r.watershed

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

r.watershed

Shane Carey
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Helmut Kudrnovsky
Shane Carey wrote
> Hi,
>
> I am trying to extract river network from a 5m DEM with some success using
> r.watershed. Has anyone tested this algorithm on high resolution LiDAR
> data
> for example - 1meter DTM and what kind of results have they obtained?

in the NC sample dateset there is lidar derived 1m x 1m DEM:

r.info map=elev_lid792_1m@PERMANENT                                            

+----------------------------------------------------------------------------+
 | Map:      elev_lid792_1m@PERMANENT       Date: Wed Feb 14 19:45:24 2007  
|
 | Mapset:   PERMANENT                      Login of Creator: helena        
|
 | Location: nc_spm_08_grass7                                                
|
 | DataBase: D:\grassdata                                                    
|
 | Title:    Rural area: Lidar-based 1m DEM                                  
|
 | Timestamp: none                                                          
|

|----------------------------------------------------------------------------|
 |                                                                          
|
 |   Type of Map:  raster               Number of Categories: 255            
|
 |   Data Type:    FCELL                                                    
|
 |   Rows:         750                                                      
|
 |   Columns:      700                                                      
|
 |   Total Cells:  525000                                                    
|
 |        Projection: Lambert Conformal Conic                                
|
 |            N:     220750    S:     220000   Res:     1                    
|
 |            E:     639000    W:     638300   Res:     1                    
|
 |   Range of data:    min = 103.7643  max = 131.598                        
|
 |                                                                          
|
 |   Data Source:                                                            
|
 |    vector file ncfl_lidar_79200WC_pts                                    
|
 |                                                                          
|
 |                                                                          
|
 |   Data Description:                                                      
|
 |    generated by v.surf.rst                                                
|
 |                                                                          
|
 |   Comments:                                                              
|
 |    giventension=300.000000, smoothing=1.000000                            
|
 |    dnorm=47.606041, dmin=1.000000, zmult=1.000000                        
|
 |    segmax=25, npmin=100, rmsdevi=0.222627                                
|
 |    zmin_data=103.513000, zmax_data=131.668000                            
|
 |    zmin_int=103.764260, zmax_int=131.598045                              
|
 |                                                                          
|
 |    v.surf.rst -t input="ncfl_lidar_79200WC_pts" layer=0 elev="elev_lid7\  
|
 |    9200_1m" slope="slope_1m" aspect="asp_1m" pcurv="pcurv_1m" tcurv="tc\  
|
 |    urv_1m" tension=300 smooth=1.0 segmax=25 npmin=100 dmin=1 dmax=2.500\  
|
 |    000 zmult=1.0                                                          
|
 |                                                                          
|

+----------------------------------------------------------------------------+

you can test the module with this DEM to see it's appropriate for your need.






-----
best regards
Helmut
--
Sent from: http://osgeo-org.1560.x6.nabble.com/Grass-Users-f3884509.html
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
best regards
Helmut
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Mark Seibel
In reply to this post by Shane Carey
Hi Shane.

I'm happy to report that I've modeled overland water flow with r.watershed for over a quarter million acres, consisting of several large project sites, at 1 meter DEM resolution. The data source was LiDAR points to make the DEMs.

At this resolution, it becomes necessary to add culverts, or other artificial flow control features, to achieve water flowing through a road. Otherwise, water is routed along roads until a lowest point is reached for crossing.

I also use r.terraflow outputs as ancillary data to help drop in culvert locations and help provide guidance in problem areas.

My geographic area is central Florida, which is very flat and full of topographic depressions known as wetlands. These depressions interrupt the stream network continuity in reality, but r.watershed does a fantastic job making a continuous drainage network model, especially in these difficult areas.

Happy Modeling!

Mark

On Tue, Dec 5, 2017, 3:49 PM Shane Carey <[hidden email]> wrote:
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Shane Carey
Hi Mark,

Thanks for your reply! Sounds great.
How did you add culverts or other artificial flow control features to achieve water flowing through roads?

I have a rivers layer and I compared it the streams I've obtained from r.watershed and r.watershed appears to not match these streams (which were accurately digitised) and I was wondering if I had a better resolution DTM would it solve this problem?

Also, why is sink filling needed for terraflow and not watershed?

Thanks for your help :-)

On Máirt 5 Noll 2017 at 23:43, Mark Seibel <[hidden email]> wrote:
Hi Shane.

I'm happy to report that I've modeled overland water flow with r.watershed for over a quarter million acres, consisting of several large project sites, at 1 meter DEM resolution. The data source was LiDAR points to make the DEMs.

At this resolution, it becomes necessary to add culverts, or other artificial flow control features, to achieve water flowing through a road. Otherwise, water is routed along roads until a lowest point is reached for crossing.

I also use r.terraflow outputs as ancillary data to help drop in culvert locations and help provide guidance in problem areas.

My geographic area is central Florida, which is very flat and full of topographic depressions known as wetlands. These depressions interrupt the stream network continuity in reality, but r.watershed does a fantastic job making a continuous drainage network model, especially in these difficult areas.

Happy Modeling!

Mark

On Tue, Dec 5, 2017, 3:49 PM Shane Carey <[hidden email]> wrote:
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Marco Alicera
How did you add culverts?!
Such a great question and I also wonder how you did. Short ago I knew about Itzï and its ability to do it with SWMM http://itzi.readthedocs.io/en/latest/tutorial.html#culvert-modelling. Looking forward to testing it
--
Marco

2017-12-06 9:24 GMT+01:00 Shane Carey <[hidden email]>:
Hi Mark,

Thanks for your reply! Sounds great.
How did you add culverts or other artificial flow control features to achieve water flowing through roads?

I have a rivers layer and I compared it the streams I've obtained from r.watershed and r.watershed appears to not match these streams (which were accurately digitised) and I was wondering if I had a better resolution DTM would it solve this problem?

Also, why is sink filling needed for terraflow and not watershed?

Thanks for your help :-)

On Máirt 5 Noll 2017 at 23:43, Mark Seibel <[hidden email]> wrote:
Hi Shane.

I'm happy to report that I've modeled overland water flow with r.watershed for over a quarter million acres, consisting of several large project sites, at 1 meter DEM resolution. The data source was LiDAR points to make the DEMs.

At this resolution, it becomes necessary to add culverts, or other artificial flow control features, to achieve water flowing through a road. Otherwise, water is routed along roads until a lowest point is reached for crossing.

I also use r.terraflow outputs as ancillary data to help drop in culvert locations and help provide guidance in problem areas.

My geographic area is central Florida, which is very flat and full of topographic depressions known as wetlands. These depressions interrupt the stream network continuity in reality, but r.watershed does a fantastic job making a continuous drainage network model, especially in these difficult areas.

Happy Modeling!

Mark

On Tue, Dec 5, 2017, 3:49 PM Shane Carey <[hidden email]> wrote:
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user


_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Shane Carey
I just used the tutorials from this page https://grasswiki.osgeo.org/wiki/Creating_watersheds to extract the stream network - however it could be more accurate so was wondering is r.terraflow a better option for me. I was using r.watershed originally

Do you know why sink filling is needed for r.terraflow and not for r.watershed?

Thanks.

On Wed, Dec 6, 2017 at 11:42 AM, Marco Alicera <[hidden email]> wrote:
How did you add culverts?!
Such a great question and I also wonder how you did. Short ago I knew about Itzï and its ability to do it with SWMM http://itzi.readthedocs.io/en/latest/tutorial.html#culvert-modelling. Looking forward to testing it
--
Marco

2017-12-06 9:24 GMT+01:00 Shane Carey <[hidden email]>:
Hi Mark,

Thanks for your reply! Sounds great.
How did you add culverts or other artificial flow control features to achieve water flowing through roads?

I have a rivers layer and I compared it the streams I've obtained from r.watershed and r.watershed appears to not match these streams (which were accurately digitised) and I was wondering if I had a better resolution DTM would it solve this problem?

Also, why is sink filling needed for terraflow and not watershed?

Thanks for your help :-)

On Máirt 5 Noll 2017 at 23:43, Mark Seibel <[hidden email]> wrote:
Hi Shane.

I'm happy to report that I've modeled overland water flow with r.watershed for over a quarter million acres, consisting of several large project sites, at 1 meter DEM resolution. The data source was LiDAR points to make the DEMs.

At this resolution, it becomes necessary to add culverts, or other artificial flow control features, to achieve water flowing through a road. Otherwise, water is routed along roads until a lowest point is reached for crossing.

I also use r.terraflow outputs as ancillary data to help drop in culvert locations and help provide guidance in problem areas.

My geographic area is central Florida, which is very flat and full of topographic depressions known as wetlands. These depressions interrupt the stream network continuity in reality, but r.watershed does a fantastic job making a continuous drainage network model, especially in these difficult areas.

Happy Modeling!

Mark

On Tue, Dec 5, 2017, 3:49 PM Shane Carey <[hidden email]> wrote:
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user




--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Mark Seibel
In reply to this post by Shane Carey
Hi Shane.


How did you add culverts or other artificial flow control features to achieve water flowing through roads?

Culverts can be added by digitizing the connection points (line) across the road from high accumulation to high accumulation on the other side of the road. It can be a long iterative process, that can be encapsulated in a script. I've wondered if machine learning could be used to automate what seems to be such repetitive and easy culvert connections.

Also, why is sink filling needed for terraflow and not watershed?

I am not into the code, but I know r.watershed use different code to produce different model results. r.watersheed seeks the lowest points in the terrain, thus internal sinks get modeled with exit points rather than internally drained. On the other hand, r.terraflow hydrologically fills the terrain, then models flow accumulation. The results are similar, but the differences can be used to add another check on reality from another model; especially in really flat, depressional terrains.

Mark

 
Thanks for your help :-)

On Máirt 5 Noll 2017 at 23:43, Mark Seibel <[hidden email]> wrote:
Hi Shane.

I'm happy to report that I've modeled overland water flow with r.watershed for over a quarter million acres, consisting of several large project sites, at 1 meter DEM resolution. The data source was LiDAR points to make the DEMs.

At this resolution, it becomes necessary to add culverts, or other artificial flow control features, to achieve water flowing through a road. Otherwise, water is routed along roads until a lowest point is reached for crossing.

I also use r.terraflow outputs as ancillary data to help drop in culvert locations and help provide guidance in problem areas.

My geographic area is central Florida, which is very flat and full of topographic depressions known as wetlands. These depressions interrupt the stream network continuity in reality, but r.watershed does a fantastic job making a continuous drainage network model, especially in these difficult areas.

Happy Modeling!

Mark

On Tue, Dec 5, 2017, 3:49 PM Shane Carey <[hidden email]> wrote:
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Mark Seibel
In reply to this post by Shane Carey
Hi.

I just used the tutorials from this page https://grasswiki.osgeo.org/wiki/Creating_watersheds to extract the stream network - however it could be more accurate so was wondering is r.terraflow a better option for me. I was using r.watershed originally

I've found r.watershed to be the most accurate, for our work. We've field verified the streams and its really solid data. I like that r.watershed doesnt alter the terrain since the LiDAR point data captures terrain nuances so well. 

Mark 


 
Do you know why sink filling is needed for r.terraflow and not for r.watershed?

Thanks.

On Wed, Dec 6, 2017 at 11:42 AM, Marco Alicera <[hidden email]> wrote:
How did you add culverts?!
Such a great question and I also wonder how you did. Short ago I knew about Itzï and its ability to do it with SWMM http://itzi.readthedocs.io/en/latest/tutorial.html#culvert-modelling. Looking forward to testing it
--
Marco

2017-12-06 9:24 GMT+01:00 Shane Carey <[hidden email]>:
Hi Mark,

Thanks for your reply! Sounds great.
How did you add culverts or other artificial flow control features to achieve water flowing through roads?

I have a rivers layer and I compared it the streams I've obtained from r.watershed and r.watershed appears to not match these streams (which were accurately digitised) and I was wondering if I had a better resolution DTM would it solve this problem?

Also, why is sink filling needed for terraflow and not watershed?

Thanks for your help :-)

On Máirt 5 Noll 2017 at 23:43, Mark Seibel <[hidden email]> wrote:
Hi Shane.

I'm happy to report that I've modeled overland water flow with r.watershed for over a quarter million acres, consisting of several large project sites, at 1 meter DEM resolution. The data source was LiDAR points to make the DEMs.

At this resolution, it becomes necessary to add culverts, or other artificial flow control features, to achieve water flowing through a road. Otherwise, water is routed along roads until a lowest point is reached for crossing.

I also use r.terraflow outputs as ancillary data to help drop in culvert locations and help provide guidance in problem areas.

My geographic area is central Florida, which is very flat and full of topographic depressions known as wetlands. These depressions interrupt the stream network continuity in reality, but r.watershed does a fantastic job making a continuous drainage network model, especially in these difficult areas.

Happy Modeling!

Mark

On Tue, Dec 5, 2017, 3:49 PM Shane Carey <[hidden email]> wrote:
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user




--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant


_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
Reply | Threaded
Open this post in threaded view
|

Re: r.watershed

Shane Carey
Yep, I like that also - for me though - you can see the big rivers in the hillshade but it is clearly not picking them out! Any ideas why this is? Is the threshold value very important do you know?

Thanks

On Wed, Dec 6, 2017 at 1:11 PM, Mark Seibel <[hidden email]> wrote:
Hi.

I just used the tutorials from this page https://grasswiki.osgeo.org/wiki/Creating_watersheds to extract the stream network - however it could be more accurate so was wondering is r.terraflow a better option for me. I was using r.watershed originally

I've found r.watershed to be the most accurate, for our work. We've field verified the streams and its really solid data. I like that r.watershed doesnt alter the terrain since the LiDAR point data captures terrain nuances so well. 

Mark 


 
Do you know why sink filling is needed for r.terraflow and not for r.watershed?

Thanks.

On Wed, Dec 6, 2017 at 11:42 AM, Marco Alicera <[hidden email]> wrote:
How did you add culverts?!
Such a great question and I also wonder how you did. Short ago I knew about Itzï and its ability to do it with SWMM http://itzi.readthedocs.io/en/latest/tutorial.html#culvert-modelling. Looking forward to testing it
--
Marco

2017-12-06 9:24 GMT+01:00 Shane Carey <[hidden email]>:
Hi Mark,

Thanks for your reply! Sounds great.
How did you add culverts or other artificial flow control features to achieve water flowing through roads?

I have a rivers layer and I compared it the streams I've obtained from r.watershed and r.watershed appears to not match these streams (which were accurately digitised) and I was wondering if I had a better resolution DTM would it solve this problem?

Also, why is sink filling needed for terraflow and not watershed?

Thanks for your help :-)

On Máirt 5 Noll 2017 at 23:43, Mark Seibel <[hidden email]> wrote:
Hi Shane.

I'm happy to report that I've modeled overland water flow with r.watershed for over a quarter million acres, consisting of several large project sites, at 1 meter DEM resolution. The data source was LiDAR points to make the DEMs.

At this resolution, it becomes necessary to add culverts, or other artificial flow control features, to achieve water flowing through a road. Otherwise, water is routed along roads until a lowest point is reached for crossing.

I also use r.terraflow outputs as ancillary data to help drop in culvert locations and help provide guidance in problem areas.

My geographic area is central Florida, which is very flat and full of topographic depressions known as wetlands. These depressions interrupt the stream network continuity in reality, but r.watershed does a fantastic job making a continuous drainage network model, especially in these difficult areas.

Happy Modeling!

Mark

On Tue, Dec 5, 2017, 3:49 PM Shane Carey <[hidden email]> wrote:
Hi,

I am trying to extract river network from a 5m DEM with some success using r.watershed. Has anyone tested this algorithm on high resolution LiDAR data for example - 1meter DTM and what kind of results have they obtained?

Thanks

--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant
_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user
--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user




--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant




--
Le gach dea ghui,
Shane Carey
GIS and Data Solutions Consultant

_______________________________________________
grass-user mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/grass-user