SDM from presence only public sightings data?

[Hannah K]

Hi all,

I am wondering if anyone could give an insight into how to visually represent a large data set of public sightings of marine mammals. I have created basic species distribution (i.e. the sightings) maps in QGis, and these can be separated seasonally and per species, however the data is spaced over many years and sightings are not consistent.

Is it worth using occurrence data to generate SDMs? I have read alot about using MaxEnt for presence only data, and previous posts in this forum.

There is no survey effort/tracks so I am assuming species richness or abundance per unit effort cannot be carried out either? The data includes date, lat/long and species only.

Objectives are to find particular areas which represent high species diversity and abundance seasonally.

Thank you for your time,

Hannah

[GIS in Ecology]

Hi Hannah,

This is a bit of a tricky one.

I am presuming that you are wishing to look at the distribution of individual species. If that is the case, what I would recommend is a relatively simple approach. Use the sightings of all species of cetaceans combined as a measure of effort distribution as this will give you an idea of where there has been at least some survey effort. Next, create a raster data layer that identifies which grid cells have any cetacean species recorded in it, and give these cells a value of zero (this will be your effort raster data layer). Once you have done this, create raster data layers for each individual species, with a value of 1 for each cell where that particular species was recorded as present and a value of zero where it was not recorded.  You can then use a raster calculator tool to add these raster data layers together to create a new raster data layer of species richness.

This can be done either for your whole data set, or for different seasons to give an idea of seasonal hotspots.

The biggest issue will be how you interpret the resulting maps. If you have a value of zero for a raster data layer for an individual species you know there was at least some survey effort in a grid cell (because other species were recorded in it). The trouble is when you have no data, you don't know whether this is because no cetacean species were recorded there, or whether it is because there was no survey effort in that cell. 

I hope this helps, and if you have any further questions, or need any clarification on any of the above, just post back on this thread.

All the best,

Colin

[Hanmwah X]

Hi Colin,

Thanks very much for your reply. Apologies I am only reading this now.

I will try the cell grids and rasters like you suggested.

I am using QGIS, what do you think of the Heatmap plugin? Do you think it is worth using to visualise hotspots from a scientific point of view? I'm not convinced of its validity nor do I completely understand the choice of radius option.. I haven't found any peer reviewed papers which have used it yet.

Regards your point '' The biggest issue will be how you interpret the resulting maps. If you have a value of zero for a raster data layer for an individual species you know there was at least some survey effort in a grid cell (because other species were recorded in it).''

I understand interpreting the maps, but is there an issue with not actually having effort when representing marine mammal occurrence?

Is there a way of modelling species distribution without representative survey effort?

Thank you, I appreciate your time on this forum.

--
-- ======================================================================
You received this message because you are subscribed to the "GIS In Ecology Forum" discussion group (http://www.gisinecology.com/GIS_in_Ecology_forum.htm).
 
To control how often you get emails from this group, go to http://groups.google.com/group/gis-in-ecology-forum/subscribe (you will need to log on to get to this page).
 
To post to this group, either log onto the group's home page or send an email to
gis-in-eco...@googlegroups.com.
 
The rules for posting to this group can be found here: http://groups.google.com/group/gis-in-ecology-forum/browse_thread/thread/df31a0822742203f#.
 
To unsubscribe from this group, email:
gis-in-ecology-f...@googlegroups.com
 
All information on this forum is provided on an 'as is' basis. GIS In Ecology is not responsible for checking the accuracy or suitability of any posting or response.
======================================================================
---
You received this message because you are subscribed to the Google Groups "GIS In Ecology Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gis-in-ecology-f...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gis-in-ecology-forum/0d64e948-8f8e-49c6-8cdf-c543699fd949%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[cdma...@gisinecology.com]

Hi Hannah,

You can use kernel density estimates (which is what the Heatmap plugin
produces) to help you get an idea of species distributions. However, in
order to create a kernel density estimate (KDE), you need to decide on
something called the bandwidth (also known as H and the search radius).
This cna be difficult to do, and is not necessarily valid for species
with a patchy distribution (which is most mairne species). It may well
be worth giving ago, to see what the maps it gives you looks like.

In terms of modelling species distirbutions without accompanying survey
effort, the way I described below does not require that you have survey
effort (it uses the presence of other species as the absence points for
the target species). This is known as pseudo-absence modelling. However,
you just need ot be careful that you don't over-interpret the results,
unless you can validate them with an independent data set (but this is
an issue for all types of species distribution modelling). You can also
use presence-only modelling, but this needs your data set to be quite
large, and again has the same issue with inteprering the resulting
distributions.

How large is your data set? And how many different species does it
contain?

All the best,

Colin

On 2019-10-17 19:40, Hanmwah X wrote:
> Hi Colin,
>
> Thanks very much for your reply. Apologies I am only reading this now.
>
>
> I will try the cell grids and rasters like you suggested.
>
> I am using QGIS, what do you think of the Heatmap plugin? Do you think
> it is worth using to visualise hotspots from a scientific point of
> view? I'm not convinced of its validity nor do I completely understand
> the choice of radius option.. I haven't found any peer reviewed papers
> which have used it yet.
>

> Regards your point '' _The biggest issue will be how you interpret the

> resulting maps. If you have a value of zero for a raster data layer
> for an individual species you know there was at least some survey

> effort in a grid cell (because other species were recorded in it)._''

>> [1].

>> For more options, visit https://groups.google.com/d/optout.
>

> https://groups.google.com/d/msgid/gis-in-ecology-forum/CADKc50rfUF1Gv-xt3ZwmT%3DYTO4DAnSmM_Su6YmK%3Dy1AVUnrBCw%40mail.gmail.com
> [2].
>
>
> Links:
> ------
> [1]
> https://groups.google.com/d/msgid/gis-in-ecology-forum/0d64e948-8f8e-49c6-8cdf-c543699fd949%40googlegroups.com?utm_medium=email&utm_source=footer
> [2]
> https://groups.google.com/d/msgid/gis-in-ecology-forum/CADKc50rfUF1Gv-xt3ZwmT%3DYTO4DAnSmM_Su6YmK%3Dy1AVUnrBCw%40mail.gmail.com?utm_medium=email&utm_source=footer

[Hanmwah X]

Hi Colin, 

I have the upwards of 13,000 public sightings, a few years of ship surveys, CODA, SCANS 1&11, ObSERVE data. SCANS and CODA etc. have effort attached to them, but I would like to incorporate these data with the public sightings (as this is the biggest data set). Could I incorporate SCANS etc. sightings data with the public sightings data and use the pseudo-absence modelling that you mentioned above? It would increase the over all data set, making it more feasible for this type of modelling? 

I appreciate your time. Thank you

To view this discussion on the web visit https://groups.google.com/d/msgid/gis-in-ecology-forum/d33a8d72b8bfe8857c9e244be20242de%40gisinecology.com.