Wednesday, December 10, 2008

Little Osage Creek Revisited

On November 12, 2008, I blogged about Little Osage Creek and the flood study I'm working on for that. After having the successful computer model of the stream geometry, and a first calculation of the height and spread of the flood waters, I have been working on 1) improving the model, and figuring out how to 2) properly map the spread of the flood and 3) reduce the height and spread of the flood waters by doing minor improvements to the stream channel.

This has taken considerable time, much more than I anticipated. Even though the model I ran a month ago ran successfully, the more I looked at it, the more I realized I had small errors in it: a culvert partially blocked by the ground data in the model that was clear in real life; a road over the culvert improperly entered; a missing cross-section; a feature of the program called ineffective flow areas not entered for some sections where it was needed, or entered at the wrong place or wrong elevation. All of these took some time to sort through and correct.

The mapping turned out to be a bigger problem, however. Just last Thursday, a CADD tech was finally assigned to me for this. I would have preferred a junior engineer, but this tech is a good man, intelligent and eager to learn, a hard worker. But, as I explained to him how to take the flood elevation and the topography and pinpoint the horizontal extents of the flood zone, I had a lot of explaining to do. He had lots of problems with it. Monday afternoon we spent over an hour trying to work on it, and had a lot of trouble with one particular area. We broke off our discussion just after 5:00 PM. It seemed that the report I had to give to the Centerton City Council the next evening was in jeopardy.

Right after that, in the quiet at my desk, I realized exactly what I had to do to correctly do the mapping. Our problems came from the West Branch running at right angles to the Main Branch. I had accounted for this by the ineffective flow areas feature. However, all the cross-sections therefore could not be extended far enough to where the flood waters hit the ground. In the ten quiet minutes at my desk, I realized I had to bend the cross-sections at the inside corner of where those two branches met at right angles. Yesterday morning I showed him my mark-up with the bent cross-sections, and showed him how to map the extent of the flood. By afternoon he had it done; we put 25 packets together, and I went off to the 7:00 PM meeting to make my report.

The mapping was temporary, however, completed only for this oral report. I now have to modify the model with these bent cross-sections. This means new geometry for part of fourteen cross-sections. This means having the tech cut those sections from the CADD program and labeling the key points. This means back to data entry on the computer model, probably two to three days of work, then re-running the model, then seeing if the temporary mapping we completed yesterday is still valid.

I should have recognized right at the start that these cross-sections should be bent. That's what happens when you lay your tools down for a few years, and try to pick them up again. At CEI I am the senior engineer who is not in upper management. My job is to train all the youngin's, and to help them solve their more difficult problems. But instructing someone in how to do something is much different than actually doing it.

This has been a good exercise for me. My tools are sharper; I'm better at using them than I ever was before; and I've enjoyed it (sort of).


Anonymous said...

Ah, GIGO strikes again.


David A. Todd said...

Well, not quite garbage either in or out. More like: I couldn't really tell where the water was going till I ran a preliminary model, after which I would need to revise the preliminary data to keep it from becoming garbage. That's what I failed to do in a timely way: recognize the need to revise and re-run; hence it became garbage.