Hi everyone,
I am currently working on LES models implemented in WRF. I found few articles presenting and studying these models [1], [2]. In these articles, filtered incompressible Navier-Stokes equations in Cartesian coordinates are presented to make the sfs stress tensor appear. However, WRF is a compressible code and equation are written in (x, y, eta, t) referential frame (not Cartesian). Skamarock et al. [3] only give the final equations for the different turbulent modeling option.
Then I have few questions :
- Is there a document explaining how WRF equations are filtered ?
- What approximations are made here ?
- Does the compressible WRF equations need a Favre/Hesselberg averaging for the prognostic variables ?
- Since there are 2nd order terms in F_Ucor (coriolis and curvature terms), shouldn't there be also an sfs stress tensor associated with these terms ?
[1] : Simon, Jason S., Bowen Zhou, Jeffrey D. Mirocha, et Fotini Katopodes Chow. « Explicit Filtering and Reconstruction to Reduce Grid Dependence in Convective Boundary Layer Simulations Using WRF-LES ». Monthly Weather Review 147, no 5 (1st May 2019): 1805‑21. https://doi.org/10.1175/MWR-D-18-0205.1.
[2] : Kirkil, Gokhan, Jeff Mirocha, Elie Bou-Zeid, Fotini Katopodes Chow, et Branko Kosović. « Implementation and Evaluation of Dynamic Subfilter-Scale Stress Models for Large-Eddy Simulation Using WRF ». Monthly Weather Review 140, no 1 (1st January 2012): 266‑84. https://doi.org/10.1175/MWR-D-11-00037.1.
[3] : Skamarock, C., B. Klemp, Jimy Dudhia, O. Gill, Zhiquan Liu, Judith Berner, Wei Wang, et al. « A Description of the Advanced Research WRF Model Version 4.1 », 2019. A Description of the Advanced Research WRF Model Version 4.3 | OpenSky.
I am currently working on LES models implemented in WRF. I found few articles presenting and studying these models [1], [2]. In these articles, filtered incompressible Navier-Stokes equations in Cartesian coordinates are presented to make the sfs stress tensor appear. However, WRF is a compressible code and equation are written in (x, y, eta, t) referential frame (not Cartesian). Skamarock et al. [3] only give the final equations for the different turbulent modeling option.
Then I have few questions :
- Is there a document explaining how WRF equations are filtered ?
- What approximations are made here ?
- Does the compressible WRF equations need a Favre/Hesselberg averaging for the prognostic variables ?
- Since there are 2nd order terms in F_Ucor (coriolis and curvature terms), shouldn't there be also an sfs stress tensor associated with these terms ?
[1] : Simon, Jason S., Bowen Zhou, Jeffrey D. Mirocha, et Fotini Katopodes Chow. « Explicit Filtering and Reconstruction to Reduce Grid Dependence in Convective Boundary Layer Simulations Using WRF-LES ». Monthly Weather Review 147, no 5 (1st May 2019): 1805‑21. https://doi.org/10.1175/MWR-D-18-0205.1.
[2] : Kirkil, Gokhan, Jeff Mirocha, Elie Bou-Zeid, Fotini Katopodes Chow, et Branko Kosović. « Implementation and Evaluation of Dynamic Subfilter-Scale Stress Models for Large-Eddy Simulation Using WRF ». Monthly Weather Review 140, no 1 (1st January 2012): 266‑84. https://doi.org/10.1175/MWR-D-11-00037.1.
[3] : Skamarock, C., B. Klemp, Jimy Dudhia, O. Gill, Zhiquan Liu, Judith Berner, Wei Wang, et al. « A Description of the Advanced Research WRF Model Version 4.1 », 2019. A Description of the Advanced Research WRF Model Version 4.3 | OpenSky.