Debunking Media Hysterics That "Seattle Heat-Waves Could Kill Hundreds"

Via Cliff Mass Weather and Climate blog,

On Friday, the front page of the Seattle Times had a terrifying story about Seattle heat waves that could kill hundreds--- in fact, over 700 per event.  And that only by rapidly cutting greenhouse gas emissions, might our fellow citizens be saved.

Courtesy of the Seattle Times

Their online headline deepened the angst, telling us that Seattle was "unprepared" for the "deadly heat waves" that were being stoked by global warming.

This story was based on a paper in the journal Science Advances:

Increasing mitigation ambition to meet the Paris Agreement’s temperature goal avoids substantial heat-related mortality in U.S. cities.    

The title of this paper gives one a hint that this is an advocacy document, not an objective look at science and risk. There is supplemental information on the paper here

As I describe in detail below, there are profound issues with that paper and with the Seattle Time article that promoted it.  And as I will discuss, such poor journalism and problematic papers have the potential to undermine progress in dealing with the actual threats accompanying global warming.

There are two aspects of the paper that have substantial problems.  First, there is the estimate of how temperature will change this century under global warming.  Secondly, there is their approach for estimating the additional deaths associated with various levels of global warming.

Let's start with the second... how they connect global warming with increased mortality.

For each city, they completed a statistical analysis of the death rates FROM ALL CAUSES versus mean daily temperature, with their best estimate of the relationship shown by blue and red lines, corresponding to mortality for cold and warm events  (see example for Seattle below).  The gray area is a measure of the uncertainty of their estimate.   The they also find a temperature of minimum mortality (MMT), which for Seattle was 18.5C (65.3F).   According to their methodology, more folks start dying when the temperature is higher, for example at 80F (26.7C).

From Climate Advances article supplementary materials (click to enlarge)

There are so many problems with their approach, it is hard to know where to start. 

First, they ignore the deaths from cold temperatures and the fact that global warming will reduce such mortality.   As documented in local media such as the Seattle Times, there are a number of homeless deaths each year due to exposure (cold weather).  But such deaths are probably eclipsed by the toll produced from roadway deaths, due to fatal accidents on icy roads.   I know of dozens of such deaths from legal cases that I and other local meteorologists have worked on--and those are the tip of the iceberg.  These deaths will be reduced as the region warms.

Second, they assume the Seattle folks will stubbornly refuse to purchase air conditioners and the local government will fail to provide more cooling centers as the region warms.   That with fifty years of warning, Seattle folks will not do what residents of every warm city in the U.S. has done--get AC.  This is ridiculous.

This study assumes Seattle residents will not buy these devices.

Third, they don't determine the cause of death in their statistical analysis, which is very relevant.   When temperature start warming up around here (yes above 65F), what do people do?  They rush out for active outdoor activities, which tragically, but infrequently, leads to loss of life. Cyclists are being killed all the time.  Folks die floating down our ice-cold rivers, falling while hiking (e.g., Rattlesnake Ridge), boating accidents, to name a few. These deaths increase with better weather.     This paper should have determined the numbers truly heat-caused deaths (e.g.,  heat stroke), but failed to do so. And then the authors linearly extrapolate the death rate to higher temperatures for which they are obviously no data (dashed line above), with no basis for doing so.  Can one believe such an extrapolation?   Very doubtful.

But it is worse than that.  Their analysis is so problematic that for several cities they get seriously silly results.  In Atlanta and San Francisco, mortality goes DOWN at very high temperatures.  And these are very different environments:  one is steamy hot and the other generally cool during the summer.  Clearly, something is wrong.

From Climate Advances article supplementary materials (click to enlarge)

And in Phoenix they found the minimum mortality when the daily average temperature was 34.5 C or
94F.    Sounds like the residents of Phoenix have entrained some lizard DNA!

From Climate Advances article supplementary materials (click to enlarge)

So we have a paper in which they ignored the reduction in cold-period mortality, assumed Seattle residents would not adapt and buy air conditioners like everyone else, did not separate our true heat-related mortality, and make simply extrapolations into temperature regimes for which there is really little clue on heat-related impacts.    The problems note above by themselves should have been red flags to the reviewers and alerted the media that this material does not belong on front page of a major newspaper. 

Unfortunately, the problems with this paper don't stop there, but includes their climate projections for Seattle and other cities.

Their climate projections are based on a single modeling system, run many times in what we call ensemble mode.  The resolution of their system was very coarse (a grid spacing of roughly 150-200 km) and unable to properly simulate coastal, inland water bodies, the coastal zone,  and the terrain effects so crucial for weather or climate prediction in many locations (particularly Seattle).

The figure below from their paper (showing the difference in warming between their high and low forcing simulations) gives you a feeling for the blocky low resolution nature of their modeling system and how inadequate it is near the coast.

From Climate Advances article (click to enlarge)

This paper uses collections of climate simulations based on different greenhouse gas forcing that warm the planet by 1.5C, 2C and 3C by the end of the century.  Then they applied a simple bias correction scheme based on the errors of their model during a short contemporary period.  There are all kinds of technical issues with what they did (I will not get into them here), but their results, presented in their supplementary information, shows the major flaws in their simulations.

To see the problems, consider their table S2 shown below.  I will do my best to explain it clearly.

This table shows their results for a number of U.S. cities.  The first column shows the observed maximum daily temperature for a short period (1987-2000).  (As an aside, the table is mislabeled, it is the observed daily mean maximum temperatures, not instantaneous maxima.  The reviewers should have caught this).  Why they use such an abbreviated contemporary period that certainly doesn't catch the extreme observed values, I do not know.   Then they show the bias-corrected, daily maximum temperatures for the end of the century for three climate projections (global warming of 1.5, 2 and 3 C).

Their observed period daily maxima shows Seattle cooler than San Francisco, which is cooler than LA.  Not unreasonable.    But their climate simulations show strange stuff. Seattle (36.2C) has more extreme warm days than Los Angeles (35.1C), and is hugely more extreme than San Francisco (31.8C).  Boston extreme warm days increase by 12.6C to 44.8C, making them more extreme than Dallas (42.1C).   Yes, Bostonians will be rushing to Dallas to escape extreme heat.

From Climate Advances article supplementary materials (click to enlarge)

I could go into more detail, but it is clear that their simulations are totally unreliable, particularly for a city like Seattle, with nearby cold water, nearby mountains, low summer humidities, and cool summer evenings.

The bottom line.

This is an advocacy paper with substantial technical issues.   That it got published shows the weaknesses of the peer-review system and a clear example to the public and the media  that they should be careful before taking a single paper like this too serious.   The fact that a major U.S. newspaper, such as the Seattle Times, rushes to put such obviously problematic work on the front page of its paper is disturbing.  This is not educating the public, but attempting to scare them and gain clicks.  One does not have to wonder why some vulnerable individuals are suffering from "climate anxiety" when local media publishes such unfounded and unsupportable predictions of massive deaths.

For those of us concerned about the serious issues of climate change, such papers and articles are a real set back.  Seattle will warm substantially by the end of the century if greenhouse gas emissions are not reduced.  There will be impacts (daily highs in the lower 80s rather than mid-70s in Seattle, reduced snowpack, bigger rain events), and the public and decision makers need accurate information about the impacts, not hype and exaggeration.    Articles and headlines like shown above, undermine the credibility of climate science and push the public to disengage from what is obviously bogus information.