ASA Connect

 View Only
Expand all | Collapse all

Climate change

  • 1.  Climate change

    Posted 06-16-2015 08:49

    Dear Colleagues:

    One of the pleasures of being an applied statistician (or if you prefer, data scientist) is being able to weigh in with some expertise on a variety of important subjects, both in our work and in public debate.  One of the most important subjects is climate change, and it's one where the statistical community seems to have been notably quiet.  Most of us are not climate scientists, it is true, but don't we have something to contribute in the analysis of temperature time series, and in the assessment of forecasting models? 

    I noticed years ago that global warming (as climate change was called then) seemed to have halted.  This observation was subsequently made by such proponents of climate action as the "New York Times" and the "Economist". They claimed that the halt was temporary, and perhaps it is, but to the best of my knowledge, none of the forecasting models predicted such a halt.  Now it is claimed that there is no halt, and NOAA, the leading agency pushing climate action in the U.S., has adjusted ("corrected") their data to support this.

    I would invite any of you who are interested to download their data on global average temperature and analyze it yourself.  It's at http://www.ncdc.noaa.gov/cag/time-series/global/globe/land_ocean/1/4/1880-2015.csv.  I did so, and in half an hour I had linear regression results.  Again, I invite you to do your own analysis and either confirm or refute what I found; namely,

    1. Going back 50 years, to 1965, global warming has definitely occurred.

    2. Over the last 15 years, global warming has stopped.

    I hope that some of you will take up this question, and look forward to learning what you conclude.

    Morris Olitsky



    ------------------------------
    Morris Olitsky
    Statistician
    USDA
    ------------------------------



  • 2.  RE: Climate change

    Posted 06-17-2015 22:46

    I completely agree with Morris Olitsky.

    Rather than ruining the ASA's professional credibility with political lobbying to increase funding for the IRS, NSA and food stamps, this is an area in which we have insight and a special expertise to contribute to the national debate.

    There are three questions: is the earth warming (GW), is it caused by or exacerbated by man (Anthropomorphic Global Warming or AGW), and, if so, how much does man contribute? (Two other important questions, "Can the US do anything about it by itself" and "Are the consequences dire if the predictions are true?", are beyond the scope of the science of statistics.)

    Almost all the knowledge on this topic come from two fields: data analysis and physics (thermodynamics and the greenhouse gas effect). Climatologists use physics to build models and data analysis to refine and assess models.

    Statisticians can and should be major contributors to the model assessment debate. I would go further than Mr. Olitsky. We should assess the data itself. Sensors have changed, the environment of their sites have changed, their number and sophistication has changed. How comparable are weather station readings from 1960 to readings of 2015 for example? What about readings from the 1850s? How comparable are satellite readings of the 70s to those of 2015? What other evidence exists (tree rings, ice core samples, etc.) and what is their accuracy, bias, reliability, especially comparing readings over 100, 200, 500, 2000 years? If we are measuring global temperatures, how good a spatial sample are these measurements to accurately extrapolate to a global measurement? What is the reliability of the models, their predictive intervals? Have any of the standard model fit assessment techniques been used?

    Next data question to me is, correlation  aside, what data evidence exists to say there is a casual relationship between greenhouse gases and warming? I understand the simple physics, but the earth is not a simple laboratory, but is an extremely complicated physical system which has undergone dramatic climate changes in the past (buried under a mile of ice and buried under a mile of molten lava). Again, statisticians could contribute a great deal to this debate,

    Also, what are the testable hypotheses? I believe that the standard models have been consistently wrong in their predictions. Whenever data don't fit the model (the 17 year hiatus), we change the data. This isn't how science is supposed to work. How do I hear that 4 of the hottest years on record occurred in the last 10 years, yet there is a hiatus, even cooling? Which is it? It seems as though the current thinking is that AGW is the null hypothesis, when in fact it should be the alternative. 

    This is an area where the ASA could uniquely provide a valuable national service by being a truly unbiased, fair, objective assessor of data, uncertainty, models and model fit, and conveying same to our Legislators and regulators. Our conveyance should not either "support" AGW or "deny" AGW, but merely give the strength of the empirical case for both views. This would be an ideal role for a professional organization. Perhaps a committee could be formed to analyze the science behind the data and models.

    I am not an expert and I am agnostic as to whether AGW exists or not. I would like to find the truth as best we understand it, free from the political fog within which this topic is always enveloped.



    ------------------------------
    Terry Meyer
    ------------------------------




  • 3.  RE: Climate change

    Posted 06-18-2015 07:52

    Global warming is happening, and there is no "hiatus."

    Your posts both make it very clear that you (like me) are not experts. The people who are experts overwhelmingly agree that global warming is happening. Do you really think that they don't understand the models that they are building and using? The only "political fog" surrounding global warming is coming from people who stand to be negatively impacted if we do something about climate change.

    The complexity of climate means that patterns over a mere 15 years are simply not as reliable -- there are too many other factors at play. In addition, any appearance of a lack of warming is generally driven by the fact that 1998 (the year most people trying to show a halt use as a starting point) was incredibly warm.

    Even given that starting point, 2014 was the warmest year on record (since 1880) and the ten warmest years have all been since 1998 (http://climate.nasa.gov/news/2221/). Even if the warming had slowed (it doesn't appear to have), we are still not recovering to previous temperature norms. CO2 emissions continue to rise (link), and, given that we know that CO2 acts as a green house gas, we can predict further warming going forward, even without complex models. The complex models can help further understand the expected course, but they are also not trying to predict the temperature in a given year. They are predicting a longer-term trend and they have largely been correct on that point.

    Despite your claim of neutrality, your posts read like a grab bag of political talking points from those who are refusing to acknowledge that the climate is changing. There simply is not a meaningful debate to be had about the science of whether or not warming is occurring. If you want to debate what should be done about this, that could be a great discussion to have.

    Phil Plait wrote a wonderful parable about the current state of the debate. I highly recommend it: http://www.slate.com/blogs/bad_astronomy/2015/06/10/cautionary_tale_a_parable_of_science_fiction.html

    ------------------------------
    Mark Peterson
    Asst. Professor
    Viterbo University
    ------------------------------




  • 4.  RE: Climate change

    Posted 06-18-2015 09:26

    I agree with Dr. Peterson that the science is pretty well settled in favor of anthropogenic climate change. (Although, I disagree with characterizing the evidence as "incontrovertible," as some are wont to do because this is empirical science and not mathematics, after all.)

    Peter Guttorp, a statistician at the University of Washington (not to be confused with Wash U in St. Louis) is very knowledgeable when it comes to the analysis of climate data; I recommend visiting his web page.

    Incidentally, I would encourage anyone reading this thread to be on the lookout for "manufactured doubt." (Although, to be clear, I do not believe anyone in this discussion here is intentionally being misleading.) We went through this before with tobacco and as someone who resides in Florida (at least for the time being) I would rather not be underwater before we decide to take appropriate action. As such, I would be in favor of a statement from the ASA acknowledging that the evidence supports the reality of anthropogenic climate change.

    ------------------------------
    Robert O''Brien
    ------------------------------




  • 5.  RE: Climate change

    Posted 06-18-2015 09:27

    Please do not start filling our inboxes with political talking points for this thread.  I can see this thread getting out of hand very quickly if people start talking data points over an issue as large as this; it certainly won't be hashed out over a discussion board.

    I do wonder if the entity of ASA should be taking an official stance on the topic of global warming.  Since the mission statement includes "advancing research and promoting sound statistical practice to inform public policy and improve human welfare," I think there is grey room for interpretation. 

    Does this mean ASA is 1) taking a public stance on issues, 2) whether the ASA should be giving it's seal of approval on studies for public policy ("done in cooperation with ASA", e.g.), or 3) whether the members of the ASA need to individually make sure they're unbiasedly involved with issues in areas of their own expertise?


    ------------------------------
    Weston McManus
    SAS Institute
    ------------------------------




  • 6.  RE: Climate change

    Posted 06-19-2015 09:34

    Colleagues,

    If one takes the data given by Mr. Olitsky and plots a ten-year moving average the shape differs dramatically  --  it suggests (to me) that temperature is increasing and has been increasing sharply since about 1971  --  supporting most current scientific opinion  --  i agree with the challenge that statisticians should lend their voices more vociferously to this debate  -- thanks --

    ------------------------------

    Joe Viscomi
    Research Analyst/Adjunct Instructor
    ------------------------------




  • 7.  RE: Climate change

    Posted 06-22-2015 18:21

    Professor Peterson:

    Parables are not convincing arguments, at least to me. I'm also surprised that you used the argument that "the ten warmest years have all been since 1998".  That pattern is consistent with a statistical model of rising temperatures for 35 years followed by a hiatus (not a cooling) in the last 15 years.  Btw, my analysis doesn't depend on using 1998 as a base year; you can use other base years in that vicinity and still show no significant trend.

    The strongest argument here is that 15 years is not necessarily reliable; I'm ready to accept the climate scientists' arguments that the hiatus will end and warming will resume--when it happens.  Like Terry Meyer,  I am "agnostic" on this and would like to learn the truth.

    ------------------------------
    Morris Olitsky
    Statistician
    USDA
    ------------------------------




  • 8.  RE: Climate change

    Posted 06-23-2015 04:24

    Mr. Olitaky,

    Please explain how you analyzed the data to claim there's 'no significant trend.' 

    And then please explain how you get from 'no significant trend' to your original claim that 'global warming has stopped.' Surely you understand the difference between these two conclusions.


    ------------------------------
    Andrew Taylor
    University of Hawaii At Manoa
    ------------------------------




  • 9.  RE: Climate change

    Posted 06-23-2015 08:32

    Mr. Olitsky, you seems to have misunderstood my argument in at least a couple places, so I want to be clear.

    I did not state that the "the ten warmest years have all been since 1998" meant the climate was still warming. Instead, I left that analysis to others, several of whom have posted very nice graphs from the data linked in this thread that you have chosen not to address (not to mention the extensive support from people in that field). Rather, I used that statement to support this claim: "Even if the warming had slowed ... , we are still not recovering to previous temperature norms" Which is nearly exactly what you said in your most recent statement -- and is still a potentially dangerous problem.

    Further, the parable was not meant to convince any one of the scientific reality of the problem. Rather, it is to demonstrate the catastrophic effects of "agnosticism" leading to failure to act. What evidence would it take to convince you that climate change is real, and that we should do something about it?

    You state that you are ready to accept the argument when climate change resumes -- last year was the hottest year ever recorded. What other evidence are you looking for?


    ------------------------------
    Mark Peterson
    Asst. Professor
    Viterbo University
    ------------------------------




  • 10.  RE: Climate change

    Posted 06-18-2015 09:48
    There _are_ statisticians working on these issues, and there's an ASA Section devoted to Statistics and the Environment. I strongly encourage interested parties to join that section and study the relevant literature in depth as an alternative to forming committees of non-experts to weigh in on a politically charged issue. 







  • 11.  RE: Climate change

    Posted 06-18-2015 16:16

    Wow! Two emails in a row both saying that the ASA needs to take a leadership role in global warming and both are clearly from people who propose (at least implicitly) that there are two sides to the argument. This is a classic example of the fallacy of the golden mean. If some scientists believe in global warming and some scientists do not, that means that the truth is somewhere in the middle.

    Some arguments have two sides and other arguments are settled. Persisting in the belief that a settled argument is unsettled leads to a waste of resources. For example, there are some scientists who believe that the HIV virus does not cause AIDS. They are welcome to express that opinion, but it would be folly to spend scarce research resources examining therapies that do not rely on the link between HIV and AIDS. There are (still!) some scientists who believe that cigarette smoking does not cause cancer. Should we entertain their doubts and re-examine all the observational studies linking smoking and cancer to find their flaws? There are some scientists who believe that all humans have an invisible energy field and that trained therapists can manipulate this field to cure diseases. Should we run rigorous and expensive clinical trials to study this?

    An interesting question is what does it take to decide that a research question is settled in a particular direction and that further research to prove the opposite is a waste of resources. Mitch Gail has a nice article in JASA that shows how the arguments about smoking and cancer was settled in spite of what the tobacco company lawyers were saying (Gail. Statistics in Action. JASA (1995) 91(433): 1-13). It's also worth noting that just about every settled question still has scientists who persist on arguing for the losing side. And these people are very intelligent. R.A. Fisher was a life long skeptic of the link between smoking and cancer.

    Is the hypothesis that greenhouse gases have caused global warming a settled hypothesis? Well, yes, but you can still find a few scientists who are beating the dead horse of global warming denialism. NASA has a nice page about the evidence and cites sea level rise, global temperature rise, warming oceans, shrinking ice sheets, declining Arctic sea ice, glacial retreat, extreme weather events, ocean acidification, and decreased snow cover. For details, see Global Climate Change: Evidence

    Climate Change: Vital Signs of the Planet remove preview
    Global Climate Change: Evidence
    The Earth's climate has changed throughout history. Just in the last 650,000 years there have been seven cycles of glacial advance and retreat, with the abrupt end of the last ice age about 7,000 years ago marking the beginning of the modern climate era - and of human civilization.
    View this on Climate Change: Vital Signs of the Planet >

    But what's worse is that both letter writers argue that ASA has not been doing enough about this. A quick Google search shows this is false. The ASA has a standing committee on climate change. For details, go to ASA Advisory Committee on Climate Change Policy (ACCCP).

    Amstat remove preview
    ASA Advisory Committee on Climate Change Policy (ACCCP)
    ASA Participates in 5th Annual Climate Science Day on Capitol Hill, February 11, 2015; Photo of Michael Stein with Congressman Bill Foster
    View this on Amstat >

    The would also find an ASA policy statement from 2007 that supports a UN report asserting that "warming of the climate system is unequivocal, as is now evident from observations of increases in global average air and ocean temperatures, widespread melting of snow and ice, and rising mean sea level. … Most of the observed increase in globally averaged temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations. … Discernible human influences now extend to other aspects of climate, including ocean warming, continental-average temperatures, temperature extremes, and wind patterns." For details, see ASA Statement on Climate Change

    Amstat remove preview
    ASA Statement on Climate Change
    Adopted 11-30-07 by the ASA Board of Directors The American Statistical Association (ASA) recently convened a workshop of leading atmospheric scientists and statisticians involved in climate change research. The goal of this workshop was to identify a consensus on the role of statistical science in current assessments of global warming and its impacts.
    View this on Amstat >

    The fact that neither letter writer was aware of these efforts is a sign of intellectual laziness. Please do a bit more research before you toss out incendiary statements.

    Every few years, there are interesting and informative presentations at the Joint Statistical Meetings about the complex statistical models associated with global warming. Maybe the letter writers should attend these talks and raise their concerns and questions there.

    But for ASA to somehow become an honest broker between two sides of an argument where one side has all the ammunition and the other side is shooting blanks makes as much sense as for the ASA to re-open the investigation of who shot John F. Kennedy.
    ------------------------------
    Stephen Simon
    Independent Statistical Consultant
    P. Mean Consulting
    ------------------------------



  • 12.  RE: Climate change

    Posted 06-18-2015 19:23

    Accusing other participants of "intellectual laziness" because they were unaware of the ASA's 2007 statement is uncharitable and completely unnecessary; it is sufficient to point to the evidence, which is compelling. This is a professional organization and we ought to act like professionals.

    ------------------------------
    Robert O''Brien
    ------------------------------




  • 13.  RE: Climate change

    Posted 06-19-2015 01:22

    WOW. Such emotion! I obviously hit a hot button (pun intended).

    I said I was not an expert. I said I was agnostic. So let me pursue it a bit.

    Let me ask those of you who replied to point me to a knowledgeable (meaning written with statisticians among the lead authors) article written in a peer-reviewed journal that goes through in detail the underlying temperature data upon which AGW relies: what sources, what measurements, where measurements were taken, when, what biases exist, what error rates, how good is the spatial sampling, what "adjustments" are made and why, etc. I am again no expert, but am curious what temperature records exist for any length of time with sufficient accuracy to judge future global temperatures (100 years hence) to fractions of a degree. (I am asking here only for temperature records.)

    Second, could those of you familiar with the AGW literature explain to me the essence of the "A" part of the AGW versus simply GW. I understand the simple physics of greenhouse gases. If carbon dioxide increasing due to fossil fuels combined with the theory of greenhouse gases is the answer, fine, but I was hoping for more: some empirical evidence that separates AGW from GW beyond simply CO2 concentration is increasing.

    Again, I'm agnostic but I am curious.

    Thank you.



    ------------------------------
    Terry Meyer
    ------------------------------




  • 14.  RE: Climate change

    Posted 06-19-2015 15:57

    Pointing out the "intellectual laziness" of the initial two participants was "charitable" and not "completely unnecessary." To only tie the intellectual laziness to a lack of awareness of ASA statements and committees was quite generous.

    As for professionalism, it is certainly not professional to insinuate that the scientists and government agencies studying climate change are unaware of complex nature of their research area or to make unsupported aspersions about the integrity of those scientists and agencies w.r.t. their data. At the same time, it is not unprofessional to use a pithy phrase to point out that someone challenging the voluminous findings in an incredibly complicated and large area of study has not done any research. At times it not "sufficient" to just point to the evidence.


    ------------------------------
    Wils Corrigan
    Statistician/Geographer
    ------------------------------





  • 15.  RE: Climate change

    Posted 06-19-2015 06:34
    Morris,

    Thank you for the link to the dataset. I agree it clearly shows a consistent increase in global temperature from the late 60's. This can be clearly see by eye. However I cannot see any tailing off in that increase in the last 15 years as you claim.

    For example if you take say a 5 year moving average of the data the averages from 2008 onwards are a) higher than at any other time in the series and b) with the exception of 2014-2015 monotonically increasing. This is clearly inconsistent with your claim that the data shows that global warming having stopped for the last 15 years.

    Regards
    Tom Parke

    Tessella Ltd




  • 16.  RE: Climate change

    Posted 06-19-2015 16:10

    I agree with all those who have pointed out the hubris of Mr. Olitsky believing he, in half an hour, can do a better analysis of the climate record than can the many experts who have been working on this for decades. Nonetheless, having seen the data Mr. Olitsky claimed showed no warming over the last 15 years, I was curious how he came to that conclusion, and took his challenge to analyze the data myself, in the straightforward way suggested by his post.

    OLS linear regression on these data, for the years 2000-2015, gives an estimate of the slope of 0.0186471 degrees Celsius per year, with a 95% CI of (0.0090375, 0.0282566); the P-value of a two-sided test is 0.001. Visual analysis of the variability around this regression, e.g. a plot of residuals vs. fits, or simply a smoother on the original scatterplot, suggests that if anything the slope is increasing: warming has been accelerating.

    Mr. Olitsky, how can you possibly claim from these data that 'global warming has stopped'?

    ------------------------------
    Andrew Taylor
    University of Hawaii At Manoa
    ------------------------------




  • 17.  RE: Climate change

    Posted 06-22-2015 14:37

    I think most people would agree that climate data are highly correlated.  In such cases, drawing appropriate conclusions is never easy, as the title of the very short paper referenced below points out. 

    Yves Nievergelt (2011) Increasing Data with a Negative Slope, The American Statistician, 65:4, 262-262,

    DOI: 10.1198/tas.2011.11114

    -----------------------------
    Ronald Christensen
    Univ of New Mexico
    ------------------------------




  • 18.  RE: Climate change

    Posted 06-21-2015 23:18

    Thank you for providing the link but it seems dead.

    Bye


    ------------------------------
    Nicola Mingotti
    Universidad Carlos III
    ------------------------------




  • 19.  RE: Climate change

    Posted 06-22-2015 15:41

    The period at the end of the sentence interferes with the link. Try this:

    http://www.ncdc.noaa.gov/cag/time-series/global/globe/land_ocean/1/4/1880-2015.csv



    ------------------------------
    Robert O''Brien
    ------------------------------