uafgeotools / mtuq

moment tensor uncertainty quantification
BSD 2-Clause "Simplified" License
67 stars 22 forks source link

Syngine tags consistency #266

Open SeismoFelix opened 4 months ago

SeismoFelix commented 4 months ago

When fetching GFs from syngine database for the PREM velocity model, the same model parameter passed to taup_model during the creation of the the object process_sw :

model = 'prem' 
process_sw = ProcessData(filter_type='Bandpass',
freq_min=1/float(freqs_sw[1]),
freq_max=1/float(freqs_sw[0]),
pick_type='taup',
taup_model=model,
window_type='surface_wave',
window_length=wl_sw,
capuaf_file=path_weights,
 )

, does not work when fetching the syngine-derived Green Functions. Here, Instead of defining model as prem, it has to be specified what kind of prem are we interested:

model = 'prem_i' 
greens = download_greens_tensors(stations, origins, model)

So, in this case the same tag used for process_sw does not apply for download_greens_tensors (unlike 'ak135').

This is because the Taup travel times library in Obspy names the PREM model generically, just 'prem'.

On the other hand, syngine database requires the specific type of prem model to fetch the green functions, either the isotropy (prem_i) or anisotropy one (prem_a).

In the same line, when in MTUQ is used the ak135 model, it is assumed the ak135f_2s. Is there a reason behind such an assumption and not being necessary to specify the model resolution (ak135f_1s, ak135f_2s, or ak135f_2s)?

Thanks!