Data for: Characterization of large-scale preferential flow across continental United States

Understanding preferential flow (PF) at large scales is critical for improving land management and groundwater (GW) quality. However, limited knowledge of this process, due to soil surface heterogeneity and observational constraints, hampers progress. In this study, we propose estimating effective PF at remote sensing footprint scale (4 – 9 km) by examining its impact on soil moisture (SM) distribution and shallow GW (SGW) table fluctuations (depth  5 m). Effective PF encompasses macropore, funnel, and finger flow pathways influencing SGW table fluctuations. We compiled daily SGW observations (2019-2021) from 19 continental US (CONUS) sites through USGS. Using inverse modeling in HYDRUS-1D, SGW data, and CHIRPS precipitation data, we inversely estimated soil hydraulic parameters of the dual porosity model (DPM) simulating vertical flow from soil surface to subsurface. Effective PF presence was inferred using three criteria: (1) daily precipitation >=  the site-specific average across multiple (calibration) years, (2) daily observed SGW table increase, and (3) daily difference between observed and DPM simulated SGW tables 50% of the site-specific RMSE. Leveraging optimized DPM parameters and associated soil texture, classified PF events, and Soil Moisture Active Passive (SMAP L3E) satellite-based SM, a Random Forest algorithm with 10-fold cross validation predicted large-scale effective PF events. Results indicate seasonal dependence, with spring having the highest occurrence of PF events. The Random Forest model achieved 98% accuracy in predicting large-scale PF events, with SMAP SM and saturated hydraulic conductivity (Ks) among the 4 most impactful variables. Our approach provides a soil hydraulic property, site characteristic, soil texture and remote sensing based generalized tool to analyze large-scale effective PF.

Methods
A total of 19 sites with diverse soil-vegetation-climate characteristics across CONUS were selected for the study (Fig. 2). Daily mean GW data at these sites were gathered from observation from the USGS National Water Information System (NWIS). Following similar approaches from Babajimopoulos et al., 2007 and Costa et al., 2023, sites with GW tables less than or equal to 5 meters were defined as shallow and GW tables deeper than 5 meters were defined as deep. Out of the 19 sites, 3 sites (Laurens, Georgia; Tioga, New York; Saline, Missouri) had deep GW table data, 12 sites (Blaine, Nebraska; Jones, North Carolina; St. Lawrence, New York; Chatham, Georgia; Red Lake, Minnesota; Walker, Texas; Kalamazoo, Michigan; Monmouth, New Jersey; St. Croix, Wisconsin; Big Horn, Montana; Pasquotank, North Carolina; Tazewell, Illinois) had SGW table data from 2019-2021, and 4 sites (Hooker, Nebraska; Garden, Nebraska; Duplin, North Carolina; Clinton, Illinois) had SGW data from 2021-2022. Precipitation data at a daily scale and 4800 m spatial resolution were collected from the Climate Hazards Group Infrared Precipitation with Station data (CHIRPS), accessed through Climate Engine. Van Genuchten soil hydraulic parameters from the soil catalog in HYDRUS-1D for various textural classes (Carsel & Parrish, 1988) were used as initial parameters in the inverse modeling of the dual porosity model (DPM) (Gerke & van Genuchten, 1993). Site-specific surface soil texture, slope, and elevation were collected from the Soil Survey Geographic Database (SSURGO) at a 250 m spatial scale for the selected study soils. In this study, we assumed vertical soil texture homogeneity, while addressing soil surface lateral heterogeneity by selecting sites across CONUS with diverse surface soil textures. NASA’s SMAP L3E product provided SM (z < 5 cm) at 2–3-day intervals with a 9-km resolution across CONUS. SM data from SMAP was accessed through the Earth Data Application for Extracting and Exploring Analysis Ready Samples (AppEEARS) database. Soil physical properties such as bulk density from the World Soil Information Service (WoSIS) database were gathered from SoilGrids with 250 m grid size. The 12 sites with SGW table data from 2019-2021 in conjunction with CHIRPS precipitation data as input were used to inversely model the soil hydraulic parameters as output using the DPM. The 4 remaining independent sites with SGW table data from 2021-2022 were used to validate the inversely modeled soil hydraulic parameters by forward modeling using DPM.