Cat 1 to Cat 5: A Scale for Atmospheric Rivers

In an article published online Tuesday for the February Bulletin of the American Meteorological Society, researchers and forecasters unveil a different type of river gauge: a scale for diagnosing the strength of atmospheric rivers (ARs). The five-category AR Cat scale is envisioned as a way to help emergency managers and others, including the general public, quickly distinguish between weaker ARs—which are mostly beneficial—and the strongest ARs, which can lead to catastrophic flooding.

The multiyear effort to develop the AR Cat scale was led by Martin (Marty) Ralph at the Center for Western Weather and Water Extremes (CW3E), part of the Scripps Institution of Oceanography at the University of California San Diego. Second author on the BAMS paper, which was released online Tuesday, is Jonathan Rutz, a meteorologist with the National Weather Service’s Western Region headquarters in Salt Lake City, Utah.

“Atmospheric river” isn’t yet a household phrase on par with “polar vortex”, but the AR concept has gained increasing traction among scientists and stakeholders. ARs are often cited in weathercasts and NWS forecast discussions, especially across the Pacific Coast states, where the phenomenon plays a huge role in delivering both desired and unwanted rainfall. One well-known example is the Pineapple Express, the Hawaii-to-West Coast flow that frequently qualifies as an AR.

Figure 1. Visible satellite image of an atmospheric river bringing copious amounts of moisture from the tropical Pacific into the Pacific Northwest on Dec. 8, 2015. Image credit: NOAA, via BLM Oregon & Washington.

In 2018, the AMS Glossary added this formal definition of “atmospheric river”:

A long, narrow, and transient corridor of strong horizontal water vapor transport that is typically associated with a low-level jet stream ahead of the cold front of an extratropical cyclone…Atmospheric rivers are the largest "rivers" of fresh water on Earth, transporting on average more than double the flow of the Amazon River.

What’s absent from this definition is numbers—in particular, how powerful and sustained the water vapor transport needs to be to qualify as a strong AR, a weak AR, or even an AR at all.

“Currently, there is no concise method for conveying the spectrum of benefits and hazards faced by communities during a particular AR event,” Ralph and colleagues wrote. “This scale is intended to serve the western United States, and other regions with significant AR climatologies, in the same way that scales for hurricanes, tornadoes, and nor’easters have served other parts of the country.”

One rating, multiple dimensions

Like its counterpart scales for tornadoes and hurricanes, the AR Scale has five categories, ranging in this case from Cat 1 (weak) to Cat 5 (exceptional). A distinctive aspect of the AR Scale is that it is keyed to location, strength, and duration. Each AR rating is calculated at a given point by the amount of integrated vapor transport, or IVT (the amount of water vapor carried each second across a line perpendicular to the atmospheric river flow) together with the duration of minimal AR conditions (meaning an IVT of at least 250 kilograms per meter per second).

Thus, instead of having a single rating at any point in time, each atmospheric river would be predicted and assessed for every relevant point in its path. The resulting AR Cat ratings would apply to the entire event rather than to the AR’s instantaneous strength.

Figure 2. The AR Cat scale, which categorizes each atmospheric river event at a given location by the peak value of integrated water vapor transport (IVT) and the duration of minimal AR conditions (IVT of at least 250 kg/m/s). Image credit: Figure 4, from Ralph et al., “A scale to characterize the strength and impacts of atmospheric rivers,” BAMS (February 2019), courtesy American Meteorological Society.

As can be seen in Figure 2, duration and IVT strength are both critical to determining an AR Cat rating. For example, imagine a moderately strong AR that was expected to bring peak IVT values of around 800 kg/m/s. If this AR were to affect a coastal location for a period of less than 24 hours, it would be rated at that location as a Cat 2 (“mostly beneficial, also hazardous”). Yet if the same AR lingered for more than 48 hours at that location, it would then qualify as a Cat 4 (“mostly hazardous, also beneficial”).

The descriptive labels convey the mixed-blessing nature of atmospheric rivers. In California, the authors noted, “ARs contribute 25%–50% of annual precipitation (including critical snowpack) in just a few days each year.”

The duration component of the AR Scale has been added to what was originally an IVT-strength-only scale (the left-hand axis of Figure 2). That initial version was introduced in 2016 and tested at CW3E during the stormy West Coast winter of 2016-17.

“Some have noted there could be an AR Cat 6, for situations where the max IVT exceeds 1250 and at least minimal AR conditions last for more than 48 hours,” Ralph said in an email. “These happen, but are very rare, and it may be something that emerges later, if needed.”

Pros and cons

One big strength of the AR Cat scale is that its variations across a grid of roughly 30 by 30 miles could be generated as a forecast element by computer models fairly easily. Because the scale incorporates both intensity and duration, it could serve as useful shorthand for the threat posed by an incoming AR.

In one retrospective analysis of ARs that affected the Russian River near Guerneville between January 1980 and November 2017, the authors noted, 3 out of the 10 Cat 5 events produced major flooding, compared to 6 out of 22 Cat 4 events and just 2 out of 78 Cat 3 events. Ralph and colleagues suggest that Cat 4 and Cat 5 AR events could be subclassified as “major” ARs, along the lines of violent tornadoes (EF4 and EF5) and major hurricanes (Cat 3, 4, and 5).

Figure 3. Two trucks cross a flooded road in Guerneville, Calif., on January 2, 1997. Recent analysis shows that an atmospheric river that lasted for over 100 hours at the central California coast, from Dec. 29, 1996, to Jan. 2, 1997, would have qualified at some locations as a Cat 5 on the proposed AR Cat scale. The associated heavy precipitation and runoff caused more than $1 billion in damages. Image credit: John G. Mabanglo/AFP/Getty Images.

Among the downsides of the AR Cat scale acknowledged by its creators:

•  The scale is not designed to capture very intense rainfall rates over short periods and small areas, which can produce devastating flash floods (such as over burn scars from recent wildfires).

•  The altitude of the rain/snow line, which can greatly affect runoff and other impacts, is not incorporated.

The new scale is only a proposal at this point, but its developers plan to continue their work over the coming months. Their goal is to produce a final AR Cat scale for operational use that will closely resemble the one outlined in their paper. CW3E already publishes IVT forecasts on its website based on output from the NAM, GFS, and GEFS models.

“The AR scale is a significant step forward, providing forecasters with a tool to distinguish between primarily beneficial and primarily hazardous storms,” said Rutz in a Scripps news release. “I anticipate that this scale will be adopted and highly used.”

Are we ready for another scale?

Given the proliferation of weather-related scales, I checked in with someone who thinks a lot about risk communication and weather: Joseph Ripberger, deputy director for research at the University of Oklahoma’s Center for Risk and Crisis Management. A political scientist, Ripberger has worked closely with meteorologists at the National Weather Center on weather-related risk and public policy.

I asked Ripberger whether he thought the public might be at risk of “scale fatigue”:

“On the contrary, my research suggests that most people want more information about risk and uncertainty in weather forecasts. Unfortunately, it can be very difficult to communicate this information. Meteorologically precise scales can generate misperceptions and confusion, whereas more generic “risk language" (i.e., low risk, moderate risk, high risk) can generate ambiguity because people interpret language in different ways.”

While he hasn’t yet studied the AR Scale himself, Ripberger said he generally advocates for a combination approach that combines probabilities with precise language to communicate risk — “e.g., there is a 50-75% chance that an AR will cause 30-50 inches of rain in next two days, which will likely cause extreme flooding in the region.” If the AR Cat scale, or something like it, were folded into an ensemble modeling system that provided probabilistic guidance, and linked to impacts-based messaging, perhaps it could help move the needle in the direction suggested by Ripberger.

As for the growth of one-to-five scales in the weather world, Ripberger saw it as “an artifact of history”—given the precedents set by Fujita and others—but also one "that happens to align with psychometric findings that it can be difficult for people to conceptually distinguish between points along more than five categories.” He pointed to research as far back as 1978 showing that having no more than five or six categories on a rating scale appears to hit a psychological sweet spot.

For an in-depth exploration of atmospheric rivers and California’s past, present, and future climate, see “The Flood that Could Change Everything” by Eric Zerkel, part of weather.com’s United States of Climate Change package. Also, see our April 2018 post, New Study: Precipitation Whiplash is Hitting California, and It’ll Get Worse.