Guest post by Javier who has a PhD in molecular biology and writes this Spanish blog.
1. Solar activity variability and its proxies
The Sun is a variable star and periodically changes its activity levels producing variations in radiation emission, magnetic field intensity, magnetic polarity, particle emissions, and surface convection. These changes affect the Earth in several ways that manifest through auroras, magnetic storms, changes in galactic (GCR) and solar cosmic rays, and a generally agreed small climate effect. Solar variability is included in some coupled general circulation climate models.
We have been observing the Sun with modern instrumentation for only a century. Prior to that solar variability can be inferred through proxies. Sunspot numbers have been reconstructed for the past 400 years, and before that we have auroras catalogues that extend with less reliability for a few centuries more. To reconstruct past solar activity for the last 11,000 years we rely on solar modulation of cosmogenic isotopes generated by GCR in the atmosphere, mainly 14C and 10Be.
14C is quickly oxidized to 14CO2 and then enters the carbon cycle and is removed from the atmosphere within years or decades by the oceans and the plants. Every year trees capture 14CO2 proportionally to its abundance in the atmosphere and permanently fix it into cellulose in their growth rings, where it can be measured thousands of years later. 14C data becomes contaminated around 1875-1900 from increasing human CO2 emissions and from 1950 due to atmospheric nuclear bomb tests. 10Be is deposited from the atmosphere to the surface both through a precipitation dependent and a precipitation independent pathway. Once it falls on polar snow, it can be recovered from ice cores whose layers can be dated, and measured. As both isotopes follow independent climate pathways, modern reconstructions based on their coincidence, beryllium deposition models, carbon budget models and geomagnetic reconstructions are believed to represent essentially solar modulation and are internally coherent between both isotopes.
2. Solar variability periodicities
Analysis of periodicities present in the proxy record of solar activity provides the means to study the possible existence of a significant response from Earth’s climate to Sun’s variability. Fourier analysis of solar modulation proxy records has been performed by multiple authors with similar results. Figure 1 shows the analysis performed by McCracken et al., 2013, on 10Be, 14C, and a solar modulation reconstruction from both isotopes carried out by Steinhilber et al., 2012. The main periodicities found in the three cases are shown in the table in figure 2. They are:
The 11-year Schwabe cycle (not shown in the analysis) displays a high variability from 8 to 15 years. Probably due to this variability from its most basic cycle, all solar periodicities are also highly variable in their duration. The duration of the Schwabe cycle is anticorrelated to its amplitude (Waldmeier’s rule)
Figure 1. The Fourier amplitude spectrum for GRIP 10Be, the modelled estimate of the 14C production rate and the modulation function (in MeV) computed from the EDML and GRIP 10Be data, and the INTCAL09 14C record with the periods of the stronger spectral peaks annotated in years. Left panel, periodicities 200 < T < 3000 years for the interval 9400 – 300 BP. Right panel, for the interval 4700 – 3500 BP, which does not contain Grand Minima. Notice the absence of the 208-year period on the right panel. Source: McCracken et al., 2013.
The 65-year periodicity seems to have lows at ~1905 and ~1970 and could thus be related to the ~ 60-year periodicity found in many climate phenomena, like temperatures and AMO.
The 87-year Gleissberg cycle also displays a high variability, and has been confirmed independently in aurora records since the fifth century AD (Feynman & Fougere, 1984), and in large solar proton events analysis (McCracken et al., 2001). The Gleissberg cycle is reported to affect the hydroclimate as this periodicity is frequently found in lacustrine varved sediments and flood records (Czymzik, et al. 2016). In Central Europe, flooding events appear to increase at the lows of the Gleissberg cycle.
Figure 2. Periodicities found in solar activity proxy data. Those squared in red are discussed here. The ~ 2500-year cycle will be named here the Bray cycle to honor its discoverer. Source: McCracken et al., 2013.
The 105-year periodicity is often mistakenly reported as being part of the Gleissberg cycle. The low in this periodicity of ~ 1910 would correspond to a present low.
The 208-year de Vries cycle is only apparent in cosmogenic records as prominent peaks of isotope generation with that spacing during one thousand-year period windows centered 2500 years apart. Despite its intermittent presence it is the dominant cycle in cosmogenic records and has been traced back in beryllium ice core records to the glacial period 50 kyr ago (Wagner et al. 2001). The de Vries cycle was discovered as a correspondence between high 14C records and narrow tree rings. As tree ring widths respond to both temperatures and precipitations that also change regionally, the correspondence is stronger in certain areas, like Central Europe or the Tibetan Plateau (Raspopov et al. 2008).
The 350, 510, and 708-year periodicities will not be discussed as they are seldom present in the literature.
The 976-year Eddy cycle, also referred as the millennial cycle, is one of the most constant solar periodicities reported. The Eddy cycle most clear climate correspondence is with several peaks in the Bond series of ice rafted debris that reflect cold periods in the North Atlantic area (figure 3)
Figure 3. Correspondence of the 12,000 years smoothed and detrended record of 14C with the averaged stacked ice rafted debris in North Atlantic sediments that is a proxy for iceberg discharges. Both series have peaks corresponding to the ~ 1000-year periodicity known as the Eddy cycle, most prominently during the first half of the Holocene, when this periodicity was dominant. Source: Bond et al., 2001.
The 1126-year periodicity has been reported by Humlum et al., 2011, in the GISP2 ice core record of Greenland temperatures for the past 4000 years.
The 1300, and 1770-year periodicities will not be discussed as they are seldom present in the literature.
The last clear Holocene periodicity is the 2500-year cycle. This cycle was identified by J. Roger Bray in 1968 from a consilience of geophysical, biological, and glaciological evidence contrasted with solar activity reconstructed from sunspot naked eye observations and aurora records. This proposed solar cycle was later confirmed by Sonnet and Damon by spectral analysis of the 14C record when it became available in the late 80’s and named Hallstattzei (later Hallstatt) for a late Bronze-early Iron cultural transition in an Austrian archeological site during the cycle’s previous to last minimum, 2800 years ago. However the name breaks the tradition of naming solar cycles after their discoverer, and so I propose the name to be changed to the Bray cycle. The lows in the Bray cycle coincide with the coldest periods of the Holocene, that are marked in its latest instances by major glacial readvances:
0.4 kyr BP. Little Ice Age
2.8 kyr BP. Sub-Atlantic Minimum. Greek Dark Ages.
5.3 kyr BP. Mid-Holocene Transition. Ötzi buried in ice. Start of Neoglacial period.
7.8 kyr BP. Boreal/Atlantic transition and precipitation change.
10.3 kyr BP. Early Holocene cooling event (Björck et al. 2001)
12.8 kyr BP. Younger Dryas cooling onset.
Figure 4. Variation in 14C after removal of the long-term trend. An oscillation of the ~ 2500-year Bray cycle is superposed on the data to indicate times when periods of very low solar activity would be expected to occur (arrows). As with every solar cycle, there is some variability in the spacing that complicates mathematical analysis. Adapted to show correct cycle length from Clilverd et al., 2003.
The ~ 2500-year Bray cycle besides being detected in temperature proxy records, vegetation changes and glacial expansions, has also been identified in polar atmospheric circulation changes by O’Brien et al., 1995.
3. Simple model of solar variability
Based on the information about solar periodicities from cosmogenic isotopes and the sunspot record reconstructed since 1610 a simple model can be proposed to try to explain the periodicities and known changes in activity. Figure 5 illustrates this model. It is proposed that solar activity during the period between 1910 and 2016 constitutes the basic unit composed of 9-10 Schwabe cycles (letters A or B in figure 5) and a duration of 100-110 years. This basic unit is divided in two subunits of 50-60 years (numbers 1-4 in figure 5) by one or more Schwabe cycles of lower activity in its center.
Figure 5. Model of solar variability. The basic unit (red letters) is composed of 9 or 10 Schwabe cycles of 8-15 years for a duration of 100-110 years and is divided into two subunits of 50-60 years (red numbers) by one to three Schwabe cycles of reduced activity at its center. The Gleissberg cycle (blue) is proposed to take place between deep minima (blue circles) that tend to occur near the end and beginning of solar units. The de Vries cycle is proposed to manifest by solar activity reduction of the fourth subunit of a double unit.
The Gleissberg cycle detected in auroras and large solar proton events is tentatively stablished between deep minima like the ones that took place in 1913 between cycles 14-15 and 2009 between cycles 23-24. These deep minima constitute also minima in both auroras and large solar proton events, and sometimes more than one of these deep minima are produced sequentially like in 1900 and 1913.
Two units of 100-110 years constitute a double unit. The last subunits of double units (number 4 in figure 5) are proposed to be responsible for the de Vries cycles, with an average spacing of 208 years, when the activity of the Schwabe cycles of that subunit is greatly reduced by the modulating effect of the ~ 2500-year Bray cycle, as it happened in 1650-1700 (Maunder Minimum) and 1875-1910 (Modern Minimum).
It is then proposed that the ~ 2500-year Bray cycle acts exclusively by greatly reducing the activity of every fourth subunit of any double unit close to the lows of the cycle (red filled dots in figure 6), while leaving unaffected double units close to the highs of the cycle (red empty dots in figure 6). The combination of the de Vries and Bray cycles would then be responsible for the Maunder and Spører grand solar minima, as well as the Homer, Greek and Oort grand solar minima. The ~ 1000-year Eddy cycle appears to reduce the activity of Schwabe cycles at different times of the solar unit and is proposed to be responsible for the Wolf and Roman grand solar minima, but appears also to reduce the solar activity of the Schwabe cycles between solar units, as in the Dalton grand minimum, and probably contributing to reduce the activity during the Greek, Spører and Maunder grand minima.
Figure 6. Adjustment of the solar variability model (red curve) to the sunspots groups number for the past 400 years and the solar activity reconstructed by Steinhilber et al., 2012, for the past 3000 years. The ~ 1000-year Eddy cycle is shown as a pink sinusoidal curve. The ~ 2500-year Bray cycle is shown as a yellow sinusoidal curve with the active phase as solid line and the inactive as dashed line. The ~ 208-year de Vries cycle is shown as red filled circles during the active phase, and as red empty circles during the inactive phase. The ~ 87-year Gleissberg cycle is shown in blue. Grand solar minima are named in black, warm periods in red and cold periods in blue. Known colder periods from temperature reconstructions are highlighted in turquoise. A quiet Sun mode during grand solar minima has been proposed by several authors and shown as a black dashed line.
Figure 6 presents an attempt to match the reconstruction of solar variability during the past 3000 years by Steinhilber et al., 2012, with the simple model. Although probably wrong in many details the adjustment shows that many of the past variations in solar activity might follow simple rules determined by the periodicities found in frequency analyses.
4. On the relationship between solar variability and climate
It is clear from the examination of the relationship between solar variability and climate change during the Holocene, that the longer the solar cycle the more significant it is the worsening of the climate change experimented at its lows. Thus the lows of the ~ 2500-year Bray cycle can correspond to very significant glacier readvances and near hemispheric vegetation changes. Lows in the ~ 1000-year Eddy cycle correspond to significantly increased iceberg discharges in the North Atlantic. Lows in the 208-year de Vries cycle significantly affect tree growth in ample regions, like Central Europe and the Tibetan Plateau. The de Vries and Gleissberg cycles are also proposed to affect the intensity of the summer monsoon (Fleitmann et al., 2003; Dykoski et al., 2005) and regional precipitation patterns. When we get down to the Schwabe cycle this decreasing effect becomes so low as to almost disappear within natural variability.
So to understand the effect of solar variability on climate change it is important to recognize that it requires a very long time to accumulate. The longer the time solar activity is reduced the more profound the effect on climate. The few years of a solar minimum between Schwabe cycles do not produce enough activity reduction to affect climate significantly.
To illustrate this concept I have chosen a figure from Svalgaard & Schatten, 2016, where a black box was traced by Dr. Svalggard in one of his presentations to point that “Solar activity reached the same levels in each of the last 4 centuries” (figure 7). If we mark those high levels with red boxes and periods with lower levels with blue boxes of height proportional to the missing activity we observe that high activity level red boxes tend to correspond to periods of temperature increase and lower activity blue boxes tend to correspond to periods when temperatures don’t increase. So it is clear that the recovery from LIA has taken place mainly during periods of solar activity similar to 20th century levels. This increase in temperatures was, according to Moberg et al., 2005, of 0.6°K for the Northern Hemisphere land between 1675 and 1950, much higher than the assumed 0.1°K that solar variability can contribute according to current understanding.
Figure 7. The recovery from the Little Ice Age in terms of solar activity illustrates the possible long term incremental effect of solar variability. Although every century has periods of high (or normal) solar activity, labelled as red boxes, there is a clear tendency towards a decrease in the periods of reduced activity (blue boxes). The result could play like a staircase where the red steps increase temperature (violet curve with the low-frequency component in blue) and the blue steps decrease or stall temperatures. If temperatures have not reached equilibrium, every high activity period increases them further, and the low-frequency component (blue curve) shows two steps up with the third one being modern warming (not available for the proxies used). Sources: Svalgaard & Schatten, 2016; Moberg et al., 2005.
However it has to be remembered that it is the climate who decides the response to the solar radiation and not the Sun. The planet received essentially the same amount of energy from the Sun during the Last Glacial Maximum and during the Holocene Climate Optimum, yet the climate was completely different. And it is not only the different distribution of that energy according to latitude and season, because we know that 65°N summer insolation anomaly was +36 W/m2 105 kyr ago versus only +30 W/m2 10 kyr ago, and in the first case the planet was in glacial conditions, while in the second it was in interglacial conditions and we do not know why. So if the planet is capable of giving a very different climate response to the same amount of solar energy and latitudinal insolation we should not discard heedlessly the possibility that small prolonged changes in solar output are also capable of eliciting significant climate responses.
It is also unwarranted to assume that the natural warming post-LIA ended in 1950, in the middle of a high solar activity period, as we have no evidence for that. Surely the post-LIA natural warming has to have an end, but we don’t know where to place it, and therefore we don’t know if or how much it has contributed to modern warming since 1975.
If we were to assume that indeed post-LIA natural warming has continued at about the same rate, 0.6°K in 275 years for northern hemisphere land, then in the last 66 years it should have contributed about 0.15°K to the northern hemisphere land temperatures. But remember that this is over an average of high and low solar activity periods, and those last 66 years correspond only to a high solar activity period, so it could be even more. And if we subtract from the average warming the cooling from vulcanism, that was more intense during LIA, well, who knows, perhaps half of the recent global warming or more could be attributed to the Sun without the need to invoke an out of the ordinary solar activity for the 20th century. Of course this is based on the assumption that climate responds very slowly to small changes in solar forcing and that equilibrium has not been reached since the cooling from the LIA ended.
5. On how solar variability could affect climate
Such a complex issue as the mechanisms by which solar variability affects climate has been hotly debated for decades and cannot be illuminated here. I will try however to give a reasoned opinion based on current literature. It is clear to me that solar modulation of galactic cosmic rays (GCR) cannot be such mechanism. We tend to forget that the main modulation of GCR is done by the Earth’s geomagnetic dipole that over the course of a few thousand years experiments a variation that causes changes in GCR an order of magnitude bigger than heliospheric magnetic field changes. As the climate does not change with the Earth’s geomagnetic dipole we can discard a significant magnetic or GCR effect on climate. There might be an effect, but not significant.
The climatic effect is also probably not due to changes in radiation reaching the surface, as we know they are very small. More reasonable hypotheses are those that propose an atmospheric effect that probably propagates from the stratosphere downwards to the troposphere. The energy for climate and weather manifestations is mainly provided by the poleward heat transport, that is due to more energy entering the Earth in the tropical areas that can leave, so the surplus is transported towards the poles where more energy is leaving the Earth than entering. The intensity of the poleward heat transport depends on the latitudinal thermal gradient (LTG). LTG is more intense during cold periods and less intense during warm periods for obvious reasons. And this is why predictions about weather extremes fail miserably, as with the warming LTG decreases and the weather becomes tamer, not wilder. About two thirds of the poleward heat transport are moved by the atmosphere, and the rest by the global oceanic currents. LTG also determines the position and extension of the Inter Tropical Convergence Zone (ITCZ) on which the intensity of the monsoon depends. Recent research has shown that tropical stratosphere temperatures depend on solar activity (Gray et al., 2010), and winter climate in the Northern Hemisphere correlates to those changes (Ineson et al., 2011; Gray et al., 2013). An atmospheric top down effect of solar variability has been proposed.
Figure 8. Correspondence between residual changes in 14C (IntCal04, broken line) and varve thickness in an Eastern Finland lake sediment (solid line). Grand minima in solar activity are highlighted with their name. MCA, Medieval Climate Anomaly. Varve thickness is regulated in Lake Lehmilampi mainly by organic matter accumulation changes that correlate with seasonal temperature and precipitation changes. Source: Haltia-Hovi et al., 2007.
A variety of studies are linking solar activity to both precipitation anomalies and pressure anomalies, like an increase in winter blocking of anticyclonic conditions in the North Atlantic that favor cold conditions over Central Europe during solar activity minima (Barriopedro et al., 2008; Woollings et al., 2010). These changes, that are regional in nature and barely above natural variability in intensity during regular solar minima appear to intensify and produce a significant climate change when their action spans over prolonged periods of low solar activity, and could reflect solar induced small continuous alteration of the atmospheric LTG that could affect global climate through the integration of regional changes. There is ample energy in the system to produce the observed changes if solar variability alters its distribution. There is no need for more variability in the Sun’s output to explain the correlation between solar variability and climate change.
6. Future projections of the solar variability model
If the present analysis turns out to be correct in general terms, then we can extend the variability model towards the future, with the caveat that several periodicities in solar variability have not been analyzed nor included in the model.
At present (2016) we are finishing a ~ 100-year cycle in solar activity that constitutes the first unit of a doublet (figures 5 and 6) that should end around 2022. For the next 55 years we could expect a repetition of the past 50 years backwards. Solar cycle 25 should be like SC24, SC26 like SC23, SC27 like SC22, and SC28 like SC21. Around 2065 a mid-cycle reduction in activity like the one that took place around 1970 (SC 20) should be expected (SC 29). From 2077 to ~ 2125 solar activity corresponds to a minimum of the ~208 -year de Vries cycle, but by then the de Vries cycle would have waned due to the ~ 2500-year Bray cycle being in its upper half, so no significant reduction in solar activity is expected either. So for the major part of the next hundred years the Sun should enjoy a similar level of activity to the second half of the 20th century. A grand solar minima or even a long period of reduced solar activity might not take place for a few centuries, as the long cycles are in a very favorable disposition. We might face a solar future similar to what the Romans enjoyed between 150 BC and 350 AD. A very positive development if true.
In terms of climate, the post-LIA natural warming should end soon if it hasn’t ended already, because these recovery periods from the lows in the ~ 2500-year Bray cycle rarely last for more than 400 years. As solar centennial activity cannot increase further, we are probably contemplating an end to natural warming.
But the temperature of the planet is set by the obliquity (tilt) of its axis, not by solar variability, and Earth’s obliquity is decreasing everyday at the fastest rate in forty thousand years (figure 9). Only once in the last million years have temperatures failed to fall with falling obliquity, but when this happened (during MIS 11, 400 kyr ago), 65°N summer insolation was very high, while now it is very low.
Figure 9. Holocene climate reconstruction. Major palinological subdivisions of the Holocene (names on top) match a 2500 regular spacing (grey arches on top). The global temperature reconstruction (black curve; Marcott et al., 2013 by the differencing method with proxy published dates) has been rescaled in temperature anomaly to match biological and glaciological evidence and instrumentation temperature measurements, resulting in the Holocene Climate Optimum being about 1.2°K warmer than LIA. The general temperature trend of the Holocene follows the Earth’s axis obliquity (purple), and significant downside deviations generally match the lows of the ~ 2500-year Bray cycle of solar activity (grey boxes labelled B-1 to B-5). Significant negative climate deviations manifested also by global glacier advances (blue bars; Mayewski et al., 2004) and strong increases in iceberg detrital discharges (red curve, inverted; Bond et al., 2001) generally agree well with the lows in the ~ 2500-year Bray cycle and ~ 1000-year Eddy cycle (not shown) of solar activity.
The rebound effect of the post-LIA natural warming together with any anthropogenic warming contribution has probably taken us above the temperature that corresponds to the present inclination of the axial tilt, since by a variety of biological and physical methods our present state appears to be similar, at least in the northern hemisphere, to the conditions around the Mid-Holocene transition ~ 5000 years ago. It is possible therefore that over the next 1-2 centuries global temperatures might naturally descend by 0.2-0.4°K to reach our obliquity target, regardless of solar activity.
I would expect significant cooling and climate change from reduced solar activity around 2600 AD when the next low in the ~ 1000-year Eddy cycle is expected. Glacial inception, the onset of the next glacial period, could take place around 4000 AD when the next low in the ~ 2500-year Bray cycle acting on a much lower obliquity should set the conditions for the growth of the ice sheets.
7. Summary and conclusions.
Solar variability is not random, but follows a series of periodicities of unknown cause. Periods of low solar activity, specially those belonging to the long periodicities, with a span of several decades, coincide with known periods of significant climate worsening for the last 11,000 years. The temperature increase between 1675 and 1950 has coincided with a progressive reduction in multi-decade periods of low solar activity. A simple model of double solar activity units of 100-110 years can be adjusted to fit several periodicities and solar activity reconstructions. The model infers that the ~ 2500-year Bray cycle and the ~ 1000-year Eddy cycle have different modes of action. The Bray cycle acts through the ~ 208-year de Vries cycle. In other words, the de Vries cycle is a manifestation of the Bray cycle. The model projects that solar activity during the 21st century should be similar to solar activity enjoyed during the second half of the 20th century. Please use sunscreen when outdoors for your skin protection.
Barriopedro, D., et al. (2008). Solar modulation of Northern Hemisphere winter blocking. Journal of Geophysical Research: Atmospheres, 113(D14).
Björck, S., et al. (2001). High-resolution analyses of an early Holocene climate event may imply decreased solar forcing as an important climate trigger. Geology, 29(12), 1107-1110.
Bond, G., et al. (2001). Persistent solar influence on North Atlantic climate during the Holocene. Science, 294(5549), 2130-2136.
Bray, J. R. (1968). Glaciation and solar activity since the fifth century BC and the solar cycle. Nature, 220, 672-674.
Clilverd, M. A., et al. (2003). Solar activity levels in 2100. Astronomy & Geophysics, 44(5), 5-20.
Czymzik, M., et al. (2016). Solar modulation of flood frequency in central Europe during spring and summer on interannual to multi-centennial timescales. Climate of the Past, 12(3), 799-805.
Dykoski, C. A., et al. (2005). A high-resolution, absolute-dated Holocene and deglacial Asian monsoon record from Dongge Cave, China. Earth and Planetary Science Letters, 233(1), 71-86.
Feynman, J., & Fougere, P. F. (1984). Eighty‐eight year periodicity in solar‐terrestrial phenomena confirmed. Journal of Geophysical Research: Space Physics, 89(A5), 3023-3027.
Fleitmann, D., et al. (2003). Holocene forcing of the Indian monsoon recorded in a stalagmite from southern Oman. Science, 300(5626), 1737-1739.
Gray, L. J., et al. (2010). Solar influences on climate. Reviews of Geophysics, 48(4).
Gray, L. J., et al. (2013). A lagged response to the 11 year solar cycle in observed winter Atlantic/European weather patterns. Journal of Geophysical Research: Atmospheres, 118(24).
Haltia-Hovi, E., et al. (2007). A 2000-year record of solar forcing on varved lake sediment in eastern Finland. Quaternary Science Reviews, 26(5), 678-689.
Humlum, O., et al. (2011). Identifying natural contributions to late Holocene climate change. Global and Planetary Change, 79(1), 145-156.
Ineson, S., et al. (2011). Solar forcing of winter climate variability in the Northern Hemisphere. Nature Geoscience, 4(11), 753-757.
Marcott, S. A., et al. (2013). A reconstruction of regional and global temperature for the past 11,300 years. science, 339(6124), 1198-1201.
Mayewski, P. A., et al. (2004). Holocene climate variability. Quaternary research, 62(3), 243-255.
McCracken, K. G., et al. (2001) “Solar cosmic ray events for the period 1561–1994: 2. The Gleissberg periodicity.” J. Geophys. Res 106.21: 599-21.
McCracken, K. G., et al. (2013) “A phenomenological study of the cosmic ray variations over the past 9400 years, and their implications regarding solar activity and the solar dynamo.” Solar Physics 286.2: 609-627.
Moberg, A., et al. (2005). Highly variable Northern Hemisphere temperatures reconstructed from low-and high-resolution proxy data. Nature, 433(7026), 613-617.
O’Brien, S. R., et al. (1995). Complexity of Holocene Climate as Reconstructed from a Greenland Ice Core. Science, 270(5244), 1962-1964.
Raspopov, O. M., et al. (2008). The influence of the de Vries (∼ 200-year) solar cycle on climate variations: Results from the Central Asian Mountains and their global link. Palaeogeography, Palaeoclimatology, Palaeoecology, 259(1), 6-16.
Steinhilber, F., et al. (2012). 9,400 years of cosmic radiation and solar activity from ice cores and tree rings. Proceedings of the National Academy of Sciences, 109(16), 5967-5971.
Svalgaard, L., & Schatten, K. H. (2016). Reconstruction of the sunspot group number: the backbone method. Solar Physics in press. arXiv preprint arXiv:1506.00755.
Wagner, G., et al. (2001). Presence of the Solar de Vries Cycle (~ 205 years) during the Last Ice Age. Geophysical Research Letters, 28, 303-306.
Woollings, T., et al. (2010). Enhanced signature of solar variability in Eurasian winter climate. Geophys. Res. Lett, 37, L20805.