Comparison with Land Cover Status and Trends Data

Science Center Objects

The topographic changes detected in this study usually have a corresponding land cover change. To examine this relationship more closely, features in the topographic change inventory were compared with data from an ongoing study of the status and trends of land cover in the United States. The status and trends project uses a sampling-based approach in which land cover maps and change products are produced for five dates for 10x10-kilometer sample blocks to derive regional land cover statistics. A total of 2,693 sample blocks are distributed throughout the Level III ecoregions. These sample blocks provide the spatial framework for the project. The Central Appalachians ecoregion, which contains the most topographic change polygons of any ecoregion, was selected for detailed comparison with the land cover trends data.

The figure below shows the 13 sample blocks that contain 61 individual topographic change polygons, which sum to an area just less than 6 square kilometers. These change polygons were intersected with the land cover maps and change maps for each of the sample blocks.


Sample blocks from the land cover trends study that contain change polygons

Sample blocks from the land cover trends study that contain change polygons. These blocks were used to compare land cover change with topographic change in the Central Appalachians ecoregion.


The next two figures show an example of the multidate land cover maps for one of the sample blocks.


Example of the multidate land cover maps for one of the sample blocks

Example of the multidate land cover maps for one of the sample blocks.


The most recent land cover map for each sample block is from either 1999 or 2000. As recorded on those land cover maps, 86 percent of the area covered by topographic change polygons is mapped as the mining land cover class, 13 percent of the area as the grassland/shrubland class, and 1 percent as the forest land cover class. In looking at land cover change from 1990 to 2000 over the topographic change polygons within the sample blocks, the following is noted: the land cover changed from forest to mining for 58 percent of the area, 12 percent of the area changed from mining to grassland, and 26 percent remained as mining.


An example of the forest to mining conversion is seen in the Landsat image (bottom image) in the next figure.


Example of the forest to mining conversion is seen in the Landsat image (bottom image)


The mining to grassland conversion likely represents postmining revegetation associated with reclamation, an example of which can be seen in the foreground of the perspective view in this figure.


Perspective views (looking toward the north) of the sample block data

Perspective views (looking toward the north) of the sample block data. The 1974 land cover map has been draped over the NED (top). The 1999 land cover map has been draped over the SRTM data. Note the overall flattening of the landscape in the large mountaintop mining area on the right. The arrow marks a typical valley fill as found in this region.


Another metric reported in the land cover trends data is the number of land cover changes that have occurred over a given location between the years 1973 and 2000. For the area covered by topographic change features, 74 percent of the total experienced one land cover change (most likely forest to mining or mining to grassland), 16 percent had two land cover changes (most likely forest to mining to grassland), and 4 percent had three changes. While the coincident area of topographic change polygons and land cover trends maps is limited, the observed land cover characteristics within the polygons support the observation that in this ecoregion specific land cover changes appear to be highly correlated with topographic changes. Specifically, mining is the predominant land cover class associated with the change polygons according to the latest land cover maps, and the majority of the area experiencing topographic change had a shift in land cover from forest to mining between 1990 and 2000. These observations are supported by the land cover trends project report for this ecoregion, which concludes that most of the land cover changes in this region are related to coal mining.


It could be useful to extend the comparison of topographic change features with the land cover trends data. However, outside of the Central Appalachian ecoregion there is little overlap among the change polygons and the sample blocks. A total of 138 sample blocks contain topographic change polygons, which is about 5 percent of all the sample blocks. These 138 blocks contain 275 topographic change polygons, which represents slightly more than 5 percent of the total number of polygons in the inventory. It should not be surprising that few change polygons intersect the sample blocks. Even though there are more than 5,000 polygons distributed throughout the study area, topographic changes are scarce on the landscape as their total area is less than 0.02 percent of the area of the conterminous United States. Also, the trends sampling scheme was designed to capture land cover variability across broad areas, so there is no direct reason why many topographic change areas would be included. For these reasons, the lack of coincidence between the mapped topographic change features from this study and the land cover trends data probably precludes derivation of reliable national statistics on the correspondence of topographic and land cover change. However, the analysis summarized above for the Central Appalachians ecoregion shows that there is useful information in comparing topographic change and land cover change. Perhaps the best use of topographic change information is to aid in interpreting land cover changes in local settings. In this case, the SRTM NED difference grids could be thresholded with care to ensure that all topographic change features of interest are included before the subsequent comparison with land cover maps.