Our final task (Part B) was to perform a Corridor analysis to model the potential movement of black bears between two existing areas of the Coronado National Forest in Arizona. We were provided three criteria layers suitable for black bear habitat: roads, elevation and landcover. I used the Euclidean Distance tool to create a straight line raster surface for the roads layer and created 3 distance classes (0-100 m, 100-500m, > 500m). I manually classified the elevation layer to 5 classes (0-800m, 800-1200m, 1200-2000m, 2000-2500,m > 2500m). In the symbology tab I displayed the unique values of the landcover data. I reclassified each of the criteria layers to assign their provided, respective suitability values ranging from 1 to 10 with 10 being the most suitable.
To combine the three criteria I used the Weighted Overlay tool assigning the following relative weights: landcover 60%, elevation 20% and distance to roads 20%. I used the Raster Calculator tool to "invert" the suitability layer to create a cost surface layer. Higher habitat suitability can imply lower/preferred travel cost. The raster calculator formula for the cost surface layer was "10 - Weighted Overlay Output Raster".
Next, I created two Cost Distance surfaces. Using the Cost Distance tool I created two separate Cost Distance layers. One Cost Distance layer used the Coronado 1 polygon and the cost surface layer and the other Cost Distance layer used the Coronado 2 polygon and the cost surface layer. I then used the Corridor tool and created a Corridor Cost surface using the two Cost Distance layers as input. When attempting to narrow my Corridor Cost surface layer for my map I realized there might be a problem with my Cost Distance layers. I read the Help page on the Cost Distance tool and read about problems with Polygon layers being used as input source data. So I went back and converted the Polygon Coronado layers to Raster layers and then reused the Cost Distance tool to create two new Cost Distance layers and then reused the Corridor tool.
When narrowing down the classes to display the corridor and determine what factors to use it was really a matter of trial and error. Factors that were too large created a corridor that was too wide encompassing almost all of the Coronado 2 layer. Factors that were too small created a corridor that resembled disjointed polygons.
The most optimal corridor is displayed in dark maroon and the less optimal paths are in lighter shades of the maroon spectrum. This was a tough lab. The model pictures on pages 214-215 and 258 helped keep me on track with Part B of the lab. These analysis tools are very picky. You have to be really careful in the tool windows with the various field options. Even when naming the output raster sometimes I would remove the default .tif extension if renaming the output to something more meaningful and this created a different type of raster (grid vs. tiff). I also was confused in the Weighted Overlay tool window with the Scale Values. When not using a default scale and using the From/To/By fields the Scale Values may change. I had to make sure the scale values were the same as the Field Values since we had already reclassified using the Suitability Values. I was also confused by the 0 suitability value provided for some Landcover values.
I also didn't understand how the links provided in the lab assignment on page 11 helped because they mainly discussed using an extension CorridorDesigner that we don't have access to. I'm sure that tool would be helpful but the articles didn't explain how to help choose factors for narrowing ones corridor.
Overall the confusion helped me because I had to redo many of the steps in the lab so it reinforced the steps necessary to complete the assignment. Not sure my answer is correct but at least I understand the process.
 |
| Least-Cost Corridor Analysis of Potential Black Bear Movement in Coronado National Forest |
No comments:
Post a Comment