Many advocates for action on climate change, including the IPCC assessments and recent documentaries have promoted a view that global warming will continue through the 21st century, with global warming defined as a steady increase in global average temperatures. This prediction of warming is based on the output of multi-decadal general circulation models and is primarily due to the radiative forcing effect of anthropogenic emissions of CO2. In such models only relatively minor year-to-year variations in global average temperatures are forecast in the upward trend, except when major volcanic eruptions cause short-term (up to a few years) of global cooling. For example, see these projections of the most recent IPCC — none of the models has an obvious multi-year (i.e., >2) decrease in global average temperatures over the next century.
Such predictions represent a huge gamble with public and policymaker opinion. If more-or-less steady global warming does not occur as forecast by these models, not only will professional reputations be at risk, but the need to reduce threats to the wide spectrum of serious and legitimate environmental concerns (including the human release of greenhouse gases) will be questioned by some as having been oversold. For better or worse, a failure to accurately predict the changes in the global average surface temperature, global average tropospheric temperature, ocean average heat content change, or Arctic sea ice coverage would raise questions on the reliance of global climate models for accurate prediction on multi-decadal time scales. Surprises or experience that evolve outside the bounds of model output would likely raise questions even among some of those who have so far accepted the IPCC reports as a balanced presentation of climate science. (for a perspective different than the IPCC on applications of climate models see this).
The National Research Council published a report in 2002 entitled “Abrupt Climate Change: Inevitable Surprises” (of which RP2 was a committee member). The report raised the issues of surprises in the climate system. One of the surprises (to many) may be that the global climate models are simply unable to accurately predict the variability and trends in the climate metrics that have been adopted to communicate human-caused climate change to policymakers. Among the climate metrics with the most public visibility are the long term trends in global average surface temperature, global average tropospheric temperature, summer arctic sea ice areal coverage, and ocean heat content.
There is some emerging empirical evidence to suggest, however, that the concerns expressed here are worth consideration. The recent dramatic cooling of the average heat content of the upper oceans, and thus a significant negative radiative imbalance of the climate system for at least a two year period, that was mentioned in the Climate Science weblog posting of July 27, 2006, should be a wake-up call to the climate community that the focus on predictive modeling as the framework to communicate to policymakers on climate policy has serious issues as to its ability to accurately predict the behavior of the climate system. No climate model that we are aware of has anticipated such a significant cooling, nor is able to reproduce such a significant negative radiative imbalance. Meaningless distinctions between “projections” and “predictions” will be unlikely to convince consumers of climate models to overlook experience that does not jibe with modeled output.
There is no greater danger to support for action on important issues of human impacts on the environment than an overselling of what climate science can provide. If the climate behaves in ways that are unexpected or surprising it will be more than just credibility that is lost. Advocates for action should think carefully when gambling with the unknown predictive abilities of climate models. The human influence on the climate system is real, but the climate may not always cooperate.
Looking back, I see that your source for a dramatic ocean cooling is rather weak. Is there anything publiched on this yet? WOuldn’t it be a rather good idea to wait for some firmer evidence. I find it somewhat surprising, because it doesn’t really show up in the global T record.
As for > 2y cooling trends in the predictions… why should there be? There aren’t in reality, as http://mustelid.blogspot.com/2005/09/probably-not-betting-on-climate-with.html shows, in another context.
Comment by William Connolley — August 8, 2006 @ 12:26 pm
William - All I can communicate at this time, is to stay tuned. The issue that so much is riding on the multi-decadal global climate predictions should be a real concern to everyone.
Comment by Roger Pielke Sr. — August 8, 2006 @ 1:02 pm
SSTs here in No Cal under an ENSO neutral to La Nina regime typically never dipped below 56 Deg. F. The past few years however I have gotten anecdotal reports of temps in the low 50s. I am alarmed.
Comment by Steve Sadlov — August 8, 2006 @ 1:21 pm
How many GCM’s predicted the sharp temperature decline after the 1998 peak?
Roger, in the past you have had excellent posts explaining that GCM’s are only useful for studying processes and not for predicting the future. It is unfortunate that many highly intelligent people are unaware of how short-term and limited computer model predictive skill is in reality.
Comment by Reid — August 8, 2006 @ 1:35 pm
Roger - thats a pretty unimpressive source to build a post around. I think you’d rip other people to shreds for doing that. Also, I rather doubt the global ocean is well enough surveyed to calculate inter-annual trends in ocean heat content (yet another reason for not using it :-).
I notice you’ve ignored my point about > 2y trends not being observed, so why should they be predicted?
Comment by William Connolley — August 8, 2006 @ 1:54 pm
Reid - why do you expect the (AO)GCM’s to predict temperatures year-by-year?
Comment by William Connolley — August 8, 2006 @ 2:52 pm
William - there is substance behind the post which will be webloged in the coming several weeks. In any case, even irrespective of the actual data, policymakers are being asked to make major decisions based on inadequately evaluated multi-decadal climate model projections. The reasons to take actions, however, are much broader than climate model output alone. This is a big (and in our view unneeded) gamble. Are you really that confident on their skill?
Comment by Roger Pielke Sr. — August 8, 2006 @ 3:02 pm
Roger - thats pretty evasive. You wrote In such models only relatively minor year-to-year variations in global average temperatures are forecast in the upward trend, except when major volcanic eruptions cause short-term (up to a few years) of global cooling. For example, see these projections of the most recent IPCC — none of the models has an obvious multi-year (i.e., >2) decrease in global average temperatures over the next century. I’ve pointed out that in the obs, there is no sign of > 2 yr decreasing trends. So what are we to make of what you wrote? Are you, perhaps, complimenting the models for accurately replicating the obs? From the tone of the rest, this seems unlikely. So what exactly are you saying?
Comment by William Connolley — August 8, 2006 @ 3:10 pm
William - there will be more information to follow on specific climate metrics. However, with respect to ocean heat storage, since the “currency” of global warming is appropriately Joules and the ocean is the primary store to monitor for heat storage changes, this is where we should be focusing our efforts to monitor climate system heat changes, and thus, to diagnose the planetary radiative imbalance, as I wrote in Pielke Sr., R.A., 2003: Heat storage within the Earth system. Bull. Amer. Meteor. Soc., 84, 331-335.
The paper Willis, J.K., D. Roemmich, and B. Cornuelle, 2004: Interannual variability in upper ocean heat content, temperature, and thermosteric expansion on global scales. J. Geophys. Res., 109, C12036, doi: 10.1029/2003JC002260 provide a convincing case for the accuracy and spatial representativeness of the upper ocean heat data. As stated in that paper
“Satellite altimetric height was combined with approximately 1,000,000 in situ temperature profiles to produce global estimates of upper ocean heat content,
temperature, and thermosteric sea level variability on interannual timescales….Ongoing projects, such as the Jason/TOPEX series of satellite altimeters and the Argo float program, provide a critical foundation for characterizing variability on regional, basin, and global scales and quantifying the oceans’ role as part of the climate system.”
This data is clearly sufficiently robust that modeling comparisons have been made with this information; e.g.
Barnett, T.P., D.W. Pierce, and R. Schnur, 2001: Detection of anthropogenic climate change in the world’s oceans. Science, 292, 270-274.
Hansen et al, 2005 Earth’s Energy Imbalance: Confirmation and Implications Science 3 June 2005: 1431-1435 DOI: 10.1126/science.1110252
A cooling in the ocean heat content would certainly raise issues with the application of the ability of the global climate models to explain such an observation, as well as result in concern with respect to their forecasting skill.
Comment by Roger Pielke Sr. — August 8, 2006 @ 3:58 pm
William-
I think you miss the point of this post. Lets say that sometimes in the next 20 years the climate system evolves in a way not being predicted by the models (pick your variable, pick your trend, this is hypothetical).
Wouldn’t you agree that such climate behavior would cause some loss of credibility to those who have used climate model predictions as the basis for arguing for particular political actions?
You have written that there is a 1 in 9 chance of cooling over the next 10 years! How do you think the political debate on climate change would look in year 9 or 10 of such a cooling trend? The IPCC provides no relization in its ensembles that suggests any such possibility. In fact, according to studies I am aware of cooling over the next 10 years is impossible (show me a model that produces a realization with cooling over the next 10 years, much less 1 out of every 9). What we are saying is that 11% (or whatever) is a big number and suggests that there is greater uncertainty than is currently repreented in public representations of climate model outputs.
If uncertainties are indeed larger, then this is a big gamble to take with policy justified by predictive models, especially when such policies are better and more robustly justified by other means.
Comment by Roger Pielke, Jr. — August 8, 2006 @ 4:01 pm
I have posted several short discussions of software Verification and Validation issues a few times at Professor Pielke Sr.’s blog, and on other blogs. I continue to hope that more complete discussions of significant software engineering issues will follow. So far that has not happened. I have made other direct contacts too, but also without success. These contacts I will not discuss in a casual public forum such as blogs.
My research is conducted mainly on the Web. It is entirely possible that I have missed what I’ve been looking for. But no one has yet provided the information that I know is needed in order to provide creditability to ‘projections’/'predictions’ from any and all computer software.
I have a few questions for Professors Pielke and any others who have answers. I consider these issues to be essentially show-stoppers as far as use of the results of any of the AOLBCGCM codes and all supporting codes use in all aspects of climate-change analyses, for either (1) archival peer-reviewed publications, (2) providing true insight into the phenomena and processes that are modeled, and most importantly (3) for decision-making relative to public policies. Any and all professional software developers would absolutely require that all of the issues to be mentioned below be sufficiently addressed and documented before using any software for applications in the analyses areas for which it was designed.
In no particular order, as each of the following is very important, can anyone provide documented information about:
(1). Audited Software Quality Assurance (SQA) Plans for any of the computer software that is used in all aspects of climate-change analyses.
(2) Documentation of Maintenance under audited and approved SQA procedures of the ‘frozen’ versions that are used for production-level applications.
(3) Documentation of the Qualifications of the users of the software to apply the software to the analyses that they perform.
(4) Documentation of independent Verification that the source coding is correct relative to the code-specification documents.
(5) Documentation of independent Verification that the equations in the code are solved correctly and the order of convergence of the solutions of the discrete equations to the continuous equations has been determined.
(6) Sufficient information from which the software and its applications and results can be independently replicated by personnel not associated with the software.
(7) It is my impression that use of ensemble averages of several computer calculations that are based on deterministic models and equations is unique to the climate-change community in all of science and engineering. I can be easily corrected on this point if anyone can provide a reference that shows that the procedure is used in any other applications. (The use of monte carlo methods to solve the model equations is not the same thing). The use of ensemble averaging and the resulting graphs of the results makes it very difficult to gain an understanding of the calculated results; rough long-term trends are about all that can be discerned from the plots.
(8) Documentation that shows that the codes always calculate physically realistic numbers. For example, the time-rate-of-change of temperature, say, is always consistent with the energy equations and is not the results of numerical instabilities or other numerical solution methods problems.
(9) Documentation in which the mathematical properties (characteristics, proper boundary condition specifications, well- (or ill-) posedness, etc.) of all the continuous equations used in a code have been determined. Do attractors exist, for example.
(10) Documentation in which it has been shown analytically that the system of continuous equations used in any AOLBCGCM model has the chaotic properties that seem to be invoked by association and not by complete analysis. Strange-looking output from computer codes does not prove that the system of continuous equations possess chaotic characteristics. Output from computer codes might very likely be results of modeling problems, mistakes, solution errors, and/or numerical instabilities.
Invoking/appealing-to an analogy to the Lorenz continuous equations is not appropriate for any other model systems. The Lorenz model equations are a severely truncated approximation of an already overly simplified model. The wide range of physical time constants and potential phase errors in the numerical solutions almost guarantees that aperiodic behavior will be calculated.
Especially true considering the next item.
(11) Documentation in which it has been determined that the discrete equations and numerical solution method are consistent and stable and thus the convergence of the solution of the discrete equations to the continuous equations is assured. Actually I understand that the large AOLBCGCM codes are known to be unable to demonstrate independence of the discrete approximations used in the numerical solution methods. The calculated results are in fact known to be functions of the spatial and temporal representations used in the numerical solutions. This characteristic proves that convergence cannot be demonstrated. Consistency and stability remain open questions.
(12) Documentation in which it is shown that the models/codes/calculations have been Validated for applications to the analyses for which it has been designed.
All software, each and every one, that is used for analyses of applications the results of which might influence decisions that affect the health and safety of the public will have addressed all these issues in detail.
If my understanding of the status of these critical issues is correct I can only conclude:
(1) The software used in the climate-change community does not meet the most fundamental requirements of software used in almost all other areas of science and engineering. Almost none of the basic elements of accepted software design and applications for production-level software are applied to climate-change software.
(2) The calculated results cannot be accepted as correct and valid additions to the peer-reviewed literature of technical journals.
(3) The software should never be used in attempts to predict the effects of changes in public policy (fuel sources for energy production, say) on the climate; neither short- or long-range.
(4) The calculated results are highly likely not correct relative to physical reality.
I will say that item (11) is in fact a totally unacceptable characteristic for any engineering and scientific software. The results from any codes that have this property would be rejected for publication by many professional engineering organizations I can be easily corrected if anyone can point me to calculated results from any other areas of science and engineering in which the fact that the numerical methods are known to be not converged is accepted as being, well, acceptable practice. Buildings, airplanes, bridges, elevators, flight-control systems, nothing in fact, are never designed under this approach.
Actually all professional software development projects require far more than the information that I discuss above. Any textbook on software development can be consulted for a more complete listing and detailed discussions. Almost all the large complex AOLBCGCM codes have evolved over decades from software that was significantly more simple than the present versions. These codes have not been designed and built ‘from scratch’ on a ‘clean piece of paper’. Newly-built software, designed and constructed under SQA plans and associated procedures require very significantly more documentation and independent review, Verification, and Validation than that mentioned above.
All calculations from programs in which inherently complex physical phenomena and processes occurring within complex geometries are the focus are generally considered to be suspect and usually taken with a grain of salt. The modeling and calculation of climate change over the spatial and temporal scales for a planet present major challenges to all aspects of mathematical modeling and computer calculations. The number of important systems involved along with the inherent complexity of the phenomena and processes, interacting on multiple time scales over extreme spatial and temporal extents, represent possibly unprecedented challenges. V&V and SQA are absolutely necessary under these situations and are standard operating procedures (SOP) for all major software development projects. There are also absolutely necessary and should be SOP for software the calculated results of which are submitted for publication in archival journals. For calculated results that form the basis of policies that affect the health and safety of the public, the requirements for application of these processes are codified in the laws of the country.
In the absence of established, formal V&V and SQA processes and procedures, calculated results are very much less certain to contain true information and thus do not in fact provide knowledge. Even when these processes are rigorously applied to computer software, mistakes and errors still survive, although to a very much less degree than in the absence of the procedures. For the case of the very complex codes and applications associated with climate change, the chances of mistakes and errors are much larger than for less complex analyses. Consider situations that have occurred within the climate change community that sometimes surface in the literature. Data reduction software is an example.
The potential for mistakes and numerical errors in software in the absence of independent, formal, V&V and SQA procedures are among the reasons that engineering journals have implemented editorial policies. Apparently the organizations have decided that publication of a paper that has not been demonstrated to correctly solve the equations and to be based on the correct equations for its intended applications, has a high potential to not represent physical reality. Additionally, apparently they consider that the consequences of publishing such results will not contribute to advances in understanding and knowledge.
Regulatory agencies that are responsible decisions that affect the health and safety of the public will never, and I’m aware that I should never use always and never, make policy decisions based on computer software that has not been Verified and Validated and maintained and applied under SQA procedures. Observational data will always, of course, be more important than computer calculations alone.
If the science community does not begin to implement formal procedures, and at the same time base information on software that is not maintained and applied formally, regulatory agencies will ignore the calculations. And it is not just the large, complex AOLBGCM codes that will eventually be required to implement the processes and procedures. All software, and software users, will have to be demonstrated to be Qualified for applications to the analyses for which the software has been designed. Additionally, it is not just the ‘main’ routines in the large codes that will require these, but also the initial and boundary conditions, the pre- and post-post processing routines ( run-time options, grid generation, processing of the calculated results, etc.), the qualifications of the users, and the procedures for installing the software onto a user’s computer system. That is, all aspects of the codes, users, applications, and results.
Comment by Dan Hughes — August 8, 2006 @ 4:24 pm
Dan- thank you for the through discussion. I invite readers of the Climate Science weblog to reply to your request. My comments are below;
With respect to multi-day weather prediction, there have been millions of predictions which have been evaluated, with the model parameterizations tuned so as to provide the most accurate forecasts. This has been the primary validation approach.
However, with multi-decadal climate predictions, such repeated tests are, of course, not possible. Thus the climate modeling community has tested against historical (and even paleo-record) data. Even knowing the answers, however, the models have not accurately predicted climate variability and trends on the regional skill, as has been documented on the Climate Science weblog.
I also recommend model evaluation procedures in my book (in Chapter 12);
Pielke, R.A., Sr., 2002: Mesoscale meteorological modeling. 2nd Edition, Academic Press, San Diego, CA, 676 pp.
Comment by Roger Pielke Sr. — August 8, 2006 @ 4:46 pm
Excellent comment Dan.
Don’t expect the high priests of AGW to voluntarily open the curtain on their Wizard of Oz GCM’s. A NAS panel such as the one that looked into Mann’s statistics will be needed. The panel should be composed of computer and mathematics experts, not climate scientists.
Comment by Reid — August 8, 2006 @ 5:52 pm
I am a professional software developer working on large scale commercial systems. I can tell you that if we had to provide the massive amount of documentation required by Dan Hughes we would never get anything done. Furthermore, from experience on military projects that require intensive documentation, I can tell you that it actually interferes with quality by diverting resources from development and testing.
A general circulation model is not a life support system. Minor errors, if they exist, do not invalidate the entire model. Most important, unlike most commercial and military applications, there are many independantly developed GCM’s, which is the best possible ways to catch discrepencies.
There may be many interesting issues with climate models, but I do not believe that software quality is one of them.
Comment by Blair Dowden — August 8, 2006 @ 6:02 pm
The post by Pielke, père et fils, raises at least two important issues. First, given that the results of climate models are, at best, inexact at regional – and more importantly (from the point of view of impacts on humanity and the rest of nature) at local – scales, how does one devise policies to address climate change concerns? [I should note that even if climate models gave perfect results at these smaller scales, we would have to contend with the uncertainties generated by the techniques (more scenarios and models) used to generate future impacts.]
One answer to this issue is, of course, to reduce emissions. But that’s, at best, a long term solution. What’s to be done in the mean time? In the short-to-medium term – i.e., over the next several decades, if one considers the inertia of both the climate system and the energy infrastructure — that would do nothing to alleviate problems caused either by climate change or under the current climate. Therefore, if one believes that climate change exacerbated, if not caused, last year’s hurricanes in the U.S. or the European heatwave of 2003, the scale of damage would have been virtually unchanged had the Kyoto Protocol, for instance, gone into effect 20 years earlier, or the U.S. and Australia had embraced it enthusiastically. So what can be done in the next few decades that would respond to climate change but would not depend on the location-specific details of model projections of the impacts of climate change?
The second answer lies within the question itself: namely, we should pursue strategies and measures today whose implementation and success is independent of the precise site-specific projections from scenarios and climate models, as noted here a long time ago. The list of such measures is limited only by our imagination. Many are listed in a forthcoming publication titled, Integrated Strategies to Reduce Vulnerability and Advance Adaptation, Mitigation, and Sustainable Development. For example, regarding measures that would help cope with the impacts of climate change on agricultural productivity (and, therefore, hunger), they include, for example, undertaking research into and developing cultivars to increase yields under marginal climatic and soil conditions that exist today but could become more common under climate change (e.g., higher CO2; higher temperatures; at the higher latitudes, longer growing seasons; at lower latitudes, heat-shortened seasons; low soil moisture in some areas, too much water in others, or soils with high salinity, alkalinity or acidity; and so forth). Work on such cultivars can proceed today even in the absence of accurate projections of the impacts of climate change. If the science behind such cultivars is developed ahead of significant changes in the climate – as is likely — this science can be used to boost agricultural productivity under current climatic conditions.
This example also suggests a broader approach but whose success is also independent of accuracy/skill of climate and impact models to project impacts at sub-regional scales. This broader approach would be to reduce society’s vulnerabilities to current climate-sensitive problems, particularly if they are urgent today and could be exacerbated by climate change. Such problems include hunger, malaria and other climate-sensitive diseases, and the threats to life and property from extreme weather events. The tools, technologies, human and social capital, and institutions developed to cope with these problems today will be similar, if not the very same, as those used to address the same problems tomorrow if they are aggravated by climate change. A malaria vaccine, for instance, if developed, would likely be equally effective whether the malaria is caused by non-climate-change-related factors or by climate change. And that’s true for most other measures to treat or prevent malaria. Similarly, measures to reduce the human and economic toll associated with extreme events will work whether the events are caused by current climate or climate change. This is so for all other climate-sensitive problems, as well.
I have shown elsewhere – see A Climate Policy for the Short and Medium Term: Stabilization or Adaptation? — that over the next few decades such an approach would not only provide greater benefits, it would provide them much more economically than would any mitigation scheme. The basic reason for this is that such an approach would address a larger universe of problems than climate change alone. Going back to the malaria example, this means that it would address the total malaria problem, whether it’s the portion caused by climate change or by non-climate-change-related factors. On the other hand, mitigation approaches would only address the former portion of the problem. The latter will invariably be smaller than the total problem (unless climate change alleviates malaria). No less important, we will be solving current problems that affect current generations, while also preparing future generations to cope with such climate change as will inevitably occur.
A second reason is that for most climate-sensitive problems, if global impact assessments are to be trusted, it seems that for the next several decades the contribution of non-climate-change-related factors to these problems exceeds that of climate change. [For a more nuanced discussion, see the previous link.]
The second issue that the Pielke post raises is the one about surprises. Where is it shown that all climate change surprises necessarily have to be negative for humanity? Since there is no such theorem, and there are indeed many unknowns about the climate system, the Pielkes are right in sounding a caution.
Incidentally, the vulnerability-reduction approaches outlined above would be justified and are robust whether or not climate changes, or climate surprises are negative or positive for humanity.
Comment by Indur Goklany — August 8, 2006 @ 10:40 pm
Re the upper ocean heat data, I want to chime in on both sides of the debate
Roger (Sr): In #9 you say that “This data is clearly sufficiently robust that modeling comparisons have been made with this information”. Then you list a couple of papers that do so. I don’t think either of these papers paid too much attention to 2-year trends!
William: In #5 you say “I rather doubt the global ocean is well enough surveyed to calculate inter-annual trends in ocean heat content”. In my briefly considered opinion, there is one data source that could generate reasonable inter-annual trends, namely satellite altimeter data. This showed global sea level rising rather quickly from 1993 to 2000 and I wouldn’t be at all surprised if that increase had slowed or reversed. (It’s certainly done so in the NZ region.) So there *may* be something real here.
Re the possibility that the Earth is acting in a way that the models hadn’t predicted, I must say I’m pretty relaxed about that. Let’s wait a few more years and see, eh?
Comment by Mark Hadfield — August 9, 2006 @ 12:14 am
Roger - yes I do seem to rather miss your point. We have a “cooling” since 1998 - it doesn’t seem to be doing any harm to the debate. As to the model runs… see http://www.antarctica.ac.uk/met/wmc/sresa1b_ukmohadcm3.png which shows times when the same thing happens in the future. Thats for the UKMO hadcm3 run sresa1b scenario.
Are you missing the point that *ensemble* runs smooth out such behaviour, since fluctuations average out?
I would say that if the real world followed that model, and the fluctuation it shows around 2044-54 occurred, the warming by then would be so obvious that the cooling period would not cause any great problem.
Comment by William Connolley — August 9, 2006 @ 2:49 am
I can offer an alternative view to #14; I’m a research scientist developing software models to simulate state-of-the-art remote sensing systems (my particular area of expertise is synthetic aperture radar), and I would argue that Dan Hughes’ suggestions are sensible.
We follow a similar strategy to that outlined by Dan, and whilst it is time-consuming and occasionally tedious aspect of the work, it is essential if the models are to provide reliable results. It is a question, of course, of efficiency: is it more efficient to put in the trudge work early on and catch the errors there, or to sort the errors out after publication and a widespread belief that the answers are correct? Less haste can mean more speed in learning. At my organisation, we also have a policy of trying to avoid washing our dirty linen in full public view, but that is another story…
There is a key difference between the climate models and the models I develop, which echoes RP Sr.’s comments above; the models I develop simulate the events of a number of seconds, which are then easily validated against the real world. Simulations of decades cannot be so easily assessed. But this, more than anything, imposes a need for more stringent analysis up front.
By definition, any research work has an element of the unknown about it - and I suspect this is the difference between the requirements of large scale commercial software and research models. As there is an unknowable result until the model is finished, any answer could be a valid answer - so you have to have great confidence in the code leading to that answer to be able to trust it. In my experience, minor errors may have negligible effect, or they can completely invalidate the result (especially in a model which includes any kind of feedback - even linear feedback mechanisms). I know this, I’ve seen it happen first hand. You don’t know until you quantify the behaviour of the model, errors and all, which is what (I believe) Dan Hughes is looking for. This won’t catch all the errors - but it should catch the important ones.
What this type of process can’t capture is where our knowledge is lacking. Changes to feedback mechanisms can have a large impact on results; this is a situation where an increase the amount of knowledge we have could actually result in a widening of the confidence intervals of future projections. This would exist as a systematic bias in the models, and virtually impossible to guard against.
Overpromotion of a single strand of evidence carries credibility risks that can be politically exploited; the hockey stick saga being one example.
Comment by Spence — August 9, 2006 @ 4:14 am
I am also a professional software developer and I am currently working for a major bank on a project to make the software development system Sarbannes-Oxley compliant.
What Dan Hughes mentions is merely a variant of modern Risk Management strategies and model validation, along with some auditing of software changes.
These climate models are used to determine government policy and the spending of greater sums of money than many banks deal with aon a day-to-day basis. Loosing a few million dollars through faulty software is trivial compared to wasting billions on mitigation strategies that the models wrongly say are necessary.
If my information is correct, the code verification and management situation is even worse with climate models than my software for financial transactions. The complexity of the software has meant that functions and subroutines have been copied from one model to another rather than have each group of programmers re-write them. Software re-usability is currently in vogue but imagine what happens if the original module contains coding errors.
Comment by John McLean — August 9, 2006 @ 7:06 am
Mark - Regarding
“I don’t think either of these papers paid too much attention to 2-year trends!”,
one of the major advantages of assessing ocean heat content changes is that with accurate, spatial representative sampling of the amount of Joules in the climate system in a time slice, if the Joules decreased, this is a real loss of heat from the system. The models need to account for any such a loss even on 2 year time scales, if the loss is large enough to be significant.
As I write in my paper Pielke Sr., R.A., 2003: Heat storage within the Earth system. Bull. Amer. Meteor. Soc., 84, 331-335.
“A snapshot at any time documents the accumulated
heat content and its change since the last assessment.
Unlike temperature, at some specific level of the ocean, land, or the atmosphere, in which there is a time lag in its response to radiative forcing, there are no time lags associated with heat changes.”
Comment by Roger Pielke Sr. — August 9, 2006 @ 7:58 am
William - Re #17 Thank you for sharing the plot and for your continued valuable dialogue on the weblog.
It would be valuable to also present model forecasts (realizations and ensemble average) of the predicted changes in ocean heat content in Joules for the same time period. Then we can see the magnitude of any negative, short term excursions in the climate system heat content.
Comment by Roger Pielke Sr. — August 9, 2006 @ 8:02 am
As a professional software developer myself, I have to agree with Blair’s comment that the V&V and SQC proceedures described by Dan are impractical for this application. I’ve particapated in several very large software development projects, some of which that were originally specified to use similar, although slightly less obsessive, QC and V&V proceedures. In all cases, the tight controls were dropped very early in each project as software development was all but totally blocked while both time and money were rapidly burning.
I disagree with Blair, though, on his comment that software quality is not an issue for GCMs. To be honest, I also disagree GCMs are not as critcal as normal commercial or military software. The application of GCMs to economic and political decissions are tremedously costly. If those resources are being improperly applied, the cost to human life is very real and potentially dangerous.
I would like to suggest that the correct approach to GCM V&V is to insist that software used be clearly organized and ‘packaged’ so that it may both be understood and any executable programs be may be rebuilt from the original source code. Obviously too, the software should be internally commented, build files and build instructions should be included, and run instructions should be supplied.
The above paragraph describes what is needed at a minimum for executable programs. The purpose of software is to manipulate data and produce a useful result. When a specific result is claimed from a software product, the input data and a description of its use must also be included in the final verifcation package.
GCM may represent an interesting problem for software run time verification as these programs may simply require too much computer power to casually dupicate an output run. That does not mean, however, that well organized, clean source code, data sets, an operational documentation is not called for.
Too much money and political energy is spent on GCM output to ignore software quality.
Comment by Gary Wescom — August 9, 2006 @ 8:03 am
William, it would also be interesting to see the linked plot re-plotted using twice the resolution of the data. For example, if the linked plot is using 2-year increments, re-plot it using 1-year increments. Actually it would be very interesting to see an every-time-step plot. Whatever the edit frequency of the plotted data, all the curves should overlay.
Comment by Dan Hughes — August 9, 2006 @ 8:32 am
Oh, gosh, SSTs seem to be lower lately.
Could it be the GCM models failed to predict the extraordinary economic growth of India and China and their rapid increase of coal combustion.
China contributes enough aerosols to the global atmosphere thatair quality of the US west coast and the Arctic are measuring an impact.
In 2005, China emitted 26 million tons of S. Mt. Pinatubo released about 10 million tons of S. Hansen claculated a radiative cooling of 4.5 W/m2 caused by 6 TG S, the amount of S that remained in the stratopshere as sulfate six months after the initial eruption (10 MM tons S.)
Though India’s coal is low S on average, (0.8 percent roughly), its ash content is about 40 to 47%. Thus its S emissions were about 2.7 MM tons S in 2005 but soot emissions were about 22 million tons. A recently conducted Indian Ocean Experiment suggests that the presence of soot carbon in the atmosphere over the Northern Indian Ocean hinders its natural heating porocess by about 15 percent. How now, Asian brown cloud!
Add US S emissions of about 10+ MMT annually; and, for good measure throw in the west-bound Northern African dust storms being collected in the the Southeast US.
Seems the models missed those contributions to a recent cooling of the oceans.
Paul Harvey shares page two of his stories. You should as well.
Comment by John L. McCormick — August 9, 2006 @ 9:37 am
WRT Mark Hadfield in # 16, he would be surprised. Mean sea level has continued to rise since 2000
http://sealevel.colorado.edu/
and links therein.
If anything you could make an argument that the rate is increasing slightly.
Comment by Eli Rabett — August 9, 2006 @ 9:59 am
#23 - sorry, I don’t understand you. I’ve plotted annual data. If I plotted monthly means - or 6 monthly means - then of course the curves wouldn’t overlay. Monthly means would show more variability, and a seasonal cycle, of course. Every timestep would be pointless.
I think you are confusing software verification *of the model* with *of the output*. But in fact I don’t really know what you’re getting at in #23.
#21 - Roger - it was you that started off pointing to sfc T plots, not me. And its you thats interested in ocean heat, not me. I’m wondering if you’re going to issue a correction to your post, now your point (that the IPCC models don’t show future cooling episodes) is shown to be wrong.
Various, on software standards: I think the source codes of some of the US models are available for inspection, if you want to. HadCM3 source and doc is online, but locked: http://www.cgam.nerc.ac.uk/um/index.php. Some other comments, based on von S’s testimony, at http://scienceblogs.com/stoat/2006/07/von_ss_testimony.php
Comment by William Connolley — August 9, 2006 @ 11:08 am
John L. McCormick-
We certainly didn’t go into all of the reasons why climate models might not accurately anticipate climate behavior precisely (though I am sure that you can find a discussion of such things here on my father’s blog).
But from your comment I take it that you would then agree that model-based predictions have not in fact bounded the set of possible near-term futures. Which is of course exactly our point!
Comment by Roger Pielke, Jr. — August 9, 2006 @ 11:09 am
William-
The plot that you show does not show a ten-year cooling period in the next ten, which is what you proposed has a 1 in 9 probability.
Can you show any climate model realization that provides such a result over the next ten years? If not, then I take it you are a “skeptic” because the models don’t faithfully represent the full range of uncertainties that you forsee?
Comment by Roger Pielke, Jr. — August 9, 2006 @ 11:15 am
Regarding the software quality discussion: actually, is the code for the various GCMs open-source and publically available? Seeing as public funds are probably paying for a good chunk of the effort, i’d hope so.
Now, that might engender a whole new range of commentary in a wider community - plus may increase quality
-t
Comment by todd drake — August 9, 2006 @ 11:48 am
Re: #24 An intersting comment. With this explosion of coal cumbustion, also comes increased emmissions of CO2, CH4,and N2O. The last time I checked, coal was a hydrocarbon, and these were GHG’s. Coal produces twice as much CO2/BTU than does combustion of natural gas. With natural gas, we don’t get the cooling effect from the organic carbon and SO2 however. Energy - we can’t live with it, we can’t live without it. Tough choices.
Comment by Bryan Sralla — August 9, 2006 @ 2:53 pm
Re 25: “WRT Mark Hadfield in # 16, he would be surprised. Mean sea level has continued to rise since 2000″.
I said I wouldn’t be surprised if global sea level had fallen or stopped increasing in the last year or 2. I didn’t say I would be surprised if it hadn’t. But surprised or not, I’m grateful for the link. Isn’t the WWW wonderful?
Comment by Mark Hadfield — August 9, 2006 @ 4:26 pm
RE# 27,
Dr. Pielke, you read too much into my comment.
As I read the post:
[none of the models has an obvious multi-year (i.e., >2) decrease in global average temperatures over the next century.
Such predictions represent a huge gamble with public and policymaker opinion. If more-or-less steady global warming does not occur as forecast by these models, not only will professional reputations be at risk, but the need to reduce threats to the wide spectrum of serious and legitimate environmental concerns (including the human release of greenhouse gases) will be questioned by some as having been oversold.]
I have some confidence newly industrializing nations will embark upon and work to tighten existing aerosol emission regulations (aside from CO2) for health and economic reasons. Hope springs eternal. While advanced economies will further tighen emission regs including PMs and Hg.
In that world, perhaps this intermittant decrease in SST, due to albedo effect of atmospheric aerosols, will reverse and the models will come in closer alignment with observed warming in the out years. Dr. Hansen sees aerosol reduction as a necessary (if counterproductive perhaps)for all nations. Damned if we do and damned if we dont.
Comment by John McCormick — August 9, 2006 @ 4:30 pm
Re 20, Roger there are 2 major flaws (actual or potential) in what you’re saying here.
The first is that it remains questionable that the alleged “recent dramatic cooling of the average heat content of the upper oceans” is real. A few years ago, it would have been unthinkable to assert that you could measure changes in the ocean heat content between one year and the next. Now, with new data sources like Argo and the satellite altimeters, it might be possible to measure this. But let’s see the published data.
The more serious flaw is in the statement that “The models need to account for any such a loss even on 2 year time scales”. Wrong! A model does not need to account for everything. What do you mean by “account” anyway? Do GCMs “account” for observed interannual variability in global mean surface temperature? (Here I mean, for example, atmosphere-only or ocean-atmosphere GCMs, run with historical forcings for the 20th century.) Obviously, they don’t reproduce it in detail, in the sense that they predicted that 1998 (for example) would be anomalously warm. Do they account for it in the sense that the model’s interannual variability is similar to the real world’s? I think the answer to that question is “fairly well”. The models do show variability from year to year in global mean surface temperature and the amplitude of this variability isn’t unreasonable.
Now, do the models generate realistic levels of interannual variability in ocean heat content? Thats a damned interesting question and in a decade or so we might be able to answer it. First, we’ll need good estimates of the interannual variability in ocean heat content. I don’t believe we have those yet, but I’m open to contradiction by someone who knows about this. Second, we’d really like the time series to be a bit longer: the satellite altimeters have been up there for a little more than a decade and the Argo float programme is still spinning up. Third, we need to remember that quite a few different schemes have been used to represent the ocean in GCMs and some of them are not intended to be particularly realistic.
So, like I said, damned interesting question, but I don’t think we have the answer yet.
Comment by Mark Hadfield — August 9, 2006 @ 4:47 pm
Me again…
Roger (Sr), can you point us to a source of more information on the alleged cooling of the upper ocean from 2003 to 2005? I tracked down the WWW page for the 2006 Ocean Sciences meeting and downloaded the “Program and Abstracts” CD as an ISO image…
http://www.agu.org/meetings/cdiso/OS06.iso.zip
I burned it onto a CD and have been scanning thru it. Below is the nearest thing I can find to the presentation you describe. The abstract doesn’t mention a recent global cooling…
OS35B-15
TI: Analysis of Global Heat Content
AU: * John, L M
EM: John.Lyman@noaa.gov
AF: University of Hawaii/JIMAR, NOAA/PMEL, 7600 Sand Point WY NE Bldg. 3, Seattle, WA 98115 United States
AU: Johnson, G C
EM: Gregory.C.Johnson@noaa.gov
AF: NOAA/Pacific Marine Environmental Laboratory, 7600 Sand Point WY NE Bldg. 3, Seattle, WA 98115 United States
AU: Willis, J K
EM: Joshua.K.Willis@jpl.nasa.gov
AF: Jet Propulsion Laboratory, Pasadena, M/S 300-323, 4800 Oak Grove Dr, Pasadena, CA 91109 United States
AB: The ocean has 3 orders of magnitude more heat capacity than the atmosphere. Quantifying and analyzing how ocean heat content changes in time and space allows insight into climate variations over a wide range of time-scales. Yearly average maps of global heat content anomalies in the upper 750 m are made at quarterly intervals from 1993 through 2004 using a combination of in situ ocean thermal data and gridded sea surface height anomaly maps from altimeter data from multiple satellites (http://www.aviso.oceanobs.com/). Analysis methods follow Willis et al. (J. Geophys. Res, Vol. 109, C12036, doi:10.1029/23JC002260, 2004). Maps of global heat content anomalies, 12-year heat content changes, and annual heat content changes are discussed along with a time-series of global average heat content anomaly estimates. Detailed error analysis shows a significant decrease in the global sampling error of ocean heat content from 2002 to 2004 as the number of well-distributed in situ measurements increases due to the growth of the Argo Program array of profiling CTD floats (http://www.argo.net/). The map of 12-year change in the ocean heat content contains two main regions with pointwise trends in heat content that are significant above the interannual variability at the 95% level. One region (of significant warming) is located in the north Atlantic and the other region (of significant cooling) is located in the north Pacific. There are other less significant regions that clearly contribute to globally averaged ocean warming of about 0.91 (±0.15) W m-2 over the last decade, such as several large patches of heating along 40°S.
Comment by Mark Hadfield — August 9, 2006 @ 8:47 pm
Never mind. Here’s something relevant:
http://copes.ipsl.jussieu.fr/Workshops/SeaLevel/Posters/3_14_Willis.pdf
which refers to
Lyman, J., J. K. Willis, and G. Johnson, 2006. Recent cooling of the upper-ocean, Geophys. Res. Let.,
submitted.
Comment by Mark Hadfield — August 9, 2006 @ 8:55 pm
Hi Mark- thank you for your several comments. I will make a few comments here, but will post a weblog in more detail on this subject as soon as I can discuss the latest paper on this subject.
The models do need to account for heat loss if it is large enough to be signficant. A loss of 10% of the heat content that was accumulated over the previous 50 years would be significant, for example, in my view. What would be a magnitude that you would conclude is significant?
The question regarding the accuracy of the data, of course, is a very relevant issue. The fact that Hansen, Barnett and others have used the ocean heat data to compare with their model results certainly indicate that they have concluded that the data is robust. The Willis et al 2004 paper presents a convincing case that the ocean heat trend data can be used for monitoring interannual variations.
Comment by Roger Pielke Sr. — August 9, 2006 @ 9:33 pm
William - Regarding #26, the cooling that is shown in the model results that you link to is small compared with the longer term trend.
“As we state on the weblog, No climate model that we are aware of has anticipated such a significant cooling, nor is able to reproduce such a significant negative radiative imbalance.”
The actual observed global average upper ocean cooling is significant in terms of the long term trend. More details on this in a Climate Science weblog soon.
Comment by Roger Pielke Sr. — August 9, 2006 @ 9:40 pm
To both Pielkes:
When you talk about “professional reputations” and “credibility,” are you talking about climate science as a whole, or do you mean the individuals who tout certain models and scenarios?
If you mean the former (the field as a whole), then I have no further questions.
If you mean the latter (individuals), then who are the judges here of reputation and credibility? That is, who is going to decide if one’s professional reputation and credibility have been damaged? As I see it, you could mean “other climate scientists,” “policy-makers,” or “the public at-large.” I would be skeptical about the last two choices; so that leaves climate scientists.
Obviously, where there are obvious divides (say, e.g., Real Climate folks versus World Climate Report folks), then any sort of mistake by one side will be used by the other to argue against credibility. That already happens, right? But, budding climate scientists like me hope that most of us can get along in spite of our differences (the Climate Specialty Group of the Association of American Geographers is a good model for this).
So, I guess I am saying that reputations and credibility do not have to be issues, unless you want them to be. And, given (both of) your active roles in the climate change debate, and the fact that each of you has a web presence and engages the media, then if you if you wish for these to be issues, they will be.
Comment by Kenneth Blumenfeld — August 9, 2006 @ 10:34 pm
For those interested in seeing the source code for US climate models:
NASA GISS ModelE: http://www.giss.nasa.gov/tools/modelE/
NCAR CCSM: http://www.ccsm.ucar.edu/models/
GFDL AM2: https://fms.gfdl.noaa.gov/
Documentation for building the models and input data sets should all be available. Have at it - after all, your tax dollars paid for it.
Comment by Eric Wilcox — August 10, 2006 @ 8:16 am
Those thinking that this is going to be a better year for hurricanes might be interested in this
“Super Typhoon Saomai, the strongest to threaten China in over 50 years, slammed into the southeast coast on Thursday killing at least two people, injuring over 80 and forcing more than 1.5 million from their homes.
Saomai, one of three storms to hit East Asia in the past few days, made landfall in Zhejiang province at 0925 GMT, hitting Cangnan county just after officials there declared a state of emergency, ”
I guess those SSTs are really low these days.
And, oh yes John McCormick, climate models don’t forcast economic development, that is WGII territory.
Comment by Eli Rabett — August 10, 2006 @ 8:55 am
Kenneth- Thank you for your Comment.
The professional reputations and credibility of those who lead the international and national assessments will be questioned if the actual climate evolves in a manner which was not anticipated by those reports. Policymakers, I anticipate, would disregard (or at least devalue) further statements by these individuals. My experience with the first CCSP Report (which I resigned from and have weblogged on Climate Science)is that the exclusion of the diversity of views makes that Report less valuable to policymakers, as it fails to accurately review the subject of the Report.
For young scientists, the primary goals should be the publication of quality peer reviewed papers, effective and balanced teaching (if in an academic program), and mentoring younger colleagues. This is how we will advance climate science.
Comment by Roger Pielke Sr. — August 10, 2006 @ 8:59 am
Kenneth-
Let me answer your question with an anecdote. At the House hearing that I testified at last month, Tom Karl appeared on the panel before me. He was asked by the Committee chairman, and I am paraphrasing, “Can climate models accurately predict the future or are there still significant uncertainties in those predictions?”
Tom Karl said a few things — hemmed and hawed for a few minutes — without saying much, clearly trying to convey a sense that models were not perfect but also struggling not to suggest that the uncertainties were too large. It was not an inspired reply. After a break for a vote, Karl asked the chairman if he could try to answer the earlier question on models again. This time he simply said, and this is very close to a direct quote, “Climate models are good enough to base policy on.”
Statements like this serve several functions. First, they tie the need for action to climate models. This is simply unnecessary, as the basis for action is much broader than predictive models (see my own testimony for discussion). Second, such pronouncements, for better or worse, tie credibility of the community, and the people who make public representations on behalf of the community, to actual model performance.
With respect to both points above, if models are not in fact truth machines, then this results in risks for climate policy and climate science. If you want to consider a practical example of these sorts of dynamics, consider public earthquake predictions (e.g., Parkfield) and their effects. See:
Joanne M. Nigg, 2000. Predicting Earthquakes: Science, Pseudoscience, and Public Policy Paradox. in Daniel Sarewitz, Roger A. Pielke, Jr., and Radford Byerly, Jr. (Eds.) Prediction: Science, Decision Making, and the Future of Nature. (Washington, D.C.: Island Press. 2000): 135-156.
Thanks!
Comment by Roger Pielke, Jr. — August 10, 2006 @ 10:18 am
Several have mentioned that source listings for some of the AOLBCGCM codes are available on the Web. This is very true. And it is equally true that in theory the source coding could be used to Verify the coding. However, in order for this to be a useful exercise we need some kind of specification of what was intended to be coded into the code. In the absence of this information we cannot develop objective metrics for judging that the coding is correct.
The level of detail needed for Verification of the coding is generally many times greater than that typically available for many software products. Because the objective is Verification of the coding a specification of all the equations in the code is needed. For legacy software that has evolved over decades of time, this information is usually contained in a theory and numerical methods manual in which the continuous equations and the discrete approximations to these and the numerical solution methods used to solve the discrete equations are described in detail. A computer code manual in which the structure of the code is describe in sufficient detail that independent outside interests can understand the source code would also be helpful in any attempts to Verify the coding. I have not been successful in finding such manuals for AOLBCGCM codes.
As someone mentioned, as taxpayer-funded software, this documentation should in fact be readily available. It is not.
Comment by Dan Hughes — August 10, 2006 @ 12:30 pm
Thank you William for your consideration of my request in #23 above. I had not completely thought through exactly what was plotted in the linked graph. I think the plotted values are some kind of area-weighting of daily (Tmin + Tmax)/2 for each day and grid point, with additional averaging to get a yearly value. A short outline of exactly how the yearly GAT is calculated would be nice.
It would be of interest to see the range of the yearly variability calculated by the code because this is a known quantity from the instrumental record and certain modeling approaches. The range of the daily variability is another known quantity and it would be interesting to also see that quantity plotted for a few days. What would be really neat would be a plot with sufficient resolution to see both the daily and year cycles over a one year, say. The daily cycle should appear as a 24-hour cycle riding on the yearly cycle.
Thanks again.
Comment by Dan Hughes — August 10, 2006 @ 12:41 pm
Eli,
1. I take your comment “I guess those SSTs are really low these days” as an attempt at irony. If so, it seems to me you are in danger of falling into a very familiar trap, which unfortunately seems to ensnare many people in the climate change business, although I’m not sure that it happens with any lesser frequency in other fields. These days it’s just seems more evident in the climate change business.
The co-existence of two phenomena doesn’t prove that one caused the other, or vice versa. What you also need, as a start, is to look into is whether SSTs were as high (or higher) whenever the last typhoon of equal magnitude/ferocity hit. One thing we do know, for what it’s worth, is that regardless of the SSTs, GHG concentrations were not as high as they are today. So there is more to climate phenomena than emissions of well-mixed GHGs, as Professor Pielke, Sr., constantly reminds us.
2. You are correct: climate models do not forecast economic development. But they use emission profiles generated from assumptions about economic development, population growth and technological change. [In fact, the IPCC's SRES did not forecast or endogenously generate any of these factors. They merely assumed growth and/or changes in these factors over the next 100 yrs.] Regardless of who generates these estimates or how they are generated, the accuracy of the results of climate models are going to be affected by any errors/uncertainties in trends in economic development, population and technological change used to estimate emissions and emission pathways — unless of course, climate is not very sensitive to emissions over the time horizon being examined. Yes, one can’t blame climate modelers if faulty results occur because of errors and omissions in emissions, but the results of their models are inescapably colored by them. And one should be cognizant of that.
Comment by Indur Goklany — August 10, 2006 @ 1:18 pm
I have yet to get any information at all about the lack of grid (or series truncation) independence in AOLGCM codes. It seems to me that this situation should not ever be present in numerical calculations. The discrete equations have not been solved; the numbers printed do not represent solutions of the discrete equations. The lack of grid independence indicates that numerical errors are in fact present in the numbers.
No other science or engineering applications of numerical solution methods tolerate this situation; it is always unacceptable. An example of an exception to my assessment will be greatly appreciated.
Comment by Dan Hughes — August 10, 2006 @ 2:38 pm
I have now looked at the conference poster
http://copes.ipsl.jussieu.fr/Workshops/SeaLevel/Posters/3_14_Willis.pdf
and paper
http://www.pmel.noaa.gov/~lyman/Pdf/heat_2006.pdf
(The paper is marked “Draft - do not distribute”, but it is available on the first author’s WWW page and it’s been accepted, so I presume it’s fine to discuss it.)
The authors assert that they can track interannual variations in ocean heat content with sufficient accuracy that the cooling trend from 2003 to 2005 is real. They also mention an earlier cooling event from 1980 to 1983 and say “Most climate models … do not contain unforced decadal variability of this magnitude”. Of course, these points are open to further debate, but there’s at least a suggestion that there’s *something* a little surprising going on.
An important point that I don’t think’s been mentioned by Roger is that the trend that’s been estimated is the trend in *upper* ocean heat content or (as it’s described in the poster) globally averaged, 0/750 m thermosteric sea level. In the poster, the authors point out (as Eli did earlier in this thread) that sea level has continued to rise. So either there has been a compensating increase in heat content below 750 m or there has been a substantial increase in the rate of freshwater accumulation. The authors prefer the latter explanation and estimate that the increase in ocean volume due to freshwater input was 0.7 ± 1.0 mm/yr from 1993 to 1999 and 2.9 ± 1.1 mm/yr from 1999 to 2005.
So the sceptics can say “the oceans are cooling” and the alarmists can say “the glaciers are melting” (presuming that they are the source of this alleged fresh water). The Rogers can say “the models don’t work” and the rest of us can say “how interesting”. Something for everybody!
Comment by Mark Hadfield — August 10, 2006 @ 3:05 pm
Eli,
Your post demonstrates the problem with the concept that water temperature has an overriding influence on tropical cyclone development and intensity in tropical basins.
The SST anomalies all along the track of Saomai were slightly below climatological averages.
Take a look at:
Sea Surface Temperature Anomalies and
Hurricane Tracks
and see for yourself.
Category 4 and 5 tropical cyclones can form anytime the water temperature is above a certain threshold (I believe the research indicates 84 degrees F.), but they usually don’t. Typically, the water temperatures in the Western Pacific, Gulf of Mexico and Caribbean, exceed this value for much of the respective seasons.
The factors required to support a category 5 storm are numerous and usually short lived. Water temperature is generally not a limiting factor in many areas, and making the suggestion that water temperatures must be unusually warm to support a category 5 storm, or a very active season, are incorrect.
Comment by Jim Clarke — August 10, 2006 @ 4:46 pm
William Connolley writes, “As for > 2y cooling trends in the predictions…why should there be? There aren’t in reality,…”
You mean other than the cooling trend from the mid-1940s to the mid-1970s?
What’s that? You say that the cooling trend from the 1940s to the 1970s was caused by anthropogenic sulfur dioxide emissions (e.g., rather than the sun)?
As Dana Carvey (aka the Church Lady) would say: “How conveeeenient!”
Comment by Mark Bahner — August 12, 2006 @ 8:35 pm
“As for > 2y cooling trends in the predictions. why should there be?”
Some potential reasons (the actual reason might be some combination of these reasons, rather than one reason alone):
1) Perhaps the sun’s influence is much stronger than represented by the IPCC. (How many solar physicists have bought into the IPCC’s assessment of solar forcing versus GHGs?) And perhaps the sun will begin to quiet down (by some accounts, it’s as strong as it’s ever been in the last 8000 years).
2) Perhaps black carbon is an extremely large positive forcing…much larger than assessed by the IPCC. If so, as black carbon emissions drop rapidly in the next 10-30 years, the rate of positive forcing would be significantly
reduced. (Of course, this will be offset by the fact that sulfur dioxide emissions will probably also drop rapidly in the next 10-30 years.)
3) Perhaps methane is a much larger forcing agent than the IPCC thinks. It’s also possible that methane atmospheric concentrations will drop over the coming decades…the IPCC Third Assessment Report’s “projections” notwithstanding!
Comment by Mark Bahner — August 12, 2006 @ 9:39 pm
Reid inquires How many GCM’s predicted the sharp temperature decline after the 1998 peak?
The answer, of course, is Jim Hansen in his ~1988 version of the GISS GCM
Comment by Eli Rabett — August 14, 2006 @ 8:47 pm
“Reid inquires How many GCM’s predicted the sharp temperature decline after the 1998 peak?
The answer, of course, is Jim Hansen in his ~1988 version of the GISS GCM”
Does Scenario C show a temperature decline after 1998? No.
Does Scenario B show a temperature decline after 1998? No.
Does Scenario A–the Scenario that everyone agrees has shown to be the most unrealistic scenario based on results to date–show a temperature decline after 1998? Yes.
One out of three…and the most unrealistic scenario to boot.
Comment by Mark Bahner — August 15, 2006 @ 6:47 pm
RE: #24 - I can confirm from direct observations the darkening of the skies in China (and beyond) from not only extensive use of coal (not only for power but for heat and industrial firms use for processing and blast furnaces)but also the dust. In California our skies are sometimes dulled by it even after its trasit across the Pacific.
Comment by Steve Sadlov — August 16, 2006 @ 4:35 pm
See how good the model was Mark, even the worst version got it right….
Comment by Eli Rabett — August 16, 2006 @ 8:27 pm
Eli Rabett writes, “See how good the model was Mark, even the worst version got it right…”
Ho ho ho! Good one, Eli! It’s all the more hilarious, because I bet you’re serious.
Basically, James Hansen presented 3 squiggly lines. Two out of the three squiggly lines (the ones that otherwise were closest to actual measured temperatures) went UP after 1998. One went down. So you cherry pick the one that goes up, and claim that as evidence of how good James Hansen’s model is.
Let me know when you see the Virgin Mary in one of his squiggly lines.
Comment by Mark Bahner — August 20, 2006 @ 8:24 am
Mean global surface temperatures have not increased since 1998. Australian hydrology can demonstrate an underlying process that influences global temperatures over periods of two to three decades.
Every two or three decades sea surface temperatures in the Pacific warm or cool. The changing sea surface temperatures influence atmospheric movements in the Tropical Convergence Zone resulting in periods of more frequent and intense El Nino and, alternatively, periods of more frequent and intense La Nina. This is seen in statistical analysis of Australian flooding. A period of more frequent and intense La Nina is seen in more rainfall in Australia in the period between 1946 and 1976. This is also a period of cooler Pacific temperatures and a global cooling trend. A period of declining Australian rainfall associated with more frequent and intense El Nino is seen between 1977 and 1998.
The episodic and multi decadal phenomenon seems related to solar variability - although this implies greater climate sensitivity to solar variability in ways that people are just starting to explore.
What seems certain is that cooler Pacific sea surface temperatures bring more rainfall to Australia and moderate global temperature. A ‘cool phase’ of Pacific sea surface temperatures may have commenced in 1998.
Comment by Robert Ellison — November 6, 2006 @ 6:32 pm
RE: #56 - In Electrical Engineering / Physics parlance, the PDO appears to modulate the ENSO. During a warm PDO phase, El Ninos are memorable events. During cold PDO phases, La Ninas tend to be more memorable. I can use my own observations here in Northern California as examples. Being a bit over 40, I would trust my own observations from about age 11 onward. The mid 70s were at the tail end of PDO cold phase here, and sure enough, there was a somewhat droughty version of it in place at that time. We also had a very significant low elevation snow event in February 1976. By 1978 we were transitioning out of drought and late 1978 - 1979 we had a drought breaking El Nino. Through the 80s, winters tended to be milder in temperature. We had a mini El Nino in 1981 then a full blown one 1982 - 1984. From 1985 through 1991, the state was in a split mode, where the north had generally rainy winters while the south had drought. But it was not cold droughtyness. By 1992 we were in another El Nino which lasted into 1993. 1994 was a dry year. 1995 we had a mini El Nino, 1996 was normal to wet, 1997 - 1998 was a fairly strong El Nino rain wise, temperature wise of course we know that it made a global imprint. Significantly, during 1998 we appeared to enter into the beginning of a cold phase PDO, and true to form, just prior to the Winter Solstice we had the most significant low elevation snow event since 1976. 1999 was very neutral. 2000 into 2001 was a mini El Nino. 2002 was unremarkable. 2003 - mid 2005 we had a perisistent Siberia Express set up with tremendous snows in the high country. 2005 - 2006 Winter the precip got a late start but it was very cold, Feb - March 2006 we had an unprecented series of low elevation snow events. To date, fall - winter 2006 has been a story of early onset and a seeming contention between a more classic El Nino synoptic set up and the Siberia Express set ups so common thus far in the naughties. We just had a couple of weak zonal flow events that were cooler and less rainy than a “normal” El Nino with an obvious shift to a more northerly flow currently in process. If that becomes persistent and gets into a proper Siberia Express, then it will confound the notion of what El Nino “ought” to be doing.
Comment by Steve Sadlov — November 6, 2006 @ 8:43 pm
[...] However, skeptics like Roger Pielke jumped at the opportunity to use this as evidence that climate models are not to be trusted and anthropogenic global warming should be reconsidered: The recent dramatic cooling of the average heat content of the upper oceans, and thus a significant negative radiative imbalance of the climate system for at least a two year period, that was mentioned in the Climate Science weblog posting of July 27, 2006, should be a wake-up call to the climate community that the focus on predictive modeling as the framework to communicate to policymakers on climate policy has serious issues as to its ability to accurately predict the behavior of the climate system. No climate model that we are aware of has anticipated such a significant cooling, nor is able to reproduce such a significant negative radiative imbalance. [...]
Pingback by Ocean Cooling « Reasic — April 20, 2007 @ 1:01 pm