Caution: Read the metadata carefully before using data sets
What is the Georgia Land Use Trends Project?
The GLUT project, a series of land cover datasets (1974, 1985, 1991, 2001, and 2005) derived from satellite imagery, provides information for thirteen land cover categories. Datasets were developed to allow for a uniform analysis of land cover change across Georgia.
How are these data sets different from the 1998 land cover?
The 1998 land cover was developed as part of the Georgia GAP Project. Both GAP and GLUT were developed by the same group of individuals at the University of Georgia’s Natural Resources and Spatial Analysis Lab (NARSAL). An eighteen class land cover was used as the basis for the forty-four class vegetation map developed for GAP. The methodology used to derive the 1998 land cover varies from that used for the GLUT project and if used in a change analysis might provide confusing results. Therefore, users should not include our 1998 land cover product in analyzing change.
Where can I get copies of these datasets?
What format is used for Clearinghouse datasets?
Datasets are raster using the native Leica Geosystems Imagine format (.img). This format is compatible with most common GIS and image processing software packages. Because they are raster and made up of millions of grid cells (pixels), users need access to Spatial Analyst to further analyze data in ArcView.
Is data available as shapefiles?
No. Since raster format maintains the integrity of the individual pixel, converting to a vector format may result in a file with hundreds of thousands of single cell polygons (30m X 30m). We do not recommend converting the data unless you are familiar with additional processing methods.
Why can’t I see my house when I view the data?
Land cover maps were derived from Landsat satellite images that have a ground resolution of 60 meters (approximately 180 feet) for the 1974 and 1985 data and 30 meters (approximately 100 feet) for the 1991, 1998, 2001 and 2005 datasets. At this resolution, the minimum size that can be ‘seen’ is about 1/4 – 1/2 acre. Sensors on the satellite ‘see’ a lot of different types of imagery while collecting data for developed areas. This creates mixels or mixed pixels which don’t show individual objects such as the trees, grass and house that make up urban landscapes.
How was the spatial resolution of each dataset determined?
Data was derived from Landsat satellite images. The 1974 and 1985 land covers uses older Landsat MSS imagery with a resolution of 60 meters. All other datasets were derived from either Landsat TM or ETM+ with a resolution of 30 meters. We recommend that users not analyze any area less than one acre in size.
Why doesn’t the land cover data go back further in time?
Datasets were derived from Landsat satellite images. The first Landsat satellite was launched in 1972 and provided the first large area coverage available across the globe. To get the full extent of Georgia, the largest state east of the Mississippi, we ‘stitched’ together fourteen different Landsat images.
How were these data sets created?
Landsat sensors collect earth surface reflectance data using sensors that are calibrated for different bands of the electromagnetic spectrum. These bands include visible light (red, green, blue) and bands within the infrared and thermal parts of the spectrum. All objects on earth reflect light differently and produce what we refer to as ‘signatures’. Image processing software helps us to use these signatures to map and model what is on the earth’s surface (see the metadata that accompanies each data set for further information). We often combine images taken at different seasons over the same area to help us interpret and map the landscape.
What is the actual date for each land cover map?
We use a composite of images for an area (winter, spring, summer and/or fall), at times creating new maps for multiple years. It is occasionally difficult to collect cloud free images over parts of Georgia, resulting in a year or more wait for the right image. Therefore, these dates represent a composite that may have images one or two years older or younger than the actual year of the map. Because of the variation in dates for anyone dataset, this data should not be used for real estate or legal applications.
What is impervious surface?
Man-made materials found in developed landscapes including rooftops, driveways, sidewalks, decks and other materials that prevent water from infiltrating into the ground.
Are there any natural impervious surfaces?
Yes. Rock outcrops are natural areas of impervious surface.
What is a data set?
Impervious surface maps represent the percent of each grid cell (pixel) that is made up of impervious surface. Using a modeling technique, we can determine the density of impervious surface for each pixel. We classified data to twenty classes in 5% increments. So class one represents ranges from 1%-5% impervious, class two ranges from 6%-10% impervious, etc. We recommend you use the midpoint of the data range to represent the percent impervious, for example class one would be 3%, class two would be 8%, etc.
How do I calculate area?
Each pixel has an area of 900 square meters. Multiply the area by the percentage of impervious. For example, if a pixel is from class one you would multiply 900 by .03 which equals 27 square meters of impervious surface. Once you have the total area in square meters you can use the following conversions to calculate area. 1 square meter = 0.0001 hectares OR 1 square meter = 0.000247 acres.
Why is the data in 5% increments rather than 1% increments?
Since these are modeled results, the error in each layer is greater than 1%; therefore, we reclassified the data to improve the accuracy of the results.
Can I overlay multiple dates to calculate change for an area?
You can calculate change for individual pixels; however, we recommend calculating change for a particular area instead. Even though the data layers are aligned to keep pixel overlap, there is some error. Causes of this error include shifting of the satellite along its path from year to year, shifts in the datasets due to reprojection and resampling the resolution of the datasets.
What are we measuring?
This data set represents the percent of tree canopy found in an individual pixel. Think of this as a measure of green space.
What is this dataset?
These maps represent the percent of each grid cell (pixel) that is made up of tree canopy. Using a modeling technique, we can determine the density of tree canopy for each individual pixel. As with the impervious surface density data, we reclassified the data to twenty classes of 5% increments.
How do I calculate area?
Each pixel has an area of 900 square meters. Multiply the area by the percentage of tree canopy. For example, if a pixel is from class one you would multiply 900 by .03 which equals 27 square meters of impervious surface. Once you have the total area in square meters you can use the following conversions to calculate area. 1 square meter = 0.0001 hectares; OR 1 square meter = 0.000247 acres.