I have read that parts of the levee are 100% earthworks, while other parts have lower earthworks with a concrete wall at the peak.
Given the resolution of the photos, would a much thinner concrete wall show up, or would it disappear because it's so much thinner than earthworks?
Do the areas with concrete tops correspond to the missing areas on the map?
The concrete wall levees do show up in the sat imagery. Along the south side of east Orleans Parish you can see one, and out of the frame to the left, there is another that is visible from alongside the Industrial Canal.
Even though the concrete wall is smaller than a single ten meter pixel in the DEM dataset, it does affect the pixel's heigth value, representing an average of the total area covered by the 10 meter square. This is an effect I've seen before in Landsat visible imagery. 4 meter wide roads and trails are clearly visible in the Landsat 7 datasets which only have a resolution of 28.5 meters per pixel. Even though the road is not wide enough to color an entire pixel, its color is enough to change the average value of that pixel from those immediately adjacent to it on either side.
Because it is a linear feature, the pattern is clear, where a single 4 meter by 4 meter feature would more likely be dismissed as an aberration.
First of all, they aren't photos, but rather radar images from the Shuttle Imaging Radar.
However the resolution question is still valid. It would depend on how the low resolution images were generated. If they were generated from the same radar returns, but with lower resolution processing, then the thin (<< 10 meters width) probably would not show up. However if the lower resolution images were generated from the classified higher resolution images, then it it would depend on the nature of that resolution reduction processing. If the took the highest elevation in the lower resolution (10 m) , that is larger, cell made up of 100 of the higher resolution (1 m) cells, then you would see the height of the small walls. Assuming of course that they can be seen in the 1 m resolution images. OTOH, if the averaged, or other wise smoothed, the heights from the higher resolution imagery, the effect would be similar to the lower resolution processing of the radar signals postulated above. (Similar, not the same!)
I note that in the center image, calibrated to 6 feet above sea level, the southern portion of the industrial canal levee/wall is not visible, lending some credence to the theory that the walls are not visible, since we know from ground images that it's a concrete wall at least the upper portion is.
However all those seawalls and levees that are visible in the 6 foot image, should also be visible in the last image, if they are high enough. But It's not clear without a good topo map, whether the levees/walls marked as invisible in the 8 foot image are actually visible in the 6 foot image, or are we instead seeing land or levees that are more than 6 but less than 8 feet ASL, but with concrete walls not visible by themselves.