Atmospheric River Storm: ARkStorm

Content Cover Image

Flood map of greater Los Angeles. Source: USGS.

ARkStorm Scenario

The U.S. Geological Survey's (USGS) Multi Hazards Demonstration Project (MHDP) has prepared its second full scenario, called ARkStorm. The scenario addresses massive U.S. West Coast storms analogous to those that devastated California in 1861–62. Storms of this magnitude are projected to become more frequent and intense as a result of climate change.

The MHDP assembled experts from the National Oceanic and Atmospheric Administration (NOAA), USGS, Scripps Institute of Oceanography, the State of California, California Geological Survey, the University of Colorado, Federal Emergency Management Agency (FEMA), the National Center for Atmospheric Research (NCAR), California Department of Water Resources, California Emergency Management Agency (CalEMA) and other organizations to design the large, but scientifically plausible, hypothetical storm scenario that would provide emergency responders, resource managers, and the public a realistic assessment of what is historically possible.

The ARkStorm storm is patterned after the 1861–62 historical events but uses modern modeling methods and data from large storms in 1969 and 1986. The ARkStorm draws heat and moisture from the tropical Pacific, forming a series of Atmospheric Rivers (ARs) that approach the ferocity of hurricanes and then slam into the U.S. West Coast over several weeks. Atmospheric Rivers are relatively narrow regions in the atmosphere that are responsible for most of the horizontal transport of water vapor outside of the tropics.

Using sophisticated weather models and expert analysis, precipitation, snowlines, wind, and pressure data the modelers will characterize the resulting floods, landslides, and coastal erosion and inundation. These hazards will then be translated into the infrastructural, environmental, agricultural, social, and economic impacts. Consideration was given to catastrophic disruptions to water supplies resulting from impacts on groundwater pumping, seawater intrusion, water supply degradation, and land subsidence.

In contrast to the recent U.S. East and Gulf Coast hurricanes, only recently have scientific and technological advances documented the ferocity and strength of possible future West Coast storms. ARkStorm is intended to elevate the visibility of the very real threats to human life, property, and ecosystems posed by extreme storms on the U.S. West Coast. This enhanced visibility will help increase the preparedness of the emergency management community and the public to such storms.

Background on Massive Winter Storms

caption Click for Larger Image. Beginning on Christmas Eve, 1861, and continuing into early 1862, an extreme series of storms lasting 45 days struck California. The storms caused severe flooding, turning the Sacramento Valley into an inland sea, forcing the State Capital to be moved from Sacramento to San Francisco for a time, and requiring Governor Leland Stanford to take a rowboat to his inauguration. William Brewer, author of “Up and down California,” wrote on January 19, 1862, “The great central valley of the state is under water—the Sacramento and San Joaquin valleys—a region 250 to 300 miles long and an average of at least twenty miles wide, or probably three to three and a half millions of acres!” In southern California lakes were formed in the Mojave Desert and the Los Angeles Basin. The Santa Ana River tripled its highest-ever estimated discharge, cutting arroyos into the southern California landscape and obliterating the ironically named Agua Mansa (Smooth Water), then the largest community between New Mexico and Los Angeles. The storms wiped out nearly a third of the taxable land in California, leaving the State bankrupt.

The 1861-62 series of storms were probably the largest and longest California storms on record. However, geological evidence suggests that earlier, prehistoric floods were likely even bigger. There is no scientific evidence to suggest that such extreme storms could not happen again. However, despite the historical and prehistorical evidence for extreme winter storms on the West Coast, the potential for these extreme events has not attracted public concern, as have hurricanes. The storms of 1861-62 happened long before living memory, and the hazards associated with such extreme winter storms have not tested modern infrastructure nor the preparedness of the emergency management community.

Building the ARkStorm Scenario

Atmospheric Rivers: Wind, Rain, and Waves

On October 14, 2009 an atmospheric river channeled water vapor from a decaying typhoon over the western North Pacific, across nearly the entire width of the ocean basin, to deposit copious rains over the central coast of California, M. Dettinger.


The nontechnical term “Pineapple Express” is popularly used to describe the meteorological phenomenon that causes moisture to be drawn from the Pacific Ocean near the equator and transported to the U.S. West Coast with firehose-like ferocity. Atmospheric rivers are embedded within much broader atmospheric storms referred to technically as “extratropical cyclones” (ECs). ECs are the winter-time analogue to hurricanes, but have much different structure. Also, they gain their energy largely from the pole-to-equator temperature contrast, unlike hurricanes, which draw their energy from ocean surface heat content. ARs are the business end of ECs because where the AR hits the mountains it can create extreme precipitation, flooding and high winds. In terms of impacts, an AR is to the broader EC it is embedded within, as the hurricane eyewall is to the broader hurricane of which it is a part. The importance and structure of ARs has become recognized recently through new satellite data and field experiments.

The atmospheric mechanisms behind the storms of 1861-62 are unknown; however, the storms were likely the result of an intense atmospheric river, or a series of atmospheric rivers, striking the U.S. West Coast. With the right preconditions, just one intense atmospheric river hitting the Sierra Nevada mountain range east of Sacramento could bring devastation to the Central Valley of California. An independent panel wrote in October 2007 to California’s Department of Water Resources, “California’s Central Valley faces significant flood risks. Many experts feel that the Central Valley is the next big disaster waiting to happen. This fast-growing region in the country’s most populous state, the Central Valley encompasses the floodplains of two major rivers—the Sacramento and the San Joaquin—as well as additional rivers and tributaries that drain the Sierra Nevada. Expanding urban centers lie in floodplains where flooding could result in extensive loss of life and billions in damages.”

Because there is not yet a suitable scale for atmospheric rivers, the storm scenario is named “ARkStorm” to represent an atmospheric river (AR) with a value of 1,000 on a scale of atmospheric rivers to be determined by atmospheric scientists. The scenario storm then will be an “AR 1,000,” and other U.S. West Coast storms could be scaled in comparison. One of the outcomes of this project currently being is the furthering of a common nomenclature for West Coast Winter Storms. For more information on Atmospheric Rivers, visit NOAA’s Physical Science Divisions page on the topic.

Key Findings

Though the ARkStorm Scenario report is online, below are some of the key findings of the project:

1. Megastorms are California’s other “big one.” A severe California winter storm could realistically flood thousands of square miles of urban and agricultural land, result in thousands of landslides, disrupt lifelines throughout the state for days or weeks, and cost on the order of $725 billion. This figure is more than three times that estimated for the ShakeOut scenario earthquake, that has roughly the same annual occurrence probability as an ARkStorm-like event. The $725 billion figure comprises approximately $400 billion in property damage and $325 billion in business-interruption losses. An event like the ARkStorm could require the evacuation of 1,500,000 people. Because the flood depths in some areas could realistically be on the order of 10-20 ft, without effective evacuation there could be substantial loss of life.

2. An ARkStorm would be a statewide disaster. Extensive flooding is deemed realistic in the California Central Valley, San Francisco Bayshore, San Diego, Los Angeles and Orange Counties, several coastal communities, and various riverine communities around the state. Both because of its large geographic size and the state’s economic interdependencies, an ARkStorm would affect all California counties and all economic sectors.

3. An ARkStorm could produce an economic catastrophe. 25% of buildings in the state could experience some degree of flooding in a single severe storm. Only perhaps 12% of California property is insured, so millions of building owners may have limited or no ability to pay for repairs. That degree of damage would threaten California with a long-term reduction in economic activity, and raise insurance rates statewide — perhaps nationwide or more — afterwards.

4. An ARkStorm is plausible, perhaps inevitable. Such storms have happened in California’s historic record (1861-62), but 1861-62 is not a freak event, not the last time the state will experience such a severe storm, and not the worst case. The geologic record shows 6 megastorms more severe than 1861-1862 in California in the last 1800 years, and there is no reason to believe similar events won’t occur again.

5. The ARkStorm is to some extent predictable. Unlike for earthquakes, we have the capability to partially predict key aspects of the geophysical phenomena that would create damages in the days before an ARkStorm strikes. Enhancing the accuracy, lead time, and the particular measures that these systems can estimate is a great challenge scientifically and practically.

6. Californian flood protection is not designed for an ARkStorm-like event. Much has been done to protect the state from future flooding, but the state’s flood-protection system is not perfect. The existing systems are designed among other things to protect major urban areas from fairly rare, extreme flooding. The level of protection varies: some places are protected from flooding that only occurs on average once every 75 years; others, on average every 200 years. But the levees are not intended to prevent all flooding, such as the 500-year streamflows that are deemed realistic throughout much of the state in ARkStorm.

7. Planning for ARkStorm would complement planning for earthquakes. The ShakeOut exercise has become an annual activity in California, with more than 7 million people participating each year. Many of the same emergency preparations are useful for a severe winter storm: laying in emergency food and water, shelter preparations, exercising emergency corporate communications, testing mutual aid agreements, and so on.

The Experts Behind the ARkStorm Scenario

USGS Multi-Hazards Demonstration Project
Dr. Lucy Jones, Chief Scientist
Sue Perry, Staff Scientist
John Bwarie, Strategy & Communications Officer

ARkStorm Coordinator, Dale Alan Cox, USGS

Atmospheric Scenario: F. Martin Ralph, NOAA Environmental Systems Research Laboratory, and Michael Dettinger, USGS/Scripps

A team of atmospheric scientists with expertise in West Coast storms associated with U.S. Federal and State agencies and academic institutions have created a model for use in devising this storm scenario and future scenarios. The ArkStorm scenario, the first of the scenarios, adapts information from up to three recent U.S. West Coast storms for which data exist to produce preconditioning information, such as rainfall and soil saturation. The experts then modeled the windspeeds, rainfall, flooding, and other meteorological and hydrological inputs that would realistically accompany two large storms, similar to the 1861-62 events, which would affect both northern and southern California.

Coastal Model: Patrick Barnard, USGS, and Dan Hoover, USGS

Concurrent to the design of the ARkStorm meteorological scenario, a team of coastal marine scientists developed for ARkStorm a state-of-the-art model to assess coastal vulnerability using data (wind, pressure, and so forth) from the hypothetical storm. Ultimately, the effort will lead to real-time understanding and prediction of coastal flooding, inundation, erosion, wave heights, current strength, and cliff failure on the entire Southern California coast. The output of the model is used in the ARkStorm scenario to determine plausible consequences of the hypothetical storm. Ultimately, the model output will be used beyond Southern California and incorporate real-time data inputs for use in a real-time warning system to be used by emergency managers, lifeline continuity operators, and resource managers.

ARkStorm Coastal Publication:
Barnard, P.L., O’Reilly, Bill, van Ormondt, Maarten, Elias, Edwin, Ruggiero, Peter, Erikson, L.H., Hapke, Cheryl, Collins, B.D., Guza, R.T., Adams, P.N., and Thomas, J.T., 2009, The framework of a coastal hazards model; a tool for predicting the impact of severe storms: U.S. Geological Survey Open-File Report 2009-1073, 21 p.

Landslides: Chris Wills, California Geological Survey, and Jon Stock, USGS

To advance our understanding of landslide susceptibility, the USGS Multi-Hazards Demonstration Project has assembled a team of experts to relate storm size to landslide potential. To do this, the project acquired, digitized, and quantified maps and aerial photographs of landslides from past storms to estimate landslide potential. The team created a database of stratigraphy, soils, and slope for southern California and portions of northern California that can be used to determine landslide and debris flow susceptibility associated with storms like the scenario storm. This builds on estimates of susceptibility to earthquake-triggered landslides prepared for the ShakeOut, with additional maps and analysis of debris-flow susceptibility and maps showing areas likely to be inundated by flooding and debris flows. Output of the analysis and expert opinion will be used to inform the emergency response, physical damage, and other aspects of the scenario. To demonstrate the future of landslide hazard science, the project has developed next-generation landslide hazard maps as part of the ARkStorm scenario by using high-resolution digital topography and new mapping techniques in a focus area.

Flood Hazards: Justin Ferris, USGS, William Croyle, State of California Department of Water Resources, Flood Forecast Center, Keith Porter, University of Colorado and Kathleen Schaefer, FEMA Region IX

The state-of-the-art meteorological model created by the ARkStorm atmospherics team was used to estimate pressures, windspeeds, temperatures, and precipitation time series in nested grid with resolution up to 2 km. The windspeed time series are being used directly for ARkStorm damage modeling, but it was impractical to input the precipitation time series into a statewide hydrological and hydraulic model to produce floodmaps. Instead, ARkStorm is using FEMA’s digital flood rate insurance maps (dFIRMS) as a representation of areas vulnerable to flooding under the ARkStorm scenario. For each of 16 watersheds, experts estimated whether ARkStorm corresponds to 100-yr or 500-yr flooding, considering the ratio of runoff intensity from ARkStorm to a simulation of the historical (WY1916-2003) record, using a variable infiltration capacity (VIC) model. Since not all of a watershed is subject to the same return-period extent of flooding in a single storm, ARkStorm treats a fraction of the mapped floodplain as actually flooded the scenario. To estimate the fraction, experts examine the merged 100-yr and 500-yr flood boundaries, together with an inventory of levees. They identify levee attributes indicative of vulnerability, and for each category and geographic area, estimate a reasonable fraction of those levees that fail either by overtopping, geotechnical failure, or hypothetical (deliberate) breaching, or they estimate a reasonable fraction of the potentially inundated areas that are flooded. The experts then select individual levees to fail and the areas they protect are considered flooded. Some flooding can occur in protected areas whose levees do not fail because of the potential inability to remove stormwater.

Geological evidence suggests that floods even greater than that of 1861-62 have occurred throughout California. Paleoflood hydrology is the science that investigates the geologic evidence of the movement of water and sediment in rivers before the time of hydrologic measurement. Within the flood hazard team, experts in paleoflood science documented geomorphic, biotic, and other information to reconstruct the occurrence and periodicity of these great floods. Sites or study areas were selected on the basis of the utility of the data to model validation and estimating the 1862 flood.

Physical Damages: Keith Porter, University of Colorado

Once primary and secondary hazards are defined, a team of engineers use this information to estimate the storms’ impact in terms of physical damage, repair costs, and restoration time for buildings and other infrastructure such as dams, levees, harbors, bridges, roads, water supply systems, and electric power. To accomplish this, the ARkStorm project identified and solicited the participation of researchers, government employees, and other engineering professionals involved in the design, construction, operation, maintenance, or risk assessment of these systems. The experts performed the engineering risk assessments and reported their findings, which included a historical review of storm impacts to the system of interest, a summary of the facilities currently exposed to risk, one realistic assessment of damage and repair implications, a brief discussion of practical risk-mitigation measures, and, where appropriate, a summary of limitations or pressing needs for further study.

Environmental and Health Impacts: Geoff Plumlee, USGS, and Charles Alpers, USGS

The rainfall, flooding, winds, and physical damage to infrastructure from an extreme storm would likely also result in adverse physical, chemical, and ecological impacts on the environment in northern and southern California, including the possible extirpation of species. For example, the storm would likely lead to widespread erosion, transport, and redistribution of soils, sediments, and rock materials that would affect riverine, lacustrine, and coastal environments. Storm damage to buildings, infrastructure, and industrial facilities (such as sewage treatment plants, petroleum refineries, mines, and chemical manufacturing plants) could release debris, contaminants, and microbial pathogens into the environment. Erosion or flooding of agricultural lands could lead to extensive loss or contamination of arable soils and leaching of chemicals into the runoff. Water supplies used for human consumption, livestock consumption, or agricultural irrigation could become contaminated by a wide variety of contaminants or pathogens. Following the storm, sediments, debris, and contaminants redistributed by landslides or flood waters could then dry out and become available for further redistribution by human disturbance and wind transport.

The storm scenario integrated input from a broad spectrum of experts to examine the potential for a wide variety of possible environmental impacts, such as (1) effects from flood-induced erosion and sedimentation on riverine, lacustrine, and coastal marine environments; (2) effects from infrastructure or industrial facilities damaged by the storm that lead to releases of contaminants or debris into the environment; (3) effects on the urban/suburban environment; (4) effects on agriculture through loss or contamination of arable soils; (5) effects on water quality, both in the environment and in water supplies used for human consumption, livestock consumption, and agricultural irrigation; (6) effects on species habitat and species viability; and (7) effects of dusts released following the storm from landslides, flood deposits, and other redistributed materials. The storm scenario also allows recovery managers to better understand the types and nature of materials that would be in need of cleanup and disposal, thereby enabling more effective planning for poststorm cleanup and recovery.

In addition to the acute physical threats to safety posed by the storm, the possibility also exists for adverse health effects on humans and ecosystems. These could include, for example, potential outbreaks of disease from consumption of contaminated drinking water, or outbreaks of valley fever resulting from exposure to dusts from flood deposits containing the soil fungus Coccidiodes immitis. Through collaborations with appropriate public health and environmental agencies, the scenario  modeled potential exposures to storm-related environmental contaminants and pathogens, potential health risks posed by these exposures, and ways to mitigate these exposures and health risks.

Emergency Response: Mitchell Miller, California Emergency Management Agency

Like the ShakeOut Earthquake Scenario, the ArkStorm scenario is scheduled to serve as the basis for local and state emergency response drills. The Golden Guardian drill associated with the ShakeOut Earthquake Scenario was the largest ever, with more than 3,000 emergency responders and managers participating. While the ShakeOut earthquake was a sudden onset event largely contained to southern California, the ARkStorm event, or series of events, would cover both southern and northern California and would extend over a period of weeks, or perhaps months. Emergency response would work closely with weather and flood managers to provide command and control to mobilize responders and the resource needed to protect citizens from peril. The emergency response section helped guide the creation of the ARkStorm scenario to maximize its usefulness in the statewide and local emergency planning and response exercises.

Forecast: Dave Reynolds, National Weather Service

Unlike the catastrophic event in the ShakeOut Earthquake Scenario, a winter storm is largely predictable. Although uncertainty is an element of every forecast, the predictability of a winter storm of this magnitude would give the public and emergency management community some time to prepare. The ARkStorm scenario provides a useful mechanism to exercise and test the use of forecasts in emergency management decisionmaking. The ARkStorm-size storm would give land-use planners and policymakers a scientifically credible framework on which to base their decisions.

Policy: Kenneth Topping, President, Topping Associates International and California State Polytechnic University, San Luis Obispo

Rather than detailing all policy concerns related to all consequences of the ARkStorm, the policy section drew out policy connections between the ArkStorm scenario and current planning and operational procedures based on existing policy. The ArkStorm scenario focuses on how this exercise can be used by agencies to (1) qualify for Federal Emergency Management Agency (FEMA) and State disaster recovery funds (for example, Public Aassistance, California Disaster Assistance Act) and (2) satisfy FEMA, State, and local emergency operations plans and preparation. The ArkStorm scenario also focuses on how a storm of this magnitude raises awareness, thereby raising the question of what pre- and post-disaster hazard mitigation planning and funding resources (for example, National Flood Insurance Program, Pre-Disaster Mitigation, Hazard Mitigation Grant Program) need to be sought in light of this scenario. Such funding could be applied to such activities as moving local emergency operations centers and equipment out of a flood plain, guiding development away from flood plains, adding access-evacuation points for single-access areas inundated or isolated by flooding, and strengthening existing flood works. Another focus is how the scenario could be a standard reference point for application of future FEMA guidance in California in State and local multihazard mitigation plans under the Disaster Mitigation Act of 2000. There are detriments to creating a scenario larger than a 200-year storm. The policy section identifies the importance to local hazard mitigation planning of connecting this superstorm scenario to the 200-year flood-plain mapping.

Economics: Anne Wein, USGS, and Adam Rose, University of Southern California

The ARkStorm scenario examines the economic costs and impacts associated with a storm analagous to the series of storms that pummeled California in 1861 and1862. While the population of California was only about 500,000 people in 1862, nearly a third of the taxable land was lost in those storms. Modern engineering of flood channels, dams, and levees have saved many lives and much property since 1862; however, the population of California, the built environment, and the economy have grown since 1862, putting more at risk. The storms of 1997 and 1998 caused nearly $1.1 billion in losses in California alone. The potential for loss in the Sacramento area alone could exceed $25 billion. In addition, there will be business interruption losses from lifeline outages, building damages, and agricultural losses.

Storm Visualization: James Done, National Center for Atmospheric Research

To elucidate the nature of the meteorological component of the ARkStorm scenario, an analysis and visualization of the atmospheric data has been conducted with a focus on the California region. An animated 3-dimensional visualization of storm flow structure conveys the overall storm structure and life-cycle, showing low-level flow carrying moisture-laden air from low-latitudes up into the Western US, characteristic of the ‘Pineapple Express’. The analysis and visualization then focused on the temporal and spatial variability of very strong winds and high rainfall. To convey a greater clarity of the potential impact of a winter storm, wind and rainfall data are mapped on top of surface satellite data, infrastructure/lifeline data and/or exposure data where available. In addition to providing a thorough overview of the details of the winter storm and elucidating the threats to human life, property and ecosystems, the visualizations aid other components of the ARkStorm Scenario including secondary hazards such as landslides and flooding.

The Ark Project: A Winter Storm Campaign for USGS: Mariana Amatullo, DesignMatters, Art Center College of Design, Elena Salij, Art Center College of Design, and David Droga, Droga5

Art Center students (l to r) Scott Liao, David Styler, Leslie Lucey, Scott Reedy, Brett Brown, and Ana Ramos present The Ark Project brand identity and preparedness campaign.

Advertising and design students from the Art Center College of Design in Pasadena, California, are working with the Multi-Hazards Demonstration Project to brand and develop a media campaign to raise awareness of natural hazards in general, with specific emphasis on the ARkStorm scenario. The class of eight semester students were selected to participate in this effort, which was led by Elena Salij, Chairman, Advertising Department and internationally acclaimed David Droga, creator or the international advertising agency, Droga5. The class addressed the awareness problems associated with U.S. West Coast storms, worked to incorporate multiple sources of hazard information to include earthquakes, weather, debris flows, landslides, coastal concerns, pollution, and others. The class not only branded the scenario, but create a campaign to promote it— to bring it out of obscurity and onto every desktop.

Art Center College of Design: DesignMatters
The Ark Project: A Winter Storm Campaign for USGS

Media, Videos, and Graphics

  • Media, videos and graphics related to ARkStorm and massive winter storms can be found here.





(2012). Atmospheric River Storm: ARkStorm. Retrieved from http://www.eoearth.org/view/article/51cbf0c17896bb431f6a2eeb


To add a comment, please Log In.