United States · 1850–2025 · county and 30 m pixel data
Clearing the land
How the United States cleared its forests for farms, and what grew back. Built from county returns of the Census of Agriculture and from four decades of Landsat-based land-cover records.
A century of clearing, counted farm by farm
From 1850 the federal census asked every farm how much of its land was improved: cleared and used for crops, meadow or pasture. Summed across counties, improved land grew from acres in 1850 to in 1920.
The census counted farms, not forests. Its woodland figures cover only woodland inside farms. They cannot show the forest left outside farms. The 1925 census also changed the categories, replacing improved land with cropland and pasture, so the series shows a seam between 1920 and 1925.
Farm land in the United States, 1850–1950
Sum of county returns. Dotted segments cross censuses that did not report that item at county level.
Show as a table
Source: IPUMS NHGIS, University of Minnesota, county agriculture tables 1850–1950. Totals are sums of county rows joined to historical boundaries, and differ slightly from published national totals.
Improved farmland, % of county land
Source: IPUMS NHGIS (Version 21.0), county agriculture tables and historical county boundaries (2008 TIGER/Line basis). Equal-area projection (CONUS Albers). Land area: census-reported where available, otherwise boundary area minus open water (NLCD 1985). Alaska and Hawaii are excluded.
Forty years of loss and regrowth from orbit
Landsat satellites have imaged the country continuously since 1972. Three products derived from it will be mapped here: Annual NLCD land cover (USGS, Collection 1.2), the Forest Service's Landscape Change Monitoring System (LCMS v2025.11), and Hansen et al. Global Forest Change (v1.12).
How the census maps were made
Measures by census year
| Years | Improved land | Farm woodland |
|---|---|---|
| 1850, 1860 | Census "improved land" | Not reported |
| 1870, 1880 | Census "improved land" | Reported |
| 1890, 1900 | Census "improved land" | Not reported |
| 1910, 1920 | Census "improved land" | Reported |
| 1925–1935 | Cropland + plowable pasture | Woodland pastured + not pastured |
| 1940 | Not published at county level | Not published at county level |
| 1950 | Total cropland (includes cropland used only for pasture) | Reported |
Land area
Each county-year uses, in order of preference: (1) the land area the census reported that year, if it is within a factor of two of the boundary's area; (2) the census land area from the nearest year in which the county's boundary is the same size within 0.5%; (3) the boundary area minus open water from NLCD 1985 (class 11), averaged to 240 m. The second rule matters in coastal Louisiana, where marsh counted as land in 1920 is open water in 1985.
Corrections and exclusions
The NHGIS boundary files for 1910, 1920 and 1930 place most of Concordia Parish, Louisiana inside West Feliciana Parish. Both parishes were unchanged from 1900 to 1940, so their 1940 boundaries are used for those years. Reporting units with no boundary (some Indian reservations in 1900 and 1910, multi-county units, "all other counties" rows) are left off the map. In 1900 these held 5.5 million acres in farms, 0.7% of the national total. The 1925 returns are mapped on 1930 boundaries, because several Florida counties created in 1921–25 are missing from the 1920 file.
Reproducibility
Pipeline code: pipeline/census/ (extract request, harmonization, tiling). Join rates and every flagged county are listed in qc_report.txt.
Citation
Jonathan Schroeder, David Van Riper, Steven Manson, Grace Cooper, Zachary Krause, Tracy Kugler, Tsu Zhu, and Steven Ruggles. IPUMS National Historical Geographic Information System: Version 21.0 [dataset]. Minneapolis, MN: IPUMS. 2026. doi.org/10.18128/D050.V21.0