pnnl / e4d Goto Github PK

Need to enable the 'average' option for starting values in the velocity tomography code.

Sensitivity (i.e diag(JTJ) files are output when an inversion completes in the sigma file format. That file is missing the number of columns on the second line of the first row, so that bx and px crash when you try to load them.

When using 'average' in e4d.inp, a value of Inf is reported. The inversion then proceeds with an Inf value and the inversion produces nonsense.
Associated files to recreate issue: e4d.inp, 100H20200415_1100.srv, 100H_2.cfg,100H.out

Need to enable output of a single row of the Jacobian matrix for travel time tomography so that the Fresnel volume for a particular measurements can be visualized.

Check output of JTJ for velocity tomography. Verify it's working ... seems to be a problem.

Hi,thanks for sharing this libaray.

I am intrested in the seismic tomography.

Can this package do some 2D tomograph inversion like cross-hole/refract/reflect inversion?

Good evening,
In the compilation instructions for bx, the link for downloading exodusii doesn't work (it redirects to the sourceforge homepage).

I have tried installing exodus as part of seacas (https://github.com/sandialabs/seacas) but either I need to specify different options in the makefile for that to work, or it's not the right version, because when I try to make bx I get:

gfortran subroutines.o exo_subs.o  build_exo.o -I/opt/ssi/bin/seacas/TPL/netcdf/netcdf-c/include -I/opt/ssi/bin/seacas/include -lm /opt/ssi/bin/seacas/lib/libexoIIv2for.a /opt/ssi/bin/seacas/lib/libexodus.a -L/opt/ssi/bin/seacas/TPL/netcdf/netcdf-c/lib -lnetcdf -o bx  
/usr/bin/ld: /opt/ssi/bin/seacas/lib/libexodus.a(ex_utils.c.o): undefined reference to symbol 'H5get_libversion'
/usr/bin/ld: /usr/lib/libhdf5.so.200: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status
make: *** [Makefile:13: bx] Error 1

please advise as to whether a new link is available or if there's other options I need to incldue.
Thank you!

need to put the imaginary component in the sip file created in tank mode

implement anisotropic/directional smoothing in a user specified direction (i.e. to help with 3D inversion of 2D data)

Need to implement file comparisons for night QA checks.

Fix rms to be based in percent error instead of sqrt(chi^2)