0

In ArcGIS 10, I have several maps describing both habitat characteristics and climate. I also have a raster map of presence/absence for a species that I wish to examine using my other maps. All my maps are in raster format. I wish to merge all of the information from all of my individual maps into one table, that I can use for regression analysis, testing what features of the habitat and climate my species depend on.

I have first ensured that all of my maps have the exact same extent and shape (using the Clip tool).

I have also created a fishnet (my polygon), and wish to calculate the mean of each of my single maps within each of the grids of the fishnet (except for my species, where I wish to find the majority value for each cell). I used the Zonal statistics as table to extract the majority value of my species dataset into the fishnet. In order to check that the data was properly transferred to the table, I used the Join Field Tool to join my zonal table to the fishnet polygon. To check, I used Select by attribute and choose all instances where my species is present (value 1). The problem is that the chosen part of the dataset does not correspond with the actual species distribution (see attached picture). The picture shows the selected part of my fishnet (blue) compared to the actual distribution (presence = black, absence =gray). If done correctly, I assume all black and blue dots should overlap.

I have also attached a picture of my model and one of the specifics in the Zonal statistics as table Tool.

I hope someone can tell me what I have to do in order to join all of my dataset into one table in a correct way.

Map of distribution

Model

Zonal

2

Assuming your species binary raster is represented by 0 (absent) and 1 (present)

MAJORITY — Determines the value that occurs most often of all cells in the value raster that belong to the same zone as the output cell.

MAXIMUM — Determines the largest value of all cells in the value raster that belong to the same zone as the output cell.

Since it seems you are only looking to check if there is one cell in the raster that indicates the species is present. The maximum value than would be 1, whereas when none of the cells have the species present, the maximum would be 0.

After some discussions and learning more details about the intended outcome, it seems that the better approach for this part of the analysis is to create centroid points for the fishnet polygons, then use the Extract Multivalues to Points tool on these points to write the raster values to a table. This table can then be joined with the others.

  • I intended to use majority over maximum because in that way, I would ensure that suboptimal areas with only few presence point were not included as a presence. In that way, only grids where my species is absent more often than present will be included. But if changing it to maximum solves my problem, I am willing to use that function instead :) I will then just have to rewrite the aim of my analysis. – LHAndersen May 27 '16 at 13:45
  • 1
    That's why I'm asking for more details. It would depend on how the size of your fishnet polygons relates to the raster cell size. – PhilippNagel May 27 '16 at 13:49
  • The raster cell size is 0.1x0.1 and the fishnet have grids of 0.045484x0.045484. So since the size of my fishnet is smaller than that of my raster, so I might as well use maximum. Therefore, I tried the same model as above, only specifying maximum instead of majority, but the areas did still not overlap correctly. – LHAndersen May 27 '16 at 14:02
  • 2
    So then it seems that for this part of your analysis, you are not really looking for "statistics" in the sense that ArcGIS tools use the term, but rather just the presence of your species as the dependent variable for your regression. I would suggest you create centroid points of your fishnet grid and then use the "Extract Multivalues to Points" tool. It will simple write the raster value at the point location to a table. – PhilippNagel May 27 '16 at 14:14
  • Thank you so much! I tried using the Extract multivalues to point Tool, and it worked just perfectly! You saved my day :) – LHAndersen May 27 '16 at 14:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.