Bankfull and Mean-flow Channel Geometry Estimation through a Hybrid Multi-Regression and Machine Learning Algorithms across the CONtiguous United States (CONUS)

This dataset includes estimated river channel geometry attributes, specifically width and depth, under bankfull and mean-flow conditions across the CONtiguous United States (CONUS). The method utilized for providing these estimations are based on a hybrid approach coupling Multi Linear Regression (MLR) and eXtreme Gradient Boosting Regression (XGBR). This dataset can be linked to the National Hydrograohy Dataset Plus (NHDPlusV2.1) through Common identifier of the NHD feature, known as COMID.