search

lasiadhi / USACE-Bathymetry

0 stars 2 forks source link

Linear regression to extrapolate k #104

Closed gmuldoon closed 8 years ago

gmuldoon commented 8 years ago

Impute missing values of k for the length of the h profile of interest.

gmuldoon commented 8 years ago

Doing simple linear regression:

k = 0.000111 * x + 0.049

Data is linear with R^2 = 0.753. At the boundary condition, k = 0.049 which is a reasonable value. Checking on estimate of k at the boundary point as provided by measurements of T at the boundary. (gTT/2pi = 2pi/k )

Linear regression implies k = 0.049 at the boundary point (x=0). This sounds reasonable. Checking against k value derived from measurements at the boundary point.