(83) |
(84) |
(85) |
Instead of representing kx2+ky2 in the most obvious way, let us represent it in a manner consistant with the finite-difference way we expressed the data . Recall that which is a Fourier domain way of saying that difference equations tend to differential equations at low frequencies. Likewise a symmetric second time derivative has a finite-difference representation proportional to (-2+Z+1/Z) and in a two-dimensional space, a finite-difference representation of the Laplacian operator is proportional to (-4+X+1/X+Y+1/Y) where and .
Fourier solutions have their own peculiarities (periodic boundary conditions) which are not always appropriate in practice, but having these solutions available is often a nice place to start from when solving a problem that cannot be solved in Fourier space. For example, suppose we feel some data values are bad and we would like to throw out the regression equations involving the bad data points. We could define a weighting matrix starting from an identity matrix and replacing some of the ones by zeros. This defines .Now our regression (83) becomes
(86) |
With the Vesuvius data how might we construct the weight ? We have available the signal strength (which we have not used). We could let the weight be proportional to signal strength. We also have available the curl, which should vanish. Its non-vanishing is an indicator of questionable data which could be weighted down relative to other data.
The laboratory exercise is new this year so it may contain some unexpected difficulties. We're not sure it leads to clear solutions either. Anyway, you are given the Vesuvius data and all the programs in the book. Additionally, you are given a Fourier solver that produces the analytic solution. Please inspect both the Fourier solver and the solution it gets. Go to the web to see what pictures you can find of Vesuvius. Notice the radial drainage patterns on the amplitude of the original complex numbers. It is a little disturbing that we don't see these drainage patterns on the phase data (or maybe you can see them a little?). Any thoughts you have on that issue are certainly welcome. Any other thoughts you have on this lab are certainly welcome. This data is fun so we'd like to get this lab better focused for next year.