Hi fellow WRF(DA) users,
I am trying to assimilate composites of radar reflectivity. I have succesfully converted my own radar data (after some filtering, quality control, etc) to the ASCII format that WRFDA reads. For the moment I am only using reflectivity data and not radial velocity. Also, to isolate the effect of radar measurements, I have turned off the option to assimilate any other conventional observations. I am using the indirect method by Wang et al (2013), which assimilates mixing ratios computed from radar reflectivity rather than the actual reflectivity values. The WRFDA routine reads in my data just fine and recognizes and converts 56996 radar reflectivity observations to 'rrn (kg/kg)' observations. The statistics file shows some changes in 'qrn' (qrain?) on vertical model level 7-10, in the order of 1e-4 kg/kg which seems plausible since the composite reflectivity measurements are all at 1500m, which is close to the geopotential height of ~1455m of vertical model level 9.
However if I read both the wrfvar_output and first guess (fg, input to WRFDA) netcdf files and take the difference for output variables like 'qrain' (which was the only variable that changed according to the statistics file) and 'qvapour', the differences are all 0 and do not show any increments around model level 7-10 as was the case in statistics file.
Does anyone have an idea what could cause this issue and how to resolve it? I have checked multiple times that I am taking differences of the correct files (fg and wrfvar output), have run the DA routine multiple times but the problem still exists.
Some info that might be useful:
- WRFPLUSV3 and WRFDA 3.7.1
- cv_options=7 (also generated background errors for my domain using this control variable option)
- cloud_cv_options=3 (compiled WRFDA with the cloud control variable option to be able to make use of indirect reflectivity assimilation)
Any help is appreciated a lot!