389
with a few "stations" in the south and covering the last 500 years. This
allows them not only to determine time scales of variability but is a first estimate of spatial correlation and patterns of natural variability as recorded
in high-resolution proxy data time series of sufficient length. Mann et al.
(1995) find cycles in the interdecadal (15-35 years) and century (50-150
years) range that are significant. As with the summer temperatures derived from tree ring widths of Fennoscandinavia, these variations are not
stable throughout time. Century scale cycles are particularly strong during
the mid 17 and 18 century, a time where fluctuations to colder conditions
have been reported extensively and are referred to as "Little Ice Age"
(Bradley and Jones 1995). The spatial distribution of phase relations indicates that the variability is mostly confined to the North Atlantic region
with a clear phase difference of 45°-135° between the eastern and western
side of the Atlantic basin. Mann et al. (1995) interpret this as the result
of in-phase and out-of-phase variability. The former may be due to the
basin-wide transport of heat by the meridional overturning circulation and
its century scale variability (Mikolajewicz and Maier-Reimer 1990; Mysak
et al. 1993) while the latter is consistent with a mechanism described by
Delworth et al. (1993) (see below).
It appears from this that the North Atlantic region is one of the pace
makers of climate variability. However, one should note that still only few
proxy data come from regions other than Europe or North America, and
that therefore our view may be biased. Future emphasis must be given
to the retrieval and analysis of high-resolution paleorecords of the tropical
regions and the southern hemisphere. In the meantime, models are the only
tool that can help us understand mechanisms of variability on the decadalto-century time scale. Also, they are suitable in pointing to locations of
increased variability where proxy records, if made available, should most
likely exhibit variability.
3 Models and Mechanisms
As mentioned above, the weakness of the hypothesis of a purely solar origin
of decadal-to-century variability lies in the fact that the sensitivity of the
climate system to changes in the shortwave irradiance is about a factor of
5 smaller than that for longwave emission changes including water vapour
feedback (Cess et al. (1989) give a mean sensitivity of 0.68 K/Wm- 2 ). By
monitoring the most recent solar cycle it was found that corresponding solar
with a few "stations" in the south and covering the last 500 years. This
allows them not only to determine time scales of variability but is a first estimate of spatial correlation and patterns of natural variability as recorded
in high-resolution proxy data time series of sufficient length. Mann et al.
(1995) find cycles in the interdecadal (15-35 years) and century (50-150
years) range that are significant. As with the summer temperatures derived from tree ring widths of Fennoscandinavia, these variations are not
stable throughout time. Century scale cycles are particularly strong during
the mid 17 and 18 century, a time where fluctuations to colder conditions
have been reported extensively and are referred to as "Little Ice Age"
(Bradley and Jones 1995). The spatial distribution of phase relations indicates that the variability is mostly confined to the North Atlantic region
with a clear phase difference of 45°-135° between the eastern and western
side of the Atlantic basin. Mann et al. (1995) interpret this as the result
of in-phase and out-of-phase variability. The former may be due to the
basin-wide transport of heat by the meridional overturning circulation and
its century scale variability (Mikolajewicz and Maier-Reimer 1990; Mysak
et al. 1993) while the latter is consistent with a mechanism described by
Delworth et al. (1993) (see below).
It appears from this that the North Atlantic region is one of the pace
makers of climate variability. However, one should note that still only few
proxy data come from regions other than Europe or North America, and
that therefore our view may be biased. Future emphasis must be given
to the retrieval and analysis of high-resolution paleorecords of the tropical
regions and the southern hemisphere. In the meantime, models are the only
tool that can help us understand mechanisms of variability on the decadalto-century time scale. Also, they are suitable in pointing to locations of
increased variability where proxy records, if made available, should most
likely exhibit variability.
3 Models and Mechanisms
As mentioned above, the weakness of the hypothesis of a purely solar origin
of decadal-to-century variability lies in the fact that the sensitivity of the
climate system to changes in the shortwave irradiance is about a factor of
5 smaller than that for longwave emission changes including water vapour
feedback (Cess et al. (1989) give a mean sensitivity of 0.68 K/Wm- 2 ). By
monitoring the most recent solar cycle it was found that corresponding solar