August 31, 2022

How Serious is the ARk Storm (Catastrophic Flooding) Threat Along the West Coast?

The media was abuzz last week with stories about a greatly increased threat of a massive flooding event along the West Coast, termed an ARk Storm.  The immediate stimulus of this apocalyptic vision was a recent paper by Xingying Huang and Daniel Swain in the journal Science Advances.

This article suggests that global warming has already substantially increased the probability of West Coast catastrophic flooding events and that the potential for such flooding will be profoundly enhanced by the end of the century.

In this blog, I will describe some serious problems with this study, which greatly overstates the threat.  And I will show you the actual trends of heavy, prolonged precipitation.

What is an ARk Storm?

This term denotes a record-breaking, extended, unimaginably heavy precipitation event along the West Coast (mainly California) that produces catastrophic flooding.  It represents a dual play on words. First, the story of Noah's ark in the bible, with global flooding after extended (40 days and nights!) heavy rain. And second, an atmospheric river (an AR) event that has an extended (like 1000 years or "k") return time.

Atmospheric rivers are plumes of moisture that are associated with virtually all of the heavy precipitation events along the West Coast (see a satellite picture of one below).  Also called a Pineapple Express event in our region. Think of them as a meteorological "firehose" that dumps huge amounts of rain when it hits the terrain of the western U.S.

The only ARk storm in the historical record occurred during December 1861/January 1862 in California, over a period of about 30 days.  2-3 feet of rain fell over much of central California, with even more in some mountain locations, due to roughly a half-dozen extreme atmospheric river events.  This heavy rain was accompanied by above-normal temperatures, melting snowpack,  and a relatively wet period prior.

Larger areas of the interior of California were flooded (see estimated flooding extent below). There was extensive loss of life and damage to buildings and farms.

This kind of extreme, extended flooding event in California has occurred many times before, with evidence provided by layers of sediments in the coastal zone.   Such events appear to occur every few hundred years.

The Huang/Swain paper, based on global climate model simulations, downscaled to consider local variations using a high-resolution regional model, claims that the probability of such extreme events has already been greatly enhanced by global warming and will become far, far more probable later this century.

As you can imagine, such end-of-the-world clickbait is honey to the media bears and there have been dozens of stories on this paper during the past several weeks.  Unfortunately, there were serious problems with the paper and reality does not appear to agree with its estimates.

Let me demonstrate this to you below.

As I noted above, the Huang/Swain research is based not on observations but on an ensemble (collection) of climate model simulations.  Regrettably, they made a serious error by using climate models forced by a far too aggressive global warming scenario (RCP8.5), which is generally considered by the peer-reviewed literature to be unrealistic (too much warming).  Models driven by RCP8.5 also tend to poorly simulate the critical El Nino/La Nina cycle (also called ENSO).

RCP8.5, which makes the most draconian assumptions about fossil fuel use,  produces about twice as much warming as is reasonable.  And since atmospheric water vapor content goes up exponentially with temperature,  the use of the wrong scenario (with too much warming) is a very serious problem looking a precipitation impacts.  They should have used RCP4.5.

Another problem with this paper is their claim that the most extreme precipitation events in California occur in El Nino years.  This is inconsistent with observations:  neutral years (neither El Nino nor La Nina) are the most extreme precipitation years in that state. El Nino years are associated with more run-of-the-mill heavy precipitation events in southern California.

Inconsistency with observations

According to their model results, the frequency and amplitude of big, multi-week precipitation events should already have been increasing (see a reprint of their Figure 5a below).

Observations do not show that.   There is no evidence of an increase in heavy precipitation events or even changes in annual precipitation in California.     Which implies something is wrong with their model simulations.

                                      From Huang and Swain 2022

Let me prove this to you.    

Below is the annual precipitation in San Francisco going back to 1853.    You can see the enormous annual precipitation for 1862.    Note that there is no upward trend during the next 150 years!!  In fact, it is slightly down (there is a linear trend line in red).

Professor John Christy, of the University of Alabama Huntsville and the Alabama State Climatologist, has put together the best long-period precipitation data set in existence for the West Coast.  As shown in the graphic below that he provided to me recently (below), there is no increasing trend in either 30-day extreme precipitation events over the past 130 years.  The same is true of two-week extremes.   During the past 50 years, global warming should have become significant.
There is no increasing trend of wettest 30-day events along the West Coast.  Graphics courtesy of Professor John Christy

A More Realistic View

The Huang/Swain paper makes big claims about global warming driving major increases in the potential frequency of huge ARkstorm precipitation events in California.

Unfortunately, there are major problems with their paper, including the assumption of an unrealistic increase in greenhouse gases.  Their simulations compare poorly against observed trends that show no increases in ARk-like or heavy rain events.

So what do I think is the truth of the matter? (and I have published extensively in the peer-reviewed literature on the impacts of global warming on heavy precipitation along the West Coast)

Let us base our projections on assuming that greenhouse gases will follow the more realistic RCP4.5 scenarios.  With advancing energy technology, the RCP2.6 scenario, might be even better.

Assuming RCP 4.5, annual precipitation will not change significantly for the West Coast during the remainder of the century, probably with a small upward trend.   Global warming will result in the very heaviest atmospheric river events increasing by 10-20%.  So all else being the same, BY THE END OF THE CENTURY, the potential for ARk storms will increase by that much.  

But remember, it takes more than more moisture to produce an ARk storm: you need not only record-breaking atmospheric rivers but to have multiple ones, all hitting the same area.  In short, there must be many extremely unusual "fire hoses" occurring in a short period and hitting the same geography.

And there is more: to have a record ARk flooding event you need to have a large pre-existing, ready to melt, snowpack in the mountains, which will lessen under global warming.  And an antecedent wet period.

No wonder it is rare to get an ARk event!  A lot of moving pieces.  And for most of the pieces, there is no reason to expect enhancement by global warming.

Furthermore, California has a massive storage capacity for water in its reservoir system, something that did not exist 100 years ago.  This can help buffer the next ARk storm.

And there is one more consideration:   reality is not following the predicted extreme precipitation increase projected by the climate models. Thus, it is quite possible that the models are overdoing the impacts of global warming.     I am a modeler, read the papers, and have been at endless seminars on climate model performance.  Trust me, these models have major problems and many deficiences are in the area of clouds and precipitation.


  1. You suggest we assume the RCP4.5 or even RCP2.6 emissions scenario.

    Reality is tracking the RCP8.5, or worse, for emissions. The latest numbers are worse yet, diverging widely and wildly from RCP4.5.

    Should we disbelieve our lying eyes?

    1. Get2Surf...this is not correct. All the scenarios are essentially the same up until NOW. The question is the future. RCP8.5 makes assumptions about coal use, renewable, and population that are clearly wrong. No increased use of renewables for example. Continuous increased use of coal through the end of the century. No technological advances...cliff

    2. Please check the actual numbers. Look in the IPCC 5th Assessment Report, Working Group 1, Annex II, table AII.2.1a for the emissions used in each scenario. I'm not providing the URL, since I don't know how this site handles links, but you should be able to find the above document easily on the IPCC website.

      RCP8.5 assumed CO2 emissions* of 41.7 gigatons in 2020. Actual for 2021 (2020 should be treated as a low outlier due to the pandemic) was 36.3 gigatons.

      The closest scenario for actual 2021 emissions is RCP4.5. RCP6 actually has lower 2020 emissions, but that is because that scenario assumed a rapid slowdown in emissions growth, but for that slow growth to continue for another 6 decades. It was a sort of "be less worse, but don't actually improve" scenario. RCP4.5, in contrast, has significant immediate term emissions growth, but a peak around the year 2040, followed by moderate decline. On the other hand, RCP2.6 has its peak now, and I don't think that is realistic.

      Looking forward, RCP8.5 has global emissions continuing to accelerate until late in the century, when it becomes difficult to find more carbon regardless of how badly we want it. All the work and money we currently invest in finding new oil, natural gas, and coal to extract has to be significantly increased to fulfill RCP8.5. We have to want to burn fossils fuel so badly that people stop being price sensitive. If you tried to reduce how much you drive when gas prices recently spiked to $5 per gallon, you're one of the reasons RCP8.5 is unrealistic.

      In the 90's the global trend was indeed one of exponential increase, but around 2005, there was an inflection, and the rate of emissions increase started to slow. From 1995 to 2005, the average rate of increase in CO2 emissions was 2.4% each year. From 2009 to 2019, the average rate of emissions growth was 1.7%. Over the last 5 years in particular, it has been less than 1%. We do need a couple more years to see how much of the slowing was temporary due to the pandemic, but the data runs through 2021, when activity levels had begun to normalize. That inflection is the most significant contradiction with RCP8.5, more so than the shortfall of current emissions.

      * The other greenhouse gasses in the annex are not scaled to CO2-equivalent, but CO2 is by far the primary variable between scenarios, so for simplicity I will settle for CO2 for this thread. Multiply by 3.67 to convert from TgC to gigatonnes CO2.

    3. Ignoring the study in question, which I agree is way over hyped, the RCP scenarios set in 2005 have not been the same until now, though they are not terribly different, and we have been following RCP8.5 as Got2surf notes. See this PNAS study: It also shows that if we follow "business as intended" with emissions reductions, emissions are still closer to RCP8.5 than RCP4.5 in 2030 and 2050. Beyond 2050, I agree that RCP8.5 likely will be way off. Yes, it is possible that emissions come down more than current plans for reduction by 2050, but I would not call that a safe assumption at all given history to this point. There will also be additional scenarios that will be run with CMIP6 where RCPs will be replaced by SSPs.

  2. There can also be other events besides global warming and ENSO phases that can affect weather patterns.It should be noted that the 1861-62 winter was also influenced by a massive volcanic eruption in Africa( in May 1861) that ejected huge amounts of material into the atmosphere.That winter was the coldest in recorded history in the Northwest,too.

  3. Speaking of flooding and 100 years ago, as we approach the centennial of the great 1927 Mississippi river system flood I would be interested in learning more about the underlying weather events that triggered that flooding. Reading about the rain and the 8 months of Midwest flooding along with attempts to engineer the river systems is fascinating, but I found little on what triggered the massive rains.

    1. I read historical accounts of this event, and the best guess is that there was an extreme closed - off low that didn't move an inch for many months, as you stated. It also should be noted that when the floodwaters finally receded, record heat waves exploded in the Midwest and for much of the East Coast. I also remember a similar event occuring in 1992, when I had to travel around the midwest for my work. Massive flooding from the Mississipi caused extreme dislocations, and it camethisclose to breaching the levee at St. Louis.

  4. Hi Cliff, why don't you have one of the authors on your podcast for an in-depth discussion? I would find that conversation very intriguing.

    1. I very much like this idea!

    2. I would LOVE to do this. By the way, if you enjoy a good climate change debate, here is a recent one:

    3. Thanks for posting this! The more exposure it gets the better!

  5. A great analysis amidst the noise of the media hype resulting from seemingly truthful paper. As always, your critique and appraisal merits your credentials. Unfortunately that does not go for all weather scientists, and unfortunately the public gets more and more confused on what is truth and not.

  6. It must be hard to always being the person using actual data to refute doomsday journalism. Like swimming upstream. Love the podcast and books. Northwest weather is interesting!

  7. The two most serious dangers we face here in the US Northwest are: (1) our general lack of preparedness for dealing with a 9.0 Big One from a full rupture of the Cascadia Fault off the coast; and (2) a growing shortage of electricity creeping up on us over the next decade caused by the premature retirement without adequate replacement of the coal-fired and gas-fired generation capacity which serves the region. Everything else is a no never mind.

  8. I like Sean's idea about Cliff doing a podcast with someone who disagrees with him. That is a fantastic idea. We want to learn from the debate and then judge for ourselves.

  9. "This kind of extreme, extended flooding event in California has occurred many times before, with evidence provided by layers of sediments in the coastal zone. Such events appear to occur every few hundred years."

    Yet another fact conveniently ignored by the climate alarmists. When the first European settlers came to California they couldn't understand why there were so few indigenous people there, until they realized that the amount of fresh water rescources were minimal. CA has always been a boom or bust weather scenario, and it's experienced more widespread and long lasting droughts compared to copious amounts of precipitation.

  10. "RCP8.5, which makes the most draconian assumptions about fossil fuel use, produces about twice as much warming as is reasonable. And since atmospheric water vapor content goes up exponentially with temperature, the use of the wrong scenario (with too much warming) is a very serious problem looking a precipitation impacts. They should have used RCP4.5.
    With advancing energy technology, the RCP2.6 scenario, might be even better."

    I don't know how we could even afford to cause RCP8.5 to transpire. It's already pretty difficult to produce the amount of fossil we currently consume. How is the world going to more than double production and keep it up for decades as increasingly difficult to exploit sources are progressively developed and depleted? What are prices going to be when we miss the days when it was as "easy" as fighting ocean currents with GPS guidance to stabilize a deepwater drilling platform that has to use remote control robots to deploy wellheads and other required infrastructure at costs exceeding $100 million per well, or leveling entire mountains with skyscraper sized bucket wheel excavators?

    On the other hand, RCP2.6 is overly optimistic under our current trends. It assumes we reach peak global emissions roughly now, halve global emissions by 2050, and reach net zero by 2080. The Americas and Europe may achieve that (and have achieved peak emissions). Asia and Africa are still increasing rapidly and should be expected to peak decades behind the other regions.

    The recent trends in energy production and use do give me good hope we'll end up with global emissions trending lower than RCP4.5, however. If I had to throw a dart for where I think we're headed, it would probably require a new scenario on the board that effectively is "RCP4.0".

    As a reminder that often gets overlooked: US greenhouse gas emissions peaked in 2007. They have declined 12% from that peak through 2019 (2020 was an outlier year). Per capita emissions actually peaked in the 1970's and have declined almost 30%, but population growth has partially offset that.

  11. Hello,
    I was wondering, would extreme Atmospheric Rivers to our south in California create a harsh winter in the NW due to the jet being so far south? I recall Fort Vancouver had an average temperature of 21.3 or something like that during the month of January 1862. Do you have any data or insights about that winter in the Pacific NW, and what future ARkstorms in CA would bring here?
    Thank you.

  12. Cliff, I don't have the qualifications or reason to disagree with the facts of this blog, but I'm more worried about your messaging. You often provide great insight and opposing arguments to popular studies, which is essential to healthy scientific debate. But you sometimes say things like "Trust me, these models have major problems..." after listing a couple credentials, which I don't think is a productive or helpful way to address the problems you write about. Instead, you could write a separate blog post about why you think the models are flawed. Thanks for taking the time to read this comment

  13. For the record, this issue was made into a video by the USGS in 2011. That is not a typo!

  14. Cliff - I always enjoy your per review of recent papers that laypeople can understand. I would enjoy hearing what your take on the recent Greenland paper that's based on observational data instead of modeling data, per one of the authors. I'm sure your are aware of the paper; it indicates that 10-inches of ocean rise is already locked in as the Greenland ice sheet adjusts to the temperature increase that has already happened.

  15. Cliff, regarding your previous post about contacting you with examples of press censorship, do you have a preferred email address? Not sure if your .edu email is appropriate for blog related discussions.

  16. It looks like RCP8 predicted methane emissions (Climatic Change volume 109, Article number: 213 (2011)) is tracking closer to actuals, 1909 ppb in 2022, than the other models. In fact it is already higher than RCP4 and 3 predict as a maximum this century and RCP6 only predicts a max. of 2000. How significant is this going to be if this trend continues?


Please make sure your comments are civil. Name calling and personal attacks are not appropriate.

More Rain for the Northwest is Good News for Wildfires

After a very pleasant dry spell, another rainy period is ahead for the western side of the region and the Cascades on Friday and Saturday.  ...