What is a microburst?
Fujita’s downburst family, how wet and dry microbursts form, what Delta 191 changed, and why the signature so often sits below the WSR-88D beam.
A microburst is a small, intense downburst: sinking air from a thunderstorm or shower that hits the ground and spreads outward in damaging straight-line winds over an area no more than about 4 km (2.5 miles) across. It is too small and brief for most radars to sample well, which is why it remains hard to warn.
The squall lines, bow echoes and derechos guide covers wind systems hundreds of kilometres long. This page covers the smallest unit of that damage: one downdraft reaching the surface.
Downburst, macroburst, microburst
The vocabulary is T. Theodore Fujita’s. After analysing the 1975 Eastern Air Lines Flight 66 crash at New York’s JFK airport, he proposed that a small, violent thunderstorm outflow, not yet observed by anyone, had brought the aircraft down. In his 1981 paper in the Journal of the Atmospheric Sciences he split downbursts by size at 4 km.
| Term | Size of outflow | Peak winds last |
|---|---|---|
| Downburst | < 1 to 400 km | — |
| Macroburst | > 4 km along a side | 5 to 20 minutes |
| Microburst | < 4 km along a side | 2 to 5 minutes |
Those figures are from the NWS Warning Decision Training Division glossary, which draws on the AMS Glossary of Meteorology. It adds that a microburst’s velocity difference across the divergence centre exceeds 10 m/s. The duration is of peak winds, not of the whole event: NSSL describes microbursts as lasting five to ten minutes, with winds sometimes above 100 mph, and NCAR’s JAWS observations put the typical life cycle, from downdraft to dissipation, at 15–20 minutes.
How a microburst forms
A downdraft accelerates when the air in it is colder and heavier than its surroundings. NWS Louisville’s training note on damaging downbursts names three processes that do that work:
- Precipitation loading. A large, heavy core of rain and hail drags air down with it as it falls.
- Evaporative cooling. Rain falling into dry air evaporates and chills it. The same note stresses that dry air near 700 mb matters more than dry air near 500 mb, where it is too cold to evaporate much.
- Melting. Hail and graupel melting below the freezing level absorb heat and cool the air further.
When that cold air reaches the ground, it spreads out in every direction. That outward rush is the microburst.
Wet, dry and hybrid
The JAWS project near Denver in 1982 found microbursts under virga shafts from high-based clouds; the MIST experiment in the Southeast in 1986 studied the rain-heavy kind. The WDTD glossary defines the classes:
- Wet microbursts come with heavy rain at the surface. An outward bulge at the bottom of the rain shaft, a “rain foot,” is the common visual sign.
- Dry microbursts produce little or no rain at the ground. A ring of blowing dust beneath virga is the common sign. They favor deep, dry sub-cloud layers and high cloud bases, typical of the High Plains and intermountain West.
- Hybrid microbursts share traits of both.
For wet microbursts, Atkins and Wakimoto (1991) found a difference of more than 20 K between the highest near-surface equivalent potential temperature and the mid-level minimum was a good discriminator of microburst days in the MIST data.
Delta 191 and the aviation response
On 2 August 1985, Delta Air Lines Flight 191, a Lockheed L-1011, flew through the rain shaft of a thunderstorm on final approach to runway 17L at Dallas/Fort Worth and crashed short of the runway. The FAA’s account records 137 deaths, including the driver of a car on the highway it struck. It puts the microburst’s horizontal wind differential at 73 knots: headwind first, then a downdraft, then a tailwind the airplane could not escape. The NTSB (report AAR-86/05) cited the crew’s decision to continue into a storm with visible lightning, inadequate wind shear training, and “the lack of definitive, real-time wind shear hazard information.”
The response came on three fronts:
- LLWAS was already there. The FAA developed the Low-Level Windshear Alert System, a network of anemometer poles, in 1977 after Eastern 66. The FAA concluded that the DFW system could not have given Flight 191 a timely warning.
- TDWR was accelerated. A 2022 BAMS history by McCarthy and colleagues records that the FAA formally started the Terminal Doppler Weather Radar acquisition in 1985, demonstrated automatic microburst warnings at Denver in 1988, and bought 45 production radars. The FAA dates TDWR service from 1994.
- Aircraft were equipped. In 1988 the FAA required airborne wind shear warning or detection systems on air carrier turbine aircraft; forward-looking predictive systems entered airline service in 1994.
The FAA states that no wind shear accidents have occurred at any TDWR-protected airport since commissioning. The last wind shear accident it lists, USAir 1016 at Charlotte on 2 July 1994, happened before that airport’s TDWR was operating.
What it leaves on the ground
NWS La Crosse’s damage-survey guidance describes downburst damage as often divergent: an impact point with debris fanned outward downwind, sometimes in separate footprint-like pockets. Tornado damage is more often convergent, with debris laid toward a central axis along a narrow path. The same page warns that terrain, tree type and rot can scramble either pattern, so the direction trees fell is evidence, not proof.
How a microburst appears on radar
The NWS Louisville note lists precursors that tend to come before damaging downburst winds:
- A high reflectivity core that starts higher than in nearby storms.
- Strong, deep mid-altitude radial convergence (MARC), roughly 2–6 km above ground, with inbound-plus-outbound differences of 50 knots or more often seen.
- A rapidly descending reflectivity core.
- To a lesser degree, mid-level rotation and strong storm-top divergence.
The microburst itself is a divergence signature at the lowest tilt: inbound velocities on the side nearer the radar and outbound on the far side, with no rotation between them. Wilson et al. (1984), from JAWS Doppler data, found the outflow begins spreading below 1 km, with the strongest velocity difference about 75 m above ground.
That is the problem for the WSR-88D. By a 4/3-earth calculation (beam centre only, ignoring antenna height), the 0.5° beam is about 580 m up at 50 km range and passes 1 km near 76 km. Beyond that, the beam can overshoot the outflow entirely. The NWS Fort Worth office shows this with a 9 June 2015 Dallas microburst that produced a 58-knot gust: the Love Field TDWR showed a classic divergence in one-minute data, while the KFWS WSR-88D, farther away and updating every few minutes, showed a less defined pattern.
TDWR is built for this. NCEI lists a 5 cm wavelength, range gates of 0.067 nautical miles (about 120 m) and one-minute updates at the lowest elevation in hazardous mode. The trade is a short 90 km Doppler range and a maximum unambiguous velocity of 20–30 knots, so strong outflow folds readily.
Why microbursts are hard to warn
Witt and colleagues at NSSL, in a 2013 AMS radar conference paper, note that microbursts usually form and dissipate within 20 minutes and can develop in under 5. The fastest standard WSR-88D volume scan they cite takes about 4.2 minutes. A WSR-88D downburst algorithm they summarize, the DDPDA, had a median lead time of 5.5 minutes at 20–45 km from the radar, and 0 minutes at 45–80 km. The Louisville note adds that separating a microburst-producing cell from an ordinary one is difficult, and that heavy evaporation can occur without visibly changing the echo. In a bowing line, those downbursts are organized by a rear-inflow jet; see bow echoes on radar.
WeatherStory Radar decodes reflectivity and velocity from raw Level II and gives the beam height of any gate you click.