SPECFEM-C

This form of SPECFEM-X is for the development of SPECFEM-C prior to any integration of SPECFEM-X with SPECFEM Kokkos. It incorporates the ability to model sea-level change based on the rate-dependent formulation of Crawford, et al., 2018.

Open areas in the code/uncertainties

• Calculating SL Area - do we project the area to the vertical defined as orthogonal to the SL surface, or nah?

To do list:

• Non-dimensionalise the ice loads
• Add the final weak form term (ice contibution)
• Write ocean function updater using ice and theta comparison
• When calling set_petsc_stiffness_SL do we need to parse the SL matrices as args? arent they global?

Manual updates:

• Re-write SL input file to be consistent with ice file format
• Important rewrite of format for traction input at single GLL point - now ielmt, iface, gll on face (this assumes that traction is on a surface face)
• icerate file

Time looping

• Stiffness matrix needs to be updated at each timestep...not just the first 1 (elastic) or 2 (viscoelastic)
• Need to store nodalu, nodalphi, nodalsl etc at each timestep (not to be overwritten)
• Incorporate ice load into the RHS
• Add in time-marching scheme to estimate ${\phi_{t+1}}$, $m_{t+1}$, $u_{t+1}$, and $\theta_{t+1}$,

Clarify with HNG

• Do we need to incorporate SL terms into the true Kmat so that it influences the bcnodalv calculations?
• Why resload=load-bodyload
• apply_nonzero_bc - double check exactly what is happening here
• Why do we keep setting lots of the loads' first elements to zero. e.g. bodyload(0) = ZERO
• Do we need to split up the memory integral $$\int_{M_s} 2\mu_0\Big[\dot{\mathbf{m}} : \tilde{\mathbf{m}} + \frac{1}{\tau} (\mathbf{d} - \mathbf{m}):(\tilde{\mathbf{d}} - \tilde{\mathbf{m}}) \Big] dV$$ so that $$\int_{M_s} 2\mu_0\Big[ \frac{1}{\tau} (\mathbf{d} - \mathbf{m}):(\tilde{\mathbf{d}} - \tilde{\mathbf{m}}) \Big] dV$$ is part of the RHS?
• What are the different loads being used?