Research Question & My Role
My undergraduate project asked whether a simplified thermal model could help examine down-dip rupture limits on the Main Himalayan Thrust. I constructed the geometry and mesh, varied inputs, compared results, and wrote the interpretation. The work used Jiangheng He’s existing PGCtherm3D solver, with support from Juan Rosas and supervision from Dr. Claire Currie.
Method
I built a simplified cross-section in GiD, supplied thermal properties and geotherms to PGCtherm3D, and used ParaView to inspect temperatures along the modelled interface. Four scenarios varied Indian-crust inputs while retaining Himalayan-side assumptions.
Findings
The report preferred a warm scenario and found a hotter alternative broadly compatible with its adopted constraints; cooler scenarios placed interpreted transitions deeper. This comparison did not uniquely establish the real temperature field or predict a future earthquake.
Limits of the Evidence
The model simplified fault geometry, assumed steady conditions, omitted frictional heating, and included solver-required layers. A revision must reconcile heat-flow units and sign conventions, define numerical convergence criteria, and separate model selection from independent validation. The original inputs have not been rerun.
What I Take Into Engineering
The transferable practice is sensitivity to assumptions: vary inputs, retain evidence, and state where a persuasive visualisation stops supporting the interpretation.
What I Would Test Next
I would create a reproducible input set, audit units, test mesh convergence and boundary placement, and compare results with independent heat-flow and seismic-depth observations. AI was not part of the documented workflow; any future organisational aid would remain subordinate to solver validation and physical interpretation.