Difference between revisions of "CVM Improvement"
(Created page with 'There is interest in whether CVM-S and CVM-H models are converging through full 3D tomography work of Chen and Tape. To look into this issue, we propose a plan to compare the mo…') |
|||
| (7 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
| − | + | To investigate whether CVM-S and CVM-H models are converging through the full 3D tomography [[F3DT]] work of Chen and Tape, we will compare a southern California region down to 50km depth (see kml file below for exact coordinates) using both models, because both models provide Vp, Vs, and density (rho) for this region. At 1km spacing, with the area approximately 200km x 300km x 50km = 3,000,000 mesh points. | |
| − | + | We will extract low resolutions (1km) meshes using CVM-S4 and CVM-SI9. We will use query by depth, using the UCVM software to remove topography from CVM-H | |
| − | We will | + | We will create difference meshes for the Vs parameter, then plot slices and possibly volumes. We will do this only if tomography results update all three of these parameters. |
| − | We | + | The CVM-H release notes say that version 5.5 (Sept 2008) was the last version of CVM-H without Carl Tape inversion results. It also says that CVM-H 6.0 was the first with Tape inversion results. We propose to use these two versions (if we can find a distribution of them) of CVM-H as the starting and improved models. This assumes we only want to examine inversion improvements. CVM-H has also recently added a basin based on a Graves paper. This is a CVM-H improvement, but not based on tomographic results. |
| − | + | We also make a difference meshes between CVM-S4 and CVM-H 5.5 starting models and CVM-SI9 and CVM-H 6.0 improved models. If the overall difference between CVM-S and CVM-H is higher for original models than for the improved models, this provides evidence that the models are converging. | |
| − | |||
| − | |||
| − | |||
| − | We also make a difference meshes between CVM-S4 and CVM-H 5.5 starting models and CVM-SI9 and CVM-H 6.0 improved models. If the overall difference between CVM-S and CVM-H is | ||
== UCVM Regions KML File == | == UCVM Regions KML File == | ||
| − | + | To compare the CVM-S and CVM-H, we will use the CVM-S coverage region as our comparison region because both models define material properties for most of this region. Here are coordinates for the comparison region in kml and lat,lon format. | |
| + | * CVM comparison region in kml format: [http://hypocenter.usc.edu/research/ucvm/UCVM_Coverage_Region.kml UCVM Regions (kml)] | ||
| + | * CVM comparison region, based on corners of CVM-S (not considering 1D Hadley Kanamori Background model) | ||
| + | ** 31.00,-116.75 | ||
| + | ** 35.00,-121.00 | ||
| + | ** 37.50,-118.25 | ||
| + | ** 33.50,-113.75 | ||
== Related Entries == | == Related Entries == | ||
*[[CVM-H]] | *[[CVM-H]] | ||
| − | *[[CVM- | + | *[[CVM-S]] |
*[[Community Velocity Model]] | *[[Community Velocity Model]] | ||
*[[F3DT]] | *[[F3DT]] | ||
Latest revision as of 20:56, 3 January 2012
To investigate whether CVM-S and CVM-H models are converging through the full 3D tomography F3DT work of Chen and Tape, we will compare a southern California region down to 50km depth (see kml file below for exact coordinates) using both models, because both models provide Vp, Vs, and density (rho) for this region. At 1km spacing, with the area approximately 200km x 300km x 50km = 3,000,000 mesh points.
We will extract low resolutions (1km) meshes using CVM-S4 and CVM-SI9. We will use query by depth, using the UCVM software to remove topography from CVM-H
We will create difference meshes for the Vs parameter, then plot slices and possibly volumes. We will do this only if tomography results update all three of these parameters.
The CVM-H release notes say that version 5.5 (Sept 2008) was the last version of CVM-H without Carl Tape inversion results. It also says that CVM-H 6.0 was the first with Tape inversion results. We propose to use these two versions (if we can find a distribution of them) of CVM-H as the starting and improved models. This assumes we only want to examine inversion improvements. CVM-H has also recently added a basin based on a Graves paper. This is a CVM-H improvement, but not based on tomographic results.
We also make a difference meshes between CVM-S4 and CVM-H 5.5 starting models and CVM-SI9 and CVM-H 6.0 improved models. If the overall difference between CVM-S and CVM-H is higher for original models than for the improved models, this provides evidence that the models are converging.
UCVM Regions KML File
To compare the CVM-S and CVM-H, we will use the CVM-S coverage region as our comparison region because both models define material properties for most of this region. Here are coordinates for the comparison region in kml and lat,lon format.
- CVM comparison region in kml format: UCVM Regions (kml)
- CVM comparison region, based on corners of CVM-S (not considering 1D Hadley Kanamori Background model)
- 31.00,-116.75
- 35.00,-121.00
- 37.50,-118.25
- 33.50,-113.75
