Is It a Bird?

Five problems, one conclusion

Where do you draw the line?

A radar receiver always hears something: its own electrical noise, plus echoes from the ground, weather and wildlife. Detection means deciding that a particular echo is strong enough to be a real target. That threshold sets the trade-off. Set it too low and "more real targets will be detected, but at the expense of increased numbers of false alarms." Set it high and you get the opposite.Wikipedia

Because the background changes from place to place, over hills, in rain or near a town, a fixed threshold produces too many alarms in some areas and misses targets in others. Most modern radars therefore use a constant false alarm rate (CFAR) detector. It estimates the local background from the cells around each one under test and sets the threshold just above it. Guard cells next to the test cell are left out, so a target's own energy does not raise its own bar.Wikipedia

CFAR keeps the alarm rate steady, but it does so by raising the bar wherever the background is busy. A small drone flying through clutter, exactly where Part 3 says it will be, faces the highest bar of all.

The false alarm that flies

Clutter can be mapped and filtered because it mostly stays put. Birds don't. They move, they fly at drone heights, and each one returns about as much energy as a small drone (Part 2). A single flock can put hundreds of drone-sized echoes in front of a radar at once.

That is a known, under-solved problem. A Sandia National Laboratories publication notes that radar can detect drones at more than 1 km, but that nuisance and false alarms from natural sources, birds in particular, have not been adequately dealt with.Sandia National Laboratories Research presented at SPIE in 2023 describes counter-drone radars being swamped by bird tracks. It reports that machine learning on track behaviour (how a target moves over time) can recognise target drones with better than 90% sensitivity and specificity.SPIE

Better than 90% is impressive in a trial, but a sensor that misjudges one target in ten will still miss drones and raise false alarms when hundreds of birds pass by. Classification helps, and it works best with a second, independent piece of evidence.

Every bird is a test the radar has to pass. Over a flock, thousands of tests a minute.

Layering sensors that fail differently

Each sensor family fails in a different way, which is exactly why they work together:

The U.S. military's own counter-drone vehicles are built this way. The M-LIDS system pairs a Ku-band radar with a 30 mm gun and Coyote interceptor drones on one vehicle.Wikimedia Commons The U.S. Defense Innovation Unit's 2025 low-cost counter-drone sensing challenge named acoustic, passive RF, radar, optical, infrared and multi-sensor technologies together, with a target of cutting ownership costs by 50–80%.The Defense Post

Ukraine's air defence shows the payoff at national scale. Its passive acoustic networks are fused with radar data. According to United24, about 20% of targets entering Ukrainian airspace are tracked by acoustic data alone, which is coverage radar was not providing.United24 Media

What five weeks of radar physics add up to

Across the five parts, the same conclusion keeps returning:

1. Small targets shrink radar range with the fourth root of echo size (Part 1).
2. Small drones are bird-sized echoes, and plastic or foam makes them smaller still (Part 2).
3. Flying low hides them behind the Earth, terrain and ground clutter (Part 3).
4. Hovering can defeat Doppler filtering, leaving only a faint rotor signature (Part 4).
5. Thresholds cannot separate them from birds without another source of evidence (Part 5).

None of this makes radar obsolete. It remains the best sensor for range, speed and all-weather tracking, and nothing else gives the same air picture. But against small, cheap, low drones, radar is one layer of a stack, not the whole defence.