Inundation Mapping Breaking at Road Crossings (1D Steady)

[Ski907]

I’m relatively new to using RAS Mapper to create inundation maps from 1D steady simulation results on Lidar DEMs, as this terrain data has recently become available in my region.

I’m having an issue on steep reaches at road crossings/bridges where the inundation polygons do not extend over the roadways (when they should) because of a large difference between water surface elevations upstream to downstream. The software seems to more or less linearly interpolate between my upstream and downstream section across the bridge location, intersecting the roadway terrain without resulting in inundation, even though the model results shows the road and bridge being overtopped.

Is there a way to resolve this issue in RAS without manually editing the polygons in ArcGIS to match the model results?

The bridge below is overtopped by about 1.5 feet in the model, but clearly not at all in the resulting inundation map. The solid black lines are my XS. The light black lines are 1 foot contours on the WSE results.

[Lonnie A]

It is as you state due to the linear interpolation between U/S and D/S cross-sections. You might try moving your bounding cross-sections closer to the road which will reduce the elevation difference some.

The other thing you could look at is the continuity of flow. It looks like you have a lot of flow approaching and overtopping the road in the right overbank but on the downstream side the section has most of the flow back in the channel. You might need to adjust ineffective and/or n-values to try to force more flow to be in the right overbank on the downstream side. The difficulty you are having here is that more than likely the right overbank on the downstream side has a higher WSE than in the channel which you can’t model in a 1D model.

[dueckrandy]

This is a problem you will encounter to varying degrees at almost every road crossing you ever model. As mentioned in the previous response, the problem is the linear interpolation of the water surface between the bounding cross sections. The greater the drop in HGL from u/s to d/s, the more pronounced this problem will be. A potential solution to this is to make manual adjustments to the mapped floodplain in GIS. You can do this by copying the bounding cross sections and moving them inward to roughly the top of slope on either side. Then set the WSEL of both of those cross sections to the model results for the u/s bounding cross section. This assumes that the WSEL passing over the roadway is constant, which I believe is a reasonable approximation for mapping purposes in many cases. Then use GIS tools to re-map the area through the crossing and make the required adjustments. If you want to get really detailed and accurately model the flow over a complicated roadway flowpath, you could build a separate model with additional cross sections over the roadway that excludes any culvert or bridge openings and sends the weir flow amount through those sections. Depending on the complexity of the situation, you can build really elaborate models with lateral weirs, separate calcs, rating curves, etc. From your screenshot, it looks like a pretty simply situation. The accuracy of mapping needed at a smaller roadway crossing is typically not very high, since it does not usually affect flood insurance for anyone. Just my 2 cents.

[Anonymous]

Thanks for the replies folks- as I feared there is apparently no secret menu with options to tweak the inundation mapping in these circumstances. I was hoping since my bridge section was georeferenced when created in HECGeo-RAS that there may be some option to extend the WSEL from the upstream bounding XS to at least the centerline of the bridge section before starting linear interpolation to the downstream bounding section.

We are editing the road overtopping in the inundation shape files manually in ArcGIS. This is for a screening level study of infrastructure vulnerability (which includes bridge approach washouts) on a watershed level scale, so we have dozens of these situations to deal with. I like the idea of running a separate overtopping flow model- we’ll look into that, but I have a feeling the manual edits, however tedious, may be quicker based on the number of these we have. We used to do all the inundation mapping manually in ArcGIS with a topomap before Lidar and RASMapper, so I can’t really complain…

[dueckrandy]

Well, I would still complain if I had to do truly manual mapping through dozens of crossings . Since you have many of these to do, let me clarify what I meant by “manual mapping”. There is a quick way to map an entire RAS model in GIS by using exported cross sections with WSEL’s in the attributes. You can make spatial changes, extend cross-sections, add “mapping-only” cross sections, etc. Then with a few quick steps, re-map the entire model onto a topo raster. The basic steps are to produce a water surface tin, convert to raster, then compare it to the topo raster using the Raster Calculator tool to make an inundation grid raster. Then raster to poly.

[Author]

Posts