small-scale areas, and the SOCS at 0~30 cm in the study area was mainly distributed along the river with relatively flat topography, which was basically consistent with the river course. The results of the study yielded a relatively accurate characterisation of the surface soil organic carbon stock and spatial distribution in the Kamalan River basin, which can enrich the understanding of soil organic carbon stock in the perennial permafrost region of the Great Khingan and provide data support for simulation studies related to ecological processes.
