RCMAP: Yearly Maps of Western U.S. Rangeland Vegetation Since 1985

Rangelands — arid and semiarid land covering at least 207 million hectares of the Western United States — provide ecological, hydrological, agricultural and recreational goods and services. They are under growing pressure from climate change, management practices, altered fire regimes, energy development and recreation, and they are far too vast for field observations alone to track how they change.

What RCMAP is

The Rangeland Condition Monitoring Assessment and Projection (RCMAP) project, a partnership of the U.S. Geological Survey and the Bureau of Land Management (BLM), produces annual maps of rangeland vegetation across the West from 1985 to the present. They help land managers and scientists:

• monitor changes in vegetation composition;
• evaluate past management and target future improvements;
• locate critical wildlife habitat;
• assess landscape health and fragmentation.

Climate effects are often gradual and don't show up as a change of land cover (shrubland turning to grassland, say). RCMAP's fractional cover data — the 0–100% cover of each component in each place — capture those gradual changes.

Figure 1. Field observations, high-resolution imagery, Landsat imagery and other data train artificial intelligence and machine learning (AI/ML) models that predict annual vegetation cover since 1985. USGS.

How it works

• Training data: high-resolution image sites observed over time, BLM field data, and field data curated by the LANDFIRE program.
• Predictors: mainly Landsat imagery — one composite per year at peak growth and one of senesced (brown) conditions — plus topography and spectral indices.
• The AI/ML models learn how component cover in the field relates to those predictors, then run on each year's Landsat data; post-processing limits noise and follows how components recover after fire.
• Trends come from a linear model and a structural-change method that finds break points with a moving time window.
• Validation uses independent field data from long-term monitoring sites and checks of model fit.

Figure 2. RCMAP time series give yearly cover and trends since 1985 for each 30-meter pixel, and summarized by pastures, allotments, watersheds and other management units, in the rangelands viewer. USGS.

What the maps show

From the data available in 2022:

• Trends varied strongly by ecoregion — driven by fire history in some, by climate in others.
• Maximum temperatures and growing-season precipitation were linked to every component's trend; bare ground was the most sensitive to climate.
• Case studies have shown the data at work on the ground, with more planned with the BLM.

Rangeland vegetation tends to change gradually — with management, year-to-year climate, vegetation makeup and annual grass invasions — and RCMAP's record of where and when it changed helps explain those patterns.

Getting the data

RCMAP time series from 1985 on are available:

• to download;
• as map services (Web Map and Web Coverage Services);
• in the rangelands viewer;
• in Google Earth Engine (search "RCMAP").

Annual updates will extend the series and improve it with more training data and better methods. The series has also been used to map ecological potential — the cover expected without disturbance, so that departures from it show disturbance — and to project cover under several climate change scenarios.

Figure 3. RCMAP datasets currently available. USGS.

Sources

Based on "Rangeland Condition Monitoring Assessment and Projection (RCMAP)," U.S. Geological Survey fact sheet, citing Rigge and others (2019, 2021); a work of the United States government in the public domain.