Typhoon simulation in the nested domain is poor

[jc sun]

Hello, everyone

In a case study of typhoon simulation, I employed a three-layer nested grid model with resolutions of 27 km, 9 km, and 3 km, driven by 0.25° GDAS data. Interestingly, I found that the simulation results using the 27 km domain were the most accurate, with typhoon tracks, central pressure, and maximum wind speeds closely matching the CMA best track dataset. However, when using the 9 km and 3 km domains, the simulated typhoon intensity became significantly stronger. Since the simulation results in the nested domains were not satisfactory, I decided not to enable the feedback mechanism. I’m wondering whether I should adjust the grid domain settings or modify the parameterization schemes. So far, I have experimented with changing the microphysics scheme, which had a notable impact on the typhoon track but only limited effects on its intensity. I would greatly appreciate hearing any expert suggestions on this matter.

&physics
 physics_suite                       = 'CONUS'
 mp_physics                          = 8,8,8,
 cu_physics                          = 1,1,0,
 ra_lw_physics                       = 1,1,1,
 ra_sw_physics                       = 4,4,4,
 bl_pbl_physics                      = 1,1,1,
 sf_sfclay_physics                   = 1,1,1,
 sf_surface_physics                  = 2,2,2,
 radt                                = 12,12,12,
 bldt                                = 0,0,0,
 cudt                                = 5,0,0,
 icloud                              = 1,
 num_land_cat                        = 21,
 sf_urban_physics                    = 0,
 sst_update                          = 1,
 isftcflx                            = 1,
 surface_input_source                = 1,

 

[Ming Chen]

Typhoon intensity simulated by WRF is usually weaker than observation when the resolution is coarse. This is one reason why we recommend using high resolution to simulate typhoon intensity. For your case, one possible reason is that the intensity in your initial condition of nested domain is already weaker than observation, and the initial condition subsequently affects intensity simulated by WRF. Note that WRF physics/dynamics to a certain extent can correct bias in initial condition, however, they cannot fully remove the bias.
By the way, Tiedtke scheme is a better option for hurricane/typhoon simulation. Probably you can try.