Topo to Raster Interpolation not 'hydrologically correct'

[ChrisParker]

Hi,

I am trying to create a hydrologically correct DEM based upon UK Ordnance Survey Profile contours (polyline feature class). When I have tried to use the 'Topo to raster' tool to do this (Spatrial Analyst Tools - Interpolation - Topo to Raster) the DEM that is interpolated is not hydrologically correct as ArcGIS says it should be (http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=Hydrologically_correct_surfaces%3A_Top...). I can tell because:
- When I turn on the optional 'Output stream polyline feature' it does not represent the stream network properly
- When I apply a flow direction and flow accumulation analysis to the interpolated DEM it does not represent the stream network properly
In both cases, the stream network consists of lots of very short lengths of channel which are are separate from each other (rather than one continuous drainage network). My problem seems to be similar to that reported here:
http://gis.stackexchange.com/questions/9237/methodology-for-creating-accurate-drainage-networks-and-...

I can seem to improve how the interpolation algorithm works by:
- Making the cell size of the interpolated DEM larger
- Making the optional tolerance values in the tool dialogue larger (this allows it to try and fill larger sinks)
However, neither of these fully resolves the problem.

It is possible to subsequently modify the interpolated DEM so that it becomes hydrologically correct (using the 'fill' analysis) but this results in some completely flat areas in the DEM (which is unhelpful for the analysis that I wish to carry out on the interpolated DEM). I am hoping that if I can get the 'Topo to raster' tool to work properly it will be a little more refined in its representation.

So my question is - has anyone had experience with this issue and managed to find a way to correct it?

Thanks,

Chris

[SteveLynch]

Chris

What version of ArcGIS are you using?

Are you only using contour input?

What is the average spacing between contour lines?

What cell size are you using?

Have you tried using stream/rivers as input?

How many iterations are you doing?

Are your data in a projected coordinate system?

Thanks
Steve