Overview

Available here are the datasets producing 1.8M large river sediment concentrations derived from fused (previously MODIS) and non-fused (Landsat-5,7,8,9 & Sentinel-2) image reflectances made over 1,253 CONUS sites, as validated with over 25,000 in situ measurements across these same sites using machine learning. Satellite super-resolution data fusion ‘teaches’ a coarse resolution sensor (500m) what it would have seen if it were a fine resolution satellite (30m).  Datasets include:

trainingMatchups_2000-2023.csv - All Fusion and LS2 matchups within +/- 1 day of Water Quality Portal TSS measurements used to train a machine learning model between 2000-2023.

fusion_reflectance_raw.csv - All raw median reflectance, standard deviation, pixel count, date, and siteID data for fusion images generated between 2000-2023.

LS2_reflectance_raw.csv - All raw median reflectance, standard deviation, pixel count, date, and siteID data for LS2 images generated between 2000-2023.

fusionSed_all.csv - All TSS estimates from the RF model including matchup LS2, matchup Fusion, predicted LS2, and predicted Fusion between 2000-2023.