As lovers of the Appalachian mountains in which we live, it is very disturbing to hear about the mountaintop removal operations to extract coal from beneath their peaks.  But unless you happen to live in the directly affected communities, it may be hard to see the impact.  Even in the affected communities, the full physical impact may be difficult to see from the ground, and the long term impacts, from poisoned drinking water and flooding, to the loss of jobs may be a little harder to see directly.  Satellite imagery, however, is one way to get a birds-eye view of the direct physical impact on the landscape.

Google maps has a neat new feature which lets you create a timelapse of the last twenty years for virtually any location on Earth.  Scrolling through Appalachia provides plenty of jarring examples of the impact of mountaintop removal over the last few decades.  Here are some examples in southwest Virginia (unfortunately these maps don’t work on mobile devices):

There are plenty more, especially as you move up into Kentucky and West Virginia.

As another example, I did a project for a class a few years back where I analyzed National Land Cover Database imagery of Virginia from 1992-2006 to determine where land cover change had occurred.  These mountaintop removal projects stuck out very prominently in the output.  Below are a few examples.  The areas shaded in gray and blue have changed their landcover classification.  You can see the classification from year to year in the popup legend ².

I’m working on a dynamic map to make it easy to explore the land cover change data on your own.

1. In the above images, note the 'MODIS' land cover classification.  This is a yearly land cover dataset which has much lower resolution (500m² per pixel vs 30m²) so is less accurate, but shows some change in the second example
2. In the above images, note the ‘MODIS’ land cover classification.  This is a yearly land cover dataset which has much lower resolution (500m² per pixel vs 30m²) so is less accurate, but shows some change in the second example