Calculating AREA of Maxent output in ArcGIS

[Topas]

Hello,

I made a suitability map for wolves in Switzerland in Maxent, loaded
the output ascii file in ArcGIS, converted it into floating point
raster and everything is fine. The suitability value from 0 to
0.952... is well displayed.
In the "Symbology" tab of the layers properties I classified the
values into two classes:

1. from 0 - 0.5 indicating areas which do not suit for wolves
2. from 0.5 - 0.952... indicating areas which suit wolves

I get a map classification with most of the map red (not suitable) and
some green (suitable) patches.
Some of them are clumped, building bigger patches of suitable area,
some are single pixels of green not connected to the others.
I want to get rid of the single pixels and other small patches of only
a few pixels cause these areas are too small to be counted as suitable
wolf habitat.
Is there a way to say ArcGIS to only keep patches above a certain
threshold of total pixel size (or pixel number) and eliminate the
others. Or do you know another way to do this?

I explain it again in a biological way: I want to sort out single
suitable habitat pixels in the map to end up with only connected
suitable area pixels which form a habitat big enough for a wide -
roaming species like wolf to survive in.

I hope I explained well what I want. I am stuck here and can't figure
out how this works.

Thank you,
Tobias, Switzerland

[Heather Peacock]

Hi Tobias,

I don't know of a command or tool to automatically remove the pixels you want, however, you may be able to remove these single pixel 'habitat area' manually.  Convert the raster to a polygon (convert raster to features or raster to polygon), make sure it is in a geodatabase so you have the area calculated for you.  In the attribute table you can select the areas with the smallest area or areas below the minimum threshold you have determined for your species.  Turn on the editor tool and simply delete these polygons.  If you need to work in rasters you can convert the polygon file back to raster, but to avoid distortion make sure to unclick simplify polygons when you convert to polygons initially.

This should work, let me know if you have any problems.

Heather

--
You received this message because you are subscribed to the Google Groups "Maxent" group.
To post to this group, send email to max...@googlegroups.com.
To unsubscribe from this group, send email to maxent+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/maxent?hl=en.

[Topas]

Hi Heather,
thank you. I also found out that deleting the pixels manually is maybe
a good alternative to finish the job.
You say my polygon is in a geodatabase to let the AREA be calculated.
Do you mean an ESRI Personal Geodatabase or File Geodatabase?
How can I input my shapefile into one on these?
I'll check it tomorrow, now is time to go home and sleep. :)

Thank you again!!!

Tobias

On Nov 25, 8:36 pm, Heather Peacock <heather.peac...@gmail.com> wrote:
> Hi Tobias,
>
> I don't know of a command or tool to automatically remove the pixels you
> want, however, you may be able to remove these single pixel 'habitat area'
> manually.  Convert the raster to a polygon (convert raster to features or
> raster to polygon), make sure it is in a geodatabase so you have the area
> calculated for you.  In the attribute table you can select the areas with
> the smallest area or areas below the minimum threshold you have determined
> for your species.  Turn on the editor tool and simply delete these polygons.
>  If you need to work in rasters you can convert the polygon file back to
> raster, but to avoid distortion make sure to unclick simplify polygons when
> you convert to polygons initially.
> This should work, let me know if you have any problems.
>
> Heather
>

> > maxent+un...@googlegroups.com<maxent%2Bunsu...@googlegroups.com>
> > .

[Heather Peacock]

Yes, an ESRI file or personal geodatabase.  In ArcCatalog you can create a new gdb by right clicking the folder in which you wish to work.  I prefer to work with file geodatabases but either will work; they have more storage space.  You can either create the geodatabase first and save the shapefile to it directly from the raster to polygons tool or you can import the shapefiles into the gdb (right click then select import).

Good luck with it.

 

Heather

 

 

To unsubscribe from this group, send email to maxent+un...@googlegroups.com.

[Antonio Trabucco]

Dear Heather,
You can consider the zonal commands in arcgis. Defining the zone width to an
extent similar to the home range of wolves, and extract how many pixels are
suitable and maybe the sum of suitability pixels within the zone (zonalsum)
...... then a pixel with some suitability for species distribution may be
accounted together with the overall suitability in the zone. Take away
single isolated suitable pixels has been advisable, but as geographic
biodiversity studies intensify, I found more and more interest for those
isolated pixels (e.g. for planning of biological corridors). Take away
isolated pixels is becoming a more thoughtful process, and I would like to
hear comments.

Antonio Trabucco
Forest Ecology and Management
Division Forest, Nature and Landscape
K.U.Leuven

Thank you again!!!

Tobias

maxent+un...@googlegroups.com.

[JC Riveros]

Hi Topas,

You could use the Spatial Analyst extension of ArcGis. There are several commands to "clean" your raster dataset.

I guess you should use Region Group to filter the clusters above certain threshold.

JC

--
You received this message because you are subscribed to the Google Groups "Maxent" group.
To post to this group, send email to max...@googlegroups.com.
To unsubscribe from this group, send email to maxent+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/maxent?hl=en.

--

JC Riveros

[Topas]

Hello Antonio,
thats quite a good point you are having. Isolated patches can indeed
function as corridors for some species and there is much sense in it
not to eradicate them during the processing in ArcGIS but only as long
as it makes sense in a biological way. My resolution is 1000 x 1000m
(1sqkm). I have isolated patches consisting of only three pixels,
makes a total size of 3 sqkm. This is much much too less for a wide -
roaming species like wolf to use it as habitat but can be used as
stepping stones to move across or through not suitable or even
dangerous areas. Studies have shown (from radio - collared wolves)
that wolves use small forest patches to cover over day when they
travel through let's say more urbanized areas. So, whether to skip or
to retain pixels and small patches up to a certain size is really a
matter of which species you deal with and how it behaves etc...
Tobias

> > maxent+un...@googlegroups.com<maxent%2Bunsu...@googlegroups.com>
> > .