Nudging in WRF

[janac4]

Hi everyone,
I am familiar with WRF model for real-time simulations using GFS input data. Since NCEP produces FNL products (more accurate forecast since more observational data is assimilated in the process), I would like to use these data to create more accurate initial conditions (and also forecasts). However I'm not sure the proper methodology to work with these data is nudging or the use of WRFDA model. I am not familiar with the differences in both methodologies and would like some guidance.
Also in WRFDA USER SUPPORT: Frequently Asked Questions when they talk about "cycling" I assume it is related to using WRF without having to remove spin-up time, which seems very interesting. Could this be done with FNL data? How can all be related?
Thank you.

 

[Ming Chen]

You may use FNL as the forcing data to run WRF. The process to create initial and boundary conditions using FNL is the same as that using GFS.

I am not an expert in WRFDA. Please find helpful information in the. website here.

 

[janac4]

Thanks for your help. I was thinking about merging GFS boundary conditions with FNL data as initial conditions with nudging for the first run hours, to assimilate analysis data without installing WRFDA, but I'm not sure this workflow is possible or consistent. I assume that when I read about nudging the nudged data is the same as the input and forcing data, can it be changed?