50
Chapter 3: Climate Spectra and Stochastic Climate Models
Figure 3.13: Spectrum ofthe mass transport through the Drake Passage and
the net freshwater flux of the Southern Ocean (dashed line). The thin lines
represents the results to be expected from (3.14) with feedback with time
constants of 50 years, 100 years, and infinite. The bar represents the 95%
confidence interval. (From Mikolajewicz and Maier-Reimer, 1990).
10-6
10-7~~~~~~--~~~~~--~~~~~
10 3
10 1
PERIOD [aJ
formation rate and large changes in the surface heat fiuxes.
However, the mixed boundary conditions are artificial, neglecting the atmospheric and sea ice dynamics. Their use with fixed forcing mayaiso lead
to instabilities of the oceanic thermohaline circulation in global ocean models and to large self-sustained oscillations on decadal and century scales;
this variability persists when stochastic forcing is added (e.g., Weaver et al.
1993). As the self-sustained oscillations disappear when "restoring" boundary
conditions (relaxation for both SST and surface salinity) are used, coupled
ocean-sea ice-atmosphere models with more realistic boundary conditions are
needed to better understand the origin of the oceanic variability on long time
scales.
3.7 Long Term Climate Changes
The general features of the spectrum of climatic variability on longer time
scales have been described by Kutzbach and Bryson (1974), Stocker and
Mysak (1992), Shackleton and Imbrie (1990), and others. The spectral density mostly increases with decreasing frequency, down to the large peaks at
the orbital frequencies. Early studies with zonally averaged energy balance
climate models have suggested that stochastic fiuctuations in the meridional
Chapter 3: Climate Spectra and Stochastic Climate Models
Figure 3.13: Spectrum ofthe mass transport through the Drake Passage and
the net freshwater flux of the Southern Ocean (dashed line). The thin lines
represents the results to be expected from (3.14) with feedback with time
constants of 50 years, 100 years, and infinite. The bar represents the 95%
confidence interval. (From Mikolajewicz and Maier-Reimer, 1990).
10-6
10-7~~~~~~--~~~~~--~~~~~
10 3
10 1
PERIOD [aJ
formation rate and large changes in the surface heat fiuxes.
However, the mixed boundary conditions are artificial, neglecting the atmospheric and sea ice dynamics. Their use with fixed forcing mayaiso lead
to instabilities of the oceanic thermohaline circulation in global ocean models and to large self-sustained oscillations on decadal and century scales;
this variability persists when stochastic forcing is added (e.g., Weaver et al.
1993). As the self-sustained oscillations disappear when "restoring" boundary
conditions (relaxation for both SST and surface salinity) are used, coupled
ocean-sea ice-atmosphere models with more realistic boundary conditions are
needed to better understand the origin of the oceanic variability on long time
scales.
3.7 Long Term Climate Changes
The general features of the spectrum of climatic variability on longer time
scales have been described by Kutzbach and Bryson (1974), Stocker and
Mysak (1992), Shackleton and Imbrie (1990), and others. The spectral density mostly increases with decreasing frequency, down to the large peaks at
the orbital frequencies. Early studies with zonally averaged energy balance
climate models have suggested that stochastic fiuctuations in the meridional