Windy missed radiosonde for days tho data posted to NOAA archive
-
@hangInThere great, thank you for thorough investigation. I'll look into filtering the data from MADIS when I get some other things out of my plate
-
@Filip_K Kuantan (48657) is missing radiosonde ascent data on Windy this morning 24 Aug 2024 0z, BUT NOAA ESRL (what I discussed last time) has it, which is weird b/c Windy supposedly gets from NOAA GTS, and usually if a station is missing data it happens to both, but today it happens to Windy only. Something gives?
-
@Filip_K This problem suddenly impacts most stations on Peninsular Malaysia, a sudden deep deterioration. There's no data from all but one from this morning 2023-09-15 0Z on Windy, which supposedly sources data from NOAA GTS, whereas I can download and display data no problem from NOAA realtime archive for ALL four stations (See below).
-
@Filip_K Radiosonde ascent data from Penang (48601) and Kuantan (48657) have been missing from Windy for 6 and 4 days, respectively. They are available from NOAA's realtime database. I suggest Windy adds this realtime database as an alternative source of data.
-
@Filip_K Radiosondes for Indonesian airports at Medan, Banda Aceh, Pontianak and Ranai are no longer accessible on Windy mobile. Are they permanently excluded? If so, why? The daily 0Z and 12Z data are available in real time from ESRL database I showed previously. Below is what I plotted from Bangka Island's Pangkal Pinang airport radiosonde for 0Z today:
-
@hangInThere
It looks like all Indonesian radiosonde ascents are currently missing from all our data feeds. We'll most likely add NOAA archive as source during an import rewrite, which is currently in progress -
@Filip_K
NOAA has two separate archives: IGRA under NCEI which stores ascent data at least 1 day old, and realtime RAOBS under ESRL which has up-to-date ascent data. Make sure you poke into the latter. -
@Filip_K There's been no update on Windy to Malaysia airports' radiosonde for 3 days.
I also checked ESRL (haven't done so for a year) and was disappointed they have cut off visits from the public. -
Hi @hangInThere, sorry for delay in response. There was an issue in our ingress pipeline. All should be caught up now
-
@Filip_K Yeah I saw this morning those airports have today's Radiosondes! Missed them much.
-
@Filip_K BTW, what about those neighboring Indonesian airports like Rania, Medan, Padang, PangkalPinang, and Pontianak? Do you see a time when Windy will report their daily radiosondes?
-
@hangInThere
Unfortunately, NOAA radiosonde archive is currently not on our to-do list, so I can't give You any specific date.Also, I quickly tried getting the data from the archive and it seems it has changed. Link to the root of the archive (https://www.ncei.noaa.gov/data/integrated-global-radiosonde-archive/access/data-y2d/) doesn't seem to work anymore, although link to specific file (like https://www.ncei.noaa.gov/data/integrated-global-radiosonde-archive/access/data-y2d/MYM00048650-data-beg2021.txt.zip) does work for me.
Does it work for You?
-
@Filip_K NOAA
(1) Radiosonde archival data retrieval still works, thru the siphon package as follows (in Python):
from siphon.simplewebservice.igra2 import IGRAUpperAir # Make the request. IGRAUpperAir returns a dataframe containing the sounding data and # a dataframe with station metadata from the sounding header. df0, header = IGRAUpperAir.request_data(date, stationCode)Note that the archive holds data only up to roughly 2 days ago. So, when asked for today or yesterday's data, it returns error.
(2) Indonesian airport radiosondes have been displaying on Windy since the day after I asked, and has been consistently doing so. It's great, though only in fm35 format (fewer levels of altitude than fm94, shown for Malaysian & Singapore airports).
(3) But Philippines still no luck.
(4) Windy used to carry 2 weeks of archival radiosonde data and it was really useful for looking up day-to-day changes whenever the weather changes drastically. But since the last major change on Windy, it's been gone. Can we bring it back?
-
@hangInThere
(1) great, thank You(2) we are not filtering fm94 out so most likely only fm35 comes through data feeds we have
(3) I do see some ascents from Philippines, although quite old and sparse:
(4) That is still available. Here is an example of Indonesian Surabaya:
-
(4) Ah, I see, it's still there in desktop browser. I just checked again, not on mobile browser nor Android app.
-
@Filip_K Now the archival radiosonde data is gone from the desktop Windy too! Can you please restore it? And on the mobile Windy as well?
-
@Filip_K It just came back. Can you do the same for the mobile Windy?
-
@hangInThere
the issue with desktop is that sometimes the ascents are loaded, sometimes not. It is most likely a bug in the frontend. I've forwarded it to my colleges.As for mobile browser and app, I believe there has never been list of historical ascents available and as far as I know, it is not planned.
-
@Filip_K The mobile app always had the 14-day archival ascent data. It's what I had been using always, until the major revision of Windy Android app in 2024 wiped it away. So, if you look into your Android app before that major revision, the codes must be reusable. Try 'grep ascents -r' from the top of the code directories.
-
@hangInThere
It seems there is a bug that prevents historical ascents from being displayed. It should be fixed in the near future. Thank You for pointing it out.
