LOTOS, code for non-linear local earthquake tomographic inversion

created by Ivan Koulakov (IPGG, SB RAS, Novosibirsk, Russia)

e-mail: ivan2art@gmail.com


A tomographic algorithm, LOTOS-07 (Local Tomography Software) is designed for simultaneous inversion for P and S velocity structures and source coordinates. The LOTOS-07 algorithm can be directly applied to very different data sets without complicated tuning of parameters. It has a quite wide range of possibilities for performing different test and is quite easy to operate. LOTOS-07 code can be freely provided to any interested person by Ivan Koulakov

Detailed description of the LOTOS-07 algorithm can be downloaded as:

Microsoft Word document: LOTOS-07_description.doc (1.478 Kb)

or as a PDF document: LOTOS-07_description.pdf (1.449 Kb)




EXE version of the LOTOS-07 code. Instructions for installing and running

The EXE version of the code is created to perform the the full tomographic inversion in a most simple way.

In order to perform and execute the tomographic inversion using the EXE version of the code, the following steps should be performed:

1. Download a Packed file with the entire code: EXE_version.zip (5 071 Kb)

2. Copy this file to any location of your computer and unpack.

3. In the present form, the code is ready for performing the inversion of the Costa-Rica data set described in the BENCHMARK . In this test the synthetic anomalies composed the face of Simon Bolivar. The initial data for this test are located in "EXE_version\DATA\BOLIVAR_\inidata\" and contain two files "stat_ft.dat" and "rays.dat".

4. The reconstructed model will be produced in a folder "EXE_version\DATA\BOLIVAR_\MODEL_01\". All the parameters for inversion, grid construction, source location, visualization and other are located in a folder "EXE_version\DATA\BOLIVAR_\MODEL_01\ini_param".

5. Start execution of the program by clicking on START_LOTOS.BAT. The Costa-Rica data set includes more than 1000 events and more than 30000 picks. The entire inversion procedure includes preliminary source location (initial coordinates and origin times are presumed to be unknown), 1D velocity optimization (starting 1D model is quite different of "true" one) and five iterations of non-linear tomographic inversion. The calculation time varies depending on computer. In my 4-years-old laptop, the entire processing takes about 5 hours.

6. After each iteration, the results can be visualized. The code produces the files with P and S velocity anomalies in a horizontal section at the depth of 15 km which are located at "EXE_version\FIG_files\hor\", files dv1_ 1.grd (for the P model) and dv2_ 1.grd (for the S model). In each iteration these files are overwritten with new results. These files can be directly visualized in SURFER. One can use an example of SURFER file in bolivar.zip (1.201 Kb) " as a template. In this case, the GRD files in a resulting image should be replaced with a newly computed GRD files. If everything is correct, the inversion results should be absolutely the same as presented in the BENCHMARK description.

7. The EXE version of the code can be used for processing of any other datasets. In particular, one can perform the inversion for the other BENCHMARK datasets. The structure of data organisation, which includes the initial datasets (stat_ft.dat and rays.dat) and initial parameters, are given in zipped files:

TURKEY__.ZIP (244 Kb)
SUMATRA_.ZIP (45 Kb)
CHILE_21.ZIP (27 Kb)

In order to start the inversion for any of these datasets, one should unzip these files and copy them to "EXE_version\DATA\" folder. Then in the file "EXE_version\all_models.dat" indicate the name of the dataset and the model. For example, for "TURKEY__" dataset and "MODEL_01", this file would be as follow:

1: code of the region (any 8 characters)
2: number of iterations
3: 0 all data, 1 odd events, 2 even events
4: kode for ref. mod. optimization. 1-yes
**********************************************
TURKEY__ MODEL_01 5 0 1

Then execution of the program can be started by clicking on START_LOTOS.BAT.

The same algorithm can be used for any real or synthetic dataset.



The sources of the LOTOS-07 code:

The sources of the LOTOS-07 code with examples of datasets can be downloaded as: LOTOS_07.ZIP (25 Mb) . The sources have been cteated using FORTRAN-90 in the Compaq Visual Fortran software (version 6.6B). The program sources are stored together with the entire projects which can be opened and executed in any compatible environment.

Examples of pictures obtained for the presented datasets using the GOLDEN SURFER software can be downloaded as: PICTURES.ZIP (33.6 Mb)

In case of using any part of the LOTOS code, please refer to the following paper:

Koulakov, I., M. Bohm, G. Asch, B.-G. Luehr, A.Manzanares, K.S. Brotopuspito, Pak Fauzi, M. A. Purbawinata, N.T. Puspito, A. Ratdomopurbo, H.Kopp, W. Rabbel, E.Shevkunova (2007), P and S velocity structure of the crust and the upper mantle beneath central Java from local tomography inversion, J. Geophys. Res., 112, B08310, doi:10.1029/2006JB004712.

In case of any unconsistencies, errors, unclear issues and other problems, please do not hesitate to contact the author, Ivan Koulakov,
e-mail: ivan2art@gmail.com