Friday, October 23, 2015

New dataset - latitude

Both the Climatic Water Deficit Toolbox and the Day Length Toolbox rely on a latitude raster for some of their calculations.  I've had many people ask me where they can get a latitude raster.  Now the answer is if you're working in the western US (WA, OR, CA, NV, UT, NM, AZ, WY, CO, MT, ID, ND, SD, and TX) you can download a raster off of our lab's webpage HERE .

These data are at 1200 m cell size and are in a UTM Zone 12 NAD83 projection.

If you are working outside of this area it is still possible to create your own from lines of latitude.  First, remove the lines of longitude.  Then covert the line vertices to points taking care to ensure that the points are somewhat distributed.  Finally, a second order trend analysis in Spatial Analyst should do the trick.

Friday, October 9, 2015

New tool - Generate points around points aka generate background points or available points

Habitat modeling techniques, such as conditional logistic regression, require a random sample of background points used to quantify available habitat. One common assumption of resource selection functions is that animals have an equal probability of using background points as they do occurrence points. For animals that move across the landscape this assumption is often violated if the entire movement corridor is considered as background. This tool alleviates this concern by creating a local sample of background points around known occurrence points. There are two versions of this tool. The large version is recommended for most applications as it tiles the data resulting in faster processing times. It does this by using a parameter specifying the number of features per tile and then tiling (splitting) the dataset. The smaller datasets result in much faster run times and the results are appended to an empty output shapefile continuously.  If there is a power outage or some other interruption it should be possible to go back and determine how much of the dataset got processed and which points got processed.

The map on the left shows the original occurrence points as white dots and 10 random points around each occurrence as small black points. The new tool is available HERE . Other GIS tools are available as always from the download page on our lab's website HERE

Monday, October 5, 2015

Came across a way to pull out one raster from a multiband in Modelbuilder

Sometimes it is nice to have a quick and simple way of pulling out the single bands from a multi-band image stack without having to resort to a python script.  This stack exchange post shows how to do it in a nice, easy, and straightforward manner in ArcGIS ModelBuilder. Kudos to xyz for coming up with this answer. I was able to use this approach recently to perform principal component analysis and then continue to work with the raster downstream.

In this example the parse path tool is used to pull out the path and file name. Then Band_1 is appended in the output name of the raster calculator using the following function: "%Path%\%File%\Band_1"*1