The MR fixes the K_Tp terms in TRF and TRM. The changes have little impact, only TaskCDECOVALEX2023 benchmarks are affected. Here, their changes.
It also improves the convergence:
I also tested the latent heat implementation of TRM agains TH2M with const. gas pressure using an altered heat pipe benchmark.
Without advection (
c_p=0) it looks relative good:
With advection there are some differences:
However, total agreement is not expected in both cases as latent heat implantation as well as advection terms differ. (See https://www.sciencedirect.com/science/article/pii/S1365160923002083 for details)