[Biomod-commits] using Migration

Mon Sep 10 09:02:06 CEST 2012

Hi there,

> I am using centroids of 10x10 km quadrats (approx. 2600 quadrats total) as the input presence/absence data. I have 22 species and used pseudo-absences to model current and future distributions for 2 time horizons and 3 climate change scenarios. Coordinates are in meters, not degrees.
> I tried setting MaxMigr to 0, and also 1, 100, 1000, and both 'limited dispersal' and 'unlimited dispersal' outputs are identical. 

Weird… indeed. Would you mind sending us your data (in private email)?  
OK, we need to further check this function and update it with the new functionalities of sp, spdep and raster packages. This function is a bit out-dated now. 

> Leaving that aside, I don't understand how Biomod.RangeSize predicts pixel loses and gains with '0 dispersal' and 'unlimited dispersal'. The Compt.By.Species summary seems to make sense, but which of the 2 scenarios (0 dispersal or unlimited dispersal) does Diff.By.Pixel data represent? It almost looks like it shows the unlimited dispersal scenario, but I am mostly interested in the 'no dispersal' projections of loss and gain. I am working with reptiles and amphibians, and the 'no dispersal' scenario would be more realistic. Something in between 0 and unlimited dispersal would be ideal, and this is where I thought the Migration function comes in handy

Diff.By.Pixel is giving you both actually. 

In the case on "no dispersal", do not consider the +1 (they show you specifically the new sites available in case of dispersal). 

In summary, playing with the 4 levels of Diff.By.Pixel, you have:

Current distribution:  (-2) +  (-1)
Future distribution with no dispersal: -1
Future distribution with full dispersal: (-1)  + (+1)
Areas lost: -2
Areas potentially gained (in case of dispersal): +1
Stable areas: -1

Does it make sense? 

As soon as the Migration function is sorted out, you can then re-use the Biomod.Range.Size function. The unlimited migration in that case, will be the one you have stetted up in Migration to constrain the future distribution. 

Best

Wilfried

> 
> Thank you
> 
> Cheers,
> Viorel 
> 
> 
> 
> On Fri, Sep 7, 2012 at 1:14 AM, Wilfried Thuiller <wilfried.thuiller at ujf-grenoble.fr> wrote:
> Hi Viorel,
> 
> It has been a long time I did not have a look at this function. 
> What is the resolution of your data? 
> Just for a first try, what does it happen if you set MaxMig at 0. 
> 
> Best
> Wilfried
> 
> 
> 
> Le 6 sept. 2012 à 20:04, Viorel Popescu a écrit :
> 
>> Hi everyone,
>> 
>> I have a few questions about using Migration and how/if it feeds into the Biomod.RangeSize function. I produced current and future distributions using Ensemble.Forecasting, but now I would like to add more realism using the Migration() function and the dispersal parameter MaxMigr. I understand that the MaxMigr units have to be in decimal degrees, but no matter if MaxMigr = 1000 or MaxMigr = 0.16667, the resulting objects (Future20a1b.Migration in my case) are completely identical in terms of probability values.
>> 
>> Migration(CurrentPred = Total_consensus_Current[,,2], FutureProj = Total_consensus_Future20a1b[,,2],
>> X=LatLong[,1], Y=LatLong[,2], MaxMigr=100, Pred.Save="Future20a1b.Migration")
>> 
>> Moreover, when trying to plot the Current, Limited Dispersal and Unlimited Dispersal distributions, the latter 2 look identical (again, regardless of the value given to MaxMigr)  
>> 
>> par(mfrow=c(1,3))
>> level.plot(Resp.Var[,'Sp21'], XY=LatLong, show.scale=FALSE, title="current distribution", cex=0.8)
>> level.plot(Future20a1b.Migration[,21], XY=LatLong, show.scale=F, title="limited migration", cex=0.8)
>> level.plot(Total_consensus_Future20a1b_Bin[,21,2], XY=LatLong, show.scale=FALSE, title="unlimited migration", cex=0.8)
>> 
>> Am I doing something wrong with Migration? Also, does the information from Migration feed into the RangeSize function? If so, it is not clear to me how...
>> 
>> Thanks in advance for any advice
>> 
>> Cheers,
>> Viorel
>> 
>> 
>> 
>> -- 
>> Viorel D. Popescu, PhD
>> David H. Smith Conservation Research Fellow
>> University of California - Santa Cruz &
>> Simon Fraser University, Biological Sciences
>> 
>> _______________________________________________
>> Biomod-commits mailing list
>> Biomod-commits at lists.r-forge.r-project.org
>> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/biomod-commits
> 
> --------------------------
> Dr. Wilfried Thuiller
> Laboratoire d'Ecologie Alpine, UMR CNRS 5553
> Université Joseph Fourier
> BP53, 38041 Grenoble cedex 9, France
> tel: +33 (0)4 76 51 44 97
> fax: +33 (0)4 76 51 42 79
> 
> Email: wilfried.thuiller at ujf-grenoble.fr
> Personal website: http://www.will.chez-alice.fr
> Team website: http://www-leca.ujf-grenoble.fr/equipes/emabio.htm
> 
> ERC Starting Grant TEEMBIO project: http://www.will.chez-alice.fr/Research.html
> FP6 European EcoChange project: http://www.ecochange-project.eu
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> -- 
> Viorel D. Popescu, PhD
> David H. Smith Conservation Research Fellow
> University of California - Santa Cruz &
> Simon Fraser University, Biological Sciences
> 8888 University Drive
> Burnaby, BC V5A 1S6, Canada
> (604) 340 4228
> www.umit.maine.edu/~Dan_Popescu/
> 

--------------------------
Dr. Wilfried Thuiller
Laboratoire d'Ecologie Alpine, UMR CNRS 5553
Université Joseph Fourier
BP53, 38041 Grenoble cedex 9, France
tel: +33 (0)4 76 51 44 97
fax: +33 (0)4 76 51 42 79

Email: wilfried.thuiller at ujf-grenoble.fr
Personal website: http://www.will.chez-alice.fr
Team website: http://www-leca.ujf-grenoble.fr/equipes/emabio.htm

ERC Starting Grant TEEMBIO project: http://www.will.chez-alice.fr/Research.html
FP6 European EcoChange project: http://www.ecochange-project.eu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/biomod-commits/attachments/20120910/edad1aca/attachment.html>