2. HPGE tests for envelopes
All the input files used for this part of the work are archived in the following Zenodo archive: https://zenodo.org/records/17857528
The localisation mask is created using loc_area_antarct_isobath.inp@7e0f825 and the bathymetry of GOSI10p3.
For the version of the SO with localised multi-envelope s-coordinates (Bruciaferri et al. 2018, Bruciaferri et al. 2024) we use 4 envelopes and we reduce HPG errors with 3 iterations of the algorithm described in Appendix C of Bruciaferri et al. 2024.
The input files used for the HPGE iterations are the following:
*) MEs_4env_3550_r12_r12.inp@f49a3df
*) MEs_4env_3550_r12_r12-r07.inp@f49a3df
*) MEs_4env_3550_r12_r12-r07-r01.inp@f49a3df
The last iteration ws run using the u-dv256@340690 workflow.
The HPGE tests are run using a 2D map for the minimum value of the bottom drag coefficient created with generate_Cdmin_field.py@ab57fbe (the output file is archived in the Zenodo archive specified at the beginning of this page).
After three iterations targeting HPGE > 1 cm/s, numerical spurious currents are very low: aftre a three months long test, the maximum error is < 2 cm/s with very few isolated points located in the deep (>4000m) ocean showing maximum HPGE > 1 cm/s:
- Timeseries of the maximum (blue), 99% (yellow) and average (red) numerical spurious currents:
- Map of the maximum (in time and in the water column) error after a three months long HPG test:
