Calculating the Coefficient of Variation Grid?

[WolfeMD]

Howdy Group,

Having just run my first set of replicate models, I am examining the
plot of model standard deviation. I am thinking that the Coefficient
of Variation (Variance / Mean) would be a more useful and appropriate
plot of the uncertainty in the model.

So two questions:

1. Am I thinking of this correctly? Since each pixel will have a
different mean because of different suitability?

2. Does Maxent have a batchmode flag or some other way of outputting a
grid file that would plot CV? Or else how would I do this? I am no GIS
expert and so I doubt my ability to accomplish this at the moment.

Thanks all,

Marnin

[Marnin Wolfe]

For the record... Coefficient of Variation = (Std. Deviation / Mean)*100... not variance/mean. Sorry!

--
Marnin Wolfe
University of Pittsburgh
Department of Biological Sciences
Ecology & Evolution Program
wol...@gmail.com (or)
md...@pitt.edu
239-595-5081

[Bill Peterman]

Marnin--

This is not a difficult thing to do in ArcGIS (I cannot speak of other
GIS). First you need to convert the Maxent .asc files into ArcGIS grid
files. To do this, go into Conversion Tools>>To Raster and select
"ASCII to Raster". Simply input your .asc file, name your output, and
make sure to set the output data type to FLOAT. Once converted, the
grids should be added to your open project.

Next, you need to make sure that the Spatial Analyst extension is
turned on (Tools>>Extensions). Then go to View>>Toolbars>>Spatial
Analyst to view the toolbar (these first steps may not even be
necessary if you see a toolbar that says "Spatial Analyst" on it.
Click on the Spatial Analyst toolbar and select "Options". Here you
will set the directory that you want your file to go to as well as the
extent, cell size, and any clipping mask to reduce your output area.
Once you have these parameters set, close the options window and go
back to the Spatial Analyst toolbar and select "Raster Calculator".
Now simply type in your equation:
CV=(StDEV/Avg)*100

Good Luck--Bill

> wolf...@gmail.com (or)
> md...@pitt.edu
> 239-595-5081

[Marnin Wolfe]

Bill and Group,

Thanks so much for the input. Actually, I haven't quite figured out how to do what you suggested, but it was useful because I was able to load both my Maxent outputs and my Bioclim variables directly in to ArcGIS for the first time and have it work using the "ASCII to Raster." I hadn't tried the "FLOAT" specification. A few questions stem from this I was wondering if you could advise me further:

1. What does "FLOAT" do?

2. When I open the Spatial Analyst toolbar, all the functions you mentioned are there, but grayed out and unclickable. I can't even use the pulldown menu that says "specify layer." Any clue why this is?

3. On a related issue, I want to include the topographic data from the USGS Hydro1K dataset (http://edc.usgs.gov/products/elevation/gtopo30/hydro/europe.html). The data are in *.BIL format which shouldn't be a problem for me, but when ArcMap loads them it says it is unable to project the raster because it has an unknown coordinate system. The USGS site says the rasters are projected in "ambert Azimuthal Equal Area projection was selected for this database. (Steinwand et al, 1995)." So I tried specifying "ETRS 1989 LAEA." This plots the raster but it won't overlap the BIOCLIM variables, which I think are in WGS1984. I tried to convert the projection from LAEA to WGS84 but this failed horribly. 

Maybe this is a question for another post, but could you advise me on this?

Thanks so much for your patience!

Cheers,

Marnin

wol...@gmail.com (or)
md...@pitt.edu
239-595-5081

[Jessica Junker]

Dear Marnin

I am also working with the Hydro-1-K datasets (slope, elevation and cti)
and I projected the layers to the lat long WGS 1984 geographical system
with Envi. They work fine with my other layers. What I could perhaps do
is put those layers on a website for you to download? Let me know.
Jessi

> > wolf...@gmail.com <mailto:wolf...@gmail.com> (or)
> > md...@pitt.edu <mailto:md...@pitt.edu>

> > 239-595-5081
>
>
>
>
> --
> Marnin Wolfe
> University of Pittsburgh
> Department of Biological Sciences
> Ecology & Evolution Program

> wol...@gmail.com <mailto:wol...@gmail.com> (or)
> md...@pitt.edu <mailto:md...@pitt.edu>
> 239-595-5081
>
> >


--
Jessica Junker
Max-Planck Institute for Evolutionary Anthropology
Department of Primatology
Deutscher Platz 6
04103 Leipzig
Germany

tel: +49 341 3550 217
fax: +49 341 3550 299
cell: +49 0160 90594996
email: jessica...@eva.mpg.de

[Marnin Wolfe]

Jessi,

That would be awesome! I have no idea what I'm doing wrong but it just won't project correctly. Just to make sure, you are working with the European Hydro1K set?

Thanks so much,

Marnin

wol...@gmail.com (or)
md...@pitt.edu
239-595-5081

[Bill Peterman]

Hi Marnin--

Glad you are making some progress. The "FLOAT" vs. "INTEGER" options
when you import an ASCII file pertains to whether your ASCII file is
comprised only of integer (e.g. 0,1) or if your file is continuous
with decimal values (FLOAT). As far as Spatial Analyst not being
accessible, it sounds like you have not turned on the extension. Go to
Tools>>Extensions and then make sure that the box next to Spatial
Analyst is checked.

-Bill-

On Apr 22, 7:19 am, Marnin Wolfe <wolf...@gmail.com> wrote:
> Jessi,
> That would be awesome! I have no idea what I'm doing wrong but it just won't
> project correctly. Just to make sure, you are working with the European
> Hydro1K set?
>
> Thanks so much,
>
> Marnin
>
> On Wed, Apr 22, 2009 at 5:18 AM, Jessica Junker

> <jessica_jun...@eva.mpg.de>wrote:

> > Jessica Junker
> > Max-Planck Institute for Evolutionary Anthropology
> > Department of Primatology
> > Deutscher Platz 6
> > 04103 Leipzig
> > Germany
>
> > tel: +49 341 3550 217
> > fax: +49 341 3550 299
> > cell: +49 0160 90594996

> > email: jessica_jun...@eva.mpg.de

>
> --
> Marnin Wolfe
> University of Pittsburgh
> Department of Biological Sciences
> Ecology & Evolution Program

> wolf...@gmail.com (or)
> md...@pitt.edu
> 239-595-5081

[Jessica Junker]

Hey Marnin
Sorry, I didn't see - I am working with the Africa layers, not the
European ones.
Jessi

email: jessica...@eva.mpg.de