Daily Estimates of Chlorophyll Concentration for 82 U.S. Rivers

This data release contains modeled daily chlorophyll concentration for 82 streams and rivers across the conterminous United States. Estimates of daily chlorophyll concentration were generated using an extreme gradient boosting (XGBoost) (Chen & Guestrin, 2016) machine learning model and various measures of water quality, reach characteristics, and meteorology. Child Items in this data release contain: Site information: This child item contains station identifiers, location information, location type, stream order, and associated NHDPlusv2.1 reach common unique identifiers (ComID). The file format is tab-delimted with 1 row per station. Model inputs and outputs: This child item contains XGBoost model inputs, observed chlorophyll concentrations, model estimates of chlorophyll and SHapley Additive exPlanations (SHAP) values. SHAP values were used to assess the contribution of variables to estimated chlorophyll. Approval status: This child item contains the data source of chlorophyll data and the approval status for all USGS NWIS data at the time the data was downloaded (10/27/2021).