Imperfect predictions of the perfect storm

Weather prediction in Seattle is problematic. The old saying, "If you don't like the weather, wait five minutes" has always been more reliable than five-day forecasts. But times are changing. It is certain the Pacific Northwest climate is warming up (the records for the past century demonstrate that pretty clearly), and global warming promises even higher temperatures. Still, as with all weather-related things here, there is doubt. If we can't forecast next week or next month, how can we know what lies 50 years ahead? Answering that question is important; it has a direct bearing on how we plan and build for the future. Can we prevent another Madison Valley flash flood? To what standard do we build our drainage systems? If we replace the Highway 520 floating bridge across Lake Washington to withstand a "100-year" storm, just how big is that storm likely to be? Two of the most controversial questions to answer are: In the next few decades, will Seattle and the Pacific Northwest be dryer or wetter? And will our storms be worse? Let's tackle the storm issue first. Doesn't it seem like there have been a lot of "100-year storms" lately? I got curious about that when I was reporting on the Hanukkah Eve Storm of Dec. 14 and the flood that killed Kate Fleming in Seattle's Madison Valley neighborhood. A new report on the Madison Valley flood from Seattle Public Utilities (SPU), prepared by the engineering firm CH2M Hill, called the storm "greater" than the 100-year rainstorm that hit Seattle in August 2004. Hmmm. Two 100-year storms in less than three years. Statistically, that's possible. Nothing says a 100-year storm literally only happens once every hundred years. The calculation actually means that, based on available data (which often doesn't cover a very long time period, not even 100 years), a 100-year storm is an event that scientists predict has a 1 percent chance of happening in any given year. Sometimes major events cluster. One hundred-year floods, for example, can happen more or less frequently based on factors such as human activity and especially development (damming, dredging, etc.). I checked newspaper accounts of 100-year rainstorms in the greater Seattle area. There have been at least six "100-year" storms since 1986 (January 1986, January 1990, October 2003, August 2004, and November and December 2006). Two 100-year storms back-to-back seems believable, but six? It looks like either the calculations for "100-year" storms are far from perfect or that the recurrence of such storms is picking up. This acceleration seems to be explained by the much-publicized prediction that intense storm activity will worsen here and around the world due to global warming. When talking with SPU officials, I asked whether Seattle's storm drainage infrastructure needed to be updated to a higher standard because of the prediction that climate change would bring more severe rainstorms. The director of SPU, Chuck Clarke, said that his department was looking at that very issue and consulting with University of Washington weather experts about it. The day following that interview, on April 20, on his Friday weekend weather segment on KUOW-FM's "Weekday with Steve Scher" (a show on which I am a regular guest), University of Washington professor Cliff Mass said that predictions of more rain and worse storms in Seattle due to global warming were not proven. He was referring to a story aired earlier that morning on the station. "The jury is out," he said. He seemed surprised that the belief was so widespread and often stated as fact. "There isn't any support for it." So if Mass is right, why do so many people believe that Seattle's weather is becoming wetter and stormier? One reason might be that the city of Seattle seems to believe it. Here's what a Seattle City Light report says: Climate scientists generally agree that if we continue releasing CO2 into the atmosphere, global temperatures will continue to rise and our climate will change. But this does not mean a gradual increase in temperature. Climate models predict an increase in violent weather – severe storms, floods, and drought – with potentially devastating impacts on plant and animal populations, and adverse effects on human health. And here's what the Seattle Department of Transportation (SDOT) says in a 2005 report: Seattle has experienced two incidents of 100-year storms during the past eight years. Climate change could further increase winter precipitation, including the intensity of winter precipitation events, and cause more frequent flood and landslide events in Seattle. Floods and landslides can damage the City's transportation infrastructure and underlying utilities, and threaten public safety, mobility, and private and public property. With officialdom sounding the alarm, I asked Mass for his opinion of the quote from the SDOT report. He emailed his reply: Well, it shows you can't believe everything you read! I don't think that quote is correct. There is often a lot of misinformation about 100-yr storms. First, you CAN get two 100-year storms in close temporal proximity. ... there is a finite probablility of that. Furthermore, the statistics upon which they are based are often very shaky. ... our records are relatively short and thus there is lots of uncertainty in many estimates of return time. And finally, the recent "100 year storms" have not been in the same locations or even due to the same mechanisms. To answer your question ... yes, the city is being premature. But I would rather them build systems with more capacity than too little ... But there is some overhyping going on here. Mass says that he and two colleagues, Eric Salathe and Rick Steed, have run "high resolution simulations of the regional weather/climate of the next century" and have generated these interesting results:

1. There is little if any trend in annual precipitation.
2. November rainfall is higher, but the rest of the winter is drier – resulting in little trend in total wintertime precip.
3. Looking at the model ouput for the number of events greater than 2 centimeters in a day, there is little trend. Thus, these initial simulations do not suggest a trend in heavy precipitation days.

... [A]t this point there is little indication of more intense precipitation in the region under global warming. So at this point I think the wise thing is to say we can't answer the question well and that right now there is no indication of heavier "pineappple express" rainstorms in the area. So, according to Mass, city planners are going overboard with their rhetoric about the coming rains. But there are others at the UW who think the worry about coming storms isn't exaggerated or premature. They include climate researchers at the Climate Impacts Group (CIG) and the state's climatologist. Unlike Mass' research, the climate models CIG is using tell them that precipitation is likely to increase in the Pacific Northwest. In 1999, Philip Mote, a professor of atmospheric science and a researcher with CIG, told the Seattle Post-Intelligencer: "The models say precipitation should increase." The story was headlined, "Global warming could bring us wetter winters." Mote holds to that prediction today. If you look at the climate change models such as those in the Climate Change 2007: The Physical Science Basis produced by the Intergovernmental Panel on Climate Change under the auspices of the World Meteorological Organization and the United Nations earlier this year, you'll see that they predict an increase in "heavy precipitation events" and say to expect more winter precipitation in the Pacific Northwest. Mote, who is also Washington's state climatologist, says the only change since the P-I story is that in 1999 CIG based their concerns on only four global climate models. Now they have looked at 10, and the overall picture, he says, is pretty much the same. Mote is careful to say that no one is certain precipitation will increase here, only that "there's a very real possibility that extreme daily precipitation will increase." Policymakers, he said, should "plan for the range." Hope for the best but prepare for the worst, in other words. He elaborated in an e-mail: In CIG, we have examined scenarios of future change from 10 global models each with two different socioeconomic scenarios. In most of them, precipitation increases in fall and winter, but these increases are modest compared with the "noise" of background variability. (Not so for temperature – in every scenario, temperature in the 2020s is substantially larger than in the 1990s). However, what one decides to emphasize from these results depends on one's risk tolerance. Cliff [Mass] is correct that there is no "proof" that we will get more winter rain – this is a point on which models are divided, and I have little confidence in projected changes in precipitation. But the City [of Seattle] is also correct in saying extreme precipitation "could be more intense" – they are planning for a worst-case scenario, and models probably support that interpretation. So, in terms of how climate change will impact local weather, we seem to have not a difference of opinion so much as a difference of consensus, if that makes any sense. Consensus says that warming is happening. Global models suggest that it will get wetter and stormier – even Mass' local model suggests Novembers might be wetter – but no one is terribly confident in the details on precipitation. Mote says more extreme precip is "a very real possibility," while Mass says concerns are over-hyped. Both scientists agree that it's good to be prudent in planning. Meanwhile, researchers continue to work the models. As for the rest of us? Wait five minutes and you may get another answer. Or another 100-year storm.