Methodology & Data Sources

Last updated July 13, 2026

Every listing captured into Land Atlas is enriched once, at capture time, by querying public government and open-data services. The results are stored on your Atlas server; nothing is fetched from third parties while you browse the map. This page explains exactly where each number comes from, how the scores are computed, and what the analysis can and cannot tell you.

Data sources

Source	What it provides	Update cadence
USDA SSURGO (Soil Data Access)	Soil map units and the dominant soil component for the parcel: surface texture, pH, organic matter, available water capacity, CEC, drainage class, flood frequency, seasonal water-table depth, farmland classification, and NCCPI productivity indexes (overall, corn, soybean). Queried live against the national SSURGO database; results are area-weighted across your drawn parcel boundary, or an equal-area circle when no boundary is drawn.	Live queries; NRCS refreshes SSURGO annually (official refresh each October).
Open-Meteo ERA5 archive	Ten most recent complete years of daily minimum/maximum temperature and precipitation, reduced to farming numbers: average frost window, growing degree days (base 50 °F, ceiling 86 °F), chill hours (32–45 °F), annual rainfall, and heat-stress days (≥95 °F / ≥100 °F). Values are ERA5 reanalysis-model estimates on a ~9 km grid and are labeled as estimates, not station readings.	ERA5 is updated daily with a ~5-day lag; Land Atlas uses the last 10 complete calendar years, fetched once at capture.
FEMA NFHL	The effective flood zone at the listing pin, classified into plain English: high (Special Flood Hazard Area — A/V zones), moderate (shaded Zone X, 0.2% annual chance), minimal, or undetermined/not mapped.	Continuously updated as communities’ flood maps become effective; queried live at capture.
USFWS National Wetlands Inventory	Wetland polygons intersecting the parcel (drawn boundary or acreage circle) and their wetland types. Land Atlas flags intersection but does not calculate the polygon’s overlap acreage within the parcel.	NWI publishes data updates roughly twice a year (typically May and October).
USGS EPQS (3DEP elevation)	Elevation at the pin plus local relief, sampled from four points ~1.5 km away — what distinguishes a mountainside from a plateau.	Backed by the USGS 3DEP elevation model, improved continuously as new lidar is published.
OpenStreetMap (Overpass API)	Water proximity and frontage (water features within ~300 m = waterfront, within 2.5 km = near-water, coastline within 5 km), the nearest named water body, and infrastructure screening: substations within 25 km, transmission lines within 15 km, power plants within 25 km, and major roads within 8 km.	OSM is community-edited and updated continuously. Coverage is good but not complete — “none found” means none mapped in OSM within the search radius, not a guarantee of absence.
Nominatim (geocoding)	Converts an address to coordinates for captures that arrive without them. Every geocoded pin carries a precision label (exact, street, zip, city, or county) so you know how much to trust the placement.	Live OSM data.
US Drought Monitor	Drought category history for the listing’s area — how often and how severely the location has been in drought.	Published weekly (every Thursday) by NDMC, USDA, and NOAA.
USGS NWIS (groundwater)	Planned: groundwater levels from monitoring wells. Public USGS coverage proved too sparse for an automated per-parcel factor, so the Water Security analysis currently lists groundwater as unknown and tells you to pull state well logs instead of guessing.	Active sites report continuously; historical records vary by site.

All of these are free, public services queried directly — Land Atlas uses no licensed datasets, and deliberately excludes licensed ownership, contact, and MLS data.

How the scores work

The farmability score (0–100) blends the layer scores a listing actually has: soil 30%, water security 20%, climate 20%, terrain 10%, crop fit 10%, and land risk 10%. When a layer is missing, its weight is dropped and the rest are renormalized — a missing layer never counts as a zero. No overall number is shown unless at least one substantive layer (soil or climate) is present.

Soil score: NCCPI productivity 40%, drainage 20%, flooding & water table 15%, capability class 15%, composition (pH, organic matter, texture) 10% — renormalized when factors are missing.
Climate score: a transparent 0–100 built from the 10-year frost window, growing degree days, rainfall, and heat stress.
Water-security score: climate water balance (rainfall versus reference evapotranspiration), county-level 10-year drought history, mapped surface-water proximity, and soil water retention. Groundwater, water rights, irrigation-district status, and well yield remain explicitly unknown until verified documents are supplied.
Terrain and crop fit: terrain reflects dominant slope and local relief; crop fit is the average suitability of the three highest-ranked crops.
Land risk score: FEMA flood zone and wetlands presence, folded together with the parcel’s slope, local relief, and SSURGO flood frequency.

Labels map directly from the number: 80+ Excellent, 65+ Good, 50+ Fair, 35+ Marginal, below that Poor.

Three honesty rules apply everywhere:

Null over guesses. When a source has insufficient data, the field is empty and the score is marked partial with the missing layers listed — never a fabricated number.
Estimated vs. known. Model-derived values (ERA5 climate, sine-estimated chill hours, OSM-derived distances) are labeled as estimates; survey-derived values (SSURGO attributes, FEMA zones) are labeled with their source.
Every score shows its work. Soil, climate, and risk scores each ship with a factor-by-factor breakdown so you can see exactly how the number was built — this is screening-level intelligence for deciding which parcels deserve a closer look, not a verdict.

Limitations

Map units, not lab tests. SSURGO is a survey product (roughly 1:24,000 scale). It describes the mapped soil unit, not your exact acre — an on-site soil test can and sometimes will disagree. When no boundary is drawn, soil is sampled over a circle matching the listing’s acreage, which is an approximation of the real parcel shape.
Gridded climate, not a weather station. ERA5 values come from a ~9 km reanalysis grid. Microclimates — frost pockets, ridge-top wind, valley inversions — will not show up.
Comps are asking prices, not sales. The comparable-pricing card compares a listing’s asking price per acre against other listings you have captured in the same state at similar acreage (0.4×–2.5×). It is not sold-comp data and not an appraisal, and it says so on every card.
Cost figures are planning ranges. Any development or improvement costs shown are rough planning ranges for screening, not quotes or engineering estimates.
Some factors share source data. Crop fit is derived from soil and climate, water security also uses climate and soil, and terrain contributes to land risk. The chips are decision lenses, not statistically independent measurements; do not interpret the overall score as a confidence percentage.
OSM completeness varies. Infrastructure and water results reflect what volunteers have mapped; absence in the data is not proof of absence on the ground.
Not a professional report. Nothing here is an appraisal, a boundary survey, an engineering study, or an environmental assessment. Before purchase, verify with a licensed surveyor, a soil test, a title search, and the county — Land Atlas tells you which parcels are worth that effort.

Data freshness & contact

Enrichment runs once when a listing is captured and the results are stored, so each listing’s analysis reflects the source data as of its capture date. Re-running analysis (for example, after drawing a parcel boundary) re-queries the live sources. Because most sources update on annual or weekly cycles, stored results rarely drift far — but a listing captured long ago can be recomputed to pick up new source data.

Questions about the methodology, corrections, or requests: taureanclee@gmail.com.