Swedish radars sehem, sehud, seoer, seosu

[peterdesmet]

Cf. enram/registry#4, we are getting data from the Swedish radars sehem and seoer:

http://enram.github.io/data-repository/?prefix=se/hem/ : 2017, 2018, 2019
http://enram.github.io/data-repository/?prefix=se/oer/ : 2017, 2018, 2019

But these radars are not listed by OPERA: https://github.com/enram/dmp/blob/master/data/OPERA_RADARS_DB.json

@leijnse @CeciliaNilsson709 any information which radars this are and why they are missing in the OPERA database?

[CeciliaNilsson709]

Ok, I have no idea what is going on. Looks in the repository like they might be seovi (Örnsköldsvik) and sease(Ase) but that is just guessing. I'll check with Günther.

[peterdesmet]

You're right, it could be those radars indeed:

seovi stops having data in 2017-08-06
sease stops having data in 2017-05-13

While:

sehem: starts having data in 2017-11-13
seoer: starts having data in 2017-11-24

But why would they change the code?

[CeciliaNilsson709]

seovi is probably sehem, as it is close to location Hemlig (as well as Örnsköldsvik). Don't know why (or if) sease would be seoer. And no idea why they would change them 🤷‍♀️

[CeciliaNilsson709]

It's actually probably the other way around, sease is close to Hemse (so sehem) and seoer probably refers to Örnsköldsvik after all (ö->oe). But needs to be confirmed of course.

[CeciliaNilsson709]

Got this list of new(old) codes for the Swedish radars from Günther:

ANG (ANG)
ATV (VIL)
BAA (ARL, radar Bålsta will replace radar Arlanda)
HEM (ASE)
HUV (HUD)
KAA (KKR)
KRN (KIR, radar Kiruna will change name during upgrade in august)
LEK (LEK)
LLA (LUL)
OER (OVI)
OSD (OSU)
VAX (VAR)

[peterdesmet]

Aha, that is helpful! 👏

• Did he mention why they decided to change some of these codes?
• I assume we can just move data stored under the old code to the new code?
• Not sure about moving BAA to ARL, unless the new data is at the same location as the old one? Do they consider it the same thing?

[CeciliaNilsson709]

I am very helpful really.

• No, but guess it might have to do with upgrades 🤷‍♀️ Or they just felt like it 🤷‍♀️
• Yes, except 👇
• Bålsta is approximately 20km from Arlanda airport, so I think we should probably treat them separately?

[peterdesmet]

🙃

1. Ok
2. Ok, will create a separate issue for that
3. I agree: best to keep separate even if they don't. But would maybe propose Gunther to keep it as two radars in their list.

[peterdesmet]

Documented at #53. I notice that that we are not getting data yet under the new code for 4 of these changes (including KRN which will only be operational in August). Can you confirm with Günther that these is normal? 🧐

[adokter]

You might want to compare the lat,lon metadata of the new and old files. It could be that the physical locations of the radars are not exactly the same (different building in the same city), and that therefore they got a new code.

[peterdesmet]

I will test, compare and report back.

[peterdesmet]

Ok, here are the results:

Coordinates

two lines = pair of similar radars

radar	lat	long
sease	57.3035011292	18.400100708
sehem	57.30339813232422	18.400299072265625
sehud	61.5770988464	16.714399337800003
sehuv	61.577110290527344	16.714433670043945
seovi	63.6394996643	18.4018993378
seoer	63.63946914672851	18.40188217163086
seosu	63.294998168900015	14.7590999603
seosd	63.29499816894531	14.75909996032715
sevar	58.2555999756	12.8260002136
sevax	58.25559997558594	12.826000213623047

Code:

library(bioRad)
library(tidyverse)
metadata <- tibble()

# Change code below to select another radar

radar = "sevax"
date_min = "2017-09-01"
date_max = "2017-09-30"
directory = "my_data"

# Run code below to add radar to metadata

download_vpfiles(radar = radar, date_min = date_min, date_max = date_max, directory = directory)
vp_selection <- select_vpfiles(radar = radar, date_min = date_min, date_max = date_max, directory = directory)

vp <- read_vpfiles(vp_selection[1])

new_row <- tibble(
  radar = vp$radar,
  lat = vp$attributes$where$lat,
  long = vp$attributes$where$lon
)

metadata <- bind_rows(metadata, new_row)

Distances

radar_1	radar_2	distance_in_meter
sease	sehem	16.56898
sehud	sehuv	2.225491
seovi	seoer	3.50662
seosu	seosd	5.229753e-06
sevar	sevax	2.068746e-06

Code:

library(sf)
sf_metadata <- st_as_sf(metadata, coords = c("long", "lat"), crs = 4326)
st_distance(sf_metadata[9,], sf_metadata[10,])

Conclusion

The radars have different coordinates, but their distance are almost negligible. The biggest difference is between sease and sehem, which is 17m. It still looks like the exact same spot in the forest though:

I think we can safely move the old files under the new code directory?

[jshamoun]

Great comparison. I still suggest that someone checks directly with the Swedish meteorological Institute to be sure that something hasn’t changed. Some countries are busy replacing radars.

[CeciliaNilsson709]

It looks good to me. I can check with Günther again, but not really sure what more we want to know? Seems like for our purposes we can just move the files (except for future sebaa, as discussed above).

[jshamoun]

Cecilia, if you could just check with Gunther that there haven’t been changes to the radars and just changes to the meta data, that would be more than enough. From: Cecilia Nilsson [mailto:notifications@github.com] Sent: Thursday, 11 April 2019 10:28 PM To: enram/dmp <dmp@noreply.github.com> Cc: Shamoun-Baranes, Judy <J.Z.Shamoun-Baranes@uva.nl>; Comment <comment@noreply.github.com> Subject: Re: [enram/dmp] Swedish radars sehem, sehud, seoer, seosu (#5) It looks good to me. I can check with Günther again, but not really sure what more we want to know? Seems like for our purposes we can just move the files (except for future sebaa, as discussed above). — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://github.com/enram/dmp/issues/5#issuecomment-482298900>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AHVIzGE5EJZqtyM7PkAPT_EChx_hWTnZks5vf5rXgaJpZM4ckCtS>.

[CeciliaNilsson709]

Ok, I’ll double check with him so we can be completely sure 👍

[CeciliaNilsson709]

I sent this summary to Günther a while ago just to make sure we haven't misunderstood anything, waiting for him to reply.

Radar	Comment
ANG	No change.
VIL -> ATV	Not sending under VIL since jan 2019, but no ATV yet?
ARL	No change, will stop sending when BAA starts
BAA	Will appear later, don't change anything
ASE -> HEM	Change
HUD -> HUV	Change
KKR -> KAA	Not sending under KKR since august 2018, but no KAA yet?
KIR -> KRN	Not yet sending under this name. Change later
LEK	No change
LUL -> LLA	Not sending under LUL since august 2018, but no LLA yet?
OVI -> OER	Change
OSU-> OSD	Change
VAR -> VAX	Change

[CeciliaNilsson709]

Here are Günthers comments (in my translation):

Radar	Comment	Günther's comment
ANG	No change.
VIL → ATV	Not sending under VIL since jan 2019, but no ATV yet?	Should start sending under ATV soon
ARL	No change, will stop sending when BAA starts	ARL will stop on Monday 2019-04-29
BAA	Will appear later, don't change anything	Starts on Monday 2019-04-29, but will take awhile before the data appears. Coordinates are: 59.6110° N, 17.5833° E.
ASE → HEM	Change
HUD → HUV	Change
KKR → KAA	Not sending under KKR since august 2018, but no KAA yet?	KAA should be on server now
KIR → KRN	Not yet sending under this name. Change later	Will change during August 2019
LEK	No change
LUL → LLA	Not sending under LUL since august 2018, but no LLA yet?	LLA should be on server now
OVI → OER	Change
OSU → OSD	Change
VAR → VAX	Change

He also said: "Right now only our OPERA single-pol sweeps are being used to create bird profiles (three elevations for upgraded sites and ten for non-upgraded sites, ARL and KIR, but we will change to dual-pol sweeps (10 elevations) as soon as possible. BAA will be the first."

[adokter]

Thanks @CeciliaNilsson709. Did Gunther also explain why the name changed even when the upgraded radar stayed at the same location?

[CeciliaNilsson709]

Nope. Maybe they just wanted to change it up a bit 💃🤷‍♀️?

Seriously though, he didn't say explicitly, but it seems to be connected "upgrades", which I assume is dual-pol upgrades (but don't know why only some changed).

[peterdesmet]

Ok, I now know HOW to change this on the S3 bucket. I'll create a spreadsheet to track what needs to be done and what has been done already. 👌

[peterdesmet]

All code changes are in effect, except:

KIR → KRN	Not yet sending under this name. Change later	Will change during August 2019

It is still submitting under sekir.

Many of the new codes are not in OPERA DB.

[peterdesmet]

sekir is still submitting under that code, and not sekrn. As that is the only remaining thing and it is documented in the wiki (https://github.com/enram/data-repository/wiki/se), I will close this issue