SO hot

Cliff Mass, in his blog post about recent temperatures, asks the question “Why is the Northwest so warm?” He then proceeds to answer the question “Why is the Northwest warm?” What he doesn’t address is the question he himself asked: why SO warm? Perhaps we should say “so hot.”

We can see that it really was SO hot by comparing the monthly average temperature in the northwest this June, to that of previous Junes:


We can also see it in a histogram of June temperatures:


The lone value on the far right is for June 2015. Not only was this June 7.9 deg.F hotter than the long-term average, it was a full 2.6 deg.F hotter than the second-hottest June on record. It wasn’t just hot, it was SO hot.

To get more context on how exceptional it was, let’s plot data for all months, not just Junes. Of course we should remove the annual cycle and compute anomaly values, to compare each month to what is typical for that month. That gives this:


It makes this June (the final value) seem not exceptional at all. But this is a mistaken impression due to the fact that some months show a large anomaly simply because there’s more natural variability in winter months than summer months. We can see this in a boxplot of temperature anomaly by month:


Note how winter’s values cover such a large range compared to summer’s, so for them to show larger anomalies isn’t at all exceptional. Let’s correct for that by dividing each anomaly value by the standard deviation for that month, producing what I’ll call “scaled anomaly.” That gives this:


So June wasn’t just the most exceptionally hot June for the northwest, it was the most exceptionally hot of all 1,446 months on record.

How unlikely was it? Let’s ignore global warming for the moment, just pretend it hadn’t happened, and estimate the likelihood of such an extreme based on the data observed prior to June 2015. Fitting an extreme-value distribution suggests that this June was a once-in-1200-years event. That’s quite exceptional, certainly justifying the moniker “SO hot.”

But global warming did happen, it even affected the northwest. We can estimate its effect by fitting a smooth (modified lowess) to the anomaly data:


It’s hard to see on this scale, but northwest temperature has increased by about 2.5 deg.F since 1895.

Let’s replace anomaly values with detrended anomalies by subtracting the estimated long-term trend. Then, to see how exceptional June was, we still have to allow for the fact that winter months show so much more variation naturally than summer months. We’ll re-compute the standard deviation for each month and divide the detrended anomalies by their month’s standard deviation to compute “detrended scaled” anomalies. Now we have this:


Having removed the estimated effect of global warming, June’s scaled anomaly is no longer the most extreme on record. By the narrowest of margins it’s now the 2nd-most extreme.

Again fitting an extreme-value distribution suggests that this June was only about a once-in-90-years event. It’s hardly exceptional to find that in slightly more than 120 years of data. Essentially, when we remove the effect of global warming we see that this June in the northwest was hot, one might even say quite hot — about once-in-a-century hot — but not so extreme as to qualify for “SO hot” with a capital “SO.”

That’s what global warming has done to the Pacific northwest in the USA: made what would have been a once-in-1200-years event into a once-in-90-years event. The extremity of June’s temperatures have been made more than 10 times as likely. The reason is: us.

But Cliff Mass seems insistent on dismissing the effect of man-made climate change on this, as well as other, extreme events at every turn. He even insists that “the more extreme the weather anomaly, the less likely it is to be caused by human-induced (anthropogenic) global warming.” He just can’t seem to wrap his mind around the fact that without global warming, it would indeed have been hot in the northwest this June but it wouldn’t have been SO hot.

He even trivialized the effect of global warming with this bit of misinformation:

While the Northwest has been hot and dry, much of the eastern U.S. has been much cooler and wetter than normal. Even the Rockies have been wetter than normal. Global warming would warm them as well. This week I was at a weather conference in Chicago….it was quite chilly there at times.

Was much of the eastern U.S. really “much cooler … than normal” this June? Not according to the national climate data center:


Only three New England states were notably cooler than normal, and not one of them was anywhere near as much cooler than normal as the northwest was hotter than normal. It’s the kind of offhand remark that’s easy to spout to give the impression you want to give, but falls flat on its face when we look at the actual data.

As for Chicago being “quite chilly at times,” that’s about as “scientific” an argument as Oklahoma’s Senator Inhofe showing up on the senate floor with a snowball during winter.

19 responses to “SO hot

  1. On your graph, can you put the June value in a contrasted color, maybe with an arrow ?
    Thank you.

  2. “made what would have been a once-in-1200-years event into a once-in-90-years event.”
    Shouldn’t that be the other way round (would have been 1/90 event)?
    Or do you mean that an event of that magnitude now has an expected return interval of 90 years instead of 1200?

    [Response: The latter.]

  3. Once again, Tamino, you and Cliff probably agree more than you think. His comment replying to “Mal Adapted” (similar one to last time)…

    Mal Adapted,
    You are only partially right. It is certainly true that global warming will tend to shift the distribution of temperature variability, so that one is much more likely to beat previous records. Certainly true. But magnitude counts. This event is MUCH larger than the shift of the probability distribution. To put it another way, the 1-degree shift of the probability distribution does not explain a 10-20F anomaly. Yes, perhaps a degree of it is associated with human-induced global warming. But with or without increasing greenhouse gases there would have been a huge heat wave.

    Think of the Fukashima disaster. Global warming induced sea level rise contributed slightly, but with the tsunami disaster would have happened anyway…clif

    So, large temperature events like the one being explained could happen whether or not humans were on this planet.

    And, Cliff even agrees with you (indirectly) that AGW accounted for a degree of this event. He is not saying (as you stated in your article), that it did not cause this event at all.

    [Response: Once again, Cliff and I disagree more than you think.

    He has consistently attempted to minimize the effect of global warming on extreme weather events, and that bit at the end about “much cooler on the east coast” and Chicago having a few cool days shows just how deep in denial he is.

    His repeated downplaying of the impact of climate change makes it harder to persuade people that THIS IS NOT NORMAL, let alone the fact that this is dangerous (ask firefighters in the western U.S. and Canada). It’s going to get worse. Will this kind of heat wave have to become an every-year occurrence before Cliff Mass stops trying to make global warming sound like “no big deal”?]

    • Replying to Tamino,

      Well, heat waves of this type would have to demonstrate some sort of trend, from my own perspective. I believe the climate science, and think it will probably develop stronger trend of recurrence faster than Cliff thinks.

      I don’t think he is downplaying climate change at all, he seems to be acknowledging the AGW as being “part” of what is occurring. In any case, it happened, it ended as an “event” (for now).

      Do you think AGW is 100% of the cause of this event? I am getting that impression based on your writing here (part of the reason I’ve been commenting, to understand). Cliff, based on what I was able to read, seems to think about 10% of the cause was AGW based.

      [Response: No I don’t think it was 100% the “cause.” I wouldn’t even refer to it as the “cause” of its being hot — that happens. But it was, in my opinion, overwhelmingly the cause of its being SO hot.

      From my reading, Cliff Mass only begrudgingly admits even 10% attribution, and more important, has worked consistently to downplay the impact of climate change. Also my opinion: that’s destructive behavior, and it’s disingenuous of him to ask why it was “SO” warm, then avoid the “SO” part.]

  4. Just to clarify, by the “Pacific northwest” do you mean Washington, Oregon and Idaho? And did you remove a linear trend or the smoothed fit shown in red on the sixth figure?

    [Response: “Northwest” is the climate region so defined by the national climate data center. I removed the smoothed fit rather than a linear trend, since the trend over this time span is demonstrably nonlinear.]

    An interesting point about temperature variations being larger in the winter months than the summer ones. In hindsight this seems obvious for many temperate climate zones… when I get a chance I’ll examine the data for the regions I’ve lived and worked in. For my location, just off the top of my head I can recall a ~35°C extreme range for winter days but only a ~20°C extreme range for summer ones. I won’t speculate about how that might average out over full months or larger areas.

    • michael sweet

      In Hansen’s paper about extremely hot summers he calculates the number of standard deviations that the mean temperature has changed. He states that although the winter has warmed more than the summer (as predicted by scientists in advance), the anomaly divided by the standard deviation has changed more in the summer because the standard deviation in summer is smaller. That was for a global calculation.

  5. There may also be a trend in the variability, which is very important for changes in extremes.

    Especially for the average over the North West USA; because the number of stations has increased there may be a decrease in variability.

    A climatic change in variability should thus be studied keeping the number of stations fixed or per station.

  6. One other important point is that you are describing the daily average, not the maximum. When we get heat in the NW, it is air from the deserts to the east. The air being very dry, cools off quickly at night, making temperature minima stay relatively flat. To drive the average up, the daytime temperatures really go way up.

  7. The NCDC plot you put up shows the anomaly compared to 20th century. Weather people tend to talk about the 1980-2010 climatology, which is about 1F higher. Remove 1F from each state and you end up with … oh darn, still it’s mostly positive in the East. Nope, even that can’t rescue Cliff Mass.

    Chicago *is* very cold at times. When I lived there I always needed to keep a sweater in my office, which was typically chilled to 65 F in summer.

    • Once again with actual data:

      Below “average” (actually the median) in the northeastern states and the northern midwest.

      That was a cherrypick — I picked the rank of maximum temperature. Taking the rank of minimum temperature you instead get most of the country “much above average.”

      The Eastern half of Canada is actually cool by recent norms, and Quebec is cool by historical norms.

  8. He did this kind of thing once a month when I lived up there. He really is loath to say man-made climate change is happening. Oh, sure he’ll say something once in a while to keep the critics at bay. But still.



  9. Philippe Chantreau

    Meanwhile in Canada, they’re having about 10 times as many wildfires as they normally have this time of the year. It becomes hard to define normal when things are changing at the kind of pace we’re seeing.

  10. As I said in the Uncritical Mass comments, and what I think Tamino is getting at here, there is no longer any purely “natural” variation. There is now an AGW influence of some lesser or greater degree in every weather event. In extreme events that influence is most extreme and getting more so.

  11. I’m not fluent in discussing probabilities, but what does it mean to associate thresholding systems with probabilities distributions? When the jet stream flounders or collapses due to disproportional heating in the arctic, or when the long-stable thermohaline circulation is hit by large pulses of (fresh) meltwater, or when the moderation of winter freezing allows unprecedented increases in the population of bark beetles, how can probabilities math derived from prior behavior be applied?

    Sincerely curious,

    Spherical Cow

  12. Cliff is really good with day to day meteorology explanations. However, climate change such as the current drought he clearly does not understand at all.
    His prediction that the snow drought would be minor has been clearly demonstrated as wrong with much time left in the summer. Streamflows are at there lowest July flow levels ever in the Cascades for example and stream temperatures are record high too. This will be hard on glaciers, streams and salmon.

    • I agree. Cliff is a very good meteorologist at a very good institution. But his statements about climate change are just off-target and hopelessly misguided. It’s not just a matter of “different perspectives” or “magnitude vs. frequency” focus. Even going into the Cretaceous over 60 million years ago would not “cause” the heat wave by Cliff’s logic, and from a synoptic dynamics standpoint he’d be trivially right.

      He is doing the same thing now with ocean acidification in that comment thread, suggesting it is no big deal because the trend is smaller than natural variability. Yet somehow there’s countless publications and reports on why this is wrong.

      I’ve always thought we needed a better marriage of climate dynamics and synoptic meteorology, but in this case is turning into fated doom with unhappy partners.

      Part of the issue may be in how the different communities seem to think about causation. Synoptic meteorology has a lot of tools for diagnosing where air is rising and sinking, where the instability is, etc. Finding out where vorticity is being advected by the thermal wind, where q-vectors are converging, the phase/amplitude of the prevailing Rossby wave, etc are in the toolkit of the synoptician. Cliff is smart and not a “denier,” he understands this stuff. Yet somehow he is smart enough to deceive himself in how the statistics can change even when the amplitude of the synoptic event is much larger than the mean change.

  13. “Think of the Fukashima disaster. Global warming induced sea level rise contributed slightly, but with the tsunami disaster would have happened anyway…clif ”
    Onagawa Nuclear Plant survived – by less than a meter difference in the height of the tsunami versus its seawall. If we had had another meter of sea level rise, or if the Tohoku earthquake had waited another 75 years, would Clif admit that global warming was 100% responsible, since without global warming another disaster wouldn’t have happened? Or would he then claim that since the tsunami was 14.5 meters, and global warming added only a meter, that it’s only 100/15.5 percent responsible?

  14. I’m so glad that I found this. I’ve been reading Mass’s blog, and as a historian it’s clear to me that his arguments about global warming and its effects are epistemologically weak, e.g. arguing that PNW’s hot spell was not a result of global warming because it’s localized even though western Canada and Alaska were experiencing a heat wave at the same time. I haven’t the background to determine whether he was just doing a poor job translating the science for a popular audience or if he isn’t as knowledgeable as he believes himself to be.