How to choose GCMs from WorldClim?

[Weverton Carlos]

Hi everybody,

Can someone suggest me some criterion to choose a GCM from WorldClim?

I have ever read a lot of alternatives:

- Choosing the models that projects extreme conditions (maximum and minimum projections).

- Building a new projection using the mean of GCMs.

- Choosing the models that are more stable (with little variation when compared to the other GCMs) in my study region.

I thought of doing a cluster analysis with the models and choosing one GCM from each cluster, what do you think about that?

Thanks,

Weverton Trindade,

State University of Ponta Grossa (Brazil).

[Jamie M. Kass]

Weverton,

if you have a script for your analysis, I would suggest running one GCM from each major climate center (e.g., Hadley, MIROC, CSIRO, etc.). This way you can examine the variability in the output and comment on it. Some may be considered more accurate for some regions, and that may or may not be correct, but because these models are so complex and variation between sites can be so high, it’s best to report your findings for a few of them that span this variability. Maybe plot the standard deviation per cell of the output? A couple of papers I’ve read have done that.

Jamie Kass
PhD Candidate
City College of NY

[Weverton Carlos]

Thanks, Jamie.

I've seen some papers that suggest to use GCMs with "extreme projections", because it shows how the specie could respond to "the best" and "the worst" projections.

I think I'll follow this suggestion.