Projections of Future U.S. Land Cover in 2050 (NSF)

1990s trend: Continuation of land-use change trends from 1992 to 1997.
High crop demand: Land-use changes accounting for 10% increase in crop prices every five years relative to the 1990s trend scenario.
Forest incentives: $100/acre payment per year for land converted to forest; $100/acre tax per year for land taken out of forest.
Natural habitats: $100/acre tax per year on land converted from forest or range to crop land, pasture, or urban.
Urban containment: Prohibition on land conversion to urban in non-metropolitan counties.

create_probability_maps.py: Script to generate output probability maps given a series of spatial and econometric inputs.

transition_matrix_generator.py: Script to summarize probability results for any set of summary zones (county, state, etc.)

Each scenario zip file contains a folder named “probmaps” – this is where you will find the raw probability raster data, one image file for each of the following five land cover codes:

1 = crops
2 = pasture
3 = forest
4 = urban
5 = rangeland

Each of the raster files stores probability values (0 to 1) of the likelihood that any given pixel will transfer to that particular land cover class in 2050.
For example, the raster “us_reference_prob4.img” would show the probability of a pixel transitioning to urban in 2050 under the reference (1990s trend) scenario.

The pixel resolution of the probability maps are 1 hectare (100m x 100m) and cover the conterminous U.S..

NOTE: Since each pixel’s probability is calculated separately (neighboring pixels are not factored in), it would be inappropriate to interpret this data at the pixel level. For instance, one cannot use this data to definitively say a pixel will transition from forest to urban in 2050. However, since the probabilities are dependent on the U.S. county that it’s located in, we can use these rasters to generate regional summaries of land cover change at the county level scale or larger. (state, ecoregion, etc.)