DEM DIFFERENCING

DEM differencing is the process of subtracting one digital elevation model from another on a cell-by-cell level. Applying this technique to detect change over time, WSI can visualize and quantify erosion across the landscape.

 

chiques%20lidar%20difference%20graphic_e

2014 USGS post-Sandy LiDAR (above) and 2008 DCNR PAMAP LiDAR (below) of Chiques Creek Watershed, Lancaster County, PA.

stobers2008.jpg

1

Figure 1. This digital elevation model shows a legacy sediment terrace surface (purple area) prior to the breach of Stobers' dam in Denver, PA. 2008 USGS PAMAP LiDAR

stobers2014.jpg

2

Figure 2. This digital elevation model of the same area, created from LiDAR data collected after the dam breach, shows the stream incision that resulted from the drop in the water table. 2014 USGS post-Sandy LiDAR.

stobersdifference.jpg

3

Figure 3. This figure shows the result of the post-dam breach incision seen in Figure 2. Here, the red area represents erosion measured from "differencing" or "subtracting" the two digital elevation models (Figs. 1 & 2). 2014 USGS post-Sandy LiDAR.

With a primary focus on stream bank erosion, we have developed technology and methods to quantify stream bank volume loss at varying scales, from single landowner parcels to entire watersheds.

 

Because the difference between two elevation models shows change over a period of time, we can combine the results of DEM differencing with field measurements to determine rates of erosion.

big beaver differencing.jpg

LiDAR DEM differencing of Big Beaver Creek, PA from 2008 - 2014. The red-shaded areas represent stream bank volume loss in cubic meters due to erosion. 2014 USGS post-Sandy LiDAR.

Since the accuracy of DEM differencing depends on the resolution of the digital elevation models, we create our own centimeter-scale elevation models using aerial drone surveys with structure-from-motion technology. Drone surveys offer a highly-accurate, easily repeatable alternative to repeat aerial LiDAR.  

bsr sparse.png
bsr dense.png

A dense point cloud (left) and the digital elevation model derived from it (right) of Big Spring Run in Lancaster, PA, created by drone photogrammetry and structure-from-motion software.

2014

2008