3-D inversions and multiple-ridge fitting - PROGRESS REPORT

1
3-D inversions and multiple-ridge fitting -
PROGRESS REPORT

Deborah Haber, Brad Hindman, Mike Thompson
2
3-D inversions

• Code is running under IDL and has ability to
  use different size tiles
• Kernels are from model with polytropic
  atmosphere, and are scalable

3
early 3D inversion
30
From 16o tiles n 1, 2
and
2o tiles n 1
-30
-30
30
4
Inversion in x direction
5
Future Work Short Term

• Add ability to take errors into account
• Determine kernels from better model
• Port code to Fortran

6
Multiple Ridge Fits- Motivation

• Tiles of smaller sizes or at lower frequencies
  generally have a large amount of overlap in the
  peaks leading to poor determinations of the
  amplitude, widths, and frequencies of the waves.
• By fitting all the n-value peaks for a given k
  we are able to get more consistent results that
  allow us to fit for more n values than was
  previously possible.

7
Multi-ridge fit 16 degrees
8
Future Work

• Finish widget program for getting better
  starting parameters from azimuthally averaged
  power spectra.
• Update Fortran program to carry out multiple
  ridge fits on both azimuthally averaged data and
  for full 3d fits.
• Work on better characterizing background and what
  tolerances to allow for deciding on fit.