Small areas - does it make sense to use Maxent?
105 views
Skip to first unread message
DavidBecas
Feb 27, 2012, 5:34:28 AM2/27/12
to Maxent
Hi all,
I'm new to Maxent and would like to hear your thoughts on this.
One of my goals is to model the distribution of a species inside a
small marine reserve (~54km2), based on presence data (from acoustic
telemetry) and predictor variables such as depth, bottom type, slope
and others. Does it make any sense to use Maxent for such small areas?
I would, very much, like to know your opinion on this.
Cheers,
David
I'm new to Maxent and would like to hear your thoughts on this.
One of my goals is to model the distribution of a species inside a
small marine reserve (~54km2), based on presence data (from acoustic
telemetry) and predictor variables such as depth, bottom type, slope
and others. Does it make any sense to use Maxent for such small areas?
I would, very much, like to know your opinion on this.
Cheers,
David
Dimitris Poursanidis
Feb 27, 2012, 12:55:09 PM2/27/12
to max...@googlegroups.com
Hmmmm
--
Why not ???
If you have detailed data i think that you will have also nice results.
Dimitris
--
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.
><:::::))"> ><:::::΄΄"> ><:::::>>">
Dimitris Poursanidis
..terraSolutions..
marc fernandez
Feb 27, 2012, 1:00:04 PM2/27/12
to max...@googlegroups.com
I used Maxent in small areas, as long as your ecogeographical variables maps have a good resolution for your study area it works pretty well.
Marc
2012/2/27 Dimitris Poursanidis <dpours...@gmail.com>
David Le Maitre
Feb 28, 2012, 1:42:52 AM2/28/12
to Maxent
Hi David
Maxent should work well for this application which is really about defining the species distribution at the habitat scale spatially. At this kind of scale I would guess the distribution is patchy so if you have good data for absences you may want to consider a complimentary technique such as boosted regression trees. Using both approaches can offer different insights and interpretations.
Regards
--
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.
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.
--
This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard.
The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html.
This message has been scanned for viruses and dangerous content by MailScanner,
and is believed to be clean.
Please consider the environment before printing this email.
DavidBecas
Feb 28, 2012, 6:10:47 AM2/28/12
to Maxent
Thank you all for your comments!
The resolution of my variables is around 40m, which I think is pretty
reasonable. I don't have good data for absences therefore I believe
that using methods that require absence data won't be an option.
On Feb 28, 6:42 am, "David Le Maitre" <DlMai...@csir.co.za> wrote:
> Hi David
>
> Maxent should work well for this application which is really about defining the species distribution at the habitat scale spatially. At this kind of scale I would guess the distribution is patchy so if you have good data for absences you may want to consider a complimentary technique such as boosted regression trees. Using both approaches can offer different insights and interpretations.
>
> Regards
>
> David
>
> >>> DavidBecas <davidbe...@gmail.com> 27/02/2012 12:34 >>>
The resolution of my variables is around 40m, which I think is pretty
reasonable. I don't have good data for absences therefore I believe
that using methods that require absence data won't be an option.
On Feb 28, 6:42 am, "David Le Maitre" <DlMai...@csir.co.za> wrote:
> Hi David
>
> Maxent should work well for this application which is really about defining the species distribution at the habitat scale spatially. At this kind of scale I would guess the distribution is patchy so if you have good data for absences you may want to consider a complimentary technique such as boosted regression trees. Using both approaches can offer different insights and interpretations.
>
> Regards
>
> David
>
>
> Hi all,
>
> I'm new to Maxent and would like to hear your thoughts on this.
>
> One of my goals is to model the distribution of a species inside a
> small marine reserve (~54km2), based on presence data (from acoustic
> telemetry) and predictor variables such as depth, bottom type, slope
> and others. Does it make any sense to use Maxent for such small areas?
>
> I would, very much, like to know your opinion on this.
>
> Cheers,
>
> David
>
> --
> 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 athttp://groups.google.com/group/maxent?hl=en.
> Hi all,
>
> I'm new to Maxent and would like to hear your thoughts on this.
>
> One of my goals is to model the distribution of a species inside a
> small marine reserve (~54km2), based on presence data (from acoustic
> telemetry) and predictor variables such as depth, bottom type, slope
> and others. Does it make any sense to use Maxent for such small areas?
>
> I would, very much, like to know your opinion on this.
>
> Cheers,
>
> David
>
> --
> 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.
>
> --
> This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard.
> The full disclaimer details can be found athttp://www.csir.co.za/disclaimer.html.
> --
> This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard.
Bruce Miller
Feb 28, 2012, 6:36:04 AM2/28/12
to max...@googlegroups.com
One key issue is what everyone doing introductory "GIS 101" take home
message is Grain and Extent.
Do not lose sight of the fact that for spatial data analyses you
can/should only analyze data based on the coarsest grain.
So for example if your climate layers are at a resolution depending on
where you are could be 5 km^2 for small spaces your "data points" would
only be realistically considered at the same � 5 km range. Simply
tanking a down scaled climate layer from a global scale to have grids
even as small as 30s may not be be useful at very small scales.
message is Grain and Extent.
Do not lose sight of the fact that for spatial data analyses you
can/should only analyze data based on the coarsest grain.
So for example if your climate layers are at a resolution depending on
where you are could be 5 km^2 for small spaces your "data points" would
only be realistically considered at the same � 5 km range. Simply
tanking a down scaled climate layer from a global scale to have grids
even as small as 30s may not be be useful at very small scales.
If all of your habitat-climate layers are really at a much tighter
resolution (40 meters?) then modeling based on small spatial areas may
provide realistic answers. If only you species points are at 40 meter
resolution you still are constrained by the larger grain of your
climate/habitat etc.
Cheers from the jungles of Belize
leandro.abade
Feb 28, 2012, 9:14:23 AM2/28/12
to Maxent
Hello David,
I think you have a very good resolution to perform the analyses and
cannot foresee any problem. However, I may be wrong, but I think you
should be careful when selecting the points to run the models. If you
over feed your model with too much data, you may inflate the results
and get a very generalized outcome, which may be misleading or not
representative to your area.
Maybe someone with more experience could give some feedback on it.
Reply all
Reply to author
Forward
0 new messages