Estimate crest height when including flood defences

[rosepearson]

• Take perpendicular samples find the high point, align to this and extract the high point. introduce as a "self aligning"
• Support adding a patch with only a polygon and no elevations - remove values where polygon exists and interpolate removed values.

[JoePelmard]

Hi Rose,

As we discussed a few months ago, please find attached the script I created to correct stopbanks crest height of a coarse DEM using the corresponding finer DEM for the reference values and the NZIS shapefile to identify the location of the stopbanks.

The script works as for each of the line of the NZIS shapefile:

1. Discretize the line at a given resolution (say 4m = half the fine DEM's resolution)
2. For each discretized point, correct the location of highest elevation by identifying in the fine DEM the max elevation along a buffer segment perpendicular to the line and centered on it, and extract the corresponding elevation at this location. The buffered segment's length and discretization are parameters.
3. Correct the coarse DEM's cells crossed by the stopbank line by averaging the corresponding crest height of the line points included in the cell.

Some students are working on updating the NZIS database to correct some misalignment of the stopbanks line and to fill gaps. They have recently started working on updating the database by adding the crest height extracted from LiDAR data as an additional attribute (which is partly what we are aiming to do here). I will keep you posted once they've progressed a bit more.

dem_stopbanks_conditioning_lib.py.txt

[rosepearson]

Nice thanks @JoePelmard - I've implemented an initial stopbank crest preserver but will implement a more robust approach similar the your approach in the next few days.

That is great to hear you are working with some students to update the NZIS. Will you also be updating the NZIS polyline alignments? I think this would be very helpful.

Finally, OSM has ways with the tag 'man_made' and tag IDs of 'embankment' and 'dyke'. This could be another place to look to bring in community sourced stopbank information.

[JoePelmard]

Hi Rose,

Sweet, happy to hear things are moving forward.

Yes, the first task of the student was to adjust the NZIS polyline alignments to fit with the actual locations as found in DEMs and fill up gaps. There will still be some deficiencies, but it should be a meaningful upgrade as compared to the current version. I will come back to you after talking to the professor in charge to share that version with you.

Same with OSM, I'll add this to the topic.

Joe