search

dpaulat / supercell-wx

Supercell Wx is a free, open source advanced weather radar viewer.
https://supercell-wx.rtfd.io
MIT License
85 stars 14 forks source link

Level 2 Archive Missing Sweeps #185

Open acidhorse1972 opened 2 months ago

acidhorse1972 commented 2 months ago

RHO doesn't return all time frames when it should. Apparently it's chained to what time stamps are on the other products.

Here is a video of GR2Analyst showing more than what supercell-wx had shown: https://github.com/dpaulat/supercell-wx/assets/54511915/84d2d2c5-6de4-43f4-9039-f06e93be382b

At VT: 17:19 PT: 17:23 the RHO in the GIF displays the debris signature. At 17:25 it shows it growing. At VT: 17:25 PT: 17:27 it becomes disconnected.

RHO in supercell-wx shows 17:19 but jumps ahead over those other timestamps to (VT 17:25 PT 17:30) and misses that activity in between.

That location is at (30.918423 N , 88.278799 W) North of Georgetown AL and west of Chunchula AL.

All data came from KMOB.

dpaulat commented 2 months ago

I suspect you are seeing this because level 2 data is only loaded as complete files. The data is made available on the server in chunks, which I currently don't process.

Until I process partial chunks, you should at least be able to go back in time and view the scan that was initially not displayed. Is this also missing? If so, I could be indexing the file incorrectly.

acidhorse1972 commented 2 months ago

I suspect you are seeing this because level 2 data is only loaded as complete files. The data is made available on the server in chunks, which I currently don't process.

Until I process partial chunks, you should at least be able to go back in time and view the scan that was initially not displayed. Is this also missing? If so, I could be indexing the file incorrectly.

It seems its like this with any product. BR and BV where missing time frames that Gr2Analyst was getting from the K*** station. I'm looking at that NC Wilkes tornado and going backwards in time and there are timestamps missing.

dpaulat commented 2 months ago

I suspect you are seeing this because level 2 data is only loaded as complete files. The data is made available on the server in chunks, which I currently don't process. Until I process partial chunks, you should at least be able to go back in time and view the scan that was initially not displayed. Is this also missing? If so, I could be indexing the file incorrectly.

It seems its like this with any product. BR and BV where missing time frames that Gr2Analyst was getting from the K*** station. I'm looking at that NC Wilkes tornado and going backwards in time and there are timestamps missing.

This is what I'd expect for Level 2 only. Level 3 should refresh much faster.

acidhorse1972 commented 2 months ago

No this was data from Level 3 products

dpaulat commented 2 months ago

No this was data from Level 3 products

Good to know. I will investigate further.

dpaulat commented 2 months ago

I loaded the same set of data in Supercell Wx (KMOB, 2024-04-10, 17:30 UTC):

https://github.com/dpaulat/supercell-wx/assets/2886605/f21aed0c-f03a-4656-b814-6c675f776afd

Since the base tilt is less than 0.5 degrees, I switched each pane from tilt 2 (N0) to tilt 1 (NY). I saw each scan that was present in the GR2Analyst video you posted, although timestamps were slightly different. There was a single reflectivity product that did not properly load, due to the level 3 data on the server being corrupted.

I don't believe GR2Analyst had this problem, because my understanding is that it derives its products from level 2 data.

acidhorse1972 commented 2 months ago

RHO on yours is a Level 2 CC product but it didn't reproduce the same time frames that GR2Analyst (which derives its products from Level 2 data at https://mesonet-nexrad.agron.iastate.edu/level2/raw/ )

acidhorse1972 commented 2 months ago

Your Level 3 CC is N0C and is low resolution

dpaulat commented 2 months ago

It looks like I'm only using one sweep out of each level 2 file, instead of all of them (typically 2-3).