USDA Forest Service FIA DataMart

USDA Forest Service FIA DataMart is the Timber REITs catalog's window onto the Forest Inventory and Analysis Database (FIADB) - the United States' standing census of its own forestland. Where the markets half of this slice tracks what timberland-owning companies report, this record measures the trees themselves: extent, ownership, species composition, volume, biomass, growth, removals and mortality, collected on permanent ground plots and organized as roughly 59 table types per state.

The unit of observation is what makes it valuable. The PLOT table carries 63 columns of plot-level spine - identification, coordinates, elevation - while TREE runs to 105 columns per individual stem, from species code and diameter through net and gross cubic-foot volume to tree age. A companion estimation engine turns those plots into population estimates with sampling errors, and an urban-forestry wing extends the same plot design into cities. Datadory scores the record 9 out of 10 against a catalog mean of 7.81. Get a sample of this dataset cut to the states and tables your models need.

One row per measured thing, tables stitched together by CN keys. Three rows straight from the record's own sample:

table        : WY_PLOT
STATECD      : 1
UNITCD       : 3
COUNTYCD     : 123
PLOT         : 90042
INVYR        : 1990
LAT          : 32.850131
LON          : -85.835005
ELEV         : 980

table        : WY_SURVEY
STATECD      : 56
STATENM      : Wyoming
INVYR        : 2011
ANN_INVENTORY: Y
NOTES        : Annual 01 of 10 subcycles

table        : AL_TREE
PLT_CN       : 157499420010854
STATUSCD     : 1
SPCD         : 693
DIA          : 10.2
VOLCFNET     : 6.573

Read together, they show the shape of the whole system. The first row is a plot's identifying spine - state, survey unit and county codes stacking into a location, an inventory year, coordinates and elevation in feet. The second is survey bookkeeping, its note field recording that a state's panel rotates through ten annual subcycles. The third is the payoff: a single live stem - STATUSCD 1 - carrying species code 693, a 10.2-inch diameter and 6.573 net cubic feet of volume, tied back to its parent plot through the PLT_CN key. Every published estimate about US forests bottoms out in rows like that one.

Twenty-four fields anchor the dictionary below, each definition checked against the source's published references during the August 2026 research pass. They organize into four families. Identity and join keys (CN, and PLT_CN as it appears in TREE) stitch the tables to one another. Geography codes (STATECD, UNITCD, COUNTYCD, PLOT) stack into a location, with LAT/LON and ELEV adding physical position. Measurement lives in the TREE family (STATUSCD, SPCD, DIA, HT, VOLCFNET, VOLCFGRS). The COND family describes the ground a stem stands on: ownership group, forest type, stand age, productivity class and site index.

Columns that appear only inside particular tables or subsets fold under additional fields on request: the remainder of TREE's 105 columns, the growth-and-mortality transition family, regeneration and site-tree tables, the coarse-woody-debris and soils extensions, and the evaluation-unit bookkeeping behind population estimates. Name the ones your models need and the sample ships with them documented.

Geography - all 50 states plus territories, organized state by state so a region resolves cleanly. One constraint matters for mapping: public plot coordinates are deliberately fuzzed or swapped, and precise locations publish only for plots numbered above 40,000.

Temporal - annual inventory panels state by state, with observed collections running from early panels in 1984 through current 2022-2024 cycles. States entered annual sampling at different times, so depth varies by state rather than by decree.

Granularity - tree-level and plot-level records joined through CN keys, rolling up through counties and survey units to population-evaluation units. A single state's TREE table alone reaches 541 MB, which is what 105 columns times millions of stems looks like at rest.

API, files, or your warehouse. Daily, weekly, or hourly.

Your cadence is your call regardless of how the underlying panels move - a monthly warehouse load keeps supply dashboards fed between inventory cycles nobody has to babysit. Every delivery ships the complete field dictionary above, the sample rows and the coverage profile mapped to the states and tables you named.

• Timber-supply and biomass modeling - join plot- and tree-level tables across all 50 states into standing-volume and biomass baselines that wood-product models actually rest on.
• Timberland valuation inputs - site index, productivity class and stand age per condition class give growth-rate assumptions for discounted-cash-flow land math.
• Ownership-mix analysis - detailed ownership codes separate national forest from corporate private timberland, the comparison timberland investors keep asking for.
• Regional fiber-supply sizing - state inventory panels quantify what roundwood and sawnwood supply a region can support ahead of mill and logistics decisions.
• Citation-grade sourcing - stories, theses and filings anchored to the national forest census, defensible because the methodology is published alongside the numbers.

Data scientists and ML engineers (relevance 3/3) get tree-level features - species, diameter, height, volume - joinable across every state under one key structure; see data scientists use cases. Developers and data-product builders (3/3) get typed tabular layouts that drop straight into pipelines; see developers builders use cases. Journalists, academics and students (3/3) get the authoritative national inventory with published methodology behind every figure; see journalists academics use cases. Investors and quant researchers (2/3) read standing-volume and removals trends as supply signals behind wood-product equities; see investors quants use cases. Market researchers and consultants (2/3) size regional fiber supply from state panels; see market researchers use cases. Competitive intelligence and product teams (1/3) should calibrate expectations: this data describes forests, not the companies operating on them.

Within timber reits data, this record owns the physical layer: the trees, acres and ownership lines underneath every equity story in the sector. The neighbors own the financial layers. FAO Forest Product Statistics counts what the world's forest sector produces and trades each year across roughly 285 areas since 1961 - global flows where this record gives American stems; the head-to-head lives at FAO Forest Product Statistics vs USDA Forest Service FIA DataMart. SEC EDGAR XBRL Company Facts API - Weyerhaeuser (CIK 106535) carries disclosed timberland balances and depletion per filer - company-reported acreage worth crossing against what the inventory says is actually standing nearby. Quote-history records for Weyerhaeuser add market prices, and the same agency's wider record - FIA DataMart and Timber Products Output - adds mill-level production tables to the same inventory base. If the question is what physically exists on US forestland, this is the set that answers it.

Three things worth knowing upfront.

First, exact plot locations do not exist in this corpus. Public coordinates are fuzzed or swapped to protect landowner privacy, with full precision publishing only above plot number 40,000. Anyone promising stem-level geolocation from this data is selling something else.

Second, the schema speaks in codes. Species, ownership, forest type, productivity and status all arrive as integers that resolve against reference tables. The dictionary above documents the load-bearing ones; the complete code lists ride along with every delivery.

Third, variable definitions live in the program's documentation rather than inside the files themselves, so columns beyond the twenty-four verified above are confirmed present in the corpus but documented table by table at sampling. Say which states, tables and inventory years matter when you request a sample - the extract comes back cut to exactly that shape.