Global dimming and brightening: A review
Martin Wild
Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
There is increasing evidence that the amount of solar radiation incident at the Earth’s surface is not stable over the years but undergoes significant decadal variations. Here I review the evidence for these changes, their magnitude, their possible causes, their representation in climate models, and their potential implications for climate change. The various studies analyzing long-term records of surface radiation measurements suggest a widespread decrease in surface solar radiation between the 1950s and 1980s (“global dimming”), with a partial recovery more recently at many locations (“brightening”). There are also some indications for an “early brightening” in the first part of the 20th century. These variations are in line with independent long-term observations of sunshine duration, diurnal temperature range, pan evaporation, and, more recently, satellite-derived estimates, which add credibility to the existence of these changes and their larger-scale significance.
Current climate models, in general, tend to simulate these decadal variations to a much lesser degree. The origins of these variations are internal to the Earth’s atmosphere and not externally forced by the Sun. Variations are not only found under cloudy but also under cloud-free atmospheres, indicative of an anthropogenic contribution through changes in aerosol emissions governed by economic developments and air pollution regulations. The relative importance of aerosols, clouds, and aerosol-cloud interactions may differ depending on region and pollution level. Highlighted are further potential implications of dimming and brightening for climate change, which may affect global warming, the components and intensity of the hydrological cycle, the carbon cycle, and the cryosphere among other climate elements.
Received 14 November 2008; accepted 10 March 2009; published 27 June 2009.
Citation: Wild, M. (2009), Global dimming and brightening: A review,
J. Geophys. Res., 114, D00D16, doi:10.1029/2008JD011470.
I found this passage that parallels a lot of what I’ve been saying about data quality:
The assessment of the magnitude of these SSR (surface solar radiation) variations faces a number of challenges. One is related to data quality. Surface radiation networks with well-calibrated instrumentation and quality standards as those defined in BSRN [Ohmura et al., 1998] need to be maintained on a long-term basis and if possible expanded into underrepresented regions (see Figure 1b).
However in this figure, citing CRU surface temperature, he likely doesn’t understand what data quality issue might have contributed to the trend from 1960-2000
One of the effects of urbanization is the compression of the diurnal temperature variation. I recently was able to demonstrate this between two stations in Honolulu. One is in the middle of the Airport and had a sensor problem, the other was in a more “rural” setting about 4 miles away. Note how the ASOS station at the airport has an elevated temperature overall, but that the biggest difference occurs in the overnight lows, even when the ASOS sensor giving new record highs was “fixed”:
Urbanization affects Tmin more than Tmax. For example, here’s the nighttime UHI signature of Reno, NV that I drove as a measurement transect using a temperature datalogger:
Click for larger image
Even several hours after sunset, at 11:15PM, the UHI signature remained. The net result of urbanization is that it increases Tmin more than Tmax, and thus minimizes the diurnal range, which we see in Wild’s diurnal range graph.
Even the IPCC misses it:
Wild probably has no idea of this type of issue in the CRU data, but again it speaks to data quality which he seems to be keen on. He’s looking for a global solar signature in temperature data, something Basil Copeland and I have done, to the tune of much criticism. The signature is there, but small. But, when diurnal temperature variation is looked at, any solar signature is likely swamped by the urbanization signal. I’m not saying there is no solar component to what Wild is looking at, but it seems fairly clear that UHI/urbanization/land use change plays a significant role also.
Even rural stations can be affected by our modern society, as Dr. John Christy demonstrated in California’s central valley:
A two-year study of San Joaquin Valley nights found that summer nighttime low temperatures in six counties of California’s Central Valley climbed about 5.5 degrees Fahrenheit (approximately 3.0 C) between 1910 and 2003. The study’s results will be published in the “Journal of Climate.”
The study area included six California counties: Kings, Tulare, Fresno, Madera, Merced and Mariposa.
While nighttime temperatures have risen, there has been no change in summer nighttime temperatures in the adjacent Sierra Nevada mountains. Summer daytime temperatures in the six county area have actually cooled slightly since 1910. Those discrepancies, says Christy, might best be explained by looking at the effects of widespread irrigation.
Wild’s study is a very interesting and informative paper, I highly recommend reading the entire paper here (PDF 1.4 mb)
h/t and sincere thanks to Leif Svalgaard for bringing this paper to my attention.
Don Davis (19:16:17) :
So, the variation in the Earth’s distance from the Sun can be as big as +/-4.7%. This amounts to 18.8% variation in how much light we receive.
You have been misinformed [c.f. the danger of the Internet I spoke of]. The Earth and the Sun orbit their common center of mass, so the variation of the distance is just that resulting from the standard elliptical orbit. This follows from elementary physics, from careful numerical calculation, and from direct measurements of the TSI and [to stay on topic] the F10.7 radio flux that vary 7% over a year, and not 18.8%..
> The Earth and the Sun orbit their common center of mass, so the variation
> of the distance is just that resulting from the standard elliptical orbit.
My numbers were jumbled, but the Earth and the Sun both do orbit the solar
system’s barycenter. The corrected numbers are:
* Variation in radius of Earth’s eccentric orbit: +/-2.5e6 km (+/- 1.7%);
* Maximum radius of Sun’s orbit around the barycenter: 1.5e6km (+/- 1%).
These two numbers are well-known. Indeed, Oliver, in “Encyclopedia of world
climatology,” attributes the Solar-orbit radius calculation to Newton (p.256).
So, the total variation in sun-earth distance = +/-2.7%, and the consequent
variation in illumination = 10.8%.
> the F10.7 radio flux… vary 7% over a year, and not 18.8%.
The 10.8% variation in illuminance is a 20-year variation, so
the 7% annual variation in F10.7 flux doesn’t really bear on this.
Finally, eyeballing Wilds’ graph more carefully, I now see two periods between
1937 & 1987, so my eyeballed illuminance period should have been ~25 yr, not
20 yr. This, of course, does not match the 19.9yr Jupiter/Saturn synodic period.
Substantiating that claim would need a longer illuminance record than Wilds’.
Don Davis (12:05:24) :
My numbers were jumbled, but the Earth and the Sun both do orbit the solar system’s barycenter.
No they do not in the simple way you put it. The distance between the Sun and the Earth varies 3.3% year in and year out. There is no noticeable 20-year or other period in this. Often people cannot be convinced by logic or reason, so I’ll have to take to experiment. We have observed F10.7 for 60+ years and TSI for 30+ years and clearly and only see the 6.6% variation every year, no other period above the noise, except the 11-year solar cycle itself.
I know it’s late and you probably will not see this, but if you do, consider this website:
http://okc.mesonet.org
Watching regularly (I live in Oklahoma City), I can tell you that just after dusk is the premiere time to find the urban heat island, especially during the autumn. We have prevailing southerly winds, so the north end of the city tends to exhibit more urban heat effects than any other part (except sometimes downtown).
You can get actual data for the City micronet (unfortunately, the sensors are mounted on traffic signals in non-standard locations and non-standard heights) here:
http://www.mesonet.org/data/public/okc/
>> the Earth & the Sun both do orbit the solar system’s barycenter.
> No they do not in the simple way you put it. The distance between
> the Sun and the Earth varies 3.3% year in and year out.
dr. svalgaard,
i got this resolution of this dispute from an ex-nasa spacecraft-nav
physicist whom i know: the sun & the outer planets do indeed orbit
the barycenter. inside mars’ orbit. though, jupiter’s gravitational
gradient is roughly constant, so the inner planets all feel much the
same acceleration that the sun feels, towards jupiter. hence, while
the sun & the outer planets do orbit the solar system’s barycenter.,
the inner planets all behave approximately as if they were just
moons of the sun, orbiting the barycenter as an ensemble. finally,
saturn’s gravitational effect on the inner planers works just the same
as jupiter’s. so, this explains the flaw in my back-of-the-envelope
calculation.
Don Davis (14:52:12) :
the inner planets all behave approximately as if they were just
moons of the sun, orbiting the barycenter as an ensemble.
If you include the Sun in the ‘ensemble’ you are getting closer, but the explanation is still only a pseudo-explanation for the ‘unwashed masses’. The distinction between inner and outer planets is somewhat unphysical. Imagine you had only the Sun and the Earth, then there would be no problem, they orbit their common barycenter. Imagine the Sun was a double star with its companion at some distance [such as not to disrupt the Earth’s orbit]. Let the companion have the same mass as the Sun and also have a planet [with it and the star orbiting about their common barycenter]. Then the barycenter of the whole system would be halfway between the Sun and the star way outside the Earth’s orbit. Now add a Jupiter to each system, they will still orbit as ‘moons’ around their respective stars, and the barycenter of the whole system would still be halfway between the Sun and the star. Now slowly shrink the mass of the star and its planets. That would not upset the movements of the Earth and Jupiter that still move as ‘moons’ around the Sun. It would move the barycenter a bit closer to the Sun. Now also shrink the distance between the Sun and its star companion. In the end you can imagine that the star has become Jupiter, but at no time is there a point where the rules change. The rules for inner and outer planets are exactly the same. And at no time does any pair of the participants do anything else but orbit around the common barycenter of that pair. Anyway, if you can now see why your calculation was wrong, then the problem has been resolved, and we do need to look for 20-year modulations of TSI or F10.7, so all is well.
Leif Svalgaard (15:27:04) :
and we do not need to look for 20-year modulations of TSI or F10.7, so all is well
Re: Leif Svalgaard & Don Davis
Don, your explanation passes the 5-year-old comprehension-test (advocated above by Leif). Leif, yours does not (unless we restrict to a bright subset).
Towards a merger of Leif’s technical notes & Don’s narrative:
If ‘inner’ & ‘outer’ are physically-offensive categories, maybe “better” categories would be “MASSIVE & central”, “small & close”, and “big & far-out”. After all, we’re talking about weighted [by size & distance] averages.
My impression of what gets people confused when this topic arises: misleading pairwise focus.
Paul Vaughan (17:40:56) :
My impression of what gets people confused when this topic arises: misleading pairwise focus.
As usual, I am totally lost as to what you are trying to say.
And you have still not made any effort in explaining your misunderstanding. This is a very poor showing.
Paul Vaughan (17:40:56) :
Towards a merger of Leif’s technical notes & Don’s narrative
The oversimplification comes from the notion that some planets orbit the ‘solar system’ barycenter and others do not. In a sense none of them do, as the barycenter is just the mass-weighted average distance from an arbitrarily point the chosen origin of the coordinate system]. The gravitational attraction is between the Sun and the planet and that force is many orders of magnitude larger than that between a planet and any other planet, including Jupiter. So each planet and the Sun orbit their common barycenter
Leif Svalgaard (18:29:44) : continuation…
E.g. for the Earth the gravitational force between it and the Sun is Fse = Ms * Me/ De^2 = 333000 with units in Earth masses and distance in AU. Between the Earth and Jupiter on average Fje = Mj * Me / Dj^2 = 39 or 8,600 times smaller. It has nothing to do with ‘massive & central’ or ‘big & far-out’. Every body is a ‘moon’ of the Sun, inner or outer, bigger or close-in, or whatever.
Leif Svalgaard (18:41:59) : continuation…
So all the bodies move according to the combined mutual interactions. And seen from afar it is possible to calculate the center of gravity [barycenter] off all these bodies at any time and to refer the motion of each body with respect to that center, i.e. to choose that point as the origin of a coordinate system. Since forces from afar [e.g. the gravitational force of the Galaxy] acting an a collection of parts can be considered as acting on the center of mass, that point will be what is seen to move along its path in the Galaxy.
There is no inner-outer, massive-light, big-small, or what ever, divide. The solar system is a whole and the same rules work everywhere. I actually don’t see this as a mystery, and I have actually once explained it to a five-year old [number 2 here: http://www.leif.org/ ] and he had no problem with it.
Leif Svalgaard (18:29:44) “So each planet and the Sun orbit their common barycenter”
So exactly how many points do you have the sun oribiting in ‘this model’? [rhetorical question]
You’re making a lot more sense later when you speak of the “whole”.
– –
Leif Svalgaard (18:18:48) “This is a very poor showing.”
We are all volunteers. For example: I asked you a question 3.5 months ago & did not receive a reply.
– –
New Question:
What is your explanation for the 1930 spike in aa index?
Leif Svalgaard (19:19:38) “There is no inner-outer, massive-light, big-small, or what ever, divide.” / Leif Svalgaard (18:41:59) “It has nothing to do with ‘massive & central’ or ‘big & far-out’.”
You’re not making much sense here.
When teaching non-specialists at an introductory-level about weighted-averages, it is helpful to the students if I explain that a ‘heavily-weighted outlier’ has a large influence. [It’s no different in a spatial context.]
Obviously more precise measurement scales are used in calculations – if they are available …which they are in the example at hand – so the term ‘big’ used in the ‘WUWT lecture’ gets a more precise definition when the computing starts.
I’ve no doubt that the audience here realizes that the qualitative simplifications are for communication purposes.
Paul Vaughan (20:58:11) :
“This is a very poor showing.”
We are all volunteers. For example: I asked you a question 3.5 months ago & did not receive a reply.
Nonsense, you accused me of not taken into account the ‘central thesis’. I answered that I couldn’t see what that was, and asked you to tell me. You have spent an inordinate amount of volunteer time to avoid telling me. So, the question still stands.
What is your explanation for the 1930 spike in aa index?
The same as for the 1943, 1952, 1974, 1994, and 2003 spiles: very high-speed recurrent streams from large low-latitude coronal holes.
The solar wind speed for these years were:
1930: 521 km/s
1943: 517
1952: 518
1974: 527
1994: 518
2003: 535
The average solar wind speed is typically 426 km/s.
Similar spikes occurred in 1852, 1865, and 1886. They all share the distinction of occurring during the declining phase of the cycle.
Paul Vaughan (21:47:42) :
I’ve no doubt that the audience here realizes that the qualitative simplifications are for communication purposes.
I don’t think so. That this problem crops up all the time shows me that they do not understand the issue, and some take the simplifications to be real qualitative differences.
And what was her ‘central thesis’ that I should have paid attention to?
Leif Svalgaard (22:19:35) :
“The average solar wind speed is typically 426 km/s.
Similar spikes occurred in 1852, 1865, and 1886. They all share the distinction of occurring during the declining phase of the cycle.”
The Sun as a star emits by a factor of 1.5–2 more solar wind mass and energy
during solar minima in comparison with solar maxima years. Moreover, the overall
rising trend of the same order of magnitude during the past 30 years has been shown eg Veselovsky et al.2000 Richardson 1999
As this is connected with poloidal (i.e. open-field) component of solar magnetic field,whereas say CME peturbations of geomagnetic activity are at maximum a toroidal component (180 degrees out of phase) do we have a double driver? eg Ruzmaikin & Feynman (2001)
maksimovich (22:57:17) :
As this is connected with poloidal (i.e. open-field) component of solar magnetic field,whereas say CME peturbations of geomagnetic activity are at maximum a toroidal component (180 degrees out of phase) do we have a double driver? eg Ruzmaikin & Feynman (2001)
I don’t think so. These high-speed streams were feed by low-latitude coronal holes created from low-latitude active regions, so have little to do with the polar fields. There is a common misconception that these holes are ‘extensions’ of the the polar caps, that somehow droops down towards the equator. This is incorrect. The separation between poloidal and toroidal fields is somewhat arbitrary and often introduced for mathematical convenience, but has little significant physics behind it, the magnetic field is a ‘whole’.
Paul Vaughan (20:58:11) “We are all volunteers. For example: I asked you a question 3.5 months ago & did not receive a reply.”
Leif Svalgaard (22:19:35) “Nonsense”
You are mistaken. See here:
http://wattsupwiththat.com/2009/03/12/nasa-solicits-new-studie-on-the-current-solar-minimum/
[Paul Vaughan (12:05:05) Mar. 15, 2009]
– –
Leif Svalgaard (22:19:35) “[…] very high-speed recurrent streams from large low-latitude coronal holes.”
Thank you for conveying this useful information.
– –
Leif Svalgaard (22:26:01) “[…] some take the simplifications to be real qualitative differences.”
I will agree that for some it does not go beyond qualitative – and add that nonetheless scientists can (& do) try to help them establish accurate qualitative conceptions (within the existing constraints – e.g. available time).
– –
Re: maksimovich (22:57:17)
I remember turning up a presentation on this some time ago:
http://www.iono.noa.gr/cost724/Documents/WG1/Th_Dudok_de_Wit.pdf
From the presentation: “The method is simple yet the fit is better than any one produced so far”
Paul Vaughan (23:38:39) :
“We are all volunteers. For example: I asked you a question 3.5 months ago & did not receive a reply.”
You are mistaken. See here:
I would be curious to know what Dr. Svalgaard believes might be responsible for the 7.8 year signal detected in European temperature time series
I didn’t see a question mark in there and I don’t have an opinion. My ‘nonsense’ went to the “we are all volunteers” excuse. How many times have I asked you now what ‘the central thesis’ was? and how many times have you squirmed to avoid an answer? and now this lame excuse. Are you not a bit ashamed of yourself? So, tell me what ‘her central thesis’ was and how that should have changed my position vis-a-vis her papers. Looking forward to your next excuse for avoiding that issue.
Leif Svalgaard (23:59:00) “I don’t have an opinion.”
Double-standard.
Leif Svalgaard (18:23:34) :
Allan M R MacRae (16:58:30) :
BTW, some time ago I sent you a paper by Jan Veizer (GAC 2005) that in Figure 2 showed an apparently strong correlation between Cosmic Ray Flux and Low Cloud cover, (after Marsh and Svensmark 2003 and Marsh et al 2005). Did you get this paper? Do you accept or reject this correlation?
It doesn’t look too good, and employs a standard technique of persuasion, namely plotting some other quantities as well that are well correlated to ‘guide’ the eye. Svensmark notes that perhaps a better calibration of the spacecraft data would improve the correlation. The recent albedo data of Palle also do not support the correlation, as there is no clear solar signal.
_________________________________
Leif, I have re-examined Figure 2 of Veizer (2005) and cannot agree with your above comment. I tend to agree with Veizer, when he says:
Figure 2. Solar irradiance (SI), galactic cosmic ray (CR) flux and low cloud (LC) cover, 1983 – 2001 (adapted from Marsh and
Svensmark, 2003a and Marsh et al., 2005). Note the reversed scale for SI. Some authors (Laut, 2003) argue that the apparent
post-1995 divergence of clouds from celestial trends disqualifies the correlations. However, the discrepancy may arise from a
modified cross-calibration of satellites, following the late 1994 hiatus in polar orbit flights (Marsh and Svensmark, 2003a). A correction
for this drift (thick full line LC’) results in a good agreement for all parameters (see also Pallé et al., 2004b and Usoskin
et al., 2004).
_____
I have not reviewed Pallé.
We’ll see where this leads. Just because we do not understand the mechanisms whereby solar variation (or whatever) drive climate variability, does not mean that they do not exist.
Paul Vaughan (00:56:49) :
Leif Svalgaard (23:59:00) “I don’t have an opinion.”
Double-standard.
Because I have not read the paper.
So, tell me what ‘her central thesis’ was and how that should have changed my position vis-a-vis her papers.
Allan M R MacRae (02:27:07) :
Just because we do not understand the mechanisms whereby solar variation (or whatever) drive climate variability, does not mean that they do not exist.
I thought the whole point of the GCR-hypothesis was that they do understand and provide the mechanism. My criticism was of that particular mechanism and what goes for evidence for it.
Leif Svalgaard (07:15:05) “Because I have not read the paper.”
Now you are contradicting yourself.