兰州市南北两山土壤水分遥感反演及植被需水量估算

Remote sensing retrieval of soil moisture and estimation of vegetation water requirements in the north and south mountains of Lanzhou City

探究西北干旱区土壤水分和植被需水量动态变化特征,可为生态恢复不同阶段所需水资源量及水资源优化配置提供科学依据。以兰州市南北两山为研究区,基于Sentinel-2 L2A和Landsat 8 OLI遥感影像,结合实测土壤0~10 cm的111个数据,分别构建垂直干旱指数(Perpendicular Drought Index,PDI)、改进型垂直干旱指数(Modified Perpendicular Drought Index,MPDI)和植被调整垂直干旱指数(Vegetation-adjusted Perpendicular Drought Index,VAPDI)土壤水分反演模型,并采用4种模型指标定量决定系数(R^(2))、平均绝对误差(MAE)、平均相对误差(MRE)、均方根误差(RMSE)对模型反演的效果进行精度评价,选出最优的土壤水分反演模型并结合土壤水分限制系数,与研究区2019年林地、草地和耕地植被面积的空间数据、各站点生长季内的参考作物蒸散量,构建植被生态需水量模型,厘清研究区内土壤水分、植被需水量时空变化特征。结果表明:(1)2种数据源下的PDI、MPDI、VAPDI和实测数据之间均有着不同程度的线性负相关性,其中R^(2)分别为0.37、0.64和0.59,从评价指标的结果来看,MPDI的土壤水分回归模型的拟合决定系数最高,2种遥感数据反演的土壤水分空间分布格局具有一致性。(2)分辨率高的Sentinel-2L2A土壤水分反演更加精细,土壤水分整体呈波动增长趋势,多时段土壤水分的平均值为23.27%,呈现出降低再增加然后下降,总体增幅为74.07%。(3)兰州市南北两山4—10月植被需水量月均值也呈现先增加后下降的趋势,与土壤水分含量变化具有一致性,4—10月中7月植被需水量最大,为3.98×10^(7)m^(3),10月植被生态需水量最小,为0.97×10^(7)m^(3)。随着环境绿化工程的实施,兰州市南北两山从只能生长耐旱草本和低矮灌木的植物,逐步形成以多种类结合的群落结构。本研究可为兰州市南北两山土壤水资源合理利用及植被恢复提供参考。

To understand the dynamic change characteristics of soil moisture in the arid region of Northwest China,the relationship between vegetation water requirement and soil moisture was explored.The perpendicular drought index(PDI) was determined on the basis of Sentinel-2 L2A and Landsat 8 OLI remote sensing data in combination with 111 soil surface measurements in the 0-10 cm layer.The PDI,modified PDI(MPDI),and vegetation-adjusted PDI(VAPDI) were used to construct a soil moisture inversion model,and four quantitative indicators—determination coefficient(R^(2)),mean absolute error(MAE),mean relative error(MRE),and root mean square error(RMSE)—were used to assess the accuracy of the inversion model.The optimal soil moisture inversion model was selected and used in combination with the soil moisture limiting coefficient.Spatial data of the vegetation area of forest land,grassland,and cultivated land in the study area in 2019 and reference crop evapotranspiration data during the growing season at each station were collected,and a model of the ecological water requirement of vegetation was constructed to explore the spatiotemporal changes in soil moisture and vegetation water requirement in the study area.The results showed that(1) PDI,MPDI,and VAPDI determined using the two data sources showed a linear negative correlation with the measured data to varying degrees,and the coefficient of determination R^(2)was 0.37,0.64,and 0.59,respectively.The model evaluation indicators suggested that the soil moisture regression model of MPDI had the highest fitting coefficient of determination.The spatial distribution patterns of soil moisture obtained from the two remote sensing data were consistent.(2) The high-resolution Sentinel-2 L2A soil moisture inversion was more precise,and the overall soil moisture showed a fluctuating growth trend.The multitime average of soil moisture was 23.27%;it showed a trend of initial decrease,followed by an increase and subsequent decrease,with an overall growth rate of 74.07%.(3) The average vegetation water requirement and soil moisture content of the northern and southern mountains of Lanzhou City from April to October showed a trend of fluctuation and decline.The maximum vegetation water requirement between April and October was 3.98×10^(7)m^(3)—observed in July—and the minimum water requirement was 0.97×10^(7)m^(3)—observed in October.With the implementation of the environmental greening project,the northern and southern mountains of Lanzhou City have gradually formed a community structure of multispecies combination from only droughttolerant herbs and low shrubs.In general,this study provides a reference for the rational use of soil water resources and restoration of vegetation in the two mountains of Lanzhou.