Trouble Getting Flow Direction on Fairly Flat Land

6178
5
01-24-2013 10:00 AM
Labels (1)
WilliamWoodard
New Contributor
I'm trying to use ArcHydro to find watersheds and flow patterns in an urban area within a river's floodplain.  Within the district scale site there is a large, flat area where water collects (this urban district is surrounded by levees, so water flows and collects in this specific area).  This is my specific area of study. 

When I use the Fill Sink tool, it makes the entire area one color as if there is no elevation change.  When I use the Flow Direction tool on the incorrect Fill Sink raster, the Flow Direction output has the proper color directions (1,2,4,8,16,32,64,128) but still shows the area as completely flat.

If I run the Fill Sink tool on the DEM again, but this time with a Threshold value of 1, the output Fill Sink shows true elevation differences within the area.  However, when I use the correct-looking Fill Sink raster in a Flow Direction tool, a correct-LOOKING Flow Direction raster appears but with values 1-255.  I cannot move forward with this correct-looking data with the wrong values. 

Does anyone know how to get to Flow Direction to output the correct elevation differences?
Tags (2)
0 Kudos
5 Replies
ChristopherBlinn1
Occasional Contributor III
However, when I use the correct-looking Fill Sink raster in a Flow Direction tool, a correct-LOOKING Flow Direction raster appears but with values 1-255.


If you are getting values from 1-255 then the tool is working correctly.

If you have values outside of the domain 1,2,4,8,16,32,64,128 then there is an equal ability to flow in multiple directions.  The result will be the sum of the directions.  Ex. Flow can occur at 16 and 2, so the value is 18.

Set your symbology to unique values on your raster.  That will help visualize the flow direction as well.
0 Kudos
WilliamWoodard
New Contributor
cblinnmrc;266725 wrote:
If you are getting values from 1-255 then the tool is working correctly.


The Flow Direction raster with the values 1-255 does no move through the process to create watersheds.   When this raster is used in the Flow Accumulation tool, flow lines do not get created; instead it is a lot of little lines.  Further, when that is put into the stream definition tool, a raster output is created that is completely white with no apparent data seen.

Am I possible not using the Flow Direction raster in the Flow Accumulation tool correctly?
0 Kudos
MarkBoucher
Occasional Contributor III
Within the district scale site there is a large, flat area where water collects (this urban district is surrounded by levees, so water flows and collects in this specific area). This is my specific area of study.


Flat areas don't model well. Hilly areas do.

If I were you I would determine where the water will "leave" the leveed areas (where a pump or gate will be or currently exist). I would then create a an agreestream layer and "breach" the levee at this location with a short line. This is likely the low point behind the levee. Then, I would copy the storm drain or channel system into the agreestream layer. You wouldn't need to delineate well defined drainages in the DEM, but may need to "breach" bridges or culverts at road crossings. Then, make sure the DEM is well defined to its edge and add lines in the agreestream layers where the DEM poorly represents reality.

To make the resulting fil raster "hydrologically correct" you may need to put in an "innerwalls" layer. Use this to input polylines where you know there are subwatershed boundaries in the DEM. These could be RR tracks, property lines that you know overland flow would not pass.

If you do the above, when you process the DEM by burning the streams, building walls, and filling the sinks, you will get a hydrologically correct fil raster that will result in a flow direction (fdr) grid that is meaningful for the rest of your analysis. You will want to use the levee "breach" locations as hydrology points.

PS - I've found that if the agreestream and innnerwalls layers are in a gdb, the number of errors are reduced.
0 Kudos
MariaBravo
New Contributor
You may want to use an alternative to the Fill Sinks operation in ArcHydro that generates more realistic flowpaths in flat regions.
This is "Fill Sinks Plus". This is a Customized Command or button that can be installed in ArcMap 10.1 by using the ArcObjects.NET as distributed by ESRI. The installation of this button is detailed in http://www.lago-consulting.com/fill_sinks_plus.html. I hope this help.


I'm trying to use ArcHydro to find watersheds and flow patterns in an urban area within a river's floodplain. Within the district scale site there is a large, flat area where water collects (this urban district is surrounded by levees, so water flows and collects in this specific area). This is my specific area of study.  

When I use the Fill Sink tool, it makes the entire area one color as if there is no elevation change. When I use the Flow Direction tool on the incorrect Fill Sink raster, the Flow Direction output has the proper color directions (1,2,4,8,16,32,64,128) but still shows the area as completely flat. 

If I run the Fill Sink tool on the DEM again, but this time with a Threshold value of 1, the output Fill Sink shows true elevation differences within the area. However, when I use the correct-looking Fill Sink raster in a Flow Direction tool, a correct-LOOKING Flow Direction raster appears but with values 1-255. I cannot move forward with this correct-looking data with the wrong values.  

Does anyone know how to get to Flow Direction to output the correct elevation differences?
MarceloLago
New Contributor III

As you mentioned that your area is flat, I would recommend you to use "Fill Sink Plus" (Fill_Sinks_Plus - LAGO   Consulting   &   Services LLC) instead of the Fill Sink Tool. The latest version of "Fill Sink Plus" has been upgraded and now the user can enter existing culverts and drains as shape files, which may be useful in urban areas.

0 Kudos