The agreement between Landsat estimates and measurements of impervious made from high resolution aerial images was analyzed using a separate independent sample.  Accuracy varies across the state by image and is reported as R-Square and RMSE values.  Accuracies for the state are reported in the supplemental map file titled “impervious_statewide_1990_map_accuracy.pdf.”

Data provides complete coverage of Minnesota, USA.

RMS Error <7.5 meters or ¼ pixel

RMS Error <7.5 meters or ¼ pixel

The process is generally described as follows. A combination of spring and summer dates of Landsat data from 2000 was used to classify the land cover types into urban/developed vs. rural (agricultural, forest, and wetland) and water classes. Measurements of impervious surface area were made on high resolution aerial imagery for a random sample of areas for calibration of the Landsat classification model. A least squares regression model relating percent impervious to the spectral-radiometric response of the Landsat TM data for the summer imagery.  The particular Landsat parameter used to estimate percent impervious was "tasseled cap" greenness which is sensitive to the amount of green vegetation and therefore is inversely related to the amount of impervious surface area. The regression model was used to estimate the amount of impervious surface within the area of each pixel. After applying the regression model to the Landsat TM image the land cover classification was used to assign all non-urban areas a zero percent impervious value.  Accuracy assessment of the Landsat derived impervious surface estimates was performed with a second independent sample of measurements of impervious from the aerial imagery.  Accuracy assessment was based on the linear agreement of actual versus predicted impervious surface.

Image processing:

 

Impervious surface classification was performed with 19 Landsat TM satellite images shown in supplemental map file titled “impervious_statewide_1990_map_image_mosaic.pdf”. The imagery was rectified to the UTM coordinate and projection system to an RMS error 1/4 pixel (7.5-meters) using approximately 25 ground control points per image. The base layer used was the MNDOT Major Roads map. Nearest neighbor resampling to a 30-meter pixel size and the coordinates of the final image were adjusted to values evenly divisible by 30. Following rectification the imagery was transformed to tasseled cap values using ERDAS Imagine default processes. The tasseled cap data were stretched to unsigned 8-bit.

 

Impervious Model Calibration Data:

 

Model calibration sites were randomly selected and high resolution aerial imagery was used to determine the amount of impervious surface within each calibration site. Approximately 50 sites, distributed across the image area, were collected for each image. Imperious surface area was determined from the 1991 DOQQ's of 1-meter resolution for th 1990 time era and from 1-meter resolution 2003 FSA-NAIP imagery for the 2000 era. Percent impervious surface area was calculated for each site by hand digitizing impervious surface within the site.

Modeling Impervious Surface Area:

The measurements of impervious surface area from the calibration sites were used to develop a least squares regression model relating percent impervious to the spectral-radiometric response of the Landsat TM data for each of the 19 images. The Landsat parameter used to estimate percent impervious was tasseled cap greenness which is sensitive to the amount of green vegetation and therefore is inversely related to the amount of impervious. The regression model used was different for each Landsat image and is shown in the supplemental map file titled “impervious_statewide_1990_map_model_fits.pdf”.

Accuracy Assessment/ Inverse Calibration:

 To measure the accuracy of the Landsat-derived impervious surface estimates, an independent sample of  approximately 25 accuracy assessment sites were selected for each of the 19 Landsat images. The sites were randomly selected and compared with high resolution aerial imagery to determine the amount of impervious surface within each accuracy assessment sites. Accuracy was determined with a plot of actual vs. predicted plots and results are shown in the supplemental map file titled “impervious_statewide_1990_map_actvpred_stats.pdf”.

To remove estimation bias an inverse calibration was computed from the linear fit of the actual vs. predicted plot and applied to the impervious surface classification. Accuracy of Landsat derived impervious surface estimates was reassessed following the inverse calibration. The results are shown in supplemental map file titled “impervious_statewide_1990_map_accuracy.pdf”’

Although in most cases the inverse calibration did not significantly affect the R-Square or the RMSE, the bias is removed.  This is apparent from the reduction in the intercept and an increase in the slope.

Removing Non-Urban Areas:

This step identified non-urban areas within the Landsat image area and reclassified these areas in the impervious surface classification with a percent impervious surface value of zero. To identify non-urban areas a combination of spring and summer dates of Landsat imagery, taken in 2000, were used to classify the land cover types into urban/developed vs. rural (agricultural, forest, and wetland) and water classes. Classification accuracy for the 2000 Minnesota statewide cover type map had producers and users accuracies of 91.67% and 95.4%, respectively, for the urban class and overall classification accuracy 84.5% with a Kappa statistic of 0.808.  The final impervious surface classification was then overlaid with the land cover classification and areas identified as non-urban were reclassified with a percent impervious surface value of zero.  For the 1990 impervious classification a few further processes were performed that utilized the MNGAP 1990 covertype classification created by the Minnesota Department of Natural Resources.  These processes were used to remove those areas that were identified as urban in the 2000 classification, but were classified as “cropland” or “grassland” in the 1990 MNGAP cover type.  The reason for using the two separate classifications is as follows.  First, because one goal of the project is change detection, the area of the urban between the two dates needed to stay consistent.  Thus, the 2000 cover type was utilized as the primary identifier of urban.  Our methods, then assume that areas identified as urban in 2000, but not developed in 1990 will have a high greenness value (due to vegetative cover) in the 1990 imagery and thus will be modeled as having low to no impervious surface in the 1990 era.  Areas of bare soil in agricultural fields in 1990 that are urban in 2000 will have low greenness values in each date causing errors in the modeling of impervious surface.  Utilizing the 1990 MNGAP to remove the “cropland” and “grassland” from the areas considered as urban for 1990 reduced the occurrence of this error.

Clouds

The majority of the Landsat imagery used for mapping impervious surface was cloud and haze free, however, there were some instances where having clouds in the imagery was unavoidable.  Cloud/cloud shadow portions of the Landsat imagery were considered to have no available data.  Cloud/cloud shadows were identified and manually digitized from the landsat imagery to create a “cloud” mask.  The cloud mask was overlaid on the impervious surface classification and all pixels within the cloud mask were given a value of zero.  It should be noted that there was very little overall area where both clouds and urban overlapped.

Mines

Mines, considered urban in the 2000 Minnesota landcover classifications, needed to be identified for further processing in the impervious surface classification.  Bare soil is classified in the impervious models as having a high degree of impervious surface due to its low greenness value.  Much of the area of mines is made up of a dense mix of compacted soil and gravel roads surrounded by bare soil making separating the impervious surface from the pervious bare soil difficult.  Unfortunately, at the time that this data set was produced a data identifying the location and extent of all mines, small and large, in the Minnesota did not exist.  There was, however, a data set produced by the Minnesota Department of Natural Resources – Division of Lands and Minerals that identified the locations and extent of active mining area in the Mesabi Iron Range.   This data set was used to force the pixel values that fell within the iron mines data to an impervious value of zero.

Limiting to MCD Boundaries:

The final impervious surface classification was then clipped to match the boundaries of the MCD of interest. Geospatial data used to identify MCD boundaries were acquired from the Minnesota Legislative GIS office, <ftp://ftp.commissions.leg.state.mn.us/pub/gis/shape/mcd2003.zip>.

The pixel values represent the area (in percent) of completely impervious surface within the 30-m by 30-m pixel area. Values range from 0 to 100 percent.

This data may be used for educational and non-commercial purposes, provided proper attribution is given. Secondary distribution of the data is permitted, but not supported by the University of Minnesota. By accepting the data, the user agrees not to transmit this data or provide access to it or any part of it to another party unless the user includes with the data a copy of this disclaimer.

see website or contact info