( はてなダイアリーから移動しました)

Workshop on solar activity and climate change (Nagoya, 15 Nov. 2010): A personal summary

This is an English version of my posting here on 2010-11-18. The contents are not exactly the same.

I participated in a workshop on solar activity and climate change held in Nagoya University on 15 Nov. 2010. Its official announcement is here (in Japanese). I was not a presenter, but I gave some questions. I am going to write here a summary of my personal impressions, which should not be taken as a overall summary of the meeting.

As one of the organizers (Tetsuzo Yasunari) explained, the idea of this workshop emerged in the context of a symposium about IPCC held by the Science Council of Japan on 30 April 2010 [official announcement in Japanese in PDF format]. But the focus is different. The subject of discussions at Nagoya workshop is issues within science, and not relationship between science and society. Even within science, the focus of Nagoya workshop is pure scientific discussion about connections between the sun and climate, rather than the scientific outlook of climate change in the coming century which is the subject of the Working Group 1 of IPCC. Nevertheless, there seems to be some shadow of the initial interest related to IPCC in the fact that the time scales of climate change discussed were up to a thousand years, which is very short compared with the whole history of the earth.

There were few expressions of views skeptical to the outlook of global warming due to increase of greenhouse gases. Syun-Ichi Akasofu presented his strongly held view that the observed rise of temperature and retreat of glaciers continues since around 1800 in a linear fashion which he considers as "recovery from the Little Ice Age". He says that therefore the role of CO2 is minor. As the cause of "recovery from the Little Ice Age", he suggested solar activity, but not particular mechanisms. There were some other presenters who had said something skeptical to the role of CO2 elsewhere, but I did not notice they said so in the present meeting.

On the other hand, it seemed to me that the majority of participants consider that the relative weight of the role of solar variation on climate change is not well understood yet, and that their understanding may change following the developments of research in the future.

Even so, solar activity is not strengthening during the latter half of the 20th Century by any of the available indicators. Therefore the rise of global mean temperature since 1970s cannot reasonably explained by solar activity. If there may be an explanation, it is related to the fact that the 20th Century as a whole is an epoch with remarkably strong solar activity among the latest several thousand years (as introduced in the lecture by Kanya Kusano based on analysis of Carbon 14 by Usoskin, Solanki et al.). If we can assume that the climate system responds to forcing at a time scale larger than 50 years, we may consider that the effect of solar forcing is still to the direction of warming.

Roughly speaking, the participants were people from three disciplinary communities.

The first is members of the community of meteorology (atmospheric physics). Even though solar energy is ultimate source of driving force of nearly all meteorological phenomena, it is very often assumed in meteorology that the energy which the sun emits is constant. So those people who discuss the effects of solar variability is rather a minority among this community.

The second is members of the community of "solar-terrestrial environmental science" or "solar-terrestrial system science" as they call themselves. They deal with the rarefied atmosphere higher up than where meteorologists study, with the magnetosphere around the earth, and with the interplanetary space. Phenomena in this realm are obviously affected by the variation of the sun, in particular, of the solar magnetic fields. Many members of this community tend to consider that the lower atmosphere is also affected.

The third is scientists who reconstruct paleo-environment making use of isotopes. Such isotopes as Carbon 14 and Beryllium 10 are created by cosmic rays hitting other species of atoms, and the amount of galactic cosmic rays arriving to the earth is affected by solar magnetic fields. They can discuss relationships between paleoclimate indicators on one hand and solar activity as manifested in variations of cosmic rays on the other hand based on their analyses of tree rings and core samples of ice sheets.

Even though the three groups are sub-groups of earth science, it is few that they have discussion together. The meeting was an opportunity of cross-disciplinary exchanges of ideas.

The pathways which solar variation can affect climate are broadly classified as those through fluxes of electromagnetic waves (light waves) and those through magnetic fields and fluxes of charged particles (mainly protons).

Among the factors of electromagnetic waves, the first is the total solar irradiance (TSI, or so-called solar constant). It is observed by satellites since November 1978. It varies with the sunspot cycle (whose period is approximately 11 years in the modern era), and the energy flux is larger when there are more sunspots. (Even though sunspots emit smaller energy than the normal parts of the photosphere, brighter parts called faculae also tend to occur more frequently when the sunspot activity is high.) But this variation amounts to just +/- 0.1 % of TSI itself. There were times in the 17th Century called Maunder Minimum when very few visible sunspots were visible. TSI during that period was estimated to be considerably lower than present by Lean et al. (1995) based on analogy with other stars. But later studies of stars resulted in such conclusion that TSI during the Maunder Minimum is much nearer to the present value. Thus, it is not plausible that this pathway leads to detectable climate changes. (We also need to re-evaluate the results of numerical experiments which simulated the climate of the last 5 centuries using the solar forcing estimated by Lean et al. 1995).

Among the electromagnetic waves, the ultraviolet part varies with sunspot cycle and its relative magnitude is 3 to 8 %. Ultraviolet radiation is absorbed by ozone in the stratosphere, so that its variation generates variation of temperature in the stratosphere. (It also generates variation of concentration of ozone and other chemical species which is related to its generation and decomposition, but I skip discussions about that.) Kunihiko Kodera, who has been working on this subject, presented an overview of how the variation can affect the troposphere as well. As the stratosphere has smaller mass than the troposphere, it is difficult to have a process where the stratosphere gives energy to the troposphere. On the other hand, if the global scale fields of temperature and winds in the stratosphere (these are mutually connected by dynamics), it can affect the way of propagation of dynamic waves which originate in the troposphere. Therefore the distributions of large-scale troughs and ridges in the atmosphere can change, which brings positive anomalies of temperature to some regions and negative ones in others. It may result in some anomaly in global mean temperature, but it is just a residual sum of various positive and negative regional anomalies. We will not able to grasp this process by analysis focused on global mean temperature. [A recent review article by L.J. Gray et al. (2010) "Solar influences on climate" in Reviews of Geophysics also includes discussion of this pathway.]

One of the pathways which solar magnetic fields can affect climate is modulation of galactic cosmic rays.

Hiroko Miyahara introduced a new paper she co-authored (Y.T. Yamaguchi et al. 2010 in PNAS, doi:10.1073/pnas.1000113107; A press release from AORI, University of Tokyo, in Japanese), which compared cosmic ray intensity (based on Carbon 14) and humidity in Japan (based on Oxygen 18) of the same sample of tree rings. Cyclic variations are found even in the Maunder Minimum period where they cannot be detected by sunspots, and climate indicators seem to respond to double cycles, i.e. cycles of reversals of solar magnetic fields, rather than single cycles of solar variation. It suggests pathways involving solar magnetic fields. Among them, Miyahara considers that cosmic rays are likely intermediary, based on the shape of the paleoclimate signals. I think that we should not put too much weight on the pattern of time series of climate indicators at one location. As Takeshi Nakatsuka (another co-author of the paper and also an international committee member of IGBP PAGES) emphasized, we should look at spatio-temporal structures based on records at multiple locations.

About the pathway from cosmic rays to climate, the hypothesis by Henrik Svensmark is now well known, that ionization of air causes production of aerosol particles which can act as cloud condensation nuclei. According to discussions in the workshop, the hypothesis seems to need major revisions. First, the correlation between cosmic ray intensity (by neutron monitor) and amount of lower clouds as presented by Svensmark and his co-authors is questionable. (Atsumu Ohmura said that he had checked it. I introduced examination of cloud data by J. Norris (e.g. a book chapter listed as "Norris 2008" in ). On the other hand, Ohmura and his student recently found that the mechanism may be workable in the upper troposphere (though difficult to be checked by observational data). In the uppermost part of the troposphere, aerosol particles are generally few, and humidity is near saturation. Small nuclei can be effective there because small cloud particles can compete large particles if they are ice crystals rather than liquid droplets. Also, according to observational studies in Russia that Kimiaki Masuda (STE scientist) [He is that K. Masuda that Yamaguchi et al. referred to; that is not me] introduced, cosmic ray intensities in the lower troposphere does not have high correlation with solar activity, but those in the upper troposphere does. The major constituents of secondary cosmic rays in the upper troposphere are neutrons, unlike that they are muons in the lower troposphere.

There may be other pathways involving the solar magnetic field, such as those involving the global circuit of electric current.

Yukihiro Takahashi has found 27-day cycles in the energy of lightening all over the world as well as in indices of cumulus cloud activity around Indonesia. That cyclic variation is clearer in the active phase of the sunspot cycle. So he considers that it reflects the spin of the sun and it is modulated by solar activity. He has not specified the mechanism yet, but possibilities include ultraviolet radiation, cosmic rays and the electric circuit.

I made a question whether nitrogen oxides may have an important role. Kiyotaka Shibata (an atmospheric modeler strong in radiative processes) gave answer that it may be possible in the middle atmosphere, but there have been much less intensive studies about nitric acid than about sulfuric acid. Later Yutaka Matsumi (the director of the STE Lab and a middle atmosphere chemist) explained to me that nitric acid is usually not ready to condensate.

My crude summary is as follows.

Solar variability is probably the most important external forcing of climate change in the time frame of several hundred years until the former half of the 20th Century. As for the latter half of the 20th Century as well as for the outlook of the 21st Century, it is not the factor that can explain the major part of the changes of global mean temperature, but it has a minor role, and our understanding about its relative weight may change according to studies to come. The pathway of influence may include ultraviolet radiation, cosmic rays, and other not-yet-formulated mechanisms involving magnetic fields, and we cannot decide which is more important yet. If the pathway via cosmic rays is important, it is likely to involve upper clouds rather than lower clouds as Svensmark speculates. We should look at spatio-temporal distributions of climatic variablies, rather than focussing at time series of global mean values or those of variables at single locations.