Welcome New Members! We want to hear from you. Register, stop lurking and start posting!

artefacts on weather radar since 2017/8

Archived topics from the General Weather Discussion board.
Locked
h
hotnbothered
New User
Reactions:
Posts: 6
Joined: Wed Feb 08, 2017 2:56 pm

artefacts on weather radar since 2017/8

Post by hotnbothered »

Hi
Do forum members know anyone in the Met Bureau who is knowledgeable about the tech aspects of the rain radar?
I'd really love to hear the a detailed tech explanation as to why the Laverton radar is prone to showing circular artefacts in its colour coding of the expected precipitation.

Up until late 2017, the Laverton radar was showing some consistent white radial artefacts, and I posted a query about them in 2017.
Then in late 2017 the radar was upgraded. I gather the beam polarisation was changed, and the new radar can also detect horizontal rain velocity.
The radial artefacts disappeared: but instead, we often see annular or concentric artefacts in which the colour coding of the expected precipitation varies abruptly at certain distances from the radar. This is much worse on some days than others. As an example, the left image below is from a couple of days ago. This is taken from the normal loop radar website at http://www.bom.gov.au/products/IDR024.loop.shtml#skip, although it looks different because I have turned off the Locations, Range and Topography features. If you do this, you can download just rain intensity overlay, as a PNG file with transparency.
IDR024.T.202009241214.png
IDR024.T.202009241214.png (10.1 KiB) Viewed 2060 times

You can see that there is a sudden change in the coding of rain intensity at a fixed radius. On many rainy days you can see multiple bands, all concentric on Laverton, as shown on the next recent image (from last week).
IDR024.T.202009111349.png
IDR024.T.202009111349.png (16.49 KiB) Viewed 2060 times
This is a gross artefact. You can see that the colour coding is highly unreliable at a given ground location, and may be anywhere between light blue (~2 mm/hr) and perhaps red (120mm/hr), although the context shows this is impossible, and it is clearly not just due to cloud movement. The old radial artefact affected very little of the image, but this one, at its worst, makes most of the image unreliable. Often it looks like a periodic radial saw-tooth function has been applied to the raw data values. It seems to be worse recently, but I have noticed its presence many times since the "upgrade" in 2017, and I would really like to know why it happens, and why it hasn't been fixed after 2 years.
User avatar
hillybilly
Site Admin/Moderator
Reactions:
Posts: 4964
Joined: Thu Nov 26, 2009 7:26 am
Location: Howden Tasmania, 25m above sea level
Contact:

Re: artefacts on weather radar since 2017/8

Post by hillybilly »

It is related to scans further from the radar being higher (due to earth curvature) and more attenuated (radar energy lost as it passes through more and more rain). One has to make decisions around post processing to compensate. Different radars and different processing will give different features.

I haven’t noticed it being better or worse, but will keep an eye out. Many of the radars have been upgraded.
h
hotnbothered
New User
Reactions:
Posts: 6
Joined: Wed Feb 08, 2017 2:56 pm

Re: artefacts on weather radar since 2017/8

Post by hotnbothered »

Thanks, hillybilly. If it's due to post-processing and not a hardware limitation, It's hard to see why a smooth transform hasn't been used. Is this ad hoc & in-house, do you think, or a published algorithm?
Locked