17 Artifacts & Limitations
Radar lies sometimes. Knowing how it lies is the difference between a confident read and a dangerous mistake. This chapter covers the artifacts and limitations every radar user needs to know.
17.1 The beam rises and widens with range
The single most important limitation: the radar beam travels nearly straight while the Earth curves away, so the farther the storm, the higher the beam samples it, and the wider (lower-resolution) the sample volume becomes.
Consequences:
- Overshooting: distant storms have their low levels missed entirely. A tornado or shallow snow band can be invisible to a far radar.
- Underestimation: distant, shallow precipitation reads weaker than it is.
- Smearing: fine structure (couplets, hooks) blurs out at long range.
Mitigation: use the radar site closest to the storm, and weight close-range data more than distant data.
17.2 Velocity aliasing
Doppler radar can only unambiguously measure speeds up to its Nyquist velocity. Winds faster than that alias (or “fold”): a fast inbound velocity suddenly displays as outbound, and vice-versa, producing abrupt, unrealistic color reversals.
- How to spot it: a sharp jump from the strongest inbound color directly to the strongest outbound color, with no neutral zone between, often in a tight storm core.
- Why it matters: aliasing fakes huge velocity differences that can be mistaken for rotation, and it corrupts azimuthal shear.
- Mitigation: NeoWeatherwall can dealias velocity (including a 2-D dealiaser for Level 2 archive data) to unwrap the true winds. When viewing dealiased data, couplets and shear read true.
17.3 Range folding (the “purple haze”)
The radar can’t tell whether a returned pulse came from a nearby target or a more distant one beyond its maximum unambiguous range. Ambiguous returns get misplaced or flagged as no-data, often shown as a purple “RF” region. That region is a gap where velocity couldn’t be resolved, so treat it as missing data.
17.4 Ground clutter and non-meteorological returns
Lots of things that aren’t weather scatter microwaves:
- Ground clutter: terrain, towers, and buildings near the radar.
- Wind farms: spinning turbine blades produce noisy reflectivity and velocity.
- Biological targets: birds, bats, and insects (the source of fair-weather “clear-air” returns and the evening “roost rings”).
- Chaff, smoke, dust.
Mitigation: use dual-pol. Non-meteorological targets typically show very low correlation coefficient; see Dual-Polarization. If CC is junk, the echo probably isn’t weather.
17.5 The bright band (melting layer)
Where snow melts into rain, particles briefly become large, wet, and highly reflective, producing a ring of enhanced reflectivity (the “bright band”) on tilted scans. It can be mistaken for a ring of heavy rain. Lowered CC and a characteristic ZDR signature in the same ring give it away.
17.6 Three-body scatter & other reflectivity spikes
Very large hail can scatter energy to the ground and back, producing a three-body scatter spike (“flying eagle”), a spike of weak echo extending radially behind an intense core. It’s an artifact, but a useful one, because it flags large hail.
17.7 Beam blockage
Terrain can partially or fully block the beam in certain directions, leaving wedges of degraded or missing coverage. NeoWeatherwall’s Beam Blockage overlay (Radar Products) shows you where coverage is compromised so you don’t mistake a blocked sector for clear air.
17.8 Attenuation
Heavy rain and hail attenuate the beam. They absorb and scatter so much energy that storms behind them appear weaker than they really are. A strong cell can hide an even stronger one downrange.
17.9 Time is not frozen
A radar image is a volume scan that takes minutes to complete, and the frame you see is already a few minutes old by the time it reaches you. Fast-evolving storms can change between scans. Always:
- Check the timestamp.
- Loop frames (Animation) rather than trusting one image.
- Remember a couplet “now” on screen may be a few minutes in the past on the ground.
17.10 The bottom line
Radar is remote sensing with real physical limits. Cross-check products against each other, weight close-range data, dealias velocity, lean on dual-pol to reject junk, loop for trends, and always defer to official warnings and your own eyes-on safety.