Gold University of Minnesota M. Skip to main content.University of Minnesota. Home page.
Home > Research > Landscape Monitoring and Analysis > Land Cover and Change Classification

Land Cover and Change Classification

Satellite-based Approach to Land Cover and Change Classification

The following steps have been used to create satellite-based land cover and land cover change maps for the Twin Cities Metropolitan Area (TCMA), Minnesota. The TCMA is a 7,700 km2 area and includes a diversity of land cover classes interspersed with over 900 lakes and transected by the Mississippi, Minnesota and St. Croix rivers.

STEP 1: Six Landsat TM/ETM+ images containing the TCMA were obtained to study how land cover in the metro area changed between 1986, 1991 and 1998 (the entire TCMA is contained in one Landsat image). Two different dates, spring and summer, of imagery were acquired for each year. Part of the summer images from each year are shown below.

    

Landsat imagery zoomed into Woodbury, Minnesota in the TCMA.

STEP 2: The satellite data were geometrically corrected to match the UTM map projection.

STEP 3: Land use data from field observations and aerial photography were collected for a random sample of areas. Additional maps from the Metropolitan Council, the Minnesota Department of Natural Resources and the National Wetlands Inventory were also acquired.

STEP 4: With the land use data as reference material, “training” statistics, which describe the spectral-radiometric temporal responses of a subset of known areas, were generated and used to classify each pixel of the entire area into one of the five land cover classes.

Below is one of the final satellite-derived land cover maps of the TCMA.

 

 

A single Landsat classification map.

 

 

 

 

 

 

 

The overall classification accuracies were 95.2, 94.6 and 95.9% for the 1986, 1991, and 1998 maps, respectively.

STEP 5: Following image classification, analysts mapped and quantified the land cover changed between 1986 and 1998. A map of the major land cover types and the changes from rural to urban or developed uses is shown below. The majority of the changes were at the periphery of the major cities of Minneapolis and St. Paul, and the first ring of suburbs.

TCMA urban growth from 1986 to 1998 combined with 1998 MUSA boundary. The majority of the changes occur within the second and third ring of suburbs surrounding the cities of Minneapolis and St. Paul. Clear patterns emerge which highlight the urbanization activity that has occurred east of St. Paul along the I-94 corridor (completed in the mid 1980’s), which connects the metro area to western Wisconsin. Growth also was concentrated in a strip along the southwestern perimeter following the Minnesota River, and in intermittent patches throughout the northwestern perimeter. Although this figure only shows the changes from agriculture, forest, or wetland to urban, other, more specific changes can also be mapped.

 

 

 

The land cover change results were also quantified. This tabular format shows the totals for each land cover type and the trends between the years.

 

Summary of classification area statistics for 1986, 1991 and 1998. Agriculture, urban and forest are the three major land covers and the changes in their proportions represent the most significant changes. From 1986 to 1998, urban areas increased a total of 52,019 ha or 28.4%. Agriculture decreased 49,091 ha or 13.4% from 1986’s 365,046 ha. Forest decreased 5,089 ha, or 4.5% from 1986’s 112,145 ha.

The effects and relationships of urban growth determined by the satellite-derived change maps have also been examined, including the relationship to population growth. The growth rate is similar to the increase in population, (indicating relatively less “sprawl” in the TCMA than in some other urban areas. There is a strong relationship between new development and proximity to highways with almost 48% of the development occurring within 2 km of highways.

Land Cover Classification and Change Classification

Impervious Surface Mapping

Minnesota Statewide Land Cover Classification

Temporal Analysis of Vegetation Cover

 

In this section...

Satellite-based Approach to Land Cover and Change Classification

The Value of Landscape Change Mapping