Contribution invitée de Juerg Luterbacher, université de Berne, Suisse (traduit par Thibault de Garidel et Gilles Delaygue)
Alors que la nature anormale des tendances récentes de la température moyenne globale est souvent soulignée dans les discussions sur le changement climatique, les changements aux échelles régionales ont une signification sociétale potentiellement plus importante. Un exemple particulier est le lien possible entre le changement climatique et l’incidence des vagues de chaleur estivales [Meehl, G.A. et C. Tebaldi, Science, 305, 994-997, 2004] comme celles observées en Europe pendant l’été 2003 [ voir Schaer et al. , Nature, 427, 332-336, 2004; Stott et al. , Nature, 432, 610-614, 2004]. Des analyses préliminaires des températures moyennes annuelles de l’air pour l’Europe pour l’année 2004 montrent que cette année figure parmi les plus chaudes (mais pas aussi chaude que les années 1989, 1990, et 1999-2003) depuis que des mesures instrumentales sont disponibles (approximativement les 150 dernières années). La température moyenne de 2004 dépassait de plus de 0,8°C les températures moyennes de la période de référence 1961-1990. Les plus fortes anomalies ont été observées en Europe de l’Est et dans les pays Scandinaves. Pour chaque mois de 2004, la température était supérieure à celle de référence; pendant les mois de Février-Avril, Août, Octobre et Décembre la température dépassait de plus de 1°C la période 1961-1990. Au cours des 100 dernières années, les températures moyennes annuelles de surface en Europe ont augmenté d’environ 0,85°C. La tendance à la hausse s’est accélérée au cours des dernières décennies, avec environ 1,2°C de réchauffement au cours des 30 dernières années (1975 à 2004). En fait, les trente dernières années sont probablement la période pluridécennale la plus chaude en Europe des 500 dernières années au moins [Luterbacher, J., Dietrich, D., Xoplaki, E., Grosjean, M. et H. Wanner, Science, 303, 1499-1503, 2004], alors que la dernière décennie (1995-2004) est probablement la décennie la plus chaude, et enfin l’été 2003 celui le plus chaud. Ces conclusions sont semblables à celles tirées pour l’ensemble de l’hémisphère Nord.
Excuse my poor english !
During summer 2003, 14 000 old persons died due to very high temperature !
“All climate is local”! Perhaps this is an overused phrase, but it’s critical to remind ourselves that no-one experiences the global mean.
One of the factors that is attributed to keeping Europe’s climate warm is the continued high rate of North Atlantic Deep Water (NADW) formation. The impact on Europe of NADW formation is the continued transport of warm tropical ocean waters northward to Atlanic waters off Europe. In order for this global oceanic “conveyor belt” of heat to remain healthy, the Gulf Stream must transport water with high salinity (a result of tropical and sub-tropical summer evaporation) toward the North Atlantic, where it is cooled by the first winter outbreaks of cold air off the Greenland Ice Cap and the ice-covered polar oceans. When this salty surface ocean water is cooled sufficiently, it becomes too dense to float above the waters it overlies, so it sinks “like a rock”. The sinking leaves a “vacuum” in its wake which helps to draw more tropical water northward.
Greenhouse warming poses potential (somewhat speculative) threats to NADW formation (and thus to Europe’s continued warm climate). There are at least three factors that could contribute to this threat:
1.) If precipitation increases over the tropical oceans, more than evaporation increases, the sea water salinity could decrease.
2.) If the putative Arctic magnification of global warming prevents the cold air outbreaks from cooling the northward moving saline water, it may not cool enough to become convectively unstable. It is this convective instability that forms the deep water — surface water becomes unstably dense and sinks in a manner not unlike the way unstable summer air produces thunderstorms.
3.) Finally, if the North American hydrologic cycle is enhanced, and/or Greenland’s southern ice caps melt, the increased fresh water runoff from land areas could dilute the ocean surface water and critically reduce its salinity. (Water is more dense when it is salty — fresh water will float above salt water, so if the North Atlantic is freshened, the formation of NADW could be suppressed.)
One of the most prominently feared causes of sudden catastrophic climate change (as in the movie “The Day After Tomorrow”), is the sudden cut-off of NADW formation. The last time this apparently happened was during the “Younger-Dryas”, 12,000 years ago (or possibly briefly during the 8.5kya event). That led to a cold period in Europe which returned the continent temporarily to near peak Ice Age conditions.
What we are seeing with these latest observations regarding Europe’s climate, is that NADW formation seems “alive and well” and not in immediate danger of failing. (But this is my opinion, speculative, and just one possible explanation).
One reason for the observed continued health of the NADW formation process may be circulation changes, which are manifest in the increased prominence of the positive phase of the Arctic Oscillation (This means a stronger southwesterly surface wind over the western North Atlantic driving a stronger Gulf Stram current. The enhanced AO has, in itself, been attributed to the influence of GHG warming).
Another reason for the health of the NADW and the Global Conveyer Belt Oceanic circulation may be that the Arctic warming trend is least manifest during fall and early winter, which are the most critically active times for NADW formation. (During this time the greatest temperature contrasts occur between air and water over the North Atlantic. Strong cold Arctic outbreaks pass over the most saline waters of the year, which were produced by late summer evaporation in the tropics and sub-tropics and transported northward by the Gulf Stream).
Why the Arctic seems to be warming least during fall and early winter is unknown. In the late 1990’s, I thought I had a plausible hypothesis, but there were not enough quality boundary-layer data to test it. (As a teaser, my hypothesis proposed a negative feedback process initiated by Global Warming itself — an enhanced injection of (anthropogenically) warmed air just above the cold, dense surface air, then interacting with the turbulent-radiative processes which occur in the stably stratified [meaning there are temperature inversions] boundary-layer air as the Arctic plunges into its sunless season.) But the cause could be as simple as seasonal circulation changes.
Yes, I know I’ve thrown a huge amount of complexity into a mixing bowl here. Unfortunately this is the sort of interdisciplinary tangle that climate scientists face if we are to solve the climate puzzle. We need to be cognizant of everything from local-scale stable boundary layer micrometeorolgy and ocean unstable boundary layer turbulent processes to global oceanic and atmospheric circulation patterns such as the Arctic Oscillation and the Gulf Stream’s seasonal evolution.
No wonder climate science is such an endlessly fascinating endeavor!
I think that the geological records show that the formation of NADW is rather stable. Despite continued melting of the Laurentide ice sheet throughout the Pleistocene-Holocene transition, the only major perturbation of the NADW after Younger Dryas was the rapid drainage of Lake Ojibway at 8.2ka. Whilst this had a substantial impact on the climate of the North Atlantic region and elsewhere, I can’t find any evidence that it caused a consequent long-term reduction of NADW formation. Even with today’s highly elevated drainage from Arctic rivers, the amount and rate of freshwater flowing into critical areas of the North Atlantic is about two orders of magnitude lower than that during the 8.2ka event. There are no substantial glacial lakes forming today that could produce comparable amounts of water (although a few hundred years of melting of the Greenland ice sheet could change that situation). However, in this scenario warming would probably have been so widespread and pervasive as to render any “Younger Dryas” type of cooling irrelevant climatically.
In addition to #3, the estimated quantities involved in the 8.2ka event can be found at: http://pangea.stanford.edu/Oceans/GES205/superlakes.pdf and the increase of discharge of the largest rivers in the Arctic can be found at: http://news.nationalgeographic.com/news/2002/12/1212_021213_arcticrivers.html
in response to #1, there is 5-10 times more mortality from cold periods than from heat waves, according to an investigation in Europe. The figures are for towns, but may expanded to the population of the countries involved, or even for Europe. As can be seen, people adapt themselves to local/regional climate, but feel the impact of sudden changes either way out. See: http://bmj.bmjjournals.com/cgi/content/full/321/7262/670
Aviation-cirrus (aviation-smog, really) is still generally being underestimated as a contributor to both global/regional warming and global/regional dimming, to more rain and more drought. It is caused by airtraffic at great heights. These exhaust fumes of airplanes contain, apart from water and CO2, soot particles, microscopic aërosols and a host of chemical substances that don’t belong in the higher regions of our troposphere. It is unlikely that this major change in the vital climate-system should be without serious consequences. The burden of proof should be shifted to the aviation industry: Let them prove that their man-made layer of high-altitude smog has no undesirable consequences for weather, climate and the quality and quantity of the sun’s radiation that reaches the surface of Earth. In the discussion about radiative forcing, greenhouse effect and global/regional warming, contrails and aviation-cirrus should get more attention. My website http://www.contrails.nl shows pictures of aviation-cirrus in The Netherlands, from 1995 until now.
…why is the the mean temperature reference period 1961-1990?
Just curious as to why that time frame was chosen.
[Response: See our glossary item on the Surface Temperature Record. In particular, see our link therein to a FAQ provided by the University of East Anglia’s Climatic Research Unit (CRU). -Mike]
Was really hoping you guys will comments on this:
Countdown to Global Catastrophe:
http://news.independent.co.uk/low_res/story.jsp?story=603975&host=3&dir=507
[Response: May be interesting. Putting a number on “permissable” climate change is somewhat new… maybe we should look at the science behind doing that. Though, if its a policy matter, you might be better off asking Prometheus – William]
…well, thanks for the links, but I’ve been there many a time already.
I’ve just never seen a logical, or illogical for that matter, reason why if the number of stations increased to just over 3000 stations during the 1951-90 period, and held roughly constant, the reference period is just the years 1961-1990.
Why discard the first ten years of the period with the most stations for reference to? Afterall, according to CRU “many stations do not have complete records for the 1961-90 period”, yet they had no problem constructing elaborate methodologies to interpolate lots of historical data from those incomplete records? Why not the 1951-60 time frame as well?
GISS, didn’t seem to have a problem in relying on those first ten years when it constructed it’s first GISTEMP temperature record in 1987 with records from 1951-1980.
[Response: It doesn’t make any difference. Traditionally the most recent 30 year period is used, and so I imagine you’ll start seeing ‘with reference to the 1971-2000 mean’ appearing soon. The offset is simply the average anomaly for that period. There is a reason why we plot anomalies (rather than absolute temperatures), but the base line is arbitrary. -gavin]
Re #7, that ties in nicely with what Roger Pielke Jr recently wrote at the Prometheus science policy weblog (Follow Up On Landsea/IPCC
– see last 4-5 paragraphs):
“But more troubling than a lack of knowledge of the substance of the science of the IPCC reports is the political stance on climate taken by the head of the IPCC. […] So long as people within the IPCC leadership sees its role as political advocate rather than honest broker, and acts accordingly, we should not be surprised to see future controversies erupt in the IPCC. The solution is not to retreat into the illusion that it can deal only with science, but to openly confront the reality that its very existence is based on the need to connect science with policy.”
The political plot thickens, it would seem…
[Response: this appears to be an error or misstatement in RP’s post: I don’t think the IPCC folk do think its (the IPCCs) role to be a political advocacy. All the IPCC folk quoted were speaking personally – William]
re: #7 –
Countdown to Global Catastrophe:
http://news.independent.co.uk/low_res/story.jsp?story=603975&host=3&dir=507“
[Response: May be interesting. Putting a number on “permissable” climate change is somewhat new… maybe we should look at the science behind doing that. Though, if its a policy matter, you might be better off asking Prometheus – William]
A good start might be a piece which justifies the high climate sensitivity which is assumed by the authors of the report and also apparently by climate models. As this is an issue which is fairly crucial to the whole climate change debate I’d have thought it would have already been covered fairly thoroughly on this web-site. So far – not a dicky bird – apart from Gavin’s brief mention of ice age forcings.
I have made a previous for an article (see Post 34) here.
[Response: Climate models (GCMs) calculate their climate sensitivities, they don’t assume them. It should be easy to find stuff from IPCC; David Appell has a piece today about recent climateprediction.net results showing a range of sensititivies up to and beyond 8 oC for CO2 doubling – William]
Dr. Schmidt:
…arbitrary, eh? Well, at least it’s an answer.
Btw, this is a extremely interesting site you and your companions have started. Good luck with it.
I have heard it stated that climate models “predict” that early signs of climate warming will be more pronounced in the Northern Hemisphere and especially over land. Europe seems to bear this out. So does Alaska. Is this correct? If so, how do the models arrive at this conclusion?
Are other high latitude NH regions warming just as fast (e.g. Northern Canada)? If not, why not?
How does the Arctic Oscillation fit into the picture?
I know that’s a lot to answer. Any new information would be welcome.
[Response: There is evidence that the enhanced continental winter warming in the Northern Hemisphere, which has resulted at least in part from a recent trend towards the positive phase of the Arctic Oscillation (AO), may in fact also represent a response to anthropogenic impacts on climate. An excellent discussion of each of these issues is provided by this NASA press release which links to a supporting study published in Science by Shindell et al (1999) -Mike]
Regarding #10:
What climate models assume is a wide-ranging compendium of physical processes that are either well known but too complicated to incorporate into the climate model (for example the direct radiational effect of Carbon Dioxide on greenhouse warming is considerably *simplified* compared to the most sophisticated “line-by-line” radiation models that are available, simply because there isn’t enough computer power to make the line-by-line calculation at every location on Earth at every time step within in a GCM), or are not sufficiently well-known to treat them with complete certainty.
To fairly characterize the ClimatePrediction.net results, their web site makes the statement:
The aim of climateprediction.net is to investigate the approximations that have to be made in state-of-the-art climate models (read more about this). By running the model thousands of times (a ‘large ensemble’) we hope to find out how the model responds to slight tweaks to these approximations – slight enough to not make the approximations any less realistic. This will allow us to improve our understanding of how sensitive our models are to small changes
The point here that seems to have been overlooked is that using plausible “realistic” approximations to the climate system, ClimatePrediction.net’s model simulations have produced a climate sensitivity ranging from 2C to 11C. This is surely not a comforting result for those who place their faith in climate models.
However, in the interest of fairness, it must be noted that the most likely GCM sensitivity for the ClimatePrediction.net model is going to lie somewhere within this range, and by weighting the *most likely* approximations more than the more extreme approximations (which ClimatePrediction.net has the wonderful luxury of exploring), the ensemble result is likely to actually narrow and clarify the range of “most likely” climate sensitivity, rather than broaden and smear it.
Still, we are dealing with only one model in this discussion, and we have seen quite considerable divergence between models in some of the critical approximations — particularly in the way they parameterize the critical cloud feedback processes.
I wrote to Roger Pielke, Jr., expressing the hope that he’d respond to William Connolley’s comment within item 9 above. Dr. Pielke chose not to address Dr. Connolley’s comment here at RealClimate. He posted my question on his own blog, Prometheus, and responded indirectly in a pair of inter-related places there:
Reader Mail on Political Advocacy
http://sciencepolicy.colorado.edu/prometheus/archives/climate_change/000330reader_mail_on_polit.html
January 27: A Good Example Why Politics/IPCC Matters
http://sciencepolicy.colorado.edu/prometheus/
A comment: if participants in techno-civic climate discussion at the stature level of Roger Pielke and William Connolley disagree in an important way on an important point, as apparently they do, I hope that blog balkanization doesn’t stop them from debating it in public in some direct way.
[Response: Roger Pielke and I disagree on this point, as you’ve observed. RealClimate is doing its best to remain focussed on science not policy, so there probably won’t be a debate here. The “balkanisation” is indeed a problem. I’ll probably put up a comment on my pet blog (link here when I do it) – William
OK, I have: its here – William]