Long Profiles: Only Partial Results Displayed

[amrelsharkawy]

Description

When creating a north-south profile between 40°N-66°N along 10.5°E using global model data (subset: lon 5°-15°, lat 35°-70°), only a small portion of the profile displays results.

Expected Behavior

Full profile from 40°N to 66°N should display model data.

Current Behavior

• Only shows data between [observed range, e.g., 40°N-50°N]
• No error messages during script execution
• Data exists in specified region according to model subset can be downloaded from (https://cloud.rz.uni-kiel.de/index.php/s/4XWTsTmRm5KPSBM)

Profiles_EU.txt

Datasets_latest_AE_swath.txt

Julia Code used can be downloaded from: https://cloud.rz.uni-kiel.de/index.php/s/5LokRk3JKHpH85f

[boriskaus]

thanks for reporting!
Can you perhaps add some figures in your github issue so we see more clearly what is happening (and other users see this too)?

[amrelsharkawy]

Dear Boris,

Please find attached a plot comparing the results of three global tomographic models (Li 2008, Auer 2014, and Thrastarson 2024) alongside two regional models along the specified profile. Additionally, the plot includes a north-south receiver function section along the same profile, shown in the lower right panel.

The figure indicates that although the global models are expected to cover the entire profile, the tomography results appear truncated. I am unsure of the reason for this discrepancy and would appreciate your insights.

[mthielma]

I have run the files you provided. Currently, only a profile through the Auer model is taken (is that intended?). If I look at the resulting Profile1_VolData.vts in Paraview, there is no truncation (see image). Can you confirm this?

[mthielma]

I have now also combined this dataset with the Giacomuzzi dataset, topography data, the EUCrust_08 Moho data and the isc24_MedS dataset. Then I also added a depth range for the volume data from 660km depth to the surface. None of these things resulted in a truncation of the data in terms of latitude. I also plotted the contents of the jld2 file and I do not see any truncation. Given that this all works, I see two possibilities: 1) you potentially use an older version of GMG where this issue may have occurred or 2) your further processing and plotting steps result in this truncation.

[amrelsharkawy]

Hi Marcel,
thank you really very much. I have doubled checked all data sets and the plotting routines. They look fine. I also used the latest version of the GMG.

Your tests were indeed very helpful. But you probably have used only one tomographic model that covers the full profile.

When using only global models (or regional models that covers the profile), I could also slice the data along such long profiles without any truncations. See the plot below:

However, by using a combination of different tomographic models (vol_data: e.g. small-scale, regional scale, and global scale ones), the issue persists again. Please see the plot below for the same profiles. I had no problem with Moho or seismicity. This issue happens only with vol_data? Could you please check that again.

Is it correct that, profile slicing from models with different scales (Fig. 2) stored in one vol_data prefers to stick to the smallest dataset? If this is not the case, I would then expect in the second figure that Thrastrason et al.2024 and Priestly et al. 2024 models are covering the full length of the profile.

I am attaching here the list of Dataset I used to make the slices. All datasets are available on the repository provided by Boris (seafile). But I can send a couple of them again for tests, if needed.

Datasets_latest_AE.txt

[mthielma]

Hi Amr,
as I showed in the plot above, I have used the Giacommuzi and the SAVANI model together, one of which is global, the other one local. I do not see any isses there in the paraview visualization. Is this also the case for you?

[amrelsharkawy]

Paraview plots and and my plots are identical. This has nothing to do with the plotting.

[mthielma]

Ok. So when combining the SAVANI and Giacomuzzi model, there is no truncation?

[mthielma]

To avoid that we use different datasets, can you resend the link to the seafile folder here?

[amrelsharkawy]

Here is the link: https://seafile.rlp.net/d/0adf3b029b044099b5cc/

[mthielma]

At which point does the truncation issue occur? Let's say, I combine the SAVANI and Giacommuzzi model (where no truncation occurs), adding which model results in truncation?

[amrelsharkawy]

Hi Marcel,
this happened when mixing between small scale and large scale models. This is my observation. I can not say at which model exactly.

Now it is working fine. What I did, I created new list of models, where I added models one by one to the new list. At the end, the models are listed according to their coverage (i.e., starting from global to local scale ones). Also in the file that contain the profile' coordinates, profile no. 1 is the longest followed by the coordinates of the shortest ones. Seems this solves the issue.

[boriskaus]

Can you please provide a minimum working example (MWE) that clearly illustrates the problem. The statement "I cannot say in which model exactly" is of very limited scientific use.

The only thing I get from the discussion above is that there doesn't seem to be a problem, in which case we can close this issue.