bird insect partitioning and post-hoc precipitation filter

[adokter]

• Implements add function to partition mixtures of birds and insects #700
• implements add posthoc precipitation filter #701
• fixes Duplicated timestamps in csv change height distribution in vpts plot #697

[adokter]

@bart1 can you please review if this PR fixes your bug report on duplicate timestamps (#697) and close that issue if it has

[bart1]

Ill check Sent from Outlook for Android<https://aka.ms/AAb9ysg>

…

________________________________ From: Adriaan Dokter ***@***.***> Sent: Tuesday, March 4, 2025 7:10:29 PM To: adokter/bioRad ***@***.***> Cc: Bart Kranstauber ***@***.***>; Mention ***@***.***> Subject: Re: [adokter/bioRad] bird insect partitioning and post-hoc precipitation filter (PR #698) @bart1<https://github.com/bart1> can you please review if this PR fixes your bug report on duplicates timestamps (#697<#697>) — Reply to this email directly, view it on GitHub<#698 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAMV7BBOVC6C2YKNA6PWGTT2SXUBLAVCNFSM6AAAAABX6GQU5OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMOJYGUYDCNBSGA>. You are receiving this because you were mentioned.Message ID: ***@***.***> [adokter]adokter left a comment (adokter/bioRad#698)<#698 (comment)> @bart1<https://github.com/bart1> can you please review if this PR fixes your bug report on duplicates timestamps (#697<#697>) — Reply to this email directly, view it on GitHub<#698 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAMV7BBOVC6C2YKNA6PWGTT2SXUBLAVCNFSM6AAAAABX6GQU5OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMOJYGUYDCNBSGA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[adokter]

@bart1 decided to merge, as this PR had some critical fixes I wanted posted to master asap. Feel free to add your review here nonetheless and we can merge those changes later

[bart1]

@adokter Sorry for the delay, I just start working a bit with this, somethings I noticed below

First of all nice changes, useful to make script easier! On the topic of correcting VPTS data, maybe we could also think about filling holes in a vpts when the sdvvp drops to low during intense migration. That is a other frequent fix I used.

not equally high profiles

I noticed as.vpts requires the height profiles to be exactly equal otherwise it drops the profiles. If the height interval is equal (e.g. 200 meters) it might be nice to just NA fill the profiles? I think the aloft profiles for example not always go up to the same height:

devtools::load_all("~/bioRad/")
#> ℹ Loading bioRad
#> Welcome to bioRad version 0.8.1.9000
#> 
#> using vol2birdR version 1.0.5 (MistNet not installed)
pvolfile_source <- system.file("extdata", "volume.h5", package = "bioRad")

pvolfile <- paste0(tempdir(),"/volume.h5")
file.copy(pvolfile_source, pvolfile)
#> [1] TRUE

  vp <- calculate_vp(pvolfile) |> as.data.frame()
#> Running vol2birdSetUp
#> Warning: radial velocities will be dealiased...
  vp_low <- calculate_vp(pvolfile, n_layer = 10)|> as.data.frame()
#> Running vol2birdSetUp
#> Warning: radial velocities will be dealiased...
  
as.vpts(rbind(vp, vp))
#> Warning in validate_vpts(data): Extra fields found: day, sunrise, sunset
#>                    Regular time series of vertical profiles (class vpts)
#> 
#>            radar:  seang 
#>       # profiles:  2 
#> time range (UTC):  2015-10-18 18:00:00 - 2015-10-18 18:00:00 
#>    time step (s):  0
as.vpts(rbind(vp, vp, vp_low))
#> Warning in validate_vpts(data): Extra fields found: day, sunrise, sunset
#> Warning in as.vpts(rbind(vp, vp, vp_low)): Profiles found with different altitude layer specifications, keeping only the most
#>     common one, dropping  1 profile(s) ...
#> Warning in as.vpts(rbind(vp, vp, vp_low)): Profiles found with different number
#> of height layers: 10 20, retaining 20 only.
#>                    Regular time series of vertical profiles (class vpts)
#> 
#>            radar:  seang 
#>       # profiles:  2 
#> time range (UTC):  2015-10-18 18:00:00 - 2015-10-18 18:00:00 
#>    time step (s):  0

Here it would be nice if vp_low was just filled with NA values especially if the difference is in the high range (e.g. a mixture of profiles upto 4000 and 5000 meter)

Clean mixture

I would watch out with the argument naming u vs U, and v vs V, would it not be better to use something like V_wind. I know most of the time it won't be needed but still I could see bugs slipping in easily. I find capitalization to distinguish parameters difficult.

Filter precip

Maybe you can use something like this as a default parameter for dbz : ifelse(x$attributes$how$wavelength<7,x, y). In this way you can optimize the default for both C and S band, that should at least help with the last note in the help description.

#697

Is mostly resolved except the source_file seems to be mixed up. Not a big issue as I expect not many people to use it but still it would be good to make sure that is correct

[adokter]

Thanks @bart1, I'll start working on incorporating this feedback

[adokter]

Incorporating reviewer feedback to this PR in #704