The coming death of the oceans

I wrote a couple of years ago about the supposed increasing extinction of species, and want to spend today’s essay on a paper about the state of the oceans, or more accurately, about those who write about it. But first, a summary of just how different 2014 was in temperature terms.  All the data are now in, and they are summarised by Ole Humlum on climate4you, which I regard as the best of the dataset summaries. You can see there that 2014 was not a special year. Even those who think it was the ‘hottest ever recorded’ have trouble showing that it was. Another source, the European Centre for Medium-Range Weather Forecasts, says that 2014 was warm, but so were other years — and that the margins between them are smaller than the error involved.

I was tickled to discover that Gavin Schmidt, of NASA’s Goddard Institute for Space Science (GISS), the principal trumpeter of 2014 as the hottest year, had been quite matter of fact a year earlier about 2013, which was not, on any measure, the hottest. One more year of numbers isn’t in itself significant, he said thenWhat matters is this decade is warmer than the last decade, and that decade was warmer than the decade before. The planet is warming. The reason it’s warming is because we are pumping increasing amounts of carbon dioxide into the atmosphere. The problem for Dr Schmidt is that we are indeed pumping that stuff up, but the temperature is not rising in sympathy. You can see the problem in Humlum’s graph of CO2 and temperature, using GISS data, here.

GISS GlobalMonthlyTempSince1958 AndCO2-1

On the face of it, what we are seeing today is a kind of repeat of the period prior to 1975: the global temperature anomaly is not increasing, but CO2 is rising steadily.

OK, on to the oceans. I wrote a while ago about John Ioannidis, who decided that much published medical research needed replication, and that far too much of the ‘knowledge’ doctors relied upon was flawed. I came across a similar perspective on the frequently heard assertion that we are all going to die because we are ruining the oceans. You can read the paper here. ‘Reconsidering Ocean Calamities’ is by Carlos Duarte and seven others, most of them Australian, and it is thought important enough for Nature to have devoted an editorial to its message.

What is the message? [A]ccounts of the deterioration of the oceans stemming from the scientific community run the risk of conveying the hopeless notion to managers and the public that we are confronted with an insurmountable environmental crisis of gigantic proportions. Although emphasizing problems may be intended to propel remedial action, it may achieve the contrary, because an overly negative message may lead society into pessimism or the belief that the ocean is beyond restoration.

No one who reads the Australian press or follows news on our television could dispute that. We are almost bombarded with claims that the Great Barrier Reef is under severe threat, or on the verge of extinction, while the oceans are turning acid in front of our eyes. Duarte et al comment on both issues.

How has this happened? [W]e contend that the marine research community may not have remained sufficiently skeptical in sending and receiving information on the problems caused by human pressures in the ocean and that there is a need to revisit the process by which potential or isolated problems escalate to the status of ocean calamities… Scientists are expected to remain skeptical, questioning, or doubting or to suspend judgment until sufficient evidence and proof is offered to draw conclusions through organised skepticism… However, there is a perception that scientific skepticism has been abandoned or relaxed in many areas, which has allowed opinion, beliefs, and tenacious adherence to particular theories to play a major role in holding beliefs based on interpretations unsupported by evidence. 

The authors do not dispute that there can be what they call ‘calamities’, and over-fishing is one of them. The collapse of the Newfoundland cod fishery is perhaps the most celebrated example, but Duarte and colleagues warn that it is all too easy to jump from one real example to an assumption that all the other possible issues will be calamities too.

[O]nce hypothetical problems have risen to the status of calamities in the literature, they seem to become self-perpetuating. Indeed, the marine research community seems much better endowed with the capacity to add new calamities to the list than they are to remove them following critical scrutiny… there are flaws in the processes in place to sanction scientific evidence, such as organized skepticism, that need to be addressed to help weed out robust from weak cases for ocean calamities.

They offer a long list of reasons for this situation, but conclude: The rise of ocean calamities has generated a worldview in which a host of ecological syndromes resulting from human-driven pressures is leading to the collapse of the ocean. The addition of new problems, such as new invasive species, ocean acidification and deoxygenation, or the perils from plastic pollution, to the litany validates and strengthens this worldview, forming a more compelling case for action to reduce human pressures. 

Although reducing human pressures on the marine environment is a positive outcome, this may provide a motivation to inadvertently—or, in worst cases, deliberately—fall into the white hat bias, defined as “bias leading to distortion of information in the service of what may be perceived to be righteous ends”… Clearly, no righteous end justifies the perpetuation of scientific bias.

It’s a good read, despite the gross over-use of ‘However’ to start a sentence. The Nature editorial deals with it fairly, and recognises that Duarte et al aimed a shaft or two at Nature itself. Since Carlos Duarte has managed to get a critical paper into the holy of holies, you might wonder about his future. He wonders a bit, too. Nature quotes him as saying: There are a lot of conversations around meetings about the excess doom and gloom in our reporting of ocean health, but perhaps this is the first paper to bring these concerns out of the privacy of peer conversations. This is a silent movement, as there is a lot of peer pressure against voicing those concerns openly, so my co-authors and I expect significant heat upon us to be derived from our paper.

I hope not.


Join the discussion 30 Comments

  • Walter Starck says:

    Environmental doomscryers have a noticeable propensity for claims
    of critical concerns wherein evidence is lacking or doubtful and thus difficult
    to refute. Species extinctions, ocean acidification, accelerating SLR, GH heat
    hiding in the deep sea, overfishing, and indeed AGW and Climate Change itself
    are all examples.

  • Peter Donnan says:

    Though an irregular reader of your site, Don, I return from time to time and am always impressed with your upbeat optimism and buoyancy about the latest doom and gloom reports on global warming etc. As I mentioned once before you remind me of Shakespeare’s “ever- fixed mark that looks on tempests and is never shaken”.

    This may seem an impertinent suggestion for a future posting – indeed you may have done it already – but I’ll propose it anyway!

    Recently I was reading a comment by Colin Nicholson (2009: 392) in his book on stock market investing where he wrote:

    “Having determined a price at which to buy when building a position in a stock, the next step is to decide the price at which the investment would fail. …Determining the price at which my investment will have failed before I place a subsequent buy order is critical to my whole plan. I call this my sell stop. My sell stop is a price level at which I will have determined, written down and marked on my chart of the stock. This is the price I will sell to cut my losses…..I regard myself as a fairly unemotional investor….we are all emotional beings to some extent and I am no exception, simply on the lower end through experience….”

    In terms of those who write about climate change etc, – either for, against, agnostic etc – it would be worthwhile have a clear statement about where their views have failed, so either there is nothing to worry about or perhaps we should relocate etc.

    * So in terms of years, 2013 or 2014 may not have been anything special in terms of past patterns of heat, drought, temperate etc. What sets of data in a year would be troubling, if any?

    * Even though CO2 levels concern many, are there levels that would be disturbing in your view?

    * What rises and levels in global temperature would lead to revision or abandonment in your optimistic thinking about global warming?

    * Are there any ocean calamities that would similarly lead you to a change in thinking?

    * If the Great Barrier Reef[Ice caps, areas of land etc] simply disappeared or collapsed into rubble, would that be a factor?

    Similar questions, too, would be applicable to those who most support global warming. It is true that these are hypothetical questions rather than ones that arise strictly from the empirical evidence. The point, too, about emotionalism, is relevant because as you have stated in the past, there are all sorts of doom and gloom agendas, scare-mongering, vilification of people/views that are dissenting, engendering of guilt, even messianic prophecies of doom.

    Is it possible, as Colin Nicholson does with stocks, to clearly state a sell stop position on global warming, which in effect is an admission (whether for supporters, deniers, naysayers, agnostics etc) that one’s previously held views have failed?

    • DaveW says:

      Hi Colin,
      I think most scientists would have a ‘sell stop’ when a test of a hypothesis proved counter to prediction. True, scientists have been known to make ad hoc adjustments to keep a failed hypothesis alive, but eventually the data is expected to win out. So, for example, the hypothesis that atmospheric [CO2] is the primary driver of global temperature is inconsistent with the data and people who believed that hypothesis should be cutting their losses. Unfortunately, this debate is not primarily about testing scientific hypotheses, but promoting political agendas and and setting a ‘sell stop’ point in a political debate seems to be a difficult thing to do.

      I’m not sure how much, if any, of the almost linear rise in atmospheric [CO2] that has been recorded over the last several decades is due to the burning of fossil fuels. The hypothesis that this rise is largely fuelled by human emissions does seem reasonable though and I’d be interested if anyone here could suggest a ‘sell stop’ point for this model?

    • David says:

      Excellent question!
      The “sell stop position” as you call it is when there is a greater than 5% chance that their observed effect is in fact zero. They then fail to reject the null hypothesis.

    • Don Aitkin says:


      I have done something like this before, though I don’t remember when, but not at this scale. Each of these questions would involve a whole essay, I think, though one could do a quick swipe. Eg., I would become concerned if warming continued unabated at significant steps over twenty years, even though this has occurred already in my lifetime. But I wouldn’t assume, unless there were clear evidence to the contrary, that this was due to more CO2, since there has been no significant increase in the last 18 years even though there has been a steady increase in CO2. I would be concerned more about how we coped with the increase. As for increases in CO2, they seem to be used profitably by growing things, to our common advantage.

  • chrisl says:

    Hello Don Just out of interest ,what do you think the ph level of the ocean is? (No googling!)

    • Don Aitkin says:

      Not a good question. There is no single figure for any ocean, and pH levels can vary at the one place over 24 hours, and over a stretch of ocean as well.

      I pass by the semantic question of whether there is such a thing as ‘acidification’ of the oceans

      • chrisl says:

        Good answer Don. My point was that the ocean is not even acidic( approx 8) and the scale being logarithmic there is little chance it ever will be,

        • David says:

          That argument is too silly for words. Earthquakes are measured on a logarithmic scale (aka Richter scale) so does that tell you anything about the probability of earthquakes
          becoming bigger or smaller.

          Obviously the issue is the change in the Ph in comparison to normal. For example, the normal Ph of human blood is 7.35 to 7.45. However, doctors will still refer to a reading of 7.2 as “acidic” even though 7.2 it actually alkaline (i.e. > 7.0). So what would you think of a doctor that did not bother to check their patient’s serum Ph because it is measured on a log scale? That would down like a “turd in a punch bowl” in the Coroners court. 🙂

          • Don Aitkin says:

            As with so much else in climate, there is no reliable measure of pH in the oceans, either over time or across space. On the wording, I agree that ‘becoming more acidic’ is commonly used, even though one could say ‘less alkaline’ or ‘moving towards neutral’. It does seem to be the case that those warring about climate invariably use ‘becoming more acidic’, and a sceptical person might reasonably wonder whether or not they do so because in part it is scarier.

          • David says:

            Yes I agree with you.

          • chrisl says:

            The other issue is how does the ocean become less alkaline?Apparently by taking co2 from the air and forming Carbonic acid.Considering there is hardly any CO2 in the atmosphere(400 parts per million) and it is rising very slowly (2ppm per year) and the oceans are enormous and carbonic acid is a very weak acid and ph is logarithmic,it is hard to see what all the fuss is about.

          • Don Aitkin says:

            And how many ordinary citizens know all that, do you think?! The scare is quick and easy (and not soundly based).

          • chrisl says:

            Isn’t Co2 an amazing compound! It can trap heat, melt glaciers, turn water into acid etc etc but somehow plants thrive on it!
            Meanwhile down in Hobart it is 14 degrees at the height of summer, They must not have got the memo!

          • David says:

            “….and ph is logarithmic”

            You obviously have not absorbed, what I thought was a fairly pointed critique of your fascination with log numbers. They are just a way of counting. How we decide to count something, whether it is in log numbers, base 10, binary numbers
            or Roman numerals does not tell us anything about the nature of what we are counting. For example the population of Canberra can be converted from base 10 to a log. So what does that tell us about the population of Canberra? Nothing
            of course.
            An change in human ph from 7.4, which is normal, to 7.2 will kill you!
            And any one who as ever tried to keep marine fish will know that they have to constantly check the water’s Ph, or else the fish will die.
            So it should be obvious, even to you, that Ph is important.

          • chrisl says:

            Do you understand what logarithmic means?
            It is not quite the same as roman numerals.
            Ph of 7 is 10 times more acidic than ph of 8.
            So how much acid needs to be poured into the ocean to increase the acidity tenfold?

          • David says:

            ANY base 10 number (except 0) can be converted to a log number. The issue in the example you provide is “the 10 times” bit, which may or may be correct, NOT whether the Ph is expressed on log scale or as a base 10 number.
            For example, if I argue, “that the sky is blue”, its no more true (or false) whether I say it in French or English.

          • David says:

            So where do you get your 10 times from? For example this
            article argues that a change in Ph due to CO2 will be 0.3 to 0.4.

            “By the year 2100, various projections indicate that the oceans will have acidified by a further 0.3
            to 0.4 pH units, more than many organisms like corals can stand. This will create conditions not seen on Earth for at least 40 million years.”


            If a 0.2 drop in serum Ph will kill a person it seems that a change of 0.4 in the Ph of seawater may be a problem for some marine life that have highly porous membranes like coral.

  • David says:

    “…but so were other years — and that the margins between them are smaller than the error involved.”

    So what! That does not invalidate a trend. It is not a valid statistical test to take the difference between to data points and compare that against “the error”. As anyone remotely familiar with statistical analysis will know data analysis will produce many errors terms. Which error are you referring to?

    • Don Aitkin says:

      Instrumental error. Recent thermometers are accurate to about 0.1 degree C, and the extra warming of 2014 was at around an order of magnitude smaller. There is no trend, unless you decide that there must be one, and take a straight line between the start and the finish. As any graph of global temperature will show, something happened around 2000, after which temperatures rise and fall a bit, but they don’t go on rising and rising. Last year was like many others: in comparison to say 1950 it was warmer, but in comparison to the years after 2000 it was much the same. So, if CO2 is responsible for the warming, then there is something else out there, single or plural, that can mask it.

      • David says:

        Don you are wrong!

        Even if this statement were true

        “Instrumental error. Recent thermometers are accurate to about 0.1 degree C, and the extra warming of 2014 was at around an order of magnitude smaller.”

        it does not invalidate a trend. Go an read a text on statistical tests. In particular focus on the difference between “error” and “bias” . The error term is mathematical construct that deals with random variation in the data. By construction the expected value of the error term is zero! Statisticians deal with error by increasing the sample size.

        The whole reason we do tests of statistical significance is to deal with error, (from whatever source e.g. measurement or climatic).

        Your statistical test is voodoo statistics.

        • Don Aitkin says:

          Oh dear. This is a semantic quibble. We do not have a sample of measurements, and we cannot therefore increase the size of the sample. What we have is the average of the actual measurements. We can of course express them to the limit of the decimals involved, but that is just precision, not accuracy.

          Better still, take it up with the people at the ECMWF, whom I was quoting.

          • David says:

            Don climate scientists increase their sample size by incorporating multiple data sets (surface temperature, troposphere or sea temperature), or by using monthly
            data instead of yearly data, or by going back in time further that 1998 (the denial-ists favorite date). Professor Muller
            had 14 million data points in his analysis by incorporating all the data he could find.

            And I went to that page and I searched on the term “error” and got no hits. They don’t use the statistical test you describe.

          • David says:

            “What we have is the average of the actual measurements.”

            YES! This is exactly what a denialist will seek to do. Hide thousands of data in means. Reduce 100 years of climate trends to 12 data points starting from 1998 and “wonder” why they cant find statistically significant trends.

            Give me a break 🙂

          • Don Aitkin says:

            The average I referred to is the figure used by NOAA and NASA. It’s not mine. Go and read what they said. I’m a reporter here …

    • dlb says:

      David, I took error to be the error of the mean i.e. the so called error bars you often see on scientific graphs. This has nothing really to do with measurement error but is analogous to the standard deviation of readings across the hundreds of temperature stations world wide. If the error bars overlap then there is no statistical difference between the warmest years, think ANOVAs.

      • David says:

        Yes, if asked to guess what Don was referring to, so would I.

        But having estimated some measure of dispersion (however it might be expressed) you DO NOT simply compare that to the difference in temperatures for 2013 and 2014 to try and infer something.

        If those error bars which you refer to, cross zero, then we can say that the relationship is not statistically different from zero. But they are NOT used to test the statistical difference between one data point and the next.

        • dlb says:

          Not sure with what you mean by error bars crossing zero?
          Wouldn’t a t-test explain whether there is statistical difference in temperature between two years such as 2014 and 2013? A year being a set of observations not a data point. I would think the error bars on any such comparison would give a pretty good approximation as to whether there is a significant difference between the years.

  • Don Aitkin says:

    I’ve now learned how to adjust the size of graphs to fit the space allotted. The Humlum graph, a nice piece of work, is now on display.

Leave a Reply to Don Aitkin Cancel Reply