Researching the Wet, Wild World of Atmospheric Rivers

When a molecule of water vapor heads toward the poles, it may well be hitching a ride on an atmospheric river (AR). A growing amount of research is zeroing in on these narrow but powerful channels of airborne moisture, which are far more widespread and influential than scientists once thought. Garden-variety ARs tend to have a beneficial influence overall, but the biggest and baddest ARs can produce colossal rainfall and snowfall and major destruction. And Godzilla notwithstanding, major AR events are actually no more likely to strike California during El Niño than during La Niña.

I’m getting up to speed on ARs this week at the 2016 International Atmospheric Rivers Conference. Close to 100 forecasters and researchers are gathered here at a global epicenter of AR science: the Scripps Institution of Oceanography at the University of California, San Diego. The conference kicked off with an opening talk by Martin (Marty) Ralph, a long-time NOAA scientist who joined Scripps in 2013. Ralph is one of the leaders of the burgeoning science of ARs and the founder of the Scripps-based Center for Weather Weather and Water Extremes (CW3E). A growing amount of research at CW3E and elsewhere is aimed at producing useful AR guidance for weather forecasters and water resource managers. Many of the scientists at this meeting are also working on a book on ARs to be published in 2017.


Figure 1. This graphic shows an atmospheric river interacting with U.S. West Coast mountains and a midlatitude cyclone over the northeast Pacific on 5 February 2015. Labeled are the approximate locations of tropical moisture entering the atmospheric river and a warm conveyor belt (WCB) transporting warm, moist air just ahead of a cold front. Image credit: Adapted from NOAA/ESRL/PSD; Source: EOS Meeting Report.


What exactly is an AR?
The very definition of an atmospheric river is something of a work in progress. Although heavy rain and snow is the main concern with ARs, the events are best classified by the amount of water vapor being carrried through the air. In general, an AR is a narrow corridor transporting large amounts of water vapor. ARs typically have a fairly strong meridional component (movement along a north-south axis), although they can be oriented zonally (east-west) as well. They are closely associated with the preexisting concept of low-level jet streams, and of course they’re not literally channeled in banks like an earthbound river.

The AR label was coined by Yong Zhu and Reginald Newell in a 1998 paper published in Monthly Weather Review. Zhu and Newell estimated that about 95% of the water vapor moving across midlatitudes toward the poles was being carried in plumes that spanned just 10% of Earth’s circumference at those latitudes. (At least some of this moisture may be generated directly at midlatitudes rather than being transported there, an idea put forth by Helen Dacre [University of Reading] in a 2015 paper in the Bulletin of the American Meteorological Society.)

ARs are traditionally defined as being at least 2000 kilometers long and no wider than 1000 km, although a growing practice is simply to required that the length be at least twice the width. In midlatitudes, ARs typically have a peak value of vertically integrated water vapor transport (IVT) of at least 250 kilograms per meter per second. IVT denotes the amount of water vapor being carried across a line drawn at Earth’s surface, going up through the full depth of the atmosphere.

As with tropical cyclones, satellites are critical for detecting ARs over the ocean. At first, satellites could only provide limited detail on how much atmospheric moisture was present at each level of the atmosphere. The advent of the Special Sensor Microwave Imager (SSM/I) in 1987 allowed for the detection of integrated water vapor--the total amount above a point at Earth’s surface. This was a huge advance, opening the door to eventual real-time mapping of water vapor across the oceans as well as tracking of atmospheric rivers as they headed toward shore. Other tools such as GPS water vapor sensors are employed over land, where the SSM/I product doesn’t work as well. We now have automated routines that can infer the presence of ARs over land and sea in both historical and current datasets, and project their development days into the future. One product at CW3E projects AR development based on GFS model forecasts going out up to 180 hours.



Figure 2. Forecast of integrated water vapor transport (IVT, in kilograms per meter per second) derived from the GFS model run at 06Z (2:00 am EDT) Tuesday, August 9, 2016, and valid at 06Z (2:00 am EDT) Thursday, August 11. High values of IVT correspond to large amounts of water vapor being transported. Flash flooding is possible across parts of Minnesota late Wednesday (see below) as an AR-type moisture channel interacts with a strong upper-level trough and an associated cold front. Image credit: Center for Western Weather and Water Extremes, Scripps/UCSD.


Figure 3. Two runners watch as waves crash against the rocks near San Francisco’s Golden Gate Bridge on December 28, 2005. A series of wet winter storms associated with an atmospheric river event struck California in late December 2005 and early January 2006, bringing more than 20” of precipitation to the Sierra Nevada and widespread 24-hour rainfall totals of more than 5” on New Year’s Eve. Image credit: Justin Sullivan/Getty Images.

More than California
The iconic Pineapple Express--the southwesterly current that gained fame in the 1990s as a prolific rain and snow producer for the U.S. West Coast--is just one type of atmospheric river. As much as half of precipitation along the U.S. West Coast has been found to be AR-related. However, ARs can be found across the globe, especially if you broaden the definition so that the required IVT amounts are a certain percentage above the local norm rather than a fixed threshold worldwide. Based on this type of location-adjusted classification, ARs extend from the tropics all the way to Greenland and Antarctica, where they can account for a surprising percentage of precipitation. In a 2014 paper for Geophysical Research Letters, Irina Gorodestskaya (University of Aveiro, Portugal) reported that major spikes in snowfall across parts of East Antarctica in 2009 and 2011 were related to clusters of several AR events in each year.


Figure 4. The percentage of total annual precipitation that falls during atmospheric river events, based on an algorithm that detects AR events in climatological analyses. Across much of California as well as Tennessee and Kentucky, the percentage tops 30%. Image credit: Courtesy Bin Guan, UCLA and NASA/JPL.


Figure 5. California’s Lake Mendocino in December 2006. Image credit: Kglavin/Wikimedia Commons.

How AR forecasts might help save water ahead of drought
One of the perversities of California’s water storage system is the requirement that some large reservoirs release water in midwinter to help protect against the potential for late-season flooding. The system works beautifully as a flood prevention tool, but it’s based mainly on decades-old, by-the-book rules that take into account only the water that’s fallen, the amount being stored, and the time of year, not the long-range weather forecast or the seasonal climate outlook. California’s Lake Mendocino, built in 1958 on the Russian River north of San Francisco, has never gone over its spillway, thanks to careful management by the U.S. Army Corps of Engineers (USACE). As an AR-type series of storms dumped more than 20 inches of rain on the region in December 2013, some 25,000 acre-feet of water--more than half a typical winter’s storage--was released from the lake, even though it was far from full at that point. The rest of the winter produced less than 10 inches of rain, and the region’s drought intensified over the next year. A more flexible storage system based on weather and climate guidance might have allowed more of that much-needed water to be kept in the lake.

Lake Mendocino is now the focus of a proof-of-concept study called FIRO (Forecast-Informed Reservoir Operations) that could provide a major boost to California’s drought readiness. The goal is to apply weather and climate guidance and determine if the USACE could safely make adjustments to its mandated water-release levels, thus allowing it to keep more of the water from big AR events. “Lake Mendocino is our guinea pig,” said USACE engineer Cary Talbot. He pointed out that Congress mandates the Corps to reduce flood risk but not to protect water supply. This means that any water-saving measures must be rigorously evaluated and shown not to affect flood risk. FIRO’s goal is a big one, with a hefty forecasting challenge that’s brought in a raft of collaborating institutions. Nobody here in drought-tormented California needs to be told that the stakes are high. As Jeanine Jones (California Department of Water Resources) put it, “These days we expect our water to work a lot harder….What was good enough in the past [isn’t] good enough now.”

I’ll have more news from the AR conference in an upcoming post. For more background on ARs, see:

--this handy 2015 overview in Forbes by Marshall Shepherd (University of Georgia)
--a more detailed FAQ from NOAA’s Earth System Research Laboratory
--an even more comprehensive mini-review by Luis Gimeno (University of Vigo, Spain) and colleagues, published in 2014 in the open-access journal Frontiers in Earth Science


Figure 6. Total rainfall projected by NOAA’s Weather Prediction Center for the 7-day period from 8:00 am EDT Tuesday, August 9, 2016, to 8:00 am EDT Tuesday, August 16. Some localized amounts could exceed these projections by a considerable margin. Image credit: NOAA/NWS/WPC.

Big rains still on tap for Southwest, Gulf Coast
As we reported on Monday, this week holds the potential for major deluges along the central Gulf Coast and the southern Arizona desert. The NWS cautioned on Monday afternoon that significant river flooding would be possible later this week along parts of the Gulf Coast. Parts of the region have already notched 6” to 10” over the last several days, and additional rainfall amounts in some spots could exceed 10” by the weekend, especially near the coastline from New Orleans, LA, to Apalachicola, FL. A flood watch was in effect from Tuesday morning through Wednesday evening for much of the northern and western Gulf Coast of Florida.

Flash flood watches are in place over most of Arizona and parts of western New Mexico, with heavy rains already occurring south of Tucson on Tuesday morning. An unusually strong upper-level trough for early August will be interacting with a slug of moisture circulating into the area along the east side of Tropical Depression Javier, which is weakening as it grinds northwestward along the west coast of Baja California. Showers and thunderstorms should dump at least an inch of rain over widespread areas, with much higher amounts possible locally as a result of small-scale features impossible to predict in advance. Nighttime storms are possible with this unusually potent set-up, which would exacerbate the risk for any motorists attempting to drive through high water.

In addition, as shown in Figure 2 above, large amounts of atmospheric moisture will be converging on the Upper Midwest by late Wednesday ahead of the strong-for-August upper-level trough. The NWS has tagged parts of north central Minnesota with a moderate risk of flash flooding by Wednesday afternoon and evening.

Apart from Javier, there are no tropical cyclones or areas of immediate interest in the Atlantic or Eastern Pacific. In the Northwest Pacific, Tropical Storm Conson is gradually strengthening over open waters, while Tropical Storm Omais continues on the decline east of Japan.

We’ll have our next post by Thursday at the latest.

Bob Henson