I'll add that the determination of the gradient boundary as conceived by Arthur A. Stiles was essentially a problem in engineering hydrology. You could choose a flood with a particular frequency of occurence (somewhere between, say, the 6-month and 1-year flood) and call the limits of its floodplain the gradient boundary and not be far off. In the age of drones and LIDAR, that would probably be the most efficient way to map State-owned beds of streams and rivers adjoining or crossing very many tracts. There would be some ground surveying needed to calibrate the model, i.e. demonstrate that the model does in fact predict a floodplain limit that isn't inconsistent with the features observed on the ground, but the bulk of the work would just be aerial mapping and flood modeling.
I don't know where you are in Texas, but I assist in the High Plains Experience Course in the panhandle at Boy's Ranch (Old Tascosa) each year in late spring, and we have a gradient boundary course. Except for the part of how to actually use the Col. Stiles method to determine the qualifying bank, Kent has given you a really good (and much more concise than I could) answer.
Who are you talking to, Monte?
Monte, post: 396503, member: 11913 wrote: Except for the part of how to actually use the Col. Stiles method to determine the qualifying bank, Kent has given you a really good (and much more concise than I could) answer.
One element of the gradient boundary is how low it typically is. Gradient boundary heights of little more than a foot above the surface of the river at normal stage would not be unexpected on the Colorado, for example.
The whole theory of the gradient boundary was developed to deal with broad, flat flood plains, such as that of the bend in the Red River where oil wells in disputed ownership had been drilled along it. In the situation where a river has well-defined banks, errors in location resulting from choosing the wrong gradient height are typically fairly negligible, particularly where the sinuosity of the bank is merely approximated by a series of lines connecting points on the gradient at wide intervals.
To Andy, I was referring that post to the OP.
