Data provenance & methodology | SWATGenX

Canonical reference for flood screening, rainfall context, national inputs, and SWAT+ packaging — with steward citations and known limitations.

This page is the single hub for how SWATGenX implements screening and modeling. Product and category pages summarize outcomes; they link here for proof, dataset stewardship, and appropriate use.

Data provenance (summary)

Native resolution is what the steward documents for the source product. SWATGenX may resample or tile data for operational speed; model-ready time series and rasters are documented in pipeline code.

Dataset	Steward / reference	Native resolution / cadence (steward)	Licensing / attribution	Role in SWATGenX
NHDPlus HR	USGS NHDPlus HR (USGS cites on the order of ~27 million flowlines vs ~3 million in NHDPlus V2)	Built from 1:24,000 NHD, 10 m 3DEP-derived terrain, and WBD — per USGS product documentation	USGS federal open data norms; cite USGS + product version in publications	National hydrography backbone for delineation and screening maps
WBD / HUCs	USGS Watershed Boundary Dataset	Vector HUC polygons (official hierarchy per USGS)	USGS; cite WBD release used	HUC8/HUC12 organizational units for basin selection and reporting
PRISM	Oregon State PRISM Climate Group	Gridded climate commonly on ~4 km cells; PRISM defines a day as the 24 hours ending at 12:00 GMT — confirm in PRISM metadata for your extract	PRISM requests attribution; follow OSU PRISM citation guidance	Climate inputs for SWAT+; historical daily precip percentiles for rainfall status (see below)
NLCD	USGS Multi-Resolution Land Characteristics (MRLC)	30 m national land cover (Landsat-based; nominal cell size per MRLC product)	MRLC data use / citation policy (see MRLC site)	Land cover for HRUs and reporting
gSSURGO	USDA NRCS gSSURGO	Gridded SSURGO; typical national grids ~10 m — verify against your gSSURGO snapshot	NRCS / USDA terms; cite SSURGO/gSSURGO source	Soils for SWAT+ HRUs
NSRDB	NREL National Solar Radiation Database	NREL documents ~4 km horizontal spacing (≈0.038°) and 30-minute solar time series for CONUS-class products — confirm version in use	NREL NSRDB terms of use; cite NREL + dataset version	Solar, humidity, and wind drivers for SWAT+ weather generators where used
USGS streamflow	USGS Water Services (Instantaneous Values, Daily Values, peaks; discharge parameterCd 00060)	IV often ~15-minute measurements (transmission cadence varies); DV daily; peak records for frequency — see USGS for provisional vs approved data	USGS Water Data policies; provisional data disclaimer where applicable	Live gage context; annual peaks for flood frequency
MRMS (where shown)	NOAA MRMS	Multisensor radar–gauge products; resolution and latency per NOAA documentation	NOAA / NSSL usage terms	Multi-sensor radar/gauge precipitation products for near-real-time map context

Flood screening methodology

Annual peak streamflows are retrieved from USGS peak records. SWATGenX fits a Log-Pearson Type III distribution to the log10 of peaks and computes quantiles for return periods Q2 through Q500, following the structure of USGS Techniques and Methods 4–B5 (Bulletin 17C). Station status classes (e.g. normal, minor, moderate, major) compare recent flow to those thresholds.

Regulatory mapping: Gage-based return-period classes are screening indicators, not FEMA Flood Insurance Rate Maps, NFIP regulatory floodplains, or a substitute for licensed hydraulic engineering and HEC-RAS-class studies.
Return period (e.g. Q100): average recurrence interval language — not a prediction of the next flood year.
Limitations: short records, zero-flood years, regulation, urbanization, and climate non-stationarity affect frequency estimates. Bulletin 17C discusses uncertainty and methodological choices; users should treat outputs as triage.
Deeper flood studies often use event hydrology and hydraulic models (e.g. USACE HEC-HMS / HEC-RAS); see Flood risk modeling software and SWAT+ vs HEC-HMS.

Rainfall percentile methodology (PRISM)

For each USGS station, SWATGenX maps the gage to a PRISM grid cell (same mesh as operational PRISM extraction). It builds a historical series of daily precipitation at that cell and computes percentiles p50, p75, p90, p95, p99. The latest 24-hour accumulation is classified (e.g. extreme ≥ p99) for map status. Daily totals follow PRISM’s definition of a meteorological day (24 hours ending 12:00 GMT per PRISM documentation).

Default baseline years in batch statistics: 2000–2022 (inclusive) where those daily grids exist — see compute_rainfall_statistics_for_stations in prism_rainfall_stats.py. Percentiles are spatially specific to the PRISM cell; they are not design-storm intensities from NOAA Atlas 14.

MRMS provides complementary radar multisensor context for recent conditions; interpret radar products with NOAA documentation and known blind spots.

SWAT+ generation chain (artifacts)

SWAT+ is a restructured, object-oriented successor to SWAT with relational (SQLite) project databases managed through SWAT+ Editor. SWATGenX automates QSWAT+ (GIS/delineation and setup) and SWAT+ Editor configuration in a server pipeline, then packages outputs for download.

Typical package contents: project folder with SQLite Editor database, SWAT+ text inputs, climate time series (e.g. PRISM-derived .pcp/.tmp and related formats), HRU and routing setup, and documentation — exact layout matches the pipeline version that produced it.
After download: open in SWAT+ Editor, adjust management or parameters, run SWAT+ locally or in your HPC environment.

More workflow context: SWAT+ model generation, How it works.

How to cite outputs

Cite the underlying data stewards (USGS, PRISM, USDA NRCS, MRLC, NREL, NOAA as applicable) per each dataset’s citation guidance. Cite SWAT+ using Texas A&M / SWAT documentation appropriate to your version. For SWATGenX-delivered reports or packages, note the platform name, approximate run date, and watershed identifier (USGS site or HUC) so results can be reproduced.

Watershed modeling guide — taxonomy and definitions
Watershed modeling software — tool selection
Watershed Explorer — interactive workspace