What a radar actually measures
A ground weather radar transmits a short microwave pulse and listens for the fraction scattered back by whatever is in the sampled volume. The US network radar, the WSR-88D, works at about 10 cm wavelength from an 8.5 m antenna, giving a beam roughly 0.95 degrees wide.
The returned power becomes the reflectivity factor Z. Because Z spans something like seven orders of magnitude it is expressed logarithmically, in dBZ. The single most important property, and the one that explains most misreadings: Z is proportional to the sum of the sixth power of drop diameters. A few large drops return vastly more energy than many small ones. dBZ therefore indicates precipitation intensity and particle size. It is not a rainfall rate, and it is emphatically not turbulence, icing or lightning.
NOAA's own interpretation of the scale: roughly 20 to 40 dBZ is light precipitation, 40 to 50 moderate, 50 to 65 heavy precipitation or hail, and above 65 extremely heavy, typically water-coated hail. Below about 15 to 20 dBZ you are often looking at drizzle, virga that never reaches the ground, or something that is not weather at all.
In broad terms on this map, blues and greens are weak returns, yellow and orange moderate, red and magenta the strongest returns consistent with heavy rain or hail.
Base versus composite, and why a mosaic is neither
Base reflectivity comes from a single elevation scan, usually the lowest. It is closest to what is actually happening near the surface and it is available as soon as that tilt finishes.
Composite reflectivity takes the maximum reflectivity found at any elevation for each grid box across a complete volume scan. It answers "what is the strongest echo anywhere in this column", which is better for judging storm intensity, at the cost of looking like far more widespread precipitation than is actually falling, and of only existing once the whole volume is done.
A worldwide mosaic is neither cleanly. National networks do not all contribute the same product: some send lowest-tilt reflectivity, some a maximum or composite field, and compositing rules differ between providers. The correct statement is that this layer is a blend of national radar products, not a single uniform physical quantity, and that intensity can step discontinuously where two networks meet.
Why the picture has holes and artifacts
Every one of these is standard radar interpretation, and all of them are visible on this layer if you know to look.
- Beam blockage. Mountains and nearby structures block the lower beams, leaving permanent wedges of reduced sensitivity radiating away from a radar site.
- Cone of silence. Operational scan strategies do not look above about 19.5 degrees elevation, so an inverted cone directly over each radar is never sampled. On a mosaic that is a small hole or an artificially weak patch right over the site.
- Overshooting with range. The beam climbs away from the surface with distance, both geometrically and through atmospheric refraction. At a couple of hundred kilometers the beam center is several kilometers up. Shallow rain and snow are missed entirely at long range, and distant echoes describe the middle or top of a storm rather than anything near the ground.
- Anomalous propagation. Strong low-level temperature inversions bend the beam down into the ground and generate broad false echoes, classically on calm clear nights. It can look convincingly like widespread light rain.
- Ground clutter and biological targets. Buildings, terrain, wind farms, aircraft, chaff, birds, bats and insects all scatter. Clutter filtering is applied but is imperfect, and also removes some genuine weak returns. Networks filter differently, which is itself a source of seams.
- Bright band. Where the beam crosses the melting level, wet snowflakes return disproportionately strongly, producing a ring of exaggerated intensity around a radar. Worth knowing for aviation readers, because that is the same altitude band where icing risk concentrates.
- Attenuation and range folding. Heavy precipitation weakens the signal behind it, so one cell can hide another; second-trip echoes can be placed at the wrong range entirely.
How fresh and how sharp it really is
The layer always draws the newest available past frame, on 10 minute steps. That is not the whole latency story. A single radar's volume scan takes roughly 4 to 10 minutes depending on the scan strategy in use, and different national networks run on their own cycles of typically 5 to 15 minutes. A mosaic frame is therefore not a synchronous snapshot: two adjacent regions can be several minutes apart in validity. Add source and processing latency and the picture on screen is realistically 10 to 20 minutes behind the sky. Convective cells grow and move meaningfully on that timescale.
On sharpness, the free tier serves native tiles only to zoom 7. Above that the map is interpolating. Apparent detail at high zoom is upsampling, not data. Display smoothing is also enabled, which makes the field more legible and simultaneously smears sharp gradients such as a gust front or a storm edge.
What this layer does not tell you
Every data source has an edge. These are the ones that matter for reading this layer correctly.
- It shows precipitation echo, not rain at the ground, and never turbulence, icing or lightning.
- White or empty can mean no radar coverage rather than no precipitation. There is effectively nothing over open ocean and large gaps across much of Africa, Central Asia, the Arctic and parts of South America and Southeast Asia.
- The frame is 10 to 20 minutes old once source latency is counted, and different parts of it have different validity times.
- Native data stops at zoom 7. Anything sharper on screen is interpolation.
- Echoes very close to a radar and at long range from one are the least trustworthy parts of the picture.
- Intensity is not calibrated consistently between national networks, so comparisons across a seam are unsafe.
AeroScope is an situational awareness and research tool. Nothing here is approved for air traffic control, navigation, flight planning or collision avoidance. For operational decisions, use the authoritative source named above.