Radar guides

Weather radar colors: what they mean

Reflectivity in dBZ, the approximate rain rates behind the greens and reds, why the same color means different things in different apps, and how to read velocity, CC and ZDR color scales.

On a standard radar map, the colors stand for reflectivity: how much of the radar’s pulse bounced back, measured in dBZ, with cool colors for weak returns and warm colors for strong ones. The color is a label someone chose for a number. The number is what the radar measured.

The how to read weather radar guide covers the three kinds of data a radar returns and what the radar cannot see. This page stays on the colors: what a given dBZ implies about rain rate, where that relationship breaks, and how to read the other color scales you will meet.

What dBZ is

Reflectivity factor, Z, has awkward units (mm6/m3) and a huge range. NOAA’s JetStream reflectivity page gives the example of tiny droplets near 0.001 and hail in a severe storm near 36,000,000 in those units. Taking a logarithm compresses that into decibels, dBZ, and JetStream puts the practical scale at about −35 to +85 dBZ. Every 10 dBZ is a tenfold step in Z.

The sixth power is the key fact. Returned power scales with drop diameter to the sixth power, so a few big targets can outshine many small ones. JetStream’s illustration: a cubic metre holding 729 drops of 1 mm and a cubic metre holding a single 3 mm drop both return 29 dBZ, yet the first is raining about 22 times harder. Color tracks returned power, not water.

The conventional scale, in general terms

There is no universal standard color table. Level II radar data, as NOAA’s National Centers for Environmental Information describes it, is digital radial base data: numbers per gate, with no colors attached. Every display maps those numbers to colors with its own table.

The familiar convention is loose: blues and greens for light returns, yellow and orange in the middle, reds for strong returns, and magenta or purple and white at the top. JetStream’s general guidelines attach meaning to ranges rather than hues: 20 to 40 dBZ light precipitation, 40 to 50 moderate, 50 to 65 heavy precipitation or some hail, and above 65 extremely heavy precipitation including water-coated hail.

dBZ to rain rate: one relation, many answers

To turn Z into a rain rate R, you need a Z–R relation, and there are many. The WSR-88D default is Z = 300R1.4, which replaced the Marshall–Palmer relation Z = 200R1.6 used with the older WSR-57 network, per a 1999 NWS Tallahassee paper by Jeffrey Fournier. The same paper notes the tropical alternative offices can switch to, Z = 250R1.2, because the default underestimates tropical rain.

The table below is a standard-relation calculation from Z = 300R1.4, rounded, beside the table JetStream publishes. The JetStream column matches Marshall–Palmer to within rounding (our arithmetic), which is itself the point: two NOAA sources, two answers for the same color.

dBZZ = 300R1.4 (mm/h)Z = 300R1.4 (in/h)JetStream table (in/h)
200.50.020.01 (trace)
302.40.090.10
355.40.210.22
40120.480.45
45281.10.92
50632.51.90
531044.1—

The table stops at 53 dBZ on purpose. The WSR-88D precipitation processing described in an NWS Morristown paper by Steven Hunter ignores reflectivity above an adaptable threshold, set at 53 dBZ, on the supposition that such returns are hail rather than rain; the same paper notes Denver used 50 and Melbourne planned 55. Hunter also calls the default relation a compromise between stratiform and convective rain.

Real storms stray further. JetStream shows a 40–50 dBZ cell near Fort Worth in April 2012 that produced more than 4.5 inches of rain in 90 minutes, and Hurricane Ida at landfall on 29 August 2021 showing mostly 35–45 dBZ while 10 to 15 inches fell across southeast Louisiana. Small tropical drops return little power for a lot of water.

50+ dBZ and hail

Hail is large, and wet hail is very reflective, so the reds and purples above 50 dBZ are where hail becomes a real possibility. Reflectivity alone cannot confirm it. Differential reflectivity is the usual check: JetStream shows the 11 April 2016 Wylie, Texas supercell with reflectivity above 65 dBZ and ZDR near 0 dB in the same area, the tumbling-hail signature; that storm produced hailstones up to 5.25 inches, per NWS Fort Worth. For how hail cores sit inside a rotating storm, see the supercell, on radar.

The bright band

In widespread steady precipitation you will sometimes see a ring of stronger color around the radar that does not move like a storm. That is often the bright band. The NWS Warning Decision Training Division defines it as a narrow horizontal layer of higher reflectivity where snow melts to rain. Melting flakes are still large but coated in water, so they return far more power than the snow above or the rain below. Because the beam climbs with range, a tilt crosses that layer at roughly one distance all the way around, which draws the ring.

It is a color that means the wrong thing. Hunter’s paper cites a 5–10 dB enhancement in most studies, enough that rain rate can be overestimated up to five times, and WDTD warns that a bright band in the lowest tilts inflates precipitation estimates.

Velocity colors

Velocity uses a different convention. On NWS displays, per JetStream, green is motion toward the radar with negative values and red is motion away with positive values; brighter shades mean faster radial speed. The sign convention dates to early Doppler radars that pointed straight up, where downward motion was negative. Tables often run past green and red: JetStream’s bow echo example marks inbound winds above 90 mph in blue. What matters is not a single color but green and red packed together, which is how a mesocyclone first shows up.

CC and ZDR color tables

Dual-polarization products get their own scales, and again the values carry the meaning.

  • Correlation coefficient runs from 0 to 1. JetStream groups it as high (above 0.97) for all rain or all snow, moderate (0.80 to 0.97) for hail, water-coated hail and rain/snow or sleet/snow mixes, and low (below 0.80) for non-meteorological targets such as birds, bugs, buildings and smoke.
  • Differential reflectivity is in dB and centred on zero. Round targets such as small drops, sleet and tumbling hail sit near 0 dB; flattened large raindrops are positive; vertically oriented ice crystals can be negative, per JetStream.

WeatherStory Radar decodes reflectivity, velocity, ZDR and correlation coefficient from raw Level II and lets you click any gate for its exact value instead of judging a color.

The other guides

  • How to read weather radarWhat the colors mean, why velocity matters more than reflectivity, and what the radar physically cannot see.
  • What is NEXRAD?The US network of WSR-88D Doppler radars: how the radars scan, what Level II and Level III data are, how the network was built and upgraded, and where it cannot see.
  • Correlation coefficient on radarWhat CC measures, the values rain, snow, hail, smoke and birds actually produce, and the artifacts that push it low when nothing unusual is there.
  • How to read radar velocityWhat Doppler velocity actually measures, how the zero line reveals the wind profile, why data folds and aliases, when storm-relative velocity helps, and how to tell a rotation couplet from convergence.
  • What is virga, and why radar shows rain that never landsPrecipitation that evaporates before it reaches the ground: where it forms, why the radar beam sees it while the surface stays dry, how it links to dry microbursts, and the clues that separate it from rain at the ground.
  • What is a cumulonimbus cloud?The thunderstorm cloud, defined: how it grows out of towering cumulus, why its anvil flattens near the tropopause, what an overshooting top says about the updraft, and why radar sees the rain inside it but not the cloud.
  • Mammatus clouds: what they are and why they formPouch-like lobes hanging from the underside of a thunderstorm anvil. What the definitions actually say, the ten proposed formation mechanisms, and why radar does not see them.
  • Tornadoes on radarHook echoes, velocity couplets and debris signatures — and how confident each one should make you.
  • The supercell, on radarMesocyclone, BWER, rear-flank downdraft — what a rotating updraft does to the data.
  • Squall lines, bow echoes and derechosOrganized linear convection, why bowing means damaging wind, and what actually qualifies as a derecho.
  • Outlooks and forecast modelsWhat the SPC categories mean, and the most common way people misread a convection-allowing model.

Figures in these guides are unmodified output from WeatherStory Radar, which reads Level II straight from NOAA’s public archive at full resolution. Nothing is thresholded, smoothed or decimated.

Privacy · Support