Bow echoes on radar

A bow echo is a line or cluster of thunderstorms that bends outward into a bow or crescent shape on radar because strong outflow winds are pushing its center ahead of its ends. The shape is drawn by wind reaching the ground.

The overview in squall lines, bow echoes and derechos covers where bows sit within organized convection. This page goes one level down: how a bow is defined, how it evolves, and which parts of the radar picture carry the most information.

Where the term comes from

The bow echo was first named and described in detail by T. Theodore Fujita in his 1978 Manual of Downburst Identification for Project NIMROD, written for a field study of damaging thunderstorm winds in northern Illinois. Fujita picked out two radar echo shapes as especially associated with downbursts: the hook echo, already tied to tornadic supercells, and a bow-shaped line of cells. Morris Weisman’s 2001 review in the Bulletin of the American Meteorological Society, “Bow Echoes: A Tribute to T. T. Fujita,” credits him with the name and the conceptual model that research still builds on.

The working definition is about shape and behavior, not size. Klimowski et al. (2004), studying 273 bow echoes across the United States from 1996 to 2002, define one as a non-transient bow- or crescent-shaped radar signature with a tight reflectivity gradient on its leading edge, consistent with storms propagating along a strong outflow. They note bows have been observed from roughly 10 to 150 km across.

The life cycle: tall echo, bow, comma

Fujita’s conceptual model, as Weisman (2001) summarizes it, runs through recognizable stages:

• Tall echo. The system begins as a single large, strong cell or a short line segment, either isolated or embedded in a longer squall line.
• Bow echo. As strong surface winds develop, the cell bends into a bow-shaped segment. The strongest winds sit at the apex.
• Spearhead. At peak intensity the center of the bow can sharpen into a spearhead echo.
• Comma echo. In decline, the system often becomes comma-shaped, with a rotating head on its northern (left, relative to motion) end.

Fujita’s example from 4 July 1977 in northern Wisconsin was 80–100 km across and ran through that cycle over about six hours, leaving a near-continuous path of wind damage with no tornado found anywhere in it. Klimowski et al. (2004) found 45% of their bows grew from weakly organized cells, 40% from squall lines and 15% from supercells, and about half formed after storms merged. Bows that grew out of squall lines lasted longest, averaging 3.4 hours.

The rear-inflow jet and the notch it carves

The engine of a bow is the rear-inflow jet: a current of drier, higher-momentum mid-level air flowing into the back of the system. Weisman (2001) describes it feeding the downdrafts. Evaporation cools the air so it sinks harder, and it carries its momentum down to the surface. Where the jet is strongest, the outflow pushes the leading edge forward fastest. That forward push is the bow.

In reflectivity, the jet often shows as a rear-inflow notch: a channel or wedge of weaker echo cutting into the back of the line behind the apex, where dry air is eating into the precipitation. Earlier literature called it a weak-echo channel. Burgess and Smull (1990) documented a central Oklahoma case in which a Doppler analysis showed a rear-inflow jet stronger than 40 m/s at about 1 km above ground, collocated with exactly this kind of reflectivity minimum behind the apex of an 80 km bow.

The notch is a useful sign but a late one. Klimowski et al. (2004) point out that a rear-inflow notch can appear before obvious bowing, yet may only indicate that severe winds are imminent or already occurring.

Bookend vortices

As the rear-inflow jet pushes into the gust front, the ends of the bowing segment start to rotate: cyclonic on the northern end, anticyclonic on the southern end. These are line-end or bookend vortices, as discussed in Weisman’s 1993 simulations, and Fujita had already inferred them in 1978 from how small echoes moved around the bow’s ends. The pair matters because it funnels and strengthens the rear-inflow jet between them. In the Northern Hemisphere, the Coriolis effect favors the northern, cyclonic vortex over time. That is the vortex that grows into the rotating head of the comma stage.

Mesovortices near the apex

Bookend vortices mark the ends of the bow. Along its leading edge, a much smaller kind of circulation forms. Trapp and Weisman (2003) describe these low-level mesovortices as roughly 2–20 km across, mostly below about 1 km above ground, forming along the leading edge of squall lines and bow echoes. In their simulations the significant cyclonic mesovortices sat north of the bow apex. The mesovortex winds were often stronger and more damaging than the straight-line winds at the apex itself. Observed mesovortices are at times associated with tornadoes.

That is why tornadoes in quasi-linear convective systems (QLCSs) are hard to warn:

• They are common. Trapp et al. (2005) attributed 18% of 3,828 U.S. tornadoes from 1998–2000 to QLCSs, with a larger share in parts of the Midwest.
• They are usually weaker and more likely to occur late at night or early in the morning than tornadoes from discrete cells, according to the same study.
• They give less warning. Brotzge et al. (2013) found lower detection rates and shorter lead times for QLCS tornadoes than for supercell tornadoes. Goodnight et al. (2022) put average QLCS lead time at roughly 5–10 minutes.
• Some form with little precursor. Of 530 QLCS tornadoes that Goodnight et al. (2022) classified, 36% appeared to come from shear along the line breaking into vortices, with no clear pre-existing mesocyclone. Only about 30% of those were warned before they formed, compared with 44% of the tornadoes that had a mesocyclone first.

For how circulations like these compare with supercell rotation, see the mesocyclone and tornadoes on radar.

What velocity shows

Doppler radar measures only the component of motion toward or away from the antenna, so what you see depends on geometry.

• Bow moving toward the radar. The rear-inflow jet appears as a band of strong inbound velocity behind the apex, often lined up with the notch. Strong values at the lowest tilt, close to the radar, mean momentum is near the ground.
• Bow moving across the beam. The forward jet is mostly perpendicular to the beam and can look weak even when it is not. Bookend and mesovortex rotation still shows as inbound–outbound pairs.
• Mesovortices show as small, tight couplets along the leading edge, often north of the apex. Because they are shallow, they may appear at one elevation and not the next.

How bow echoes relate to derechos

Not every bow echo becomes a derecho. The Storm Prediction Center’s derecho facts page describes derechos as produced by bands of storms that take a curved or bowed shape. The difference is persistence and scale: a derecho is a bow-echo system that holds its organization and keeps producing damage over hundreds of miles. Robert Johns and William Hirt revived the term in 1987. The formal criteria are covered in the derecho guide. The 10 August 2020 Iowa derecho is a recent example. The National Weather Service in the Quad Cities logged an unofficial 126 mph gust near Atkins, Iowa, as the bow crossed the state.

Radar data for cases like these is publicly archived, and WeatherStory Radar reads it as full-resolution Level II straight from NOAA’s archive.