A radar artifact is any feature on a radar image caused by the sampling process, propagation effects, interference, or nonweather targets rather than by the atmospheric signal the operator wanted to measure. Artifacts are common, they are not evenly distributed, and they change with atmospheric conditions and time of day.
Common families of artifacts
Ground clutter is the most familiar. It comes from stationary surface targets such as terrain, buildings, and towers, and it typically appears as a bright, disorganized ring near the radar. Biological targets such as birds and insects add a soft, low-level echo pattern that swells at dawn and shrinks after sunset.
Anomalous propagation, often shortened to AP, happens when unusual vertical temperature or moisture structure bends the beam toward the ground. A strong surface inversion can duct the beam into terrain far from the radar, painting broad patches of fake precipitation. Second-trip echoes from range folding appear as stretched, grainy features displaced from where the real storm lives.
How forecasters identify them
Cross-checking is the core skill. A real echo generally looks similar on the same storm from a neighboring radar, on the tilt above, and on satellite or lightning imagery. An artifact usually fails at least one of those checks. AP tends to be pinned to terrain features and disappears when the inversion breaks down. Bird and insect returns cluster near the surface and vanish as the beam climbs.
Dual-polarization variables sharpen the diagnosis. The correlation coefficient drops for mixed or nonuniform targets, so biological echoes and ground clutter often stand out clearly as low correlation regions. This is one reason dual-pol was called the most significant upgrade to the national radar network since Doppler.
Where they mislead
An artifact treated as weather can drive bad decisions. A duct-broadened AP field can mimic a squall line, a range-folded second-trip echo can look like a nearby cell, and a bird bloom can hide a real thin line boundary near the surface.
The habit that avoids this is simple: before trusting any interesting feature, ask what family it might belong to, look for it on another radar or another tilt, and check the surface observations underneath. If the feature only exists on one product from one radar and does not match anything else, treat it as suspicious until it is confirmed.