Thursday, February 1, 2018

Brady or Belichick, Weather or Climate?

A few days ago, Bill Barnwell of ESPN asked the question, "Who contributed more to the Pats' dynasty, Tom Brady or Bill Belichick?"

That's a great question to ask for barroom conversation because it is impossible to answer unambiguously and thus invites vigorous argument and debate.  One can play games like how many Super Bowls would the Pats have one without Brady or without Belichick.  It could be that the answer in both cases would be zero and that Brady and Belichick together add up to greater than the sum of their contributions.  Such an interaction is called synergistic.

Although it is not a perfect analogy, I thought of similar issues related to trying to tease whether or not weather or climate contributed more to a particular weather event or season.  For example, many people have been asking me if this season's snow drought is caused by climate change, a question that I loathe to answer because it is so difficult to disentangle weather and climate given our warming world. 

In reality, weather and climate are highly coupled and quite synergistic.  Extreme events are often weather events (e.g., heat waves), but those events may be more frequent, prolonged, and/or severe due to climate change.  If I invoke the Brady or Belichick analogy, Brady throws the 4th quarter comeback touchdown pass (weather), but Belichick creates the system that makes such a play more common (climate).  Perhaps under a normal coach, Brady has 20 4th quarter comebacks, but under Belichick he has 42. 

Which brings us to this winter.  The lack of precipitation this winter is clearly related to a very anomalous upper-level flow pattern.   In addition, it has also been unusually warm.  At the Salt Lake City International Airport, the mean temperature for this December and January was the 5th highest on record. 

Source: NOAA Regional Climate Centers
So, if we wanted to simply take sides in a barroom argument based on this evidence, one could argue that yes, it's been dry and warm, but there have been droughts in the past and this years warmth is not unprecedented.  The 76/77 winter was worse, and there have been years in the past equally warm, so this could just be weather or natural climate variability.  One could also simply say this is due to persistent ridging, and thus is consistent with such variability.  On the other side would be someone who argues that global warming is affecting the jet stream and shifting global temperatures to make the likelihood of such warm weather more likely.  Hence, what we are seeing this year is a reflection of global warming.  

I don't find either of these arguments very compelling because weather and climate are so intrinsically coupled, much like Brady and Belichick.  This is also why I HATE being asked how climate change affects a weather event.  As a scientist, I am interested seeing in depth analysis of historical observations and numerical simulations to investigate this problem, which is known as extreme event attribution (see the climate.gov article Extreme Event Attribution: The Climate Versus Weather Blame Game for more information).  Such studies assess how the likelihood of an event has changed in a warmer world.  The approach is not without its warts.  A lack of historical data and limitations of models sometimes create some issues, but provides insights that are not possible if two climate scientists simply sit in a bar and take sides.  For example, we might not be able to say that this winter was caused by global warming, but instead how much more likely (or potentially less likely) this winter was given global warming.  

"Potentially less likely" might seem unlikely (did I just say that?), but much depends on the question you ask.  A warm winter like we have had is almost certainly more likely today than it was 100 years ago.  The dice are loaded as I like to say.  On the other hand, it could be that the lack of precipitation is less likely.  That might sound crazy, but most climate models run for the last IPCC assessment report increase the average wintertime precipitation over northern Utah.   Thus, what is meant by drought is important.  The answers one might get might depend on if one is examining precipitation amount or snowpack water equivalent.  They also depend on how far in the future that you look as the response sometimes requires more warming to become detectable.  

None of this is to suggest that the climate is not changing and that the ski industry will be just fine at the end of the 21st century.  Change is happening.  Weather variability will give us some good snow years in the future, and Utah's colder, higher altitude climate gives us more resiliency than other regions, but during the coming decades, we are going to see the emergence of a dramatically different snow climate.  

10 comments:

  1. In a future post, could you address the different base lines that one might use to define a drought. Current Salt Lake newspaper articles state that we are in a 1 in 30 or 3% snow year. I ran a distribution analysis based on total snowfall year-to-date from October through the end of January for alta and came up with 3.7% during 1990 to 2016. But looking at the total record from 1949 through 2016, it is bimodal distributed. Using 1945 through 1989 as a Baseline, total snowfall at Alta of 143 inches yields this is a 38.6 percentage year. Utilizing the entire record, 1945 to 2016, this is a 25% snowfall year. The analysis really doesn't change if you use water equivalent inches. What is the best way of describing the current year? Is it a 3%, a 25%, or a 38% year. Thanks Kurt.

    ReplyDelete
    Replies
    1. It's complicated. In the case of snow, much depends on what measure that you use. Snowfall amount, total snow depth, water equivalent of precipitation, or the water equivalent of snow on the ground can yield different numbers. In addition, the site at which data has been collected in that long-term Alta record and this affects the statistics as well.

      Finally, there is the percentage of average (or median) or the percentile of the distribution, which are two different things.

      For skiing and most hydrologic applications, the water equivalent of snow on the ground is probably the most important parameter. At the Snowbird SNOTEL, the water equivalent of snow on the ground is 61% of the median and 52% of the average of what has been observed on 1 February over the past 29 years. The lower number relative to average reflects the fact that a small number of really big years inflates the average.

      Alternatively, the water equivalent of the snow on the ground is the 3rd lowest on record (out of 29 years), which makes this a 10th percentile year. 2003 and 2007 actually had less water equivalent at the site than this year, but only by about a half an inch (which isn't much).

      All of this is an example of why statistics can be so confusing. I don't think there's anything nefarious going on in the news reporting for this year. It's just in a media report, the details are often left out. However, you can find many examples of nefarious use of statistics in politics and marketing. Although old, see for example the classic book "How to Lie with Statistics."

      Jim

      Delete
    2. Oops. "the site at which data has been collected in that long-term Alta record *has changed* and this affects the statistics as well".

      Delete
  2. Professor;

    Great post as a lot of locals have been wondering about this season at length. In your opinion is the lack of winter below 8,000ft a trend or a couple bad years? I have been in the valley since 91 and it seems like cold winter has been on a steady decline below 8k but I wonder if the data actually shows that. Thanks for the ongoing weather education!

    ReplyDelete
    Replies
    1. This question gets at the heart of the post to some degree. If forced to answer this question, I would say "yes."

      What I mean by this is that declines in winter snowfall and snowpack are a trend in Utah at lower elevations. We discuss this at http://wasatchweatherweenies.blogspot.com/2012/04/look-at-snowpack-trends.html.

      At the same time, there is considerable year-to-year variability due to the weather patterns that predominate in any given season.

      As such, this winter is consistent with the long term trend. The severity reflects the predominant weather for the season.

      What is needed is a year where the predominant weather for the season is cold and wet. One will come. I am counting on it.

      Jim

      Delete
  3. The point that I did not explain very well is that the period of record currently being used is 1981 to the present. This skews what we consider to be a drought year because it includes the extraordinary years between 1982 and 1986. Looking for example at the long-term level of the Great Salt Lake there are two extraordinary rises in the lake level. The first was 1865 to 1875. The second was from 1982 to 1986. By including 1982 to 1986 in the period of record, this presents a distorted distribution of snowfall in precipitation and what we would consider to be a drought event. Looking at Alta snowfall in excluding those years for October through January, the current snowfall is a 25% not a 3% event. I did not mean to imply that anything nefarious was occurring, just that the period of record presents a distorted view which can mislead the public. If the public is misled through bad presentation then there is a potential for backlash. A broader. Of record excluding the hundred-year storms from 1982 to 1986 gives a different picture. We are returning to a normal drought weather pattern, but the valleys population has exploded by about 50% since the 1970s. By having a period of record beginning in 1981, I feel people get the missing pression that the lake is somehow going to fill back up. Looking at newspaper articles from 1875 to 1930, this was the same sort of thinking that led to the Water emergency during the drought of 1934. Thanks, Kurt.

    ReplyDelete
    Replies

    1. The tree ring record can also inform this discussion. I've had a post on this in the back of my mind for a while, but need to do some investigating.

      Jim

      Delete
  4. Lookin forward to it. Kurt

    Bekker, M. F., Justin DeRose, R., Buckley, B. M., Kjelgren, R. K., and Gill, N. S. (2014). A 576-year Weber River streamflow reconstruction from tree rings for water resource risk assessment in the Wasatch Front, Utah. Journal of the American Water Resources Association, 50(5), 1338–1348.

    http://doi.org/10.1111/jawr.12191

    ReplyDelete
  5. Hey Jim - interesting post. I agree that weather and climate are intrinsically coupled, but if you agree that the climate has changed, then isn't the answer to "is XX weather event the result of climate change?" always yes? No weather event is the same now because it is coupled to climate change. Answering the question "is XX weather event more or less likely now because of climate change?" requires investigation of that particular event. The recent research on the strengthening of the North American dipole sure is compelling and scary for western skiers. Would be interested in hearing your take on it: http://weatherwest.com/archives/5982
    -Jeff Massey

    ReplyDelete
    Replies
    1. The problem isn't the answer, it's the question. It's simply not the right one.

      I think the work that you cite is interesting. It's too much to discuss though in a comment.

      Jim

      Delete