ChemLotUS: A Benchmark Dataset of Lotic Chemistry across US River Networks

ChemLotUS contains water quality information in rivers and streams carefully curated and linked to a high resolution stream network. The dataset covers the contiguous United States (CONUS) with nearly 35 million records in about 290,000 sites. 11 water constituents are included so far: Calcium (Ca), Electrical Conductivity (EC), pH, Total Suspended Solids (TSS), Turbidity (Tu), Total Organic Carbon (TOC), Dissolved Oxygen (DO), Chlorophyll a (Chl-a), Nitrate (NO3), Soluble Reactive P (SRP), and Total P (TP). Raw water quality data comes from the Water Quality Portal (WQP - https://www.waterqualitydata.us/), and network information comes from the National Hydrography Dataset High Resolution (NHD-HR - https://www.usgs.gov/national-hydrography/nhdplus-high-resolution).

The repository contains the ChemLotUS dataset, as well as the code developed to obtain it, following the structure described on the Readme1 file.
The Readme2 file contains the definitions of the columns contained in each of the csv files comprised in the dataset.

This version is preferred over the two previous ones, considering the following differences:

1. With DOI: 10.4211/hs.595e2ba483524feba22273e4dfbed190 (Apr 10, 2025): During a workshop conducted in May 5 - 9, 2025 at UF using this previous version, we identified potential outliers not identified as such in the csv files. The new version solves this issue.
2. With DOI: 10.4211/hs.266099a29929423e9cec8bdb886b6a21 (Aug 26, 2024): Noting that this is the version originally submitted to peer review, the subsequent versions follow the recommendations made by the reviewers.

More details about the ChemLotUS dataset, including the methods employed to obtain it and several example use cases, are presented in the homonymous paper "ChemLotUS: A Benchmark Dataset of Lotic Chemistry across US River Networks", published on XX YY the Journal Water Resources Research (DOI 10.1029/2024WR039355).