Mapping of aquatic species

[MarkACantrell]

I love the mapper, and assume that it uses some Maximum Entropy modeling techniques to predict species occurrences.

Mapping of aquatic species needs a little work to limit aquatic species to aquatic habitats. For example, the predicted occurrences of rockbass or creek chub results in pixels in riparian and other areas. Could the results be constrained to the stream corridor or a stream centerline?

[evilkermit]

Perhaps we could add a mask environmental layer that only gets used when looking at aquatic species? We would need a way to know which species are aquatic, though, to make sure that we only use that layer when it's appropriate.

[MarkACantrell]

That might work.

Also, check with Mark Endries [email protected].
He has been working with MaxEnt models for several hundred species across
NC, including GRSM. He may have used NHD+ data to predict aquatic species.

Some wholly aquatic species may require a new raster layer of potential
habitats for prediction.

Thanks,
Mark

Sent from my iPhone

On Aug 26, 2016, at 3:02 AM, John Duggan [email protected] wrote:

Perhaps we could add a mask environmental layer that only gets used when
looking at aquatic species? We would need a way to know which species are
aquatic, though, to make sure that we only use that layer when it's
appropriate.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#131 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AUQ4mF8pJjSPh0GMa6BcyW9jHu7mkfX7ks5qjp2bgaJpZM4Jtafw
.

[tpcolson]

Some challenges:

We can’t use NHD Plus, as we’ve remapped all the streams in the park, and
NHD Plus hasn’t caught up yet, and won’t for many years. We cannot use
current NHD Plus due to gross inaccuracies, but we can use NHD;

In very many, in fact most cases, observations of aquatic species location
information does not place the observation location within 30 meters (grid
cell of environmental layer) of a NHD reach, some, I struggle to understand
how a fish was observed halfway up a mountain near no stream! We’d have to
do some pretty extensive QA of the locations.

What I know about Maxent is….that it’s very slow on my computer!

NPS Only GRSM GIS Sites:

Park Map https://grsmgis.nps.gov/MAP/

NPS GIS Portal
http://nps.maps.arcgis.com/home/search.html?q=GRSM&t=content&sortField=modified&sortOrder=desc&focus=layers

Looking for some Publicly available GRSM GIS Data?

Open Data http://grsm.nps.opendata.arcgis.com/

ArcGIS Online
http://www.arcgis.com/home/search.html?q=GRSM%20and%20orgid%3AfBc8EJBxQRMcHlei&t=content

NPS Data Store
https://irma.nps.gov/App/Reference/Search?SearchType=N&collectionId=1336

From: MarkACantrell [mailto:[email protected]]
Sent: Friday, August 26, 2016 6:51 AM
To: nationalparkservice/npmap-species
Subject: Re: [nationalparkservice/npmap-species] Mapping of aquatic
species (#131)

That might work.

Also, check with Mark Endries [email protected].
He has been working with MaxEnt models for several hundred species across
NC, including GRSM. He may have used NHD+ data to predict aquatic species.

Some wholly aquatic species may require a new raster layer of potential
habitats for prediction.

Thanks,
Mark

Sent from my iPhone

On Aug 26, 2016, at 3:02 AM, John Duggan [email protected] wrote:

Perhaps we could add a mask environmental layer that only gets used when
looking at aquatic species? We would need a way to know which species are
aquatic, though, to make sure that we only use that layer when it's
appropriate.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<
#131 (comment)

,
or mute the thread
<
https://github.com/notifications/unsubscribe-auth/AUQ4mF8pJjSPh0GMa6BcyW9jHu7mkfX7ks5qjp2bgaJpZM4Jtafw

.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#131 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFlxmM0r-O6hU1yshEvQJ8498i9IY5c8ks5qjsUsgaJpZM4Jtafw
.

[tpcolson]

Hi Tomas,

It is my understanding that the NHDPlus lines are the same as the NHD
medium resolution. The plus in NHDPlus is the addition of attribute data.
Is it the attribute data that you are saying has "gross inaccuracies"? If
so I'd be interested in hearing your take on the inaccuracies.

When I did my modelling I simply excluded all occurrences that fell outside
of 100 m of a stream. It wasn't a perfect solution, it's likely that there
were plenty of occurrences that were incorrect but happened to fall within
100 m of a stream, but I could only do so much.

If you are interested in seeing the methods and results of the modelling
work I did we have a webpage up here:
https://www.fws.gov/asheville/htmls/Maxent/Maxent.html

Feedback has been very positive on the predictions.

What is it you are all trying to do? I'm interested in collaborating.

Thanks!

Mark

Mark Endries
USFWS
160 Zillicoa St
Asheville, NC 28801
Office: 828.258.3939 ext. 231
Mobile: 828.215.1740

On Fri, Aug 26, 2016 at 7:08 AM, Thomas Colson [email protected]
wrote:

Some challenges:

We can’t use NHD Plus, as we’ve remapped all the streams in the park, and
NHD Plus hasn’t caught up yet, and won’t for many years. We cannot use
current NHD Plus due to gross inaccuracies, but we can use NHD;

In very many, in fact most cases, observations of aquatic species location
information does not place the observation location within 30 meters (grid
cell of environmental layer) of a NHD reach, some, I struggle to understand
how a fish was observed halfway up a mountain near no stream! We’d have to
do some pretty extensive QA of the locations.

What I know about Maxent is….that it’s very slow on my computer!

NPS Only GRSM GIS Sites:

Park Map https://grsmgis.nps.gov/MAP/

NPS GIS Portal
http://nps.maps.arcgis.com/home/search.html?q=GRSM&t=content&sortField=modified&sortOrder=desc&focus=layers

Looking for some Publicly available GRSM GIS Data?

Open Data http://grsm.nps.opendata.arcgis.com/

ArcGIS Online
http://www.arcgis.com/home/search.html?q=GRSM%20and%20orgid%3AfBc8EJBxQRMcHlei&t=content

NPS Data Store
https://irma.nps.gov/App/Reference/Search?SearchType=N&collectionId=1336

From: MarkACantrell [mailto:[email protected]]
Sent: Friday, August 26, 2016 6:51 AM
To: nationalparkservice/npmap-species
Subject: Re: [nationalparkservice/npmap-species] Mapping of aquatic
species (#131)

That might work.

Also, check with Mark Endries [email protected].
He has been working with MaxEnt models for several hundred species across
NC, including GRSM. He may have used NHD+ data to predict aquatic species.

Some wholly aquatic species may require a new raster layer of potential
habitats for prediction.

Thanks,
Mark

Sent from my iPhone

On Aug 26, 2016, at 3:02 AM, John Duggan [email protected] wrote:

Perhaps we could add a mask environmental layer that only gets used when
looking at aquatic species? We would need a way to know which species are
aquatic, though, to make sure that we only use that layer when it's
appropriate.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#131
issuecomment-242659642>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/
AUQ4mF8pJjSPh0GMa6BcyW9jHu7mkfX7ks5qjp2bgaJpZM4Jtafw>
.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#131 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFlxmM0r-O6hU1yshEvQJ8498i9IY5c8ks5qjsUsgaJpZM4Jtafw
.

[tpcolson]

· Download a detailed description of the analysis that generated
the shapefile (pdf 1.43 MB)
https://www.fws.gov/asheville/pdfs/Aquatic%20Species%20Mapping%20in%20North%20Carolina%20using%20Maxent%20V2.pdf
from that page appears to be a dead link.

For the entire application, we’re feeding in every species
occurrence/location into a maxent model being run on a super computer at
Oak Ridge, and out the other end a dynamic tiled prediction map is
generated, per species. On the aquatics sides, gross inaccurate locations
could be, a trout observation say, at the very top of a mountain, for which
we have no way to QA other than visually check every observation in the
database (200,000+). On the NHDPlus, we have since remapped all of the
streams in the park, and they are in the NHD, but not the NHDPlus, which
was based on old “Blue Line Topo Stream”, which include, for example, a
stream flowing UP a ridge, across a mountain, and sideways across a valley.
I expect it will be several years before our NHD improvement hits the
NHDPlus update cycle….they do it by HUC8 on a 10 year cycle I think. If we
can self-generate watershed area and 7Q10 flow-per-reach ourselves, then we
could likely do something like you did. Perhaps run an “aquatic species
only” version of what we’re doing. Emphasis on this application is fully
automatic and hands-free modelling….our intent is for new observations to
get consumed by the model and auto-update the distribution maps.

Took a look at your output data, very cool, we’d love to collaborate and
reproduce that for all of the HUC 12’s that intersect the park.

NPS Only GRSM GIS Sites:

Park Map https://grsmgis.nps.gov/MAP/

NPS GIS Portal
http://nps.maps.arcgis.com/home/search.html?q=GRSM&t=content&sortField=modified&sortOrder=desc&focus=layers

Looking for some Publicly available GRSM GIS Data?

Open Data http://grsm.nps.opendata.arcgis.com/

ArcGIS Online
http://www.arcgis.com/home/search.html?q=GRSM%20and%20orgid%3AfBc8EJBxQRMcHlei&t=content

NPS Data Store
https://irma.nps.gov/App/Reference/Search?SearchType=N&collectionId=1336

From: Endries, Mark [mailto:[email protected]]
Sent: Friday, August 26, 2016 10:35 AM
To: Thomas Colson
Cc: nationalparkservice/npmap-species; nationalparkservice/npmap-species;
Kendra Straub; Tom Remaley
Subject: Re: [nationalparkservice/npmap-species] Mapping of aquatic
species (#131)

Hi Tomas,

It is my understanding that the NHDPlus lines are the same as the NHD
medium resolution. The plus in NHDPlus is the addition of attribute data.
Is it the attribute data that you are saying has "gross inaccuracies"? If
so I'd be interested in hearing your take on the inaccuracies.

When I did my modelling I simply excluded all occurrences that fell outside
of 100 m of a stream. It wasn't a perfect solution, it's likely that there
were plenty of occurrences that were incorrect but happened to fall within
100 m of a stream, but I could only do so much.

If you are interested in seeing the methods and results of the modelling
work I did we have a webpage up here:
https://www.fws.gov/asheville/htmls/Maxent/Maxent.html

Feedback has been very positive on the predictions.

What is it you are all trying to do? I'm interested in collaborating.

Thanks!

Mark

Mark Endries

USFWS

160 Zillicoa St

Asheville, NC 28801

Office: 828.258.3939 ext. 231

Mobile: 828.215.1740

On Fri, Aug 26, 2016 at 7:08 AM, Thomas Colson [email protected]
wrote:

Some challenges:

We can’t use NHD Plus, as we’ve remapped all the streams in the park, and
NHD Plus hasn’t caught up yet, and won’t for many years. We cannot use
current NHD Plus due to gross inaccuracies, but we can use NHD;

In very many, in fact most cases, observations of aquatic species location
information does not place the observation location within 30 meters (grid
cell of environmental layer) of a NHD reach, some, I struggle to understand
how a fish was observed halfway up a mountain near no stream! We’d have to
do some pretty extensive QA of the locations.

What I know about Maxent is….that it’s very slow on my computer!

NPS Only GRSM GIS Sites:

Park Map https://grsmgis.nps.gov/MAP/

NPS GIS Portal
http://nps.maps.arcgis.com/home/search.html?q=GRSM&t=content&sortField=modified&sortOrder=desc&focus=layers

Looking for some Publicly available GRSM GIS Data?

Open Data http://grsm.nps.opendata.arcgis.com/

ArcGIS Online
http://www.arcgis.com/home/search.html?q=GRSM%20and%20orgid%3AfBc8EJBxQRMcHlei&t=content

NPS Data Store
https://irma.nps.gov/App/Reference/Search?SearchType=N&collectionId=1336

From: MarkACantrell [mailto:[email protected]]
Sent: Friday, August 26, 2016 6:51 AM
To: nationalparkservice/npmap-species
Subject: Re: [nationalparkservice/npmap-species] Mapping of aquatic
species (#131)

That might work.

Also, check with Mark Endries [email protected].
He has been working with MaxEnt models for several hundred species across
NC, including GRSM. He may have used NHD+ data to predict aquatic species.

Some wholly aquatic species may require a new raster layer of potential
habitats for prediction.

Thanks,
Mark

Sent from my iPhone

On Aug 26, 2016, at 3:02 AM, John Duggan [email protected] wrote:

Perhaps we could add a mask environmental layer that only gets used when
looking at aquatic species? We would need a way to know which species are
aquatic, though, to make sure that we only use that layer when it's
appropriate.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<
#131 (comment)

,
or mute the thread
<
https://github.com/notifications/unsubscribe-auth/AUQ4mF8pJjSPh0GMa6BcyW9jHu7mkfX7ks5qjp2bgaJpZM4Jtafw

.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#131 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFlxmM0r-O6hU1yshEvQJ8498i9IY5c8ks5qjsUsgaJpZM4Jtafw
.

[tpcolson]

need to

1. determine which species are "aquatic";

2. QA those obs using HEM tool