Radar School › Lesson 6
LESSON 6When Radar Lies
You will stop chasing things that are not there. This is the lesson that separates competent from merely confident, and it is deliberately placed late because you cannot recognize a fake until you know what the real thing looks like.
Radar is a measurement built on a stack of assumptions, and the atmosphere breaks all of them routinely. None of this makes radar untrustworthy. It makes it a tool with a known error profile, which is a different and much more useful thing.
Nothing in this lesson is a storm. Everything here is something that looks like one and is not, which makes it the most quietly useful lesson in the course.
One beam, three lies
Everything in this section happens to the same beam from the same radar. Flip through the three modes: how the beam behaves on a normal day, and the two ways the atmosphere breaks it.
The first mode is the one to sit with. A storm 100 miles away on the lowest tilt is not being watched near the ground. The tornado is at the surface. The beam is nearly two miles above it.
- It does not move. Loop it. Echoes flicker but stay put. Real precipitation travels. The permanent blob of ground clutter hugging every radar site fails the same test.
- Velocity reads zero. Terrain is not going anywhere. This is the strongest single confirmation available.
- The pattern looks wrong. Blotchy and erratic rather than organized into cells or bands.
- Check the time. Overnight and early morning, clear skies, light wind. Classic AP conditions.
Tilt is the other half of the story
The beam illustration above showed that distance carries the beam higher. Tilt changes the angle of the beam itself.
The lowest tilt, usually around 0.5°, looks closest to the ground. Higher tilts slice progressively higher through the storm.
Distance + tilt determine how high you are sampling.
Radar apps may simply label these Tilt 1, Tilt 2, Tilt 3, and so on, with Tilt 1 being the lowest available slice.
For this course, the live radar has one radar site with tilt levels so you can try it yourself, on all three products you have already met.
Search "KMLB" or "Melbourne" in the site list, then open Products.
When velocity lies
Reflectivity's lies are mostly imposters, things that are not weather. Velocity fails differently. It has two failure modes, they look nothing alike on screen, and they mean opposite things.
Aliasing
Every radar has a maximum speed it can honestly measure. Exceed it and the reading does not max out, it wraps around to the opposite extreme. A screaming outbound wind gets reported as a screaming inbound one. It is the wagon wheel effect: wheels in old films appear to spin backward because the camera samples at a fixed rate, and the radar, comparing consecutive pulses, falls for the same trick.
A real transition from inbound to outbound has to pass through zero, so you see the intermediate colors and usually a gray zero line. Aliasing jumps from one end of the color bar to the other with none of the middle colors and no zero line.
The biggest clue is persistence. Real rotation usually lasts across several scans. Aliasing may appear for one scan and vanish on the next. One strange frame is a reason to check again, not a reason to trust it.
The reason this cannot simply be fixed is a trade. Firing radar beam pulses slowly gives echoes time to return from far away but loses fast motion. Firing rapidly catches fast motion but loses track of where distant echoes came from. You cannot have both, so modern radars alternate between rates and let software stitch the results together, which corrects most folds. Most, not all.
Range folding
The other half of that same trade. Sometimes the radar cannot place a return confidently at all, and instead of inventing an answer it marks those gates RF, the purple or gray patches on a velocity display. Those are not areas of no wind. They are areas of no information, and if one is sitting over the storm you care about, you are blind at that spot. Change tilt, or switch to a different radar site.
Aliasing shows you a velocity, but it has wrapped to the wrong end of the color scale.
RF is the radar declining to show you one at all.
Aliasing gives you a bad answer. RF tells you it does not have an answer.
Things that are not weather
Birds, bats, insects, smoke, chaff and the ground itself all produce real echoes, and the display paints them the same colors as rain. Most of it is anonymous speckle. The one worth knowing by shape is the roost ring: thousands of birds leaving a roost at dawn, climbing and spreading outward together, a ring expanding smoothly from a fixed point. No weather does that.
The shortcut that skips most of this
Everything above is detective work. Where does the echo sit, what time is it, does it move, what season is this. That is a real skill and it is worth having, because it works on any radar image from any decade.
There is also a button that just answers the question.
Dual-polarization radar compares the horizontal and vertical parts of each return. One of the products that falls out of that comparison is called correlation coefficient, usually labeled CC.
How similar are the targets inside this piece of sky?
Rain is a crowd of nearly identical drops, so CC comes back high, above about 0.95. Birds, bugs, terrain and smoke are a jumble of shapes and sizes, so CC comes back low, usually under 0.80.
Here it is on a real scan, and this one is a gift: two completely different things sitting side by side in a single frame.
So one habit is worth more than the rest of this lesson put together. Before you believe an echo, look at CC. High CC usually means similar targets. Low CC means something mixed or irregular. That might be birds, clutter or debris, but sometimes it is something happening inside real precipitation.
Change the view
The rest of this lesson's rogues, one button press apart. Start each one on reflectivity, decide what you would call it, then press CC. The fourth is not like the others.
That settles a bet the front page made. Of its four unlabelled images, two were never weather: the plume was a grass fire, and the ring was birds leaving their roost at dawn.
The lie that is real weather
The bright band is a winter effect caused by melting snowflakes. As snow begins to melt, it reflects much more strongly, which can create a ring of enhanced reflectivity centered on the radar.
On CC, that same ring often drops because the radar is seeing a mix of snow, partially melted flakes, and rain.
Low CC does not always mean fake. Sometimes it marks a real transition inside precipitation, like the melting layer.
Real or fooling you?
Ten scans, each fading between reflectivity and CC. Some are weather. Some are the radar seeing things that are not there.
A storm sits 100 miles from the radar and shows no rotation on the lowest tilt. What is the safest conclusion?
You cannot conclude much. Distance costs altitude and resolution at the same time. Absence of a signature far from the radar is not evidence of absence, which is the opposite of most people's instinct.
Blotchy echoes appear overnight on a clear calm night, they do not move between frames, and velocity reads near zero across them. What are you looking at?
Anomalous propagation. A temperature inversion bent the beam into the ground. The zero velocity is the giveaway, because terrain is not going anywhere.
Intense green sits directly against intense red with none of the middle colors and no gray zero line, and it is gone on the next scan. What is it?
Aliasing. Real transitions pass through the intermediate colors and through zero. A fold jumps the whole scale at once, and real circulations persist across scans while folds often do not.