Culvert Contributing Area on a Regraded Arterial
A stormwater engineer must estimate the contributing area to a culvert that carries an unnamed tributary under a four-lane arterial. The analyst builds a flow-direction grid from a 1 m LiDAR DEM flown leaf-off in 2019, with 10 cm vertical RMSE and 0.6 m horizontal accuracy, projected in NAD83(2011) State Plane. Three problems appear. On a 40 hectare regraded field, the model produces long parallel flow lines about 30 m apart. Flow accumulation crosses the arterial 210 m west of the surveyed culvert. The arterial was milled and repaved in 2021, lowering the crown by about 15 cm. The developer's engineer argues that because the site drains to the same culvert, the project cannot increase peak discharge, and asks for a single contributing-area number in the permit application. The county requires a defensible number with documented methods.
1. Diagnose the parallel flow lines and the misplaced crossing, and prescribe a DEM conditioning approach with its trade-offs.
Read the worked answer
Parallel lines 30 m apart on a regraded field are the signature of a near-flat surface plus a fill step. Filling sinks creates a flat plateau with no gradient, and the flow algorithm then picks arbitrary, roughly parallel paths across it. The misplaced crossing is a separate defect: flow accumulation is following the pre-2021 road surface, and a 15 cm crown change is enough to reroute which cells spill toward the culvert.
Diagnose before conditioning. Check the raw DEM for the field's real drainage: ditches, tile outlets, swales. Look at a hillshade. Compare the 2019 surface against current survey shots of the road crown and the culvert inverts.
Then condition with intent. Breaching or cutting along a known drainage line is generally preferred to blanket filling because it preserves gradient and follows observed infrastructure. Filling the entire field manufactures a flat impoundment and relocates the divide. For the arterial, burn a culvert line from surveyed upstream invert to downstream invert at the correct plan position, and rebuild or lower the road crown to current survey.
Recompute after each individual edit and log the resulting contributing area, so you can state which change moved the number and by how much.
2. Explain how to represent the culvert and how to validate the result, including drainage that the DEM cannot see.
Read the worked answer
A culvert is a hydraulic connection the bare-earth DEM cannot see, because the road embankment reads as a wall in the raster. Filling alone will dam flow at the road, spill it along the embankment, or send it down the wrong side. Burn a continuous line at the surveyed inverts with the correct pipe diameter and slope, or hydro-enforce by cutting the embankment where the pipe passes, then confirm the derived divide against field evidence.
More important in a suburban watershed: a DEM knows nothing about storm sewer inlets and pipes, which routinely move water across a surface divide. Obtain the storm sewer atlas or walk the site, and decide explicitly which inlets bypass the topographic watershed.
Validate with at least one independent check: a surveyed drainage area at the culvert, a field walk of the divide, or comparison with a calibrated HEC-HMS or HEC-RAS reach. If the model and the independent check disagree by more than roughly 10 percent, resolve the discrepancy before writing any number into a permit application. A convincing-looking raster is not validation.
3. The engineer wants one contributing-area number. What do you deliver, and how do you respond to the 'same outlet, same drainage' argument?
Read the worked answer
Deliver a range with a driver analysis, not a single value. Run a sensitivity matrix: the 1 m conditioned DEM, 3 m and 5 m resamples, with and without culvert enforcement, with and without storm sewer routing, and with the elevation bias applied. Report something like '18.4 to 21.9 hectares, driven mainly by whether the 2.3 hectare inlet on the west ditch routes through the storm sewer,' and name the assumption you consider most likely. That is defensible; '20 hectares' is not.
Push back on the hydrologic claim as well. Contributing area is necessary but not sufficient. Peak discharge depends on time of concentration, curve number, and storage. A project can leave the outlet unchanged and still raise peak flow by adding impervious cover or shortening travel time, and can lower it with detention. Identical drainage area to the same culvert does not establish no impact.
Ask the engineer for pre- and post-development hydrology and record the timing assumptions in your report. Note the 2021 repaving in the methods section as a known surface change that postdates the DEM.