Using Minimum Convex Polygon in Maxent

Using Minimum Convex Polygon in Maxent

I am currently using Maxent to model species distribution over time using climate change data and species occurrence. I am working with three environmental variables which are all set to the same geographic projections and dimensions and saved as ascii files. I am also using my species occurrence data, which is displayed in three columns: species, UTME, and UTMN in a CSV file.

I am trying to use a minimum convex hull as a bias file to account for sampling bias in my occurrence data. I created these masks, by using the "Minimum Bounding Geometry" tool in ArcGIS, I input the species occurrence data, which was point data, to create a minimum convex polygon (MCP) mask. The result was a polygon that included all of the species occurrence points. In order to load the minimum convex polygon mask into maxent I had to convert the polygon -> raster, and then raster -> ascii. The aim is to use the MCP mask during Maxent model training, so that all cells that fall within the MCP polygon have equal likelihood of being selected as background locations.

When I plug all of my data into Maxent, I get an error message that says my climate layers (ascii files) and the bias mask have different geographic dimensions. How do I get them into the same geographic dimensions if they are displaying different types of data?

I am not sure on how to fix this problem, and can't move forward with my models until I can fix it. Any thoughts?

If you are using Feature to Raster from the toolbox to convert the MCP to raster, first add one of your ascii climate layers to the table of contents, then:

  1. click the "Environments" button at the bottom of the tool
  2. Set "Processing Extent" to be "same as layer "
  3. set "Processing Extent" snap raster to the ascii climate layer
  4. Set "Raster Analysis" cell size to the ascii climate layer
  5. Set "Raster Analysis" mask to the ascii climate layer

I'm sure not all of these are required to get the correct extent on your output raster, but I find it easier to just set them all and not worry about it.

Following this, the raster grid will match the climate layers, and then you can export to ascii.

You can also quickly check that it worked by opening the ascii grid in a text editor and comparing the header lines to one of your climate layers.

Watch the video: Spatial Statistics - Local Metrics