Lateral Scattering Tests: 07/08/09

Name of file:lat_scat_test_07809.html
Names of associated files:IGNORE ALL OLD PLOTS AND FIGURES
Created on:7/08/09 (6:45 AM)
Revised on:7/09/09 (4:55 AM), 7/10/09 (1:55 AM)
Structure:
Reference:

Key Reference email(s):

A. Figures from relatively good precision phase velocities and eigenfunctions. This is the snippet of the code

      DO 5000 j = 300000, 740000
C                                                                     C
      C = DBLE( j ) / 100000.00D+00 + 0.00000001D+00

This is 6 significant figures. This takes 10 hours for each MODEL for 0.0005 Hz to 0.1000 Hz ( interaval of 0.0005 Hz). If I used the 8 sig. fig. accuracy that you suggested in your email of "Los Angeles, 3:36 PM Tuesday, June 23, 2009" 


      DO 5000 j = 30000000, 74000000
C
      C = DBLE( j ) / 10000000.00D+00 + 0.00000001D+00

Then this will take more than 20 days for each MODEL to complete. Therefore, we certainly cannot use    CODE_SH_2.for    IF loss-of-precision is the problem.

I combined both the phase velocity and eigenfunction programs, CODE_SH_1.for and CODE_SH_2.for and computed the SCATTERING INTEGRALS in FORTRAN. The results were saved as BINARY OUTPUT, and MATLAB read these integrals and plotted the figures. This time, all of the OUTPUT was in BINARY and not in ASCII.

A.1 Everything is fine and good with EXACT SOLUTIONS. NO PROBLEMS HERE.

A.2.a DIRECT SOLUTIONS. (The figure below is just the same. Only that here, I SHOW THE FULL Y-AXIS RANGE FOR THE TRANSMISSION COEFFICIENTS).

A.2.b DIRECT SOLUTION compared with EXACT SOLUTION.

A.3 Same plots. Upper panel in SEMILOG. Lower panels in NORMAL

A.4 Different plots. Upper panels v_I_i and \barPv_I_i} and lower panels K v_I_i and \bar{K v_I_i }

A.5 Figure A.2.b with the new formula. The coefficients using the DIRECT and MINIMIZED solutions are SAME.

...............................................................................
..............                       2                   ......................
......v_(T)_ii = + --------------------------------------......................
..............     1   +   \bar{K_i}\bar{v_J_i}/K_i v_J_i................ .....
...............................................................................
...............................................................................
...............................................................................
..............     1   -   K_i v_J_i/ \bar{K_i}\bar{v_J_i}.....................
......v_(R)_ii = - ---------------------------------------.....................
..............     1   +   K_i v_J_i/ \bar{K_i}\bar{v_J_i}.....................
...............................................................................

 

 

Hosted by www.Geocities.ws

1