Near Channel Suspended Sediment Prediction Dataset

High concentrations of suspended sediment (SSC) in a river can represent a critical water quality concern, reduce the storage capacity of reservoirs, and impact aquatic habitat. Within a river, SSC can be conceptualized as a function of reach-scale hydraulics translating discharge into shear stress and watershed processes that determine the types and quantities of sediment supplied to the river. To explore watershed controls on sediment supply, we utilized SSC data from over 1000 US Geological Survey gages spread across the continental United States (CONUS). We find that the geometric mean SSC spans over five orders of magnitude with clustered high and low values throughout the CONUS indicating a dependence on regional watershed properties. Here we utilize publicly available geospatial datasets (topography, soils, land use, and climate) to explore the potential dependence of mean SSC for over 100 variables. We find that catchment-wide and point-scale geospatial variables provide few explanatory univariate trends for the observed mean SSC patterns. We utilized principal components analysis to reduce the dimensionality of the exploration to a limited subset of variables. Extreme variability within mean SSC and data limitations prevents a complete prediction of SSC from geospatial data, however multiple nonlinear regression reveals that the geospatial pattern in mean SSC is primarily a function of climate (aridity), vegetation, and soil properties. Understanding SSC dependence on watershed properties represents an important step for linking watershed processes and fine-grained transport dynamics and how changes in climate and the environment may further affect sediment volumes and watershed management.

河流中悬浮泥沙浓度（Suspended Sediment Concentration, SSC）处于高值时，会构成严重的水质隐患，降低水库库容，并对水生栖息地造成不利影响。在河流系统中，悬浮泥沙浓度可被视为河段尺度水动力条件与流域过程共同作用的结果：前者将流量转化为剪切应力，后者则决定了输入河流的泥沙类型与总量。为探究流域对泥沙供给的调控作用，本研究使用了分布于美国本土（Continental United States, CONUS）境内1000余座美国地质调查局（US Geological Survey, USGS）监测站点的悬浮泥沙浓度数据。研究发现，美国本土范围内悬浮泥沙浓度的几何均值跨越五个数量级，高低值呈集群分布，表明其与区域流域属性密切相关。本研究借助公开可用的地理空间数据集（地形、土壤、土地利用与气候数据），针对逾100个变量探究其与悬浮泥沙浓度均值的潜在关联。研究发现，流域尺度与点位尺度的地理空间变量，仅能对观测到的悬浮泥沙浓度均值分布模式提供有限的单变量解释力。本研究采用主成分分析（Principal Components Analysis, PCA）将探索维度缩减至有限的变量子集。尽管悬浮泥沙浓度均值存在极强的变异性，且受限于数据可得性，无法仅通过地理空间数据实现悬浮泥沙浓度的完整预测，但多元非线性回归分析结果显示，悬浮泥沙浓度均值的地理空间分布模式主要受气候（干旱度）、植被与土壤属性的调控。明晰悬浮泥沙浓度与流域属性的关联，是连接流域过程与细颗粒泥沙输移动力学、厘清气候与环境变化如何进一步影响泥沙总量及流域管理的重要一环。