0

Good afternoon,

I have recently developed a Maxent species distribution model in R for a reptile endemic to Madagascar, using native-range presence-only data. The model performs well (as judged by AUC).

I have now projected this model onto Florida, where it identifies no suitable habitat. However, my focal taxon is well established in Florida. I did not use Florida presence data in the model because I wanted to check if the model would have a priori predicted the colonisation event; as such, I trained/validated the model using Madagascar as the model extent.

I would like to recover quantified probability values for the known Florida occurrences; at the moment, I can see in general that the predicted probabilities in the colonised area are low, but I cannot quantify them. I can find the equivalent, quantitative values for my known Madagascar occurrences (because they were included in the model) in the model output (Training points) or using evaluate in 'dismo' (Validation points).

Does anybody know of a method/package/function which I can use to find these values for my Florida points (currently just stored as a .csv).

I can share more technical details of the model if necessary; it is a fairly standard Bioclim/Maxent/presence-only data in R, using predominantly 'dismo'.

PhelsumaFL
  • 61
  • 8

1 Answers1

0

You can get them with raster::extract(). For example:

library(raster)

d <- read.csv("data.csv") #your csv here
r <- raster("maxent.tif") #your suitability raster here
vals <- extract(r, d) #these are the values of r at your locations

Regarding your issues with low predicted suitability when transferring the model to another geographic region, this can be an overfit issue. AUC of training points will be high if overfit is present, hence you should rely on another metric to assess it. For example, you can run a cross-validation (partition Madagascar records into training and testing points) and quantify the difference in AUC among training and testing points. Otherwise ORmin and ORmtp are widely used. You can check this paper for further details and how to deal with overfit in maxent models: Radosavljevic & Anderson, 2013.

Good luck!

/Emilio

E.B.
  • 76
  • 3
  • Dear Emilio, thank you for the advice. This worked perfectly. If you are happy to provide your full name and Institutional Affiliation (if applicable) I would like to thank you in the Acknowledgements section of the manuscript. Thanks again. – PhelsumaFL Jul 27 '20 at 17:56