I often see this during winter storms. If it is sleeting instead of rain or snow, it doesn’t have to be coming down hard for the radar colors to look like those of a severe thunderstorm. Why does this happen?
WAG on my part, but my first thought is that sleet probably provides a harder return than rain or snow.
Until very recently weather radar could not directly detect sleet. The radar waves had only single polarization and the return data varied only in intensity and doppler shift. Any depiction of sleet was basically a graphical guess.
Starting around 2011 (in the US) the nationwide network of WSR-88D weather radars were gradually updated to dual polarization which provides separate vertically and horizontally polarized waves. This allows measuring the aspect ratio, or length vs height of precipitation particles. This differs between rain, sleet and snow, so for the first time various types of precip could be directly measured.
One of the radar products returned by WSR-88D is “hydrometeor classification”, which is a built-in algorithm to differentiate between rain, sleet, snow, debris and biologic targets. It is not 100% accurate but gives a rough idea.
Since radar waves travel in a straight line, due to earth’s curvature, the further from the radar station the higher in the atmosphere the signal will be. At higher altitudes sleet may exist which melts to rain at lower altitudes (which may be below the radar “horizon” at that distance). This can often show large circular bands of sleet which only exist in the upper atmosphere.
More info (PDF): http://www.weather.gov/media/lmk/soo/Dual_Pol_Overview.pdf
Hail forms in thunderstorms. Perhaps it was hail and the intensity was due to the (heightwise) depth/thickness and ripeness of the cumulus nimbus clouds… hail is upwards rain … the water drops are blown up and frozen and they even cycle around ,up down, up , down… before falling to ground.
Do you have an example? I’ve never seen that. The local weather people can choose how to display whatever data they’re showing. If they’re using the same colors they use for rain storms, that’s just dumb.
Well, there’s only so many colors. There seems to be some overlap in the purple/violet section of the spectrum which is shared by severe rain storms and moderate ice/sleet. I noticed this myself the other week when we were treated to a severe storm and I was trying to guess if the bright purple mass headed my way was a ton of rain or a lot of ice (turns out it was both – we got hail)
The weather people don’t really get to pick from the radar since I assume the radar colors are largely standardized and they may not own the radar at all.
Is this why maps showing multiple radar sites at night often appear to show precip clustered around each radar site? i.e. they’re just catching fog at low altitude near the radar, and clean air at high altitude farther away?
Ah, but they can change the colors. The data from NWS radar comes as numbers. From there, radar viewing software assigns a color scale. Radar programs available to the public such as GRLevelX and Pykl3 (the only two I’m familiar with) allow you to change the way the data is displayed by customizing the color profile with a simple text file that lists the DBZ value and the RGB color you want to assign to it. There’s some examples here under “color curves” GRLevelX Tools
Actually, hail is a problem because it’s so non-reflective to radar. At lower altitudes where the hail has a sheen of water on it the water creates some return. A thigh altitudes where it’s all frozen, there’s almost zero return from hail.
The pretty colors you see on TV are heavily “improved” by artistic license working from the already heavily algorithmically enhanced pictures coming out of the radar system itself. Which are themselves a very long ways away from the raw radar reflectivity values of 1950s weather radar.
I’ll take your word on it. Problem still remains that there’s only so many colors – the example I posted uses the entire gamut, even with the double-usage of the purples. You could delve into more shades of each color but I assume you want the differences to be easily noticed.
I suppose you could come up with a unique array for every time you show a radar image but it’s probably more useful if the casual viewer can look at the TV and see “Dark green, okay so it’ll be rainy but not bad” rather than deciphering the image from scratch each day. Or, worse yet, in events of severe weather.
With one exception, there is no mapping of color to precipitation type in the data returned from the NEXRAD system. That exception is “hydrometeor classification”. In that mode it does not show reflectivity but only precip type: http://fj9672r62qdly8wm38b4do8d.wpengine.netdna-cdn.com/wp-content/uploads/2015/01/2015-04-24-23.43.54.jpg
As already explained the radar data product “reflectivity” does not and cannot specifically detect sleet. There is a mapping of reflection strength in dBZ to color: dBZ (meteorology) - Wikipedia) This radar product is purely signal strength – it does not show type of precip.
From 2011 to 2013, the US network of WSR-88D NEXRAD weather radars were upgraded to dual polarization which for the first time enabled direct detection of sleet vs snow vs rain: http://www.nws.noaa.gov/com/weatherreadynation/news/130425_dualpol.html
However, “dual pol” simply provides the raw data – differentiating these precip types requires algorithmic interpretation. That can be done by (a) accepting the built-in interpretation of the radar datastream which is “hydrometeor classification”, or (b) using custom software on the receiving end to combine various radar products on a single screen. How this is done is up to the individual end user software.
Some smaller local market TV stations simply use Gibson Ridge GRLevelX: http://www.grlevelx.com/
Larger TV stations and web sites may write custom software to interpret and blend the various radar data products. A few stations may have their own commercial weather radars.
So if a particular weather provider is showing precip type and reflectivity simultaneously on the same screen and trying to map both to a single color code – it’s because they wrote software for this. Any deficiency or ambiguity is because of their decisions in how to display this.
Most web-based and TV weather radar is simplified and does not remotely show the underlying detail available. In their attempts to simplify the display for a broad audience, some may introduce ambiguities like color codes which overlap between reflectivity and precip type. If anyone is interested in seeing the real story, get a radar app like GRLevelX, RadarScope or PYKL3.
Sophisticated radar apps like GRLevelx and RadarScope allow selection of individual radar products (reflectivity, velocity, echo tops, etc), and the tilt angle of each radar site. They can display multiple radar products two-up or four-up on the same screen, each with its own unique color code.
These can also display FAA terminal doppler radar sites. GRLevelX is Windows only, RadarScope is available on Mac, iOS and Android but unfortunately not Windows: https://radarscope.io/
We get various shades of blue for snow, greens for rain going to yellow, red, and purple. Pink denotes freezing rain and sleet.
On top of the excellent technical descriptions provided by joema and others, this seems to be a problem of the color scales that have become standard.
Green-yellow-red-brown-pink is a terrible color scheme for precipitation intensity. It’s not useful for anyone with color blindness. For people with normal color vision the transitions between color aren’t perceived evenly. E.g., going from the map linked by Jophiel, the dark green to yellow transition looks like a much bigger deal than dark green to darker green. That scheme also wastes multiple color channels that could be used to indicate different types of precipitation.
Of course, since precipitation color schemes are now traditional and expected by everyone watching the 6 o’clock local news, you don’t want to confuse Grandma by introducing new and technically superior color schemes. But if we were starting from scratch, it might make sense to assign a single hue to each class of participation, and use a simple intensity scale for each. Say, purple-blue for rain, orange-red for snow, and yellow-green for sleet. People with deuteranopia still won’t be able to perceive all the information, but they will get everything except the distinction between snow and sleet (which are probably the least consequential things to confuse). Then, we could consider whether there should be a linear mapping from some radar measurement to color, or a non-linear curve that reflects precipitation intensity according to human perception, risk to drivers, etc.
I didn’t see this earlier, but here’s what that is. National Weather Service radars operate in two modes, “precipitation mode” and “clean air mode.” In precipitation mode, the radar does a complete “volume scan” in 6 minutes. A volume scan is when the radar does a complete rotation at several different angles in order to scan almost the whole volume of air surrounding the radar. In clean air mode, it takes 10 minutes. It rotates slower, with more power and with greater sensitivity so meteorologists can pick out fainter information like light precipitation, frontal boundaries or light snow, which is hard to pick up on radar.
The reason you see all that green is that it’s not raining so the radar has been put into clean air mode and is picking up everything from birds and insects to dust in the air. (out west, they often see bats on radar as millions of them leave their caves at night to feed video) The reason you don’t see any green in the distance is that the radar isn’t powerful enough to pick them up that far away and/or due to the curvature of the earth the radar can’t see that low at a distance.
The blank circle in the middle of the green stuff is the radar’s “cone of silence” which is the result of the simple fact that the radars can’t be pointed straight up. The very center of that circle is where the radar is located.
Yes, this is a good explanation. A volume scan pattern can be seen here. Due to earth’s curvature and since the radar elevation can only scan so high and so low, there can be some gaps and distance-related artifacts:
http://weather.uprm.edu/newinterface/images/earth_curvature.jpg
This can especially be seen in the dual-polarization “hydrometeor classification” product, which often depicts circular bands of precip. This is due to the conical volume scan pattern, not because a precip boundary exists at that terrestrial distance: https://photos.smugmug.com/photos/i-3hJWMKG/0/XL/i-3hJWMKG-XL.jpg
Also some radar apps will suppress the clutter from clear air mode, so there can be a difference in how the same radar data is depicted from different apps. In RadarScope, “expert mode” disables this filtering.
Starting in 2015 the NEXRAD system was updated with a high-speed scan pattern called SAILS, which can complete a volume scan as fast as 90 seconds. SAILS is generally used only in severe weather to detect rapidly-developing storms. However an update in 2017 will enable it in clear-air mode: http://www.weather.gov/gsp/sails
Starting in 2008 some NEXRAD radar products were upgraded to “super resolution”, however I believe this is only available on Level II data which most radar apps and sites do not show. It is available in RadarScope and in GRLevel2:
Standard vs super-resolution: https://photos.smugmug.com/photos/i-ZfXjFNj/0/O/i-ZfXjFNj.jpg
Du Page College has a good radar web page, under “NEXRAD Sites”: COD: Satellite and Radar
Great site! Thanks for posting it!