Solar activity – climate relations: A critical Review.P. Stauning

Danish Meteorological Institute, Copenhagen, Denmark, (pst@dmi.dk / Phone: + 45 39157473)

Abstract. The presentation of solar activity-climate relations is extended with the most recent sunspot and global temperature data series. The extension of data series shows clearly that the changes in terrestrial temperatures are related to sources different from solar activity after ~1985. Based on analyses of data series for the years 1850-1985 it is demonstrated that, apart from an interval of positive deviation followed by a similar negative excursion in Earth’s temperatures between ~1923 and 1965, there is a strong correlation between solar activity and terrestrial temperatures delayed by 3 years.

A regression analysis between solar activity represented by the cycle-average sunspot number, SSNA, and global temperature anomalies, ΔTA , averaged over the same interval lengths, but delayed by 3 years, provides the relation ΔTA ~ 0.009 (±.002) · SSNA . Since the largest ever observed SSNA is ~90 (in 1954-1965), and the smallest possible value is ~ 0, then the total solar activity-related changes in global temperatures could amounts to no more than ±0.4°C over the past ~400 years where the sunspots have been recorded.

The shortcomings of the solar-cycle length model and of the cosmic radiation-cloud model are discussed. It is suggested that the in-cycle variations and also the longer term variations in global temperatures over the examined 135 years are mainly caused by corresponding changes in the total solar irradiance level representing the energy output from the core, but further modulated by varying energy transmission properties in the active outer regions of the Sun.

Question: Are suggested solar activity – climate relations real ?Fig. 1 spanning 1850 to 2010 presents recently updated global land/sea-surface temperature data (upper field) to represent Earth’s climate and sunspot numbers (lower field) to represent solar activity. The blue temperature and sunspot curves presents yearly averages. The red curves connect points representing min-to-min (square dots) and max-to-max (asterisks) average values over complete solar cycles. The straight lines represent coarse (subjective) trends.

From the presentation in Fig.1 we may conclude that the development in global temperatures after ~1985 is controlled by drivers other than solar activity represented by the sunspot number. Hence, we should exclude these years from the present analysis, which is not considering resent possible anthropogenic global warming scenarios.

The main obstacle for a credible solar activity-climate relation is the causality problem since the sunspots lag the global temperatures by ~15 years. In many past publications the problem has been disguised, among other, by the use of excessive smoothing.

Reid’s approach to solving the causality problem“Solar variability and its implication for the human environment” by Reid (1999) is one of the key papers on the relation between solar activity characterized by the sunspot numbers and Earth’s climate characterized by global temperatures at land and in the oceans.

Fig.3 Sunspot numbers and ocean temperatures. (from Reid, 1999)

At a first view the figures appear very convincing. The oceanic temperatures are very consistent in all three major ocean basins and also consistent with the global sea-surface temperatures (SST).

All time series are subjected to corresponding 11-years running averages. The smoothed curves for sunspot numbers and temperatures are exceptionally alike.

The major rise in solar activity starts before year 1900, well ahead of the rise in temperature starting just after 1900, to reach maximum amplitude (now at ~1945) before the middle of the century well before the global temperatures reach their maximum level (now ~1955).

However, the polynomial fitting “adjustments” serve to disguise – and not explain – the causality problem.

Relations between solar cycle length and climate.Instead of using the sunspot number as an indicator of solar activity, it was suggested by E. Friis-Christensen and K. Lassen in 1991 to use the length of the solar cycle as a parameter to measure the solar feature of importance for the changes in Earth’s climate characterized by the terrestrial temperatures. The correlation between sunspot no. and cycle length is shown in Fig. 4 at different delays (cycle shifts).

Fig. 5 (a) Sunspot numbers (+) and NH temperatures (*), (b) solar cycle length (+) (inverted) and NH temperatures (*) (from Friis-Christensen and Lassen, 1991, FCL91).

From Fig. 5a (left) it is clear that the rise during 1930-1960 in solar activity as characterized by the cycle-average sunspot numbers occurs well after the rise in temperatures during 1910 (1890) to 1940. There is, it seems, a delay of 15-20 years between the two otherwise similarly-looking characteristic features, which should exclude the temperature rise from being caused by solar activity.

However a way around this problem was found by using the cycle length instead of sunspot number as a parameter to characterize solar activity giving a ~0.5 cycle shift. The Gleisberg (1-2-2-2-1) smoothed cycle length used in the display in Fig. 5b (right) gave another ~0.5 cycle shift to provide a resulting nice fit. It is still unclear even now, two decades later, which physical parameter should be related to the cycle length.

Conclusions. Several of the previously published reports on postulated close relations between solar activity and Earth’s climate are based on clever data manipulation and neglect basic causality principles in the relations between solar activity and Earth’s climate as well as the finer details in the modulation of global temperatures with respect to cyclic solar activity variations.

• We suggest a steady variation in cycle-average global temperatures, ΔTA, = (0.009 ± 0.002)·SSNA with solar activity represented by the cycle-average sunspot number, SSNA. Further, we suggest that the solar activity-related global temperature variations are caused by variations in total solar irradiance of ~1 W/m2 over the solar cycle and 2-3 W/m2 in the longer run during the past ~150 years.

• During slowly varying conditions the cycle-averaged sunspot number provides a fair representation of solar energy output while during strong variations internal energy transmission properties modulate the solar energy output rate.

• After ~1985 the possible solar activity-related global temperature variations (0.009 deg/ssn) are in the wrong direction to explain recent global temperature enhancements.Reference: Stauning, P., Solar activity-climate relations: A different approach, J. Atm. Solar-Terr. Phys. 73, 14 pp., 2011. doi:10.1016/j.jastp.2011.06.011.

Fig. 6 displays a replot of FCL91-Fig.2 with recent temperature and solar cycle length (SCL) data included. The blue curve represent NH surface temperatures averaged over solar cycles. The black curve connects recently estimated min-to-min and max-to-max cycle lengths, while the red curve presents 1-2-2-2-1 averages of SCL.

The diversion of the temperature and SCL curves after ~1980 is evident here in contrast to the coincident final upturn in FCL91 (marked by an ellipse in Fig.5b), which was based on incomplete data (e.g. Laut and Gundermann (2000), Damon and Laut (2004)).

Fig.1

Fig.5a Fig.5b

Fig.6

Fig.3

Fig.4

Fig.8

The plot in Fig.15 presents the time-history of the deviations of the global temperature anomaly from the regression line (0.0090 deg/ssn) defined in Fig. 1. Note in the figure that the positive and negative excursions during cycles 16-19 are about equal and each has the Hale cycle length (22 years for a complete solar magnetic cycle).

In-cycle sunspot and temperature variations.Fig. 14 presents the superpositions of the yearly averages of sunspot numbers in the lower part and global temperatures in the upper part of the figure. All data are plotted relative to the mid-cycle year. The heavy red lines denote averages over cycles 10 to 21 of the sunspots and the temperature anomalies, respectively. The blue lines marked by dots display the sunspot numbers and the temperature anomalies during the largest ever recorded solar cycle #19 (1954-1965).

Discussions.From the regression displayed in Fig.1 it appears that there is a strong correlation between the global temperatures and the sunspot numbers. However, during the course of the individual cycles, as presented in Figures 10 and 14, there is little indication of a large cyclic variations in temperatures from sunspot minima to maxima during the cycle. Hence, the temperature variations could not easily be coupled to parameters that vary in concordance with the sunspot numbers without imposing excessive smoothing related, for instance, to the inertia in the global system.

However, such inertia would also have the effect of causing an added delay between the long-term changes in the solar activity characterized by the average sunspot number and Earth’s climate characterized by the global temperatures. Thereby, the causality problem in the displacement of the peak in sunspots to around 1958, i.e. ~15 years after the peak in global temperatures at ~1943 would be further exaggerated.

Here we suggest that the major part of the long-term variations in global temperatures observed over the 135 years from 1850 to 1985 are caused by corresponding long-term changes in the total energy output from the Sun, mostly through variations in the total solar irradiance (TSI).

It is further suggested that the cycle-average sunspot number is a fair indicator of the solar energy output level for slowly varying changes such as those seen in cycles 9 to 15 and 20 to 21. During faster changes in the solar energy output level such as those inferred from the global temperature variations during cycles 16 to 19 the solar energy output is possibly modulated by varying energy transmission properties in the deep solar convection region.

Fig.16

Fig.2

Fig. 2 displays cycle-average temperature anomaly values plotted against the average sunspot no. The intervals of temperature averaging have the same length as the interval used for sunspot averaging, but displaced 3 years (after sunspots). A least squares regression line, ΔTA (°C) = 0.0090 · SSNA - 0.70 has been calculated and plotted in the diagram. There is a high correlation (R = 0.77) , but the causality problem documented in Fig. 1 remains to be explained.

It is quite conceivable that the solar cycle length is an indicator of solar output, i.e., total solar irradiance (TSI), but the 1-2-2-2-1 smoothing applied only to solar activity, not to temperature, is inconsistent .

The Cosmic Ray – Climate theory.The effects of the galactic cosmic radiation (GCR) on the climate through its control of the cloudiness (e.g., Pudovkin and Veretenenko, 1995; Svensmark and Friis-Christensen, 1997; Svensmark, 2000) also fail to comply with the recorded temperature changes.

The alleged temperature effect relies on the assumed net decrease in the energy available near the surface of the Earth as the result of the reflection of solar energy by the additional cloud cover created through nucleation at ions generated by the cosmic radiation. The reflection of solar energy is supposed to dominate over the restraining effects on the energy balance from the excess cloud cover.

The GCR level is to some extent controlled by the combined shielding effect of the solar magnetic field extended into the interplanetary space by the solar wind and the Earth’s own magnetic field.

The cosmic ray level always return to the quiet level (Fig. 9). Hence the behaviour does not comply with the requirements to the main forcing process but might contribute to the small variations observed at the peak of the individual cycles (Fig.10).

Fig.8

The initial relation between cloud cover and depression of GCR is displayed in Fig. 7 while the correlation between cloud cover and GCR is shown in Fig. 8 (from Svensmark and Friis-Christensen, 1997)

Fig.7

Fig.9

Fig.10

The effects of the shielding by the Earth’s own magnetic field is evident in the variation in GCR level at the different stations in Fig.9.Over the past 400 years the core field has decreased considerably. The increase in the GCR level is shown in Fig. 11. Thus, according to the GCR-Climate theory, the global temperatures should have decreased since ”The little Ice age”.

Fig.11

An even more striking rejection of the theory is provided by Nigel March and Henrik Svensmark (2000). Their Fig.1 (here Fig.12) displays the cloud anomaly for high, middle, and low clouds and the GCR (red line) intensities.

Fig.12

Fig.13

The positive relation exists only for the low clouds, and the variation is only ± 0.6% (not ± 2% as in Fig.7) and only in the low clouds. Their Fig. 2a (here Fig. 13a) displays the GCR-Temperature correlation, which is close to 0 for most regions and not 0.5-0.9 as in Fig 8.

Fig.14

Fig.15

Thus, the energy transport from the core (Fig.16-1) via the radiation zone (2) and the interface layer assumed to carry the currents that generate the solar magnetic fields and further through the convection zone (3) to the photosphere (4) is a little faster during cycles 16-17 with relatively weak magnetic fields and impeded (slower) during cycles 18-19 with strong solar magnetic fields. Being related to transmission properties the positive and negative excursions should balance each other in the end like it is seen in Fig. 15. Thus the apparent violation of basic causality principles during cycles 16-19 is removed.