Suggestion: md.friction.coupling uses md.initialization.hyraulic_potential

[inwoo-park]

Dear ISSM developers,

Suggestion

I would like to suggest an additional copuling method in md.friction.coupling=5 which uses md.initialization.hydraulic_potential. Now, I try to use the effective pressure outputs from other external subglacial hydrology model that not implemented in ISSM. (specifically, the CUAS-MPI model: https://github.com/tudasc/CUAS-MPI)

When using the md.friction.coupling=3 for the transient simulation (or projection), the model applies a fixed effective pressure $N_{\mathrm{eff}}$ throughout the entire simulation. This approach creates a limitation because it does not account for time-varying ice pressure ($\rho_i g H$). For example, the Budd-type friction law:

$$ \tau_b = C_B N_{\rm{eff}}^{m} v_b, $$

where $N_{\rm{eff}} = \rho_i g H - P_w$ ($P_w$ is the water pressure). With md.friction.coupling=3, the effective pressure remains contant with md.friction.effecvtive_pressure during the transient period.

Therefore, I propose adding a new option, md.friction.coupling=5 with the following scheme:

0. uniform sheet
1. ice pressure only
2. water pressure assuming uniform sheet
3. Use provided effective pressure frommd.friction.effective_pressure
4. use coupled model (not implemented yet)
5. use hydraulic head in md.initialization.hydraulic_potential for computing effective pressure (Neff = rho_icegthickness - P_water)

CODE

The effective pressure calculated from md.initialization.hyraulic_potential may be computed as:

$$ P_{water} = \mathrm{md.initialization.hydraulic_potential} - \rho_w g z_b, $$

where $z_b$ is the base elevation of ice (md.geometry.base)

Then the effective pressure $N_{\mathrm{eff}}$ can be computed as:

$$ N_{\mathrm{eff}} = \max \left(\rho_i g H - P_{\mathrm{water}}, 0 \right), $$

ensuring that negative the effective pressure is not allowed.

I have created a branch hydrology-coupling here:

GitHub - inwoo-park/ISSM at hydrology-coupling.

Would it be appropriate to submit the hydrology-coupling branch to the ISSM/main branch?

Best regards,
Inwoo

[MathieuMorlighem]

Hi @inwoo-park

this is a good idea but I wonder if we are not reaching the limits of what friction is doing. We have so many options at this point it becomes hard to read and it is not user friendly anymore. I wonder if we should stop lumping some calculations of Neff in friction and transfer them to a proper hydrology model. For example, in your case, we could create a new hydrologyprescribed.m class where you provide your time series of hydraulic head and the model would just update Neff at each time step (in the C++ code). Then friction would just pick up the EffectivePressure input and do its calculation. Do you see what I mean?
Let's see what others have to say and wait until we are all happy with how to proceed

Cheers
Mathieu

[inwoo-park]

Hi @MathieuMorlighem

I agree that adding an additional md.friction.coupling option would confuse users. Rather than introducing a new option in md.friction.coupling, I will try to create hydrologyprescribed.m to assign the time-series hydraulic head dataset in another branch.

Best regards,
Inwoo

[MathieuMorlighem]

that sounds good, let me know if you need help! I think later we should transfer all the others to independent hydrology classes so this will be a good test. Thanks Inwoo!