Time Step Setting Problem of Offline and Coupled Running

[Guangwei LI]

Dear all,

       During the calibration phase, I found that the a good simulation results (soil temperature) rely on the setting of "NOAH_TIMESTEP". 

       But I'm a little confused, does the "NOAH_TIMESTEP" in  namelist.hrldas and the "time_step" in namelist.input&domains  mean the same thing ? because they look quit different: when running offline, the value is 3600s, but when running coupled, the value is 20s.

        Thanks for your help!

Guagnwei Li

[Katelyn FitzGerald]

Hi Guagnwei,

As you've noted, these variables do both control the land surface model timestep and there is sensitivity to the time step particularly for some model states.  The discrepancy between "typical" model time steps (I assume those you've referenced are from example case namelists) used within the standalone and coupled modeling contexts is driven by how these time step variables are otherwise used.

In the context of coupled runs with WRF, the time step needs to be much smaller in order to maintain model stability within the atmosphere.  The WRF documentation has some guidance regarding this.  With standalone simulations this constraint no longer exists and there is often an interest in increasing the LSM time step for computational efficiency.  That said these values can and arguably should change depending on your model configuration and particular application and if you are calibrating parameters within the land surface model you likely will want to maintain consistency between simulations. 

Hope this helps.

Katelyn

[Guangwei LI]

  Thank you for resolving my confusion！