Guest commentary on BBC documentary on “Global Dimming” aired on January 13th 2005 by Beate Liepert, LDEO, Columbia University
I haven’t yet seen the documentary. I have only read the transcript and hence was spared the pictures of the potential apocalypse and the invocation of biblical-scale famines. However, as one of the lead scientists on the topic [and who was interviewed by the BBC for the Horizon documentary (transcript, previous post)], I feel I should explain a few things about it without using religious analogies and stoking unnecessary fear.
First though, this is a nice example of the power of words: Gerry Stanhill coined the observed reduction in solar energy reaching the ground “global dimming”. He called it “global” dimming because the technical term for the radiative energy is called “global solar radiation” and it contrasts nicely with the more common “global warming”.
Secondly, there are three published studies out on long term changes in solar radiation (or “global dimming” if preferred). All use the same data sources. Solar radiation has been measured at weather stations worldwide since about 1956-57. As with many other measurements most of the data are from the Northern Hemisphere and all are taken on land. A reduction in downward solar radiation of about 4% or about 7W/m2 from 1961 to 1990 was found at stations worldwide by Gilgen et al., (1998). Gilgen et al. did a quick analysis and used all the available data with increasingly shorter records for their trend statistics. Stanhill and Cohen (2001) calculated a stronger reduction of about 8% per decade. The reason for the discrepancy might be that only 30 records were used in the latter study and it seems only the ones with the declining trend. My own analysis was based on 110 continuously recording stations worldwide from 1961 to 1990 (Liepert 2002). I confirmed Gilgen et al.’s estimate of a reduction of about 4% in three decades. Since the late 1980s a recovery seems to be occurring but the studies demonstrating this are not yet published.
Why is solar radiation changing? From observations we can separate cloud-free skies and cloudy conditions. We can hence infer clouds or atmospheric transparency as possible causes for the dimming. In my study of the US data I identified clouds as the main reason for the dimming of sunlight. Only about a fifth of the dimming could be observed during cloud-free conditions.
Why should the atmospheric transparency change at all in cloud-free conditions? V. Ramanathan explained it in the BBC documentary. Sunlight is reflected by air pollution or absorbed in the atmosphere before it reaches the ground. Field campaigns like INDOEX show this clearly (well, not “clearly” in the literal sense!). Advanced climate models include this “direct” aerosol effect and base their inputs on experiments like INDOEX.
Why should clouds change? Global warming for example. Surprised? Most climate simulations predict some “global dimming” due to the water vapor and cloud feedback of greenhouse gas forced global warming. Global warming, however, affects the entire atmosphere whereas global dimming is only a surface and near-surface phenomena. Hence global warming and global dimming are not exclusive or contradictory. (Incidentally, the decline of solar energy at the surface inferred in my study is about 60% of the increasing longwave radiation in a typical global warming climate simulation (Feichter et al. 2004)). With global warming, atmospheric moisture increases and this makes the atmosphere slightly less transparent to sunlight. Furthermore once clouds are formed, they tend to hold more water and therefore look a little darker.
Anthropogenic aerosols were however singled out in the BBC documentary. They are correctly believed to change cloud reflectivity and cloud lifetime. Scientists are currently trying to assess the magnitude of these aerosol-cloud interactions and the impact on climate. Rotstyn’s study on the Sahel drought is one example. But as Giannini et al.’s study showed, you can look look at the Sahel drought with a number of different approaches. Clouds have always posed the greatest challenge for climatologists and I regard my own research on as a contribution to this ongoing debate.
Currently, the best climate models include estimates of all these effects: anthropogenic greenhouse gas forcings, aerosols, natural solar cycles, and volcanic eruptions (see here for example). The inclusion of aerosols in climate simulations has improved the model hindcasts when tested against past climate and dimming (Wild and Liepert, 1998). The latest climate projections for the future therefore all include some estimates of aerosol changes. The scientifically interesting and new part is that introducing aerosols requires a closer scrutiny of surface energy and water budgets than was previously done. An example of this kind of analysis of a climate model in the context of “global dimming” can be seen in Liepert et al. (2004), but other model groups are currently performing similar analyses too.
Finally a comment on language. It concerns me that articles from scientists and journalists alike have a tendency to use biblical and apocalyptic terms. This might be an appropriate way to describe a baroque church in Bavaria or a painting of P.P. Rubens (for instance in my favorite museum, the Alte Pinakothek in Munich) but I would rather keep this emotive language out of scientific discussions.
Thank you for a very clearly-written and informative post. I have a few questions:
First, the transcript uses the following comment from you as the lead-in to the comments by Peter Cox predicting major effects from the decline of global dimming over the next century: “We lived in a global warming plus a Global Dimming world, and now we are taking out Global Dimming. So we end up with the global warming world, which will be much worse than we thought it will be, much hotter.” The upshot to his remarks seems to be that the 2001 IPCC maximum temperature increase estimate is wrong by nearly a factor of two. Do you agree with this, and does it appear to you that the IPCC is going to as well (and raise their estimate similarly)? Also, what is your view of Peter’s apparent predictions of major climate disruption impacts over the next twenty to thirty years?
Second, your linked 2004 article discusses the global hydrological cycle “spinning down” under the influence of global dimming. As the cycle spins back up with the waning of global dimming, will that generally mean an increase in storm activity and extremes?
This is a really well constructed article. I second the “thank you”.
Yet I am left a little confused on one point: You ask the question “Why should clouds change?” You then point out that the BBC documentary emphasizes one cause, while seeming to neglect another. They focus on anthropogenic aerosol indirect effects. But you suggest a second cause — global warming itself.
An apparent problem with that (my confusion) arises from the Liepert et al. 2004 paper. If the hydrological cycle spins down in response to the combination of increased aerosol burden and GHG warming, doesn’t that suggest a reduction in cloudiness?
In other words, the finding from your study was that most of the “dimming” occurred during cloudy conditions. So to invoke GHG warming as the cause of this, you seem to be required to argue that global warming will increase moisture in the air, and that the increased temperature and moisture conditions under which clouds exist will cause the clouds that occur to have a greater optical depth. Yet if there is less water vapor available due to the combined effect of “global dimming” and global warming, then it would seem to me that the more vapor-starved clouds that result would have to rely entirely on the aerosol indirect effect to overcome the diminished vapor availability.
The influence of global warming alone on a more moisture starved atmosphere would seem to require fewer and thinner clouds (the warmer atmosphere can hold more water vapor in suspension without requiring it to condense into cloud). Fewer and thinner clouds mean more solar radiation reaching the surface. So the burden would seem to be *completely* on the aerosol indirect effects, since they must first counteract the moisture starvation/thinner cloud effect, then go on to actually reduce the total observed solar radiation. If this argument is sound, it seems to suggest that aerosol indirect effects may actually be stronger than first thought?
Comments?
Thank you for your encouragements.
Let me explain my arguments a bit better:
1) In Feichter et al. 2004 we describe a modelling study were we run the climate model twice: One run is with estimates of present day anthropogenic and natural aerosols and green house gases (called present-day scenario ~1985). In the second simulation we reduce the aerosols to the natural concentrations (called pre-industrial ~1885) and keep the greenhouse gases constant at present day level (~1985). This is what you would assume as a scenario when we clean up aerosols. (See table 4, AP experiment, in the paper we turn the argument around and discuss the effect of aerosol increase). The result is a temperature increase of 0.8 degree Celcius in this particular model (MPI-Hamburg). Note that this model does not include ocean dynamics. Subjectively, I find 0.8 degree “a lot hotter”.
2) I agree, my 2004 paper provides counterintuitive arguments. Let me please explain it in more detail and I hope you’ll see my point:
The key is the difference between moisture storage and moisture fluxes. GHGs modify storage and fluxes, and aerosols modify fluxes.
When temperature increases in the atmosphere due to GHGs the moisture holding capacity should go up following the Clausius Clapeyron formulation. The simulated model atmosphere is indeed moister (absolute humidity goes up). It is relative humidity not absolute humidity that governs cloud formation. Relative humidity is fairly constant because it depends on other meteorological factors such as circulation (e.g. ITCZ). Once cloud formation takes place in a moister atmosphere these clouds hold more water and should be optical thicker. Optically thicker clouds reduce sun light. Reduced sunlight reduces evaporation. Now you say this should lead to moisture starvation in the atmosphere. Correct, but…
BUT there is another way to increase atmospheric moisure:
Leave moisture in the atmospheric storage for a longer time with the same or less fluxes in and out. (Lifetime of water vapor goes up from 10 to 10 1/2 days.)
The indirect effect, as it is simulated in this model, reduces precipitation efficency and helps keep rel. humidity and cloud cover constant. This is indeed as you mention important to actually tip the balance towards “reduced evaporation in a moister and warmer world”. Note the the indirect effect is highly uncertain. Hence you are right that indirect aerosol effect is important. But moisture can increase because the atmosphere can hold more due to warming. Hence it IS the combination of both.
Incidentally, even in “greenhouse gas only” simulations when you have an increase in rainfall and evaporation this increase is not as great as you’d expect from Clausius Clapeyron because of the fairly constant rel. humidity and cloud coverage. Even in “greenhouse gas only” simulations the lifetime of water vapor increases. The lifetime effect counteracts the increases in rain and evaporation due to the Clausius Claperyron. In these GHG experiments however, the lifetime effect is small compared to the Clausius Clapeyron effect.
Thus it is plausible to get a warmer, moister, darker world where it rains less.
Thank you. I understand and accept that. The GHG warmed atmosphere maintains a greater moisture reservoir through which moisture cycles more slowly than its pre-industrial counterpart. That is what your model reveals, and it is also fairly intuitive.
Still there remains a question in my mind about the relative importance of the GHG warming and the aerosol indirect effects on “global dimming”. If I understand you correctly, the model you use demonstrates that a net reduction in solar radiation (a portion of the “dimming”) is caused by the GHG warming effect alone. However there is a lack of consensus among the 10 models quoted in IPCC TAR on this issue. See Fig. 7.2; with the accompanying statement that “In response to any climate perturbation the response of cloudiness thereby introduces feedbacks whose sign and amplitude are largely unknown.” (first paragraph).
I would appreciate your perspective on this. Specifically what fraction of the “dimming” would you estimate from your work to be caused by the GHG warming? And finally, if it is fair to make this request, how much would you estimate this fraction to vary among the 10 models quoted in IPCC TAR?
The relative importance of GHG warming and indirect effect and a revised version of the IPCC TAR of the cloud feedback is exactly what I am working on. Results should be out soon. Sorry, I can’t tell more. Patience.
Beate
< smile > Reluctantly, I shall wait.
As Gavin and Mike recently posted regarding “peer review”, science normally advances in baby steps. The quantum leaps that the general public perceives are almost always the result of a long and unheralded period of “foundation building” behind the scenes.
I hope you will post here when a paper is accepted. Best wishes.
— Pete
Some more questions…
1) The emission of sulphate aerosols in Europe is drastically reduced since the mid-seventies (-70 %). The largest effect of this reduction should be found downwind of the main sources. According to the Hadcm3 model a difference of ~5 K in north Scandinavia/Russia over a 10 years time span. But there is no difference in trends, attributable to aerosols, between less contaminated area’s and the area with the highest contamination. See: aerosols
2) An investigation in Switzerland presented at the latest dimming conference on your pages (see dimming conference) shows that the reduction in surface solar insolation was compensated by increased downward LW radiation. This was attributed to the measured increase of water vapour. But should one not expect less water evaporation with decreased insolation? And what part of the SW of incoming sunlight is absorbed by water vapour, and how much (W/m2)?
[Response: due to the complexities of the system, there is no need for decreased evap (if its occurred) to result in lower WV. In fact (a bit speculative) it can be the other way round: increased WV would tend to suppress evap – William]
3) Global dimming largely is attributed to more reflection from clouds and longer lifetime of clouds as result of mainly sulphate aerosols. But the “eartshine” project points to less reflection from clouds and the general trend is less clouds (~1% globally over the last decades). Together with dimming, this points to more absorption in the atmosphere, not reflection.
[Response: ?Who says its mostly reflection? I thought it was mostly absoption – William]
4) An investigation in the Indian Ocean compared a highly contaminated area with a less contaminated one (see: Norris). The result is that there is no difference in regional cloud cover trends, neither of precipitation, with increasing contamination and that the contaminated area has more dimming, but warmed more than the less contaminated area. As the aerosol contains a large fraction of soot, this too points to more absorption, outweighing the combined direct and indirect reflective effects of sulphate and other aerosols.
5) There is increased dimming in Australia and Antarctica, where there are little or no trends in aerosols…
[Response: the Weather article, which is by Stanhill, says there has been no dimming in Australia, so I wonder what you source is for this? Of course the same article says that there is decreased pan evap in Oz, which taken together would appear to be an anomaly for the GD-causes-decreased-pan-evap theory – William]
My impression is that the warming effect of soot aerosols is underestimated and the negative effect of combined aerosols is overestimated. And indeed increased water vapour may play a role…
Your comments?
Just a quick clarification regarding those naughty clouds that prevent us from solving the greenhouse puzzle. There are three types of mechanisms that conspire to give a cloud it’s id: dynamical (air movements), thermodynamical (temperature/humidity conditions) and microphysical (droplet collisions, aerosol effects. Mix the three in different quantities and you can get clouds as different as a black snow-making giant and a cute puffy cumulus with a smile on it’s side.
To declare that a warmer atmosphere will make clouds universally thicker or thinner or cuter is a dangerous oversimplification. We need to move away from hand-waving arguments and rely on thorough data and model analysis results. Such results point to decreases in low-cloud optical thickness with warming in warm enviromnents and increases in cold ones. The former are caused by faster water loss and higher cloud bottoms with warming and the latter by increased cloud water density with warming. This just to point out that even for one cloud type there is no hard universal rule on how it behaves when things get warmed up.
Thanks William for your reaction on #7…
About comment on 3), the INDOEX report by Ramanathan (Science 2001, Aerosols, climate and the Hydrologic cycle) points to a neutral TOA (top of atmosphere) energy exchange, with 14 +/- 3 W/m2 less light reaching the surface and 14 +/- 3 W/m2 more absorption on the (high black carbon content) aerosols in the atmosphere. But they add more reflecting, by a negative (5 +/- 2 W/m2) TOA balance due to secondary effects of aerosols on cloud brigthness. That should give a cooling in the SE Asia region. But the observed changes in 4) of my comment give a warming in the part of the Indian Ocean which is most polluted. Thus in reality, any reduction of (soot) aerosols in SE Asia would induce a cooling, not a warming. That is opposite to what the Horizon program want us to believe. And it casts doubt on current global estimates of aerosol influence (and consequently CO2 influence), which probably are overestimated.
About the comment on 5), the source was the Horizon transcript, where Dr. Michael Roderick from Australian National University combined global dimming and pan evaporation. But if there is no change in dimming, no change in cloud cover and little influence of aerosols in Australia (and Antarctica), there may be no connection between these all. An alternative explanation maybe (again) a change in water vapour, but my question what, and how much, water vapour absorbs from incoming sunlight still is open…
Regarding #3 by Beate Liepert on atmospheric moisure:
Is there data that indicates atmospheric moisture is increasing
as expected by 2001 IPCC? More? Less?