Help - Attenuation Correction without KDP calculation

#1

Hello LROSE users,

I am reaching out to see if someone can assist me.

I have been trying to test the attenuation correction of DBZ using the RadxKdp tool. However, it seems mandatory to perform the KDP estimation before applying the correction. I am wondering if it is possible to skip this step, as I already have .nc files that include the KDP values. I believe it would not be appropriate to estimate parameters to generate a KDP that is “similar enough” to the already estimated values.

I would greatly appreciate any guidance if this option exists. I have reviewed the parameter files, but it is unclear to me whether skipping the KDP estimation is feasible.

Thank you in advance for your help!

#2

Hello,

My apologies for the delay in responding.

Which version of LROSE are you using? My understanding is that RadxKdp assumes that KDP needs to be calculated. However, newer versions of RadxPid and RadxRate allow a user to provide a pre-existing KDP field that will be used - both of those applications can calculate the attenuation correction for DBZ and ZDR.

However, I could see it being useful to not have to run PID and QPE when it’s not needed and still be able to calculate attenuation. I’ll reach out to the software engineers to get their thoughts.

Thanks,
Jen

1 Like
#3

Hello,

Thank you for your response.

I am using lrose-colette-20240525 on Ubuntu 20.04. I will check with RadxPid and RadxRate to see if I can extract the attenuation correction from my own KDP fields.

It would be very helpful if this functionality existed.

Thank you!

#4

Ok, please let me know the result. Mike said it might not be implemented from a user’s provided KDP, but that it should be straightforward to implement.

Once you tell RadxPid/RadxRate that KDP is available, you can compare the DBZ/DBZ_ATTEN_CORRECTED and ZDR/ZDR_ATTEN_CORRECTED fields to see if they’re different.

Jen

#5

Hello,

I am sorry for the late response.

I tried to use RadxRate stating the availability of the KDP field. Nevertheless, since I want to obtain the attenuation corrected fields, I get the following error:

ERROR: RadxRate
  Inconsistent parameters set in params file: /home/arpalas/lrose/event_analysis/params/RadxRate.params
  KDP_available is TRUE
  PID_use_attenuation_corrected_fields is TRUE
  You must compute KDP to use atten-corrected fields
Error: Could not create RadxRate object.
ERROR: RadxRate
  Inconsistent parameters set in params file: /home/arpalas/lrose/event_analysis/params/RadxRate.params
  KDP_available is TRUE
  PID_use_attenuation_corrected_fields is TRUE
  You must compute KDP to use atten-corrected fields
Error: Could not create RadxRate object.

So it makes me understand that I must compute KDP to perform attenuation correction. But it is ok, I will try comparing KDP computed with the existing.

Thank you again for your help!

PS: Now I am struggling with the spdb temperature profiles when running RadxRate. I get errors saying that the time the algorithm is looking for cannot be found. I am following this thread RadxPid not loading SPDB soundings but I have not been able to obtain the expected result, I might ask something similar in that post.

#6

Ah ok, good to confirm what is and isn’t possible. I’ll make a note to ask Mike if he can implement attenuation correction when KDP is already provided.

As for the temperature profiles, I’ll look for your post on that. We did discuss at a recent meeting bypassing the Spdb data entirely and reading soundings directly from model output, but I’m not sure how long that would take to develop. But we do recognize that that step is often one of the most challenging in the RadxRate/RadxPid pipeline.

Jen