The tornado debris signature

A tornado debris signature (TDS) is a patch of anomalously low correlation coefficient, usually with differential reflectivity near zero, sitting on top of a tornadic velocity couplet — the dual-polarization fingerprint of objects a tornado has lifted off the ground. It is the only common radar signature that indicates a tornado is doing damage rather than that one might.

The tornadoes on radar guide places the TDS at the top of a ladder that starts with the hook echo and the velocity couplet. This page goes further into the one rung that constitutes confirmation: where the thresholds come from, why they disagree, and the documented cases where the signature appeared with no tornado under it.

The physics: why debris breaks correlation

A dual-polarization radar transmits horizontally and vertically polarized energy and compares what comes back. The correlation coefficient (CC, written ρ_hv in the literature) measures how similarly the two channels behave from pulse to pulse inside one sample volume. In a volume full of similar raindrops, the answer is “almost identically”: Schuur, Ryzhkov, Burgess and Zrnić, in an AMS conference paper on these observations, note that in pure rain or dry snow ρ_hv usually falls between 0.980 and 0.997 on a well-designed radar.

Debris is the opposite of a raindrop. Shingles, insulation, leaves, splintered wood and soil vary wildly in size, shape and material, and they tumble. That heterogeneity drags CC down. The tumbling also matters for differential reflectivity (ZDR): randomly oriented scatterers have no preferred horizontal or vertical dimension, so ZDR sits near 0 dB rather than the positive values typical of flattened raindrops.

Ryzhkov et al. (2005), in the Journal of Applied Meteorology, established the signature from polarimetric data on three tornadic storms around Oklahoma City, the first of them on 3 May 1999, observed by NSSL’s Cimarron radar.

The thresholds, and why there is more than one

There is no single official number. The criteria have loosened and tightened as operational experience accumulated.

Every version starts from the same prerequisite. The current NWS guidance sheet says it plainly: first identify a valid velocity circulation at the lowest elevation tilt, then check CC and reflectivity, then look for time and height continuity. Low CC without a circulation is not a TDS.

The spread in CC thresholds is a trade-off, not a disagreement about physics. As Schultz et al. (2012) summarize it, Ryzhkov et al. speculated that light debris such as leaves and grass sits around 0.70 while structural material pushes CC lower. A strict cutoff misses weak debris; a loose one starts catching hail. Saari et al. (2014) condense the consensus to a strong couplet, reflectivity above 20 dBZ, CC below 0.7–0.8 and ZDR near 0 dB, checked for vertical continuity.

The debris ball

“Debris ball” is the reflectivity half of the picture: a small, roughly circular maximum at the tip of the hook, co-located with the couplet. On reflectivity alone it is suggestive but ambiguous: Schultz et al. (2012) show a tornado-warned supercell on 26 March 2011 whose precipitation core mimicked a debris ball. With CC available, the blob can be checked. If it is debris, CC collapses inside it; if it is precipitation, CC stays higher.

Reflectivity also works as a confidence dial. Schultz et al. (2012), in Part II, recommend treating it that way: when the other criteria are met, the higher the reflectivity inside the low-CC area, the more confident the debris call.

Confirmation, not prediction

A TDS requires debris already in the air. By the time a scan shows it, something has been damaged. That is why the NWS warning guidance treats a TDS as one route to a “radar confirmed” tornado, and why it adds nothing to lead time on a tornado that has not yet touched down. For warning ahead of the tornado, the evidence is in the mesocyclone and the couplet, not in CC.

Its operational value is in the middle of an event: tracking a rain-wrapped tornado nobody can see, confirming a tornado at night, or separating tracks. Schultz et al. (2012) describe the signature from a violent north Alabama tornado on 27 April 2011 persisting for more than an hour and a half on UAH’s ARMOR research radar, and being used by NWS Huntsville during long stretches when no ground reports were coming in.

Height and intensity

Stronger tornadoes loft debris higher, and the relationship is strong enough to use operationally, with caution.

• Entremont and Lamb (AMS Annual Meeting, 2015) examined more than 180 TDS cases and found that when debris is lofted to 10,000 feet or higher, a strong or significant tornado becomes increasingly likely.
• The NWS WDTD tornado warning guidance lists a TDS height threshold for EF2+ of 8,000–10,000 feet.
• Van Den Broeke and Jauernic (2014) reviewed every tornado reported from January 2012 to May 2013 and found a debris signature with only about 16% of them. Signatures were more common with higher EF ratings, closer range to the radar, longer paths and urban land cover, and their areal extent increased consistently with EF rating and longevity.

Schultz et al. (2012) are careful about the ceiling on this: signature diameter and height are “loosely tied” to strength, and a definitive EF-rating relationship is not possible because what was damaged, and what got lofted, varies from tornado to tornado.

False and misleading signatures

Low CC has many causes. The debris interpretation holds only when the low values sit inside the circulation, with enough signal to trust them.

• Large hail. Schultz et al. (2012) note that very large hail can produce CC of 0.70–0.80 at S band, overlapping the older debris threshold. NWS Louisville training material gives 0.70–0.95 for hail cores, typically with reflectivity above 55 dBZ. The requirement that low CC sit on the couplet is what separates the two.
• Weak echo in the inflow. CC is unreliable where signal is low. The NWS warning guidance cautions that low CC in the inflow region can look like a TDS, and asks for at least 20 dBZ. Schultz et al. (2012, Part II) document a 3 March 2012 case near Atlanta in which, after an EF-3 tornado lifted, the criteria were met repeatedly for the next hour in 15–25 dBZ echo with no damage or falling debris reported.
• Criteria met, nothing found. The same paper describes a 21 January 2012 signature in Coweta County, Georgia: CC at or below 0.35, reflectivity above 50 dBZ, a velocity signature, and vertical continuity. Two damage surveys found nothing.
• Offset from the ground track. Debris is carried by the storm-scale flow, and Part II measured low-level offsets of 1–4 km between the signature and the surveyed damage path. It also documents debris showing up after a tornado had lifted.
• Non-meteorological echo. Ground clutter and biological targets such as birds and insects also give low CC, as Schuur et al. note. They are not tied to a tornadic circulation.

A working read

Start with velocity. Find the couplet at the lowest tilt. Check CC at that spot, not somewhere nearby, and check reflectivity so you know the CC is backed by signal. Then step up through the elevation angles to see how high the low-CC column goes, and step forward through scans to see whether it moves with the circulation. For the reflectivity structure that usually surrounds it, see the hook echo; for the fundamentals of CC and ZDR, start with how to read weather radar.

WeatherStory Radar reads full-resolution Level II, including correlation coefficient and ZDR, straight from NOAA’s public archive.