Restart wrf simulation with different GFS cycle

[janac4]

Hello,
I am familiar with WRF model and have already used wrf restart files to continue previous simulations with extended GFS files of the same cycle. However, I was wondering if it is possible (both computationally and physically) to use the last wrfrst file of a simulation (that started with GFS cycle 00 UTC) to extend the forecast but using next GFS cycle files as input date (properly selected according to the new forecasting date). I am worried about accumulated variables and tendencies since I will be merging a initial state of the atmosphere with boundary conditions different from those used in the previous simulation.
Also, is it possible to create a hole new simulation with GFS files form different cycles from the beginning (without using restart files to merge them)? Is it a worse approach?
My intention is to use this mix of GFS files to avoid having to generate spin-up files in every situation, since using wrfrst files would already save a proper state of the atmosphere, so I could combine the latest GFS files without having to start again the simulation with a cold-start.
Thanks a lot.

 

[kwerner]

We do not recommend restarting WRF for a continuous run, using first-guess output from a different cycle. The reason is there may be some discrepancy between GFS data from different cycles, and this can negatively affect WRF performance, and make evaluation difficult.

It is possible to create an entirely new simulation with GFS files form different cycles from the beginning. You are free to choose data from any GFS cycle to force WRF. Unfortunately, though, this approach won't allow you to avoid spin-up.

I'm not sure this fully answers your question. If not, please provide a bit of clarification. Thanks!

 

[janac4]

Thank you so much for your response, this is exactly what I wanted to know.