Changing Midwest Assessment: Land-Cover Change Mapping
University of Michigan Environmental Spatial Analysis Laboratory
ESA Homepage


The research and results described on this webpage are part of the Changing Midwest Assessment carried out by the United States Forest Service (USFS) North Central Research Station. The purpose of the Assessment was to describe the spatial distribution, direction and intensity of the changes that have occurred on the biophysical and social landscapes of the seven-state North Central Region over the past two decades (1980 to 2000), including change in land cover, forest characteristics, plants and animals, and human demographics (

The specific goal of the University of Michigan ESALab in contributing to the project was to generate land-cover change data at a 1-km spatial resolution and to map hotspots of land-cover change between agriculture, forest, and urban cover. This posed a challenge, as no identical remote sensing or land-cover datasets existed at that resolution for both dates. Therefore we used an existing photo-digitized national land-cover classification from 1980 combined with current satellite imagery from 2000 and developed a method to assure their comparability that used a third, independent, statistical dataset (NRCS NRI) collected consistently over the two time periods. NRI statistical data plus the 1980 land-cover classification and our new 2000 AVHRR based classification were then used to determine the spatial distribution of land-cover change over the 20-year time period in the seven-state region.

The North Central Region of the United States, as defined by the USFS comprises the states of Illinois, Indiana, Iowa, Michigan, Minnesota, Missouri and Wisconsin. We focused our analysis on changes in forest, agriculture, and urban areas, and especially noted areas that have experienced urban growth and forest growth In the past century, the region has been and continues to be dominated by agriculture; it includes the majority of the productive United States corn and soybean belt that cuts diagonally through the seven-state region.  In the northwest and west, the region borders the wheat-growing region of the Great Plains.  Forests dominate the glacial landscapes in the north, which are also characterized by wetlands and lakes, and in areas of topographic relief (e.g. Ozarks and Shawnee Hills) in the south. Urban centers developed near water and rail transportation routes.  The largest urban centers in the region include Chicago, Illinois; Detroit, Michigan; Indianapolis, Indiana; St. Louis, Missouri; and Minneapolis, Minnesota.

Prior investigation of NRI statistical data indicated that agriculture was being lost while urban and forest land covers were increasing in the region between 1982-1997. Our results showed that the major trajectories and locations of rapid change within the region were 1) the conversion of agriculture and forest to high and low-density urban land-use near urban centers, and 2) the conversion of agriculture to forest in marginally productive areas, often near transition zones between predominantly agricultural and predominantly forested ecological regions.  Maps were produced to show the spatial distributions of these statistical trends.

The figures on this webpage show 1) the study region, 2) stratification units (MLRAs) used for analysis, 3) close-up representative examples of land-cover change derived from airphoto analysis, 4) methodological procedures, 5) land cover in 1980 and 2000, and 6) hotspots of land-cover change. 

The 1980 and 2000 land-cover data is available on this webpage for download in ArcGIS format. The NRI MLRA GIS may be found at:

Additional information on the regional Assessment may be found at:

Relevant publications include the journal publication abstracted below describing the methods of the land-cover change analysis and the comprehensive USFS Technical Report.

Location of the region of land-cover change analysis in the United States.

AVHRR NDVI composite from May 18 to June 1, 2000. Boundaries shown and used for ecoregional

stratification of the remote sensing classification are NRI MLRAs.

Example of Land-use change from NAPP photo interpretation for photos near Alpena, Michigan, and Fort Wayne, Indiana where a) and c) show land-use in 1980 (LUDA), and b) and d) show land-use in 2000 (“simulated LUDA”).

Overview of the land-cover and land-cover change classification and validation process.

Land cover in 1980 and 2000: a) 1980 land cover from LUDA, b) 2000 land cover from AVHRR.


The following ArcGIS datasets and associated metadata may be downloaded and used. If you plan to use the data please contact the authors of the data (see Links) and cite the two publications listed on this webpage.
Land Cover 1980
Metadata 1980
Land Cover 2000
Metadata 2000


Bergen, K.M., D.G. Brown, J. Rutherford, E. Gustafson. 2005. Change Detection with Heterogenous Data using Ecoregional Stratification,
Statistical Summaries and a Land Allocation Algorithm. Remote Sensing of Environment, 97(4): 434-446.
A ca. 1980 national-scale land-cover classification based on aerial photo interpretation was combined with 2000 AVHRR satellite imagery to derive land cover and land-cover change information for forest, urban, and agriculture categories over a seven-state region in the U.S.  To derive useful land-cover change data using a heterogeneous dataset and to validate our results, we a) stratified the classification using predefined ecoregions, b) developed statistical relationships by ecoregion between land-cover proportions derived from the 1980 national-level classification and aggregate statistical data that were available in time series for all regions in the U.S., c) classified multi-temporal AVHRR data using a process that constrained the results to the estimated proportions of land covers in ecoregions within a multi-objective land allocation (MOLA) procedure, d) interpreted land cover from a sample of aerial photographs from 2000, following the protocols used to produce the 1980 classification for use in accuracy assessment of land cover and land-cover change data, e) compared land cover and land-cover change results for the MOLA method with an unsupervised classification alone.  Overall accuracies for the 2000 MOLA and unsupervised land-cover classifications were 85% and 82% respectively. On average, the 1980-2000 land-cover change RMSEs were one order of magnitude lower using the MOLA method compared with those based on the unsupervised data.
Potts, Robert; Gustafson, Eric; Stewart, Susan I.; Thompson, Frank R.; Bergen, Kathleen; Brown, Daniel G.; Hammer, Roger; Radeloff, Volker; Bengston, David; Sauer, John; Sturtevant, Brian. 2004. The Changing Midwest Assessment: land cover, natural resources, and
people. Gen. Tech. Rep. NC-250. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Research Station. 87 p.

Documents changes in land cover, forests, selected natural resources, and human demographics and attitudes across the Midwest from roughly 1980 to 2000.

Dr. Kathleen Bergen:

Dr. Daniel Brown:

James Rutherford:

Eric Gustafson:


ESA Homepage:

Additional Changing Midwest Assesment Information:

USFS Changing Midwest Assesment Report:

Natural Resources Conservation Service(NRCS) MLRA GIS: