New volume calculations for Lower Knowles

I've got the new volume calculations 'mostly working' now. There's a new program to estimate valley walls from an initial centerline you enter with the mouse, and it iteratively calculates boundaries and a new centerline by looking at the slopes of profiles perpendicular to the centerline. I'll have more to say about this progam in the next post, as well as step-by-step instructions for going from the overall area DEM to the individual stream volume calculations.

New valley boundaries for Lower Knowles look like the following:

(Click for larger image)

Note that:

• The boundaries are much more convoluted than the previous estimates.
• The bedrock points still fall very close to the calculated thalweg.
• Although some tributaries (e.g. Bear Creek and Skate Creek) enter the valley at about the same elevation and slope, others enter at a higher slope and elevation, forming terraces which decrease the valley width.

One significant question which arises is whether these are fill terraces or strath terraces --that is, whether they represent a flood of sediment into the valley or whether they are supported on bedrock which is above the valley floor. The salient feature for this analysis is that their edges fall off at a steeper slope, and so they are not counted as part of the valley. You should be aware of this, as it may represent errors in the volume calculations (or not).

Here is a down-valley plot of the elevations:

(Click for larger image)

Below is a cross-valley plot of the descending profiles, including the bedrock base and wall slopes which were extrapolated from 67% of the maximum slope within 100 m left and right of center. Thereafter, the volume is calculated as bounded by the surface DEM at top, by the calculated bedrock base steps at bottom, and by the sloped walls at each side (separate slopes for left and right). Although this is still a crude approximation, it should help to decrease the overestimation due to the spurious thicknesses at the valley edges.

(Click for larger image)

Here is the output from the volume calculations, showing widths, thicknesses, areas, and volumes for every cell of the valley (among many other things). Here is a plot of the volume wrt valley-length coordinate, and a plot of thickness vs. width:

Here is a DEM plot of the thickness at every point in the valley (actually, some points on the boundaries of cells are omitted because the routine for determining whether a point is within a polygon excludes points on the boundary):

(Click for larger image)

Here is the same plot at a reduced z scale:

(Click for larger image)

Another question is whether these new results are appear reasonable or useful. They seem more reasonable to me than the previous estimates, which were:

Total volume:     5.591071e+06
Average width:        82.787177
Average thickness:         4.834486

         Total volume:     6.895107e+05
    Average thickness:         2.843431
Average surface width:        76.401037
   Average base width:        56.707523
       Average length:        10.002001

V' = width * length * thickness
   = (76.401037 + 56.707523) / 2.0 * (2995.596833 - 15.009817) * 2.843431
   = 564053.7471224491