Forcing mechanisms

The climate system generally can be described in terms of energy transport. The main energy source is the Sun, which emits electromagnetic radiation with wavelengths covering the full spectrum and peaking in the visible part (400-750 nm). On its way through the atmosphere to the Earth's surface, part of this radiation is reflected, scattered, or absorbed. The absorbed energy (~70 percent) is ultimately re-emitted into space with much longer wavelengths. Since the incoming solar radiation covers only half the globe (day side) and peaks at low latitudes, there are permanent energy gradients on the Earth's surface, which the climate system tries to eliminate by transporting energy through the atmosphere and the ocean (thermohaline circulation).

All the radiative processes depend strongly on the composition of the atmosphere (gases, aerosols, dust, clouds), which provide another level of complexity, as these components are strongly connected to chemical, thermal, and dynamical changes taking place in the atmosphere on vastly differing time-scales.

Any change in these complex processes can force the climate to change. We distinguish between external (orbital and solar) and internal (volcanic and ocean circulation) forcings which we will address in the following sections.

Orbital forcing

The amount of solar radiation arriving at a given point on the top of the atmosphere depends on the solar luminosity (the total amount of radiation emitted by the Sun) and the relative position of Sun and Earth in space (distance, direction of the Earth's rotational axis relative to the ecliptic). Whereas the former depends on processes taking place within the Sun, the latter is determined by the distortion of the Earth's orbit by the gravitational forces of the other planets (mainly Jupiter and Saturn) (Laskar et al. 2004). Orbital forcing affects three parameters with typical periodicities: the eccentricity (the deviation of the orbit from a circle, with periods of ~400 and ~100 kyr), the obliquity (the tilt angle of the Earth's axis with a period of ~40 kyr), and the precession of the Earth's axis (~20 kyr period). Whereas the eccentricity leads to changes in the mean annual distance between Sun and Earth, and therefore to changes in the total incoming radiation, the other two parameters affect only the relative distribution of the solar radiation, with implications for the energy transport within the climate system. At the time-scale of the Holocene,

Figure 6.1 Changes in orbital forcing during the Holocene (12 000 years BP to 3000 years in the future) for the months June (a) and December (b). The figures show clear trends depending on latitude. In June there is a strong decreasing trend in the north (a). In December, the insolation on the Southern Hemisphere first increases and then reaches its maximum between 2000 and 4000 cal. years BP before decreasing again (b).

A9e[kyrBP]

A9e[kyrBP]

an9e in Oe,

an9e

orbital forcing is slow. Nevertheless, the change in insolation over the past 12 000 years is considerable and cannot be neglected (cf. Jansen et al., this volume). Figure 6.1a shows the change in insolation in June for the Holocene period as a function of latitude. At high latitude the change in insolation amounts to 54 W m-2. In Figure 6.1b the change in insolation is shown for December. In this case, the insolation increases from the beginning of the Holocene and reaches a maximum between 4000 and 2000 years BP. The maximum change amounts to 37 Wm-2. Orbital forcing is the only forcing that is fully understood, and can be calculated not only for the past but also for the future several million years (Berger and Loutre 2004).

Solar activity and solar forcing

Since the Sun drives the climate system, solar variability is obviously a serious candidate for forcing climate change. This is especially so in view of the fact that the

Sun is a variable star showing considerable cyclic changes in its magnetic activity, as expressed, for example, in the sunspot number.

In the past, many attempts have been made to ascertain whether the solar radiation arriving at the top of the atmosphere at a distance of 1 astronomical unit from the Sun is constant (denoted by the solar constant or total solar irradiance, TSI). All the early attempts, however, failed due to the above-mentioned radiative atmospheric processes. Only after it became possible to mount radiometers on satellites outside the atmosphere was it discovered that the solar constant fluctuates in phase with the magnetic activity of the Sun. It is possible to divide the measured total solar irradiance (TSI) into three components: a background component, a darkening component controlled by the sunspots, and a brightening component related to the faculae, which overcompensate the negative effect of the sunspots and lead to a positive correlation between TSI and solar activity (Fröhlich and Lean 2004).

The instrumental data from almost the past three decades show that the change of the TSI over an 11-year Schwabe cycle is about 0.1 percent, corresponding to 0.25 W m-2 (the global mean value at the Earth's surface), an estimate that is quite small compared with the 3.7 W m-2 estimated for a doubling of CO2. The variability of the solar radiation, however, is strongly wavelength dependent and reaches values of more than 100 percent in the UV part of the spectrum. Such large changes in the spectral solar irradiance (SSI) strongly influence the photochemistry in the upper atmosphere, and in particular the ozone concentration. Model calculations show that through dynamical coupling SSI changes can cause shifts in the tropospheric circulation systems and therefore change the climate (Haigh and Blackburn 2006).

From a climate perspective, changes in forcings on decadal and shorter time-scales are less important, because many processes within the climate system occur on much longer time-scales (e.g. the thermohaline circulation, build up of ice-sheets). Therefore, the crucial questions are whether changes of TSI and SSI occur on centennial and millennial time-scales, and how large these changes are. These questions are still being debated. They can be answered in two steps: (i) how variable is the Sun's magnetic activity? (ii) how is this magnetic activity related to TSI and SSI? As we will show, the magnetic variability is indeed larger on longer time-scales. From the solar physics perspective, however, it is not yet clear if this is also the case for the TSI and the SSI. On the other hand, paleoclimate reconstructions provide growing evidence for larger changes in solar forcing than has been experienced during the past 30 years (Bond et al. 2001; Neff et al. 2001; Wang et al. 2005; Haltia-Hovi et al. 2007).

The longest historical record of solar activity is the sunspot record, which goes back to 1610 when the telescope was invented. It shows the well-known 11-year Schwabe cycle superimposed on a generally increasing trend from 1610 to the present, which is interrupted by distinct periods of low solar activity (the Maunder minimum of 1645-1715, and the Dalton minimum of 1795-1820).

To extend this record of solar activity beyond the era of direct observations we have to rely on indirect proxy data. Such data can be derived from measurements of the cosmogenic radionuclides 10Be and 14C in natural archives such as ice cores and tree rings (Beer et al. 1990; Stuiver et al. 1991; Muscheler et al. 2004). Cosmogenic radionuclides are produced continuously in the atmosphere as a result of the interaction of galactic cosmic rays with nitrogen and oxygen (Masarik and Beer 1999). The higher the cosmic ray intensity, the larger the production rate, and vice versa. The cosmic ray intensity is modulated by two magnetic effects: the solar activity and the geomagnetic field. Depending on the magnetic activity, the Sun emits plasma (solar wind) carrying magnetic fields into the heliosphere, which acts as a shield and reduces the cosmic ray intensity. The second shielding effect is due to the geomagnetic field, which prevents cosmic rays with too low an energy from penetrating the atmosphere and producing cosmogenic radionuclides. After production, the fate of the cosmogenic radionuclides depends on their geochem-ical properties. 10Be becomes attached to aerosols and is removed from the atmosphere mainly by wet deposition within 1-2 years. Ice cores are excellent archives to measure 10Be and to reconstruct its production rate in the past. On the other hand, 14C forms 14CO2 and exchanges between atmosphere, biosphere, and ocean. As a consequence of the large size of these reservoirs and the long residence times (ocean: 1-2 kyr), the amplitude of an observed 14C change in the atmosphere is considerably smaller than the corresponding change in the production rate and the 14C change is delayed. The ideal archive for the reconstruction of past 14C changes are tree rings.

Although the physics of the production processes in the atmosphere are well understood, the transport from the atmosphere into the archives is rather complex and has not yet been fully elucidated (Beer et al. 2002). Comparisons between 10Be and 14C records show that during the Holocene the production signal was dominant and system effects generally can be neglected in a first-order approach. This means that both 10Be and 14C provide an independent record of the cosmic ray intensity of the past. To extract the solar component from this signal the geomagnetic effect has to be removed. This can be achieved by taking into account the changes in the geomagnetic field intensity derived from archeomagnetic and paleomagnetic measurements. The result is a record of the solar modulation function O (Figure 6.2) (Vonmoos et al. 2006). 0 = 0 means a completely quiet Sun, a condition that has probably never occurred. 0 = 1000 MeV corresponds to an active Sun as typical for a solar cycle maximum during recent times. The O-record is characterized by a long-term trend superimposed on which are short-term fluctuations. These short-term fluctuations can be divided into cyclic and episodic features. The cyclic features show periodicities around 11 years (Schwabe cycle), 80 years (Gleissberg cycle), 205 years (DeVries or Suess cycle), and 2200 years (Halstatt cycle) (Stuiver and Braziunas 1993).

Most episodic features are strong negative spikes corresponding to so-called grand solar minima, periods when the Sun was very quiet as during the Maunder minimum. It should be noted that the record does not cover the past 300 years BP. The mean O value of the past five decades is about 700 MeV. This means that we are presently in a period of high solar activity, and that in the past there were periods with considerably lower solar activity. Similar conclusions were obtained by Solanki et al. (2004). Whether these large changes of activity are also reflected in

Figure 6.2 Reconstruction of the solar modulation function O covering the period 9300-340 cal. years BP. O depends on the intensity of the solar open magnetic field, which reduces the cosmic ray flux and therefore the production rate of cosmogenic radionuclides. This record is based on 10Be data from the GRIP ice core and shows short-term as well as long-term fluctuations. The average solar activity of the past few decades corresponds to a value of 700 MeV. The data have been low-pass filtered with cut-off frequencies of 1/(100 years) and 1/(500 years).

1000

10 000 8000

6000 4000 Age (cal yr BP)

2000

10 000 8000

6000 4000 Age (cal yr BP)

2000

the TSI and SSI is not known. Therefore we prefer not to speculate about the corresponding forcing in W m-2. There is clear evidence for larger changes from paleoclimatic records, as we will show below. Most authors have used the A14C record as an indicator of past solar activity. However, A14C reflects the deviation of the atmospheric 14C content relative to a standard value, so it also contains a long-term geomagnetic and a system component which perturbs the solar signal to some extent. Consequently, the 10Be flux and the 14C production rate are better proxies, although they still contain a geomagnetic component.

In Figure 6.3, the x-axis is based on tree ring counting. At 850 bc in Figure 6.3a, a rapid decrease of the radiocarbon age occurs followed by a plateau. The decrease corresponds to the A14C rise in Figure 6.3b (red data points). The blue curve in Figure 6.3b depicts the changes in the radiocarbon production that are needed to cause the observed A14C fluctuations. Note the difference in the amplitudes (ca. 40 percent in production, ca. 2 percent in A14C) and the lag of the A14C peak by about 20 years relative to the production. These system effects are caused by the large 14C reservoirs (ocean, atmosphere, biosphere) and the exchange between them.

Volcanic forcing and ice cores

Volcanic eruptions are the main cause of strong short-term (annual) climate forcing. The injection of large amounts of gases (SO2, CO2, H2O, N2) and dust into the atmosphere perturbs the radiative balance via enhanced absorption and scattering of solar radiation leading to a warming in the upper atmosphere and a cooling in the lower atmosphere (Figure 6.4). Due to the short tropospheric lifetime (1-3 weeks) the injection of gases must reach the stratosphere (lifetime: 1-3 years) to become globally active. The most important substance is SO2, which oxidizes quickly to H2SO4. Beside volcanic eruptions, there are other comparatively stable

S> 2700

2600

2500

2400

S> 2700

2600

2500

2400

Sphagnum imbricatum

1000

800 700 Dendro age (BC)

1000

Figure 6.3 (a) The 14C calibration curve for the period 1000-500 bc. The horizontal axis shows the calendar age. The vertical axis shows the 14C age in BP (BP: before 1950). The calibration curve is based on mainly decadal 14C measurements of exactly dated (dendrochronology) wood. Fluctuations of the curve are caused by changes of the cosmic ray intensity due to fluctuations in solar activity. (b) A14C is the deviation of the atmospheric 14C/12C ratio from a standard value in % as calculated from the data of the calibration curve. The blue curve depicts the changes in the radiocarbon production that are needed to cause the observed A14C data. The A14C curve provides information about the 14C production rate and the carbon system in the past. The interval ca. 850-750 bc represents the phase of

800 700 Dendro age (BC)

Sphagnum imbricatum

Sphagnum papillosum

Sphagnum cuspidatum

Sphagnum sect. acutifolia

extreme climate at the start of the Sub-atlantic. The "14C-clock" accelerates during that period (300 14C "years" in about 100 calendar years). The A14C curve shows a sharp rise of atmospheric radiocarbon, related to a temporary decline in solar activity. (c) Changes in species composition of Sphagnum during the Sub-boreal-Sub-atlantic transition in Engbertsdijksveen, a raised bog (after van Geel et al. 1996). Sphagnum species of the section acutifolia prefer relatively dry and warm climatic conditions. Sphagnum cuspidatum and S. papillosum prefer a relatively high water table. Sphagnum imbricatum prefers high air humidity and cool conditions. A major change started ca. 850 bc, when solar activity showed a temporary decline (a fast increase of A14C).

Figure 6.4 Schematic overview of the main atmospheric processes related to volcanic forcing (Fischer 2003, 2005). (From Robock (2000), plate 1.)

sources of SO2 such as the marine biota producing dimethylsulfide, and noneruptive volcanic emissions. Figure 6.4 is a schematic overview of the main processes involved in volcanic forcing (Fischer 2003, 2005).

Ice cores are an important source of information on volcanic eruptions. Figure 6.5 shows a record of volcanic sulfate derived from the GISP2 ice core (Zielinski and Mershon 1997). Using additional information about the volcanoes responsible for the emissions, Crowley (2000a) estimated the corresponding forcing for the past 1000 years (inset in Figure 6.5). The very sharp peaks indicate that the sulfate is rapidly removed from the atmosphere. Bay et al. (2004) linked volcanic ash layers in Siple Dome (Antarctica) with the onset of millennium-scale cooling recorded in a Greenland ice core and interpreted the results as evidence for a causal connection between volcanism and millennial climate change. The high volcanic activity during the main deglaciation of the early Holocene suggests that ice-sheet unloading and/or sea-level rise was responsible for increased volcanism during that period. One of the difficulties in quantifying the volcanic forcing is that without additional information it is impossible to distinguish between regional tropospheric and global stratospheric eruptions. Whereas one single eruption can

1000

Figure 6.5 A record of volcanic sulfate concentrations in ppb derived from H2O4 measurements on the GISP2 ice core (Zielinski and Mershon 1997) covering the past 12 000 calendar years. The inset shows an estimate of the sulfate-induced forcing in W m-2 by Crowley (2000a). Note that the forcing can be very strong, but does not last long (1-2 years).

12 000

1000 1200 1400 1600 1800 Year [AD]

2000

1000 1200 1400 1600 1800 Year [AD]

10 000

8000

4000

2000

10 000

8000

4000

2000

change weather patterns significantly over a 1-2 year period, a series of eruptions may have a long-term climatic effect. Castellano et al. (2004) compared millennial scale volcanic frequencies against SD in the EPICA-Dome-C ice core; they found no clear evidence for a close relationship between climatic change and volcanism.

Oscillations in the ocean system

An abrupt cold period occurred around 8200 calendar years BP in the North Atlantic area. It lasted for ca. 300 years and in Greenland ice-core records it is characterized by a reduction in temperature greater than 1°C, a decrease in ice accumulation rate, increasing wind speeds, and a drop in atmospheric methane levels (Wiersma and Renssen 2006, and references therein). A slowing down of the thermohaline circulation as a result of a freshwater perturbation has been proposed as the cause of the event. The slowdown resulted in a decrease of the northward heat transport in the North Atlantic Ocean, leading to pronounced cooling. The proglacial Laurentide Lakes in front of the Laurentide ice sheet were most probably the source of the freshwater pulse (Clarke et al. 2003). Model-data comparisons by Goosse et al. (2002) and Wiersma and Renssen (2006) confirm the catastrophic drainage of Laurentide Lakes as a forcing mechanism.

0 0

Post a comment