Projecting Future Landscape Change at High Spatial and Thematic Resolution

Submitted by atripp on

One of the most valuable uses of remote sensing imagery is for mapping and monitoring landscape change, and Landsat-based land cover data such as the National Land Cover Database (NLCD) and the Cropland Data Layer (CDL) from the U.S. Department of Agriculture (USDA) have been used for countless applications.  For the first time, modeled landscape projections have been produced that are more realistic, honoring patterns and trends expected based on historical and existing NLCD and CDL maps. To accomplish this, USGS scientists used Landsat-based land cover data, topographic data, soils data, and many other geospatial datasets to parameterize the USGS Forecasting Scenarios of land use (FORE-SCE) model. Future scenarios were produced for the Great Plains with an unprecedented combination of high spatial resolution (30-meter grid cells), high thematic resolution (29 land cover classes), and broad spatial coverage.  A parcel-based modeling framework using land ownership and management units for agricultural land (USDA’s Common Land Unit data) ensured a realistic representation of evolving landscape pattern. A wide-ranging suite of scenarios was used to represent uncertainties in future conditions, with 3 different climate realizations for 11 different anthropogenic land use scenarios, resulting in 33 unique scenario combinations.

Projections are complete for the south-central Great Plains and the Prairie Potholes region, with work in the upper Missouri River basin expected to be complete by the end of 2018.  The resulting projections in the Great Plains cover over 1.8 million square kilometers and represent over 4 million individual agricultural parcels. Because the projections match the spatial and thematic characteristics of NLCD and CDL, they can be seamlessly integrated into existing workflows that rely on those data, facilitating analyses of future landscape impacts on a variety of ecological and societal processes. The projections are already being used by Audubon and the Playa Lakes Joint Venture to assess landscape interactions with bird species in the Great Plains, while the University of North Dakota and USGS Northern Prairie Wildlife Research Center are examining potential future impacts on pollinators and other ecosystem services.

The left panel depicts the modeled area, with the cross-hatch area completion date in late 2018. The inset depicts 2014 and modeled 2100 land cover near Amarillo, Texas. In this scenario, agricultural expansion leads to an increase in wheat and a loss of grassland. With warmer and drier conditions anticipated in the future, shrubland encroachment into grasslands also occurs. The developed area of Amarillo itself expands substantially.


Author Name
Terry Sohl
Author Email