grids_out, _ = pydda.retrieval.get_dd_wind_field([grid_kict, grid_ktlx],
Cm=256.0, Co=1e-2, Cx=1, Cy=1,
Cz=1, Cmod=1e-5, model_fields=[“hrrr”],
refl_field=‘DBZ’, wind_tol=0.5,
max_iterations=50, filter_window=15,
filter_order=3, engine=‘scipy’)
@imd - Could you please provide more information here if you can!
- What data are you working with?
- What is the size of your grid?
Please see the pydda user guide here for more information
https://openradarscience.org/PyDDA/user_guide/retrieving_winds.html
1 Like
I am working with X-band radar data and trying to extract wind components from the radial velocity measurements. Here are the radar specifications I have:
Radar Sweep Angles: 3 total sweep angles at 4.5°, 7.0°, and 10°.
Temporal Resolution: 2.5 minutes.
Spatial Resolution: 150 meters.
Coverage Area: 200 km.
For my grid setup, I have defined the following parameters (from the user guide):
grid_limits = ((0., 15000.), (-50000., 50000.), (-50000., 50000.))
grid_shape = (31, 201, 201)
However, I’m unclear on how to determine the values for the variables cm, co, cx, xy, and cz in the context of this setup.
Could you please provide some guidance on how to proceed with these calculations? Specifically:
How can I calculate or estimate cm, co, cx, xy, and cz?
How can I calculate the wind components from the radar’s radial velocity without needing a model or additional observational data?
You could try a trial-and-error approach, then compare with the sounding or reanalysis winds until you get the desired results.
Keep the values between 0 and 1 initially, then gradually increase to two and three digits. @rcjackson can explain it better.
Are you working on only one radar?
2 Likes
I would start with the coefficients in the user guide @mgrover1 posted for these values. You will need to fine tune them from there most likely.
You can start with one Doppler radar and go from there to see what wind field you get. A single Doppler retrieval will likely be able to retrieve the along-wind component of u, v pretty well, though I would strongly recommend adding a sounding or reanalysis as a constraint.
1 Like
@syedhamidali @rcjackson I’m working with two radar systems, one with 3 sweeps(X-band) and the other with 11 sweeps (C band). Since there’s no sounding data available in my region, I typically rely on reanalysis data. However, the purpose of the work is nowcasting forecasting, so, I prefer not to use reanalysis data.
I’ve successfully run the code for individual radars and am also able to reterive wind field, but merging the two radar systems is challenging due to differing resolutions. To address this, I’m maintaining a consistent grid spacing for both radars.
Could you please suggest how to merge two radars as input in this function pydda.retrieval.get_dd_wind_field()?
Unfortunately, your X band radar configuration is suboptimal for multi Doppler wind retrievals since there are so few sweeps. What are you trying to nowcast? If it’s winds near the surface, then I would place the highest weights from your two radars to the lowest heights that are covered by both radars. You can make custom weights for each radar grid by using the weights keyword argument. With your radar configuration I would definitely suggest fine tuning the retrieval for your specific application.
1 Like
Thank you for your valuable suggestions.
My goal is to obtain wind profiles up to 10 km from the surface.
Could you please elaborate on how to perform ‘fine-tuning the retrieval’ for my specific radar configuration?
Additionally, I’d appreciate clarification on the ‘weights’ concept. Are weights assigned to variables such as cx, cy, cz, and co? If not, could you explain the weighting process and its significance in this context?"
Regards
Priya
@imd - I recommend you read through the following journal articles which describe these parameters more in detail