Climate change, highlights and glances from the 2013 IPCC WG1 report. 0: Things to Say before the Report comes
By tomorrow, the 30th September 2013, the new IPCC WG1 report about climate change will be officially released.
I imagine you all have heard of, or seen, plenty of the uproar caused by the bad news (including a post of mine, in the Questions about Future to prominent women leaders). But this has not been triggered by the True Report itself: it has only come from the draft summary for policymakers. Sure this is an important instrument, as this, and not the huge and technical Report, is where decision makers and the general public will look.
Sure, what a field practitioner could find in the new summary for policymakers sounds quite bad news.
Before to imagine actual policies (or dismiss the whole thing as "Cassandra babble"), however, I strongly feel some preparation is highly needed. I decided to write about it in this post, to the benefit of anyone interested (and, of course, my students - they "know" my approximation to English by definition ;-) ).
First and foremost: if you can, please, read the Report. This is where things are really said. The summary to policymakers may be incomplete or biased. Or seem "too prescriptive" (what many ego-driven politicians lacking technical self-confidence hardly appreciate).
The past edition of the Report was very interesting, and the one to come will be even more so - not last because it contains many updates.
Why is the Report so "dubitative"?
If you are not a scientist, and want to delve into the new Report anyway, I guess you will be quite astonished by the apparent tone. I said the summary for policymakers sounds "prescriptive". Well, I bet the Report will sound "tentative" to many of you. The Report and the summary for policymakers have been written by the same people (more or less): so, how this difference of language can be?
The reason is, the language of science never states absolute certainties. Saying about the Absolute Truth is not the task of science: it's the realm of two other (also important) parts of human knowledge, namely phylosophy and religion. What science actually does is to build models (more often than not in these days matematical, but not necessarily so) allowing their users (we people, scientists and not) to make sensible expectations about the future.
All scientific statements, in any field, have one thing in common: they are formed (grammatically!) in a way to make them falsifiable, that is checking their possible invalidity independently on its authors' wishes by repeating (or devising new) "experiments".
Notice the apparent subtlety: a statement, to be scientific, must be falsifiable, and not just "verifiable". This may seem a bit odd, but in reality it isn't. Imagine, for example, a phrase dealing with some mathematical predicate, like "in all rectangle triangle the sum of squared cathetes equals the squared hypothenuse". We all know this statement is "true" since secondary school. But let's imagine having forgotten this. How could we say it is really true, or false?
In mathematics we have many ways open. We may try "proving" it is true, using rigorous arguments (like Pythagoras').
Or, we may try proving it's false by:
A) Proving its contrary is true. (Usually, a lot of intellectual work, starting from defining very tightly what we mean by "its contrary" - which may be all but trivial).
B) Finding a counterexample. (This requires some creativity, but it's often simpler than head-first "proving the contrary is true").
Mathematics, which is not an observational science, has devised many ways to develop rigorous "proofs" (and this might mean it is not a "scientific" activity in the strictest sense: is it very "humanistic"? I personally like to think so, mathematics akin to poetry and music more than mostly imagined). But this rigor is simply not possible in, say, physics or geology, where all we know is based on measurements (which have inherent errors) and experiments (which may be incomplete).
So, all we can do to "prove" something is showing it resists repeated clever attempts of finding a counterexamples. And for this to be possible, the "provisional truth" our friend-scientist will write about will be expressed in such a way to be falsifiable.
The "truths" of science (and to some extend mathematics too), hold really true only framed within many restricting conditions (that is, for example: "If this and that and true, and assuming lady Lovelace statement about numbers in class X holds, then Y"). That is, they're highly contextual truths.
By definition, an absolute truth is not contextual. Like "God exists". Or "World climate will get warmer because of anthropic activities". In strictly logical terms these statements cannot be "proven" true or false (logicians say they are "undecidable"). You may believe, for example, that God exists (I do, incidentally). But if you want to justify this, you will have to resort to faith and feeling: sensation and logics, in these cases, may be the right tool in the box.
An absolute truth as demanding as "God exists" cannot be proven true or false because the divine transcends human reality (at least, in most, not all, religions). In this case there is not much to say.
An absolute truth like "World climate gets warmer because of human activities" is, maybe, more interesting to our case. This is the kind of statement you maybe can find in the policymakers' summary. But I bet you will never find it in the Report. Among the reason for this, one looks specially attractive to me: it does no mention to time, and seems suggesting being true forever on. Of course this cannot be: if not, because sooner or later the Sun will finish her reserve of light atoms in a five billion years, and then something will change which likely will invalidate the statement.
Another (delicate) point is "because of human activities". There is no way to "prove" human activities (or, in the lingo of the Report, "radiative forcing") are the sole responsible of climate change. Of course natural causes do overlap. As well as causes no one still is able to tell whether they are natural or not - like the ENSO (the El Niño - La Niña alternation). So, nowhere in the the Report will you find (I bet) such a categorical statement.
However, if you add all together the partial evidences (from paleoclimatology, from Anctarctica ice cores, from the retreating glaciers record, from Greenland and Anctarctics ice shelves thinning, and so on and so on) you see the likelihood the statement "World climate gets warmer because of human activities" is false is really small. So, it may well be you will find the phrase (which is not falsifiable, then non-scientific) in the summary for policymakers.
This lifts what I see as a very big problem, of communication more than making science. Laypeople craves for certainties, and in their very deep many of them would love scientists (or magicians, or popes, or dictators) tell them how things "really are". Basically, this is unaware abdication to one's own duty to be a citizen, but undeniably the shortest way to tell laypeople about what things are considered to be "true" at best of scientific knowledge is presenting them dogmatically.
Sure, we could do differently (like urging you, as I do, to read into the technical Report). But this also demands you be informed of the logical subtleties of scientific discourse, and of your responsibilities in decoding and applying (or not) them. Surely this is painful. Demanding. It asks us to debunk myths about ourself, first and foremost: like for example "Oh, I'll never be able to think like a scientist"). And, accept the burden to accept the role of active, informed World citizens.
But this would attract only the bravest, more motivated and sensitive (as you are, having resisted until this point!)
What about the "big mass"? Including the many politicians with a technical expertise not higher than primary school proud of their ignorant, selected by corrupt systems rewarding questionless fidelity more than independence of thought and passion? (In Italy we have plenty of examples sitting at our Parliament).
So, I feel using a prescriptive and dictatorial language in the policymaker summary is ugly, but in the moment I can't foresee a more effective strategy.
Which does not says we have to stick to this tone forever. "We scientists" have to explain ourselves better (our "truths" are contextual, but we feel they are also valuable to all, and useable to envision policies). But in the mean time, "laypeople" have to evolve their sense of citizenship. And education systems must support this. For men and women.
Personally I don't trust very much in "scientific popularization". Unfortunately, science is hard (and I think there is no title of merit to scientists to live with this). But, meantime, science is at grasp level of most people - especially supporting each others in connection). And any scientist, in their youth, have born as laypeople. Determination and hard work, Prussian style, overcome any difficulty.
Models and "predictions"
Much debate has been made about the success or failure of "predictions" made by climatological models in previous IPCC reports, and surely they will repeat with the next to come and all following.
But. If you look to the previous editions of the Report you'll discover climate models play a relatively important-but-minor role. Most of WG1 effort has gone in reality into collecting past and present evidences, and perfectioning measurement methods: the big bulk of all Reports in fact is made of "hard data".
Professional scientists love data. They're something objective. Allow identifying (at least tentatively) a current state and past evolution.
State and evolution are, in a sense, more important than future trends. Why? Well, the Earth climate is an extremely complex system. To date, better known than in past years. But with many aspects still waiting for discoveries and attention. A system like climate is complex precisely because we don't know for totally "sure" (the scientists' way of meaning it! ;-) ) which observable phenomena are causes, and which are effects (the two categories are not mutually exclusive, because of back-actions). There is considerable consensus, that's right. But, no "absolute certainty" - and never will, because of the way sciene advances.
Emotionally, on the other side, people is more interested to know what will it happen to them, their beloved and their descendants in the next twelve minutes, fifty years, two million years (the interest usually decreasing as time span widens - which suggests many interesting considerations about the real motives behind this passion about future, maybe more small-minded and petty than imaginable).
Models are a way to identify trends, and build scenarios allowing to put a glance into what might happen. But given most people's deep emotional need to know future exactly, untrained people is inclined to imagine models as magic oracles capable to make predictions about future climate.
Well, they are not.
The climate is the long-term average of weather and its time variation.
Weather is the outcome of phenomena whose most distinctive trait is of being chaotic, that is, "sensitive to initial conditions". Said differently: it can "predicted" for very short periods, as we all see with weather forecasts. For periods longer than, say, 128 hours very little can be said of weather in advance.
What we may tell, at most, is how likely a given weather we can imagine now will actually occur after 128 hours from now. Likelihood can be (expressed and) calculated probability any specific weather will have over time. And since probability is a number, we could take at any given instant the maximum of it, and select the corresponding "most probable" weather.
Is the time sequence of "most likely" (higher probability) weathers a "prediction"? Let's say it clearly: No!!!
It may be interesting to see what a "prediction" is in technical sense: it is a number (or, as in our case, a monstruous set of numbers) expressing something related to future accompanied by an error bar. It is not a "most likely outcome" (whose probability is not an error band).
Error bands are common to measureents (which refers to something already occurred and sensed by an instruments). Having or not a true error band is what distinguishes Real Data from numbers (please my- and other-people-students, take note: the first question we'll ask you during discussion of your graduation thesis will be "Where the error band is?" Wo/man advised, half saved.. ;-) ).
Model outcomes from the IPCC Report do not have an error band, but an estimate of uncertainty. It is definitely not the same thing. "Error" is the displacement from something measured or predicted and a "true value" (e.g. measured with a better instrument or observed on the right time). "Uncertainty estimate" is an indication which, hopefully, tells us something about how much the numbers are self-consistent - roughly. And, it is an "estimate"! That is, a creative invention which, at best of our knowledge, aims at guessing something we can't measure.
In the specific case of IPCC Report, the estimate of uncertainty is obtained from the comparison of the most likelihood situations different models provide. Fortunately, there are various climatological models based on slightly different assumptions about causal relations within atmospheric dynamics - on long term scale. The fact more than one model exists does by no mean signify a fundamental lack of consensus. On the contrary, it is a good thing, and I add this variety has been constructed in part intentionally.
Anyway, comparing different models' outcomes is not the same than comparing a measured/predicted to a "true value". The difference is not, strictly speaking, an error: it "just" (just??!) weights the degree of agreements differing assumptions might have.
Some of the scientists in IPCC Working Group 1 insist that climatological models do not yield "predictions", but rather "projections". Which are a very different thing (and which can be applied to many other mathematical models, for example atmospheric pollutant dispersion models). Is this surprising? Well, not so, once you know how these models have been derived.
Let's do a concrete example. There is a lot of consensus that all motions in the atmosphere are described by a set of coupled partial differential equations named Navier-Stokes equations, after the names of the two investigators who independently introduced them in 18th Century. Navier-Stokes equation are nice, in the sense they provide a credible model of the atmosphere. But to date there is no way to "solve" them exactly. Would this be possible, solution would include any airflow including air molecules moving back and forth under pressure of sound produced as you or me read this letter aloud. Less phylosophically, "solutions" to Navier-Stokes equations are notoriously chaotic (that is, sensitive to initial condition) and irreversibly intricate.
OK. That's what equations are. If we do nothing, the hope of getting something from them is exactly zero.
But, we could do something! Namely, we may simplify the equations, and solve the reduced version. If simplifications are made in a sensible way, then their solutions may give us some useful indication about the behaviour of the real system. In actuality, Navier-Stokes equations are simplified by dropping some of their terms, based on scale analyses which in turn suggest which terms, in given circumstances, are important, and which are not. Then, non-important terms are imagined as being zero, and a solution to the less rich equations are then searched for. Sometimes a solution is possible (most often numerically).
Because of this, Navier-Stokes equation can act as "mother" of very many different mathematical models. And this is precisely what happens, with an important highlight: models (simplified representations of reality) are designed to someone who considers some phenomena (e.g. mean airflow instead of turbulence) as "more important". If you do not know the assumptions these designers based their simplifications on, you can't really say knowing what are you doing. (In the Report, or better the myriad of scientific papers cited in it, these assumptions are more or less clearly stated, so reconstructing them is "just" a game of patience.)
All this may be condensed in some simple tiny reflections:
- Judging about the real contents of the Report is not necessarily simple: it demands possessing much background information and expertise.
- Corollary: it is highly unlikely a single person will have individually a full grasp of such a complex piece of knowledge.
- Myth-to-debunk, and corollary of the corollary: Do not expect the next "IPCC WG1 Report 2013 for Dummies".
- Reword of preceding point: Interdisciplinary knowledge all packed in a single person does simply not exist; what may exist (if people agree to) is interdisciplinarity from cooperation of individual specialists - and this applies to understanding contents, not only creating it.
In my view, all effort spent in liberating laypeople from blind faith in model "prediction" is worth. This is not a work likely to succeed if undertaken only by the scientific community: it demands an understanding of basic emotional needs (that is, deep fears) of people, especially now, when the fast pace of changing World makes all less sure they can really control its evolution. It largely involves making people more "mature", and able to accept the fact that not all in the World can be "explained rationally", nor "controlled". That numbers (especially is given by a model) should be criticized (if one if competent enough), or at least never be trusted by an act of faith.
If you feel these objectives are simple to attain, consider how engineers, just to make an example, are keen to split any work in simple slices: the whole, obtained by assembling all the results from these micro-slices, is invisible to the individuals who "created" it, and any of them will, from their own position, imagine all problems (i.e. lack of control) is the problem of someone else in the chain. Changing mindset would demand abandoning the linear, sequential mentality which to date is the standard paradigm.
Do "prediction failures" (or "confirms") really count?
Some self-appointed "skeptics" about climate change have based their position on the fact that some of the model predictions haveresulted in "failures". "No Pacific atoll has been submerged yet, after all." (At least, no atoll big and high enough on mean sea level to be inhabited and take some notice). Based on these "failures", these people claim the whole theory must be invalid.
An almost identical position starts from the "predictions" which have occurred. Based on them, some "pro-climate-change" people claim correctness of the whole conclusions.
Both these positions, apparently claiming opposite "truths", are in reality highly ideological, but not founded on sound logics. The trouble is, models do no "predictions" at all, as explained in the preceding section.
In my feeling these position, when informed to good faith and not by some more or less hidden political or entrepreneurial agenda, come from an excessive faith in the authority of numbers produced by models. A sort of, forgive me, "scientific immaturity".
Failures and confirms, then, count very little if anything at all - or, they "should".
In the meantime, what I foresee is - once again - a macroscopic communication problem. Laypeople crave for divination from future, basically because no one have explained them these divinations are possible. On the other side, scientists build models, in the attempt to test and validate their theories, and to get some glimpse of what might be an important aspect of reality.
Up now, no large effort has been made to shift public focus from model-obtained projections to facts, data, phenomena and their comprehension. And use models just as another cognitive instruments.
In conclusion, before the beginning...
I've shared some points which are quite "obvious" among professional scientists and academic scholars, but taste hardly digestible to "common" people. We "common" people, however, can (and should) understand if we apply hard enough. And, we "have to", climate change being one of the most important facets of future.
May we condense all the intricate lace which precedes in some pearls? Perhaps? Maybe I'm not the best to draw a conclusion... Yet, I try giving some of the points:
- Read the IPCC summary for policymakers, but give it the due (low) weight: the real stuff is in the technical Report.
- The technical Report is extremely interesting. But don't expect from it statements looking "strong". Consider they are strong.
- Don't be a "skeptical" (in the current climatological sense of "a priori contrary") nor an a-priori "supporter".
- When reading the Report (or any other scientific result, for what matters) interrogate yourself on your deep emotional background, fears and needs.
- Think with your head, and be thyself.
- Be opened to the fact you may not understand "all". In case, just ask. And if you know, just answer honestly.
I wish you a good waiting (until tomorrow) and reading.
We'll meet, from time to time, on some of the points.