摘要
三江源黄河源园区作为黄河的发源地,它的生态系统对整个黄河流域至关重要。针对研究区2000—2020年存在的土地退化问题,选取土壤侵蚀模数、沙化特征指数与土壤湿度指数,采用空间托普利茨逆协方差聚类法开展了3个参数的聚类分析,从而进行黄河源园区土地退化时空演化分析,并采用极致梯度提升法进行时空演化影响因素分析。结果表明:(1)土地退化现象最严重的地区是中部偏南地区与东北地区,其次为偏北地区。(2)东北地区存在水土流失与土地沙化2种问题,中部偏南地区以严重的水土流失为主,偏北地区主要表现为土地沙化问题。(3)近20 a来,土地退化指数呈现波动性下降,土地退化整体趋于好转,在偏南地区表现得更明显。(4)水土流失问题整体好转,偏北地区有土地沙化的风险。该研究对掌握黄河源园区土地退化的时空分布及发展趋势、推进黄河源园区生态保护具有重要意义。
Because Sanjiangyuan’s Yellow River Source Park flows into the Yellow River,its ecosystem is crucial to the entire Yellow River Basin of China.Of the land degradation problems observed in the study area from 2000 to 2020,this study focuses on the soil erosion modulus,sanding characteristic index and soil moisture index.To achieve a more in-depth analysis of spatial and temporal evolution of land degradation in the area,machine learning algorithms were used.The spatial Toeplitz inverse covariance-based clustering method was used to carry out the cluster analysis of three parameters as part of the spatio-temporal evolution analysis of land degradation in the Yellow River Source Park.The extreme gradient boosting method was used to analyze the factors influencing the spatio-temporal evolution.The results of the study showed that:(1)The most significant land degradation was observed in the central-southern and northeastern regions,followed by the northern region.(2)In the northeast,there are two categories of problems:soil erosion and land desertification.In the central-southern region,severe soil erosion is the main problem.In the northern region,land desertification is the main problem.(3)In the past two decades,the land degradation index has shown a fluctuating decline that land degradation overall tends to improve,and that the effects are more obvious in the remote southern region.(4)The problem of soil erosion is improving overall,and there is a risk of land sanding in the remote northern areas.This study is a critical tool for understanding the spatial and temporal trends in the distribution and development of land degradation in the Yellow River Source Park and,by extension,for the promotion of increased ecological protection in Yellow River Source Park.
作者
秦彤
李功权
范佳晨
QIN Tong;LI Gongquan;FAN Jiachen(School of Geosciences,Yangtze University,Wuhan 430000,Hubei,China;School of Geography and Environmental Sciences,Northwest Normal University,Lanzhou 730070,Gansu,China)
出处
《干旱区地理》
CSCD
北大核心
2022年第5期1490-1499,共10页
Arid Land Geography
基金
国家自然科学基金项目(42004007)资助。
关键词
土地退化
遥感
空间托普利茨逆协方差聚类
极致梯度提升
land degradation
remote sensing
spatial Toeplitz inverse covariance-based clustering
extreme gradient boosting