F. Gassmann

The predictions given so far by global circulation models (GCM) might be incomplete, because they lack a dynamical vegetation with the potential of structural changes. As an illustration for the possibility of vegetation-mediated abrupt climate changes, not described by today’s GCMs, a conceptual model called VILLIGENATOR is proposed. Its characteristics differ substantially from those of GCMs by incorporating homeostasis, critical thresholds, abrupt phase changes and hysteresis rather than a well-defined climate sensitivity. Investigations of nonlinear systems performed during the last decades reveal the VILLIGENATOR characteristics being ubiquitous for the behavior of complex systems and so, it seems wise not to exclude them as realistic possibilities also for the global climate system.

Very important insights for global climate dynamics come from ice cores, beginning with the first drillings on Greenland [1] and on Antarctica [2] in the 1980s showing surprisingly rapid climatic fluctuations during the last ice age (Heinrich events) and including the last interglacial (Eem) showing about 2 K higher temperatures than are observed today (ca. 15 ^oC in the global average). Showing no important fluctuations in these ice cores, the Eem was considered being analogous to the Holocene that was also rather stable. This picture of "turbulent" ice ages and "laminar" interglacials was made seriously uncertain by the analysis of the most recent Summit ice core from Central Greenland drilled in the GRIP (Greenland Ice core Project) between 1990 and 1992 [3, 4]. Whereas the rapid fluctuations during the glacial time were confirmed, extremely rapid changes were detected [3] also during the Eem-interglacial (135-115 kyr before present BP). According to this ice core, the longest stable warm period encompassed only 2700 years or 14 % of the Eem. The remaining 86 % of this warm time were disrupted by many sharp transitions between possibly 4 different distinct climatic states, separated from each other by roughly 2 K : A warm state (Holocene plus 2 K), a Holocene state, a Interstadial state (Holocene minus 2 K) and a glacial state (Holocene minus 4 K). Transitions between these states show no regular pattern and occurred within a few decades. An example of such an event is shown in Figure 1, where a transition took place from the warm state to the Interstadial, persisting for 750 years and was then followed by a jump back to the warm state. As causes for these transitions, variations of the solar activity or volcanic eruptions cannot be completely excluded, but the observed behavior does not seem to be a typical fingerprint of one of these processes. Were it the sun, we would expect a more oscillatory behavior and were it volcanoes, we would expect a more irregular dust content on rather small time scales. Not only are we unable to find the cause-effect chain explaining the observations, but also the observations themselves might be misleading. Only 28 km west of the European GRIP ice core, the American GISP2 (Greenland Ice Sheet Project 2) core was drilled. Whereas the two ice cores show perfect agreement over 2700 m covering about 100'000 years they do not match at all during the Eem period around 125 kyr BP. It is possible that ice flow may have altered the chronological sequences of the stratigraphy for the bottom part of one or both of the cores by producing shear folds [5].

The overall picture concerning behavioral patterns of the global climate system evolving from the ice core analyses and other information sources can tentatively be summarized as follows: It is difficult to define a "normal" state of the global climate system. Rather, several different states can be found for the same parameter values (multistability) and abrupt transitions between these states seem to be the rule. From this point of view, several crucial scientific questions arise:

• Does the global climate system have different coexisting quasistable states and if yes, how are the states nearest to the actual Holocene state characterized? Of special interest would be of course the next warmer state above today's climate.
• What is the reason for the stability of the Holocene climate being exceptional in the context of the last 250'000 years and are there critical thresholds? Especially important would be the existence of a critical atmospheric CO₂-level.
• How do climate transitions look like? One of the most important questions in this respect is the time scale of the transition process (years, decades or centuries).

The above presented VILLIGENATOR provides an example for a simple approach being able to explain different, seemingly unrelated aspects of the global climate system. Although our model has many limitations (one being that it cannot describe spatial variations in any detail), it does explain some important qualitative features of the observed Holocene and Eem climates and transparently demonstrates a mechanism being sensitive to eccentricity and so offering an explanation for the fundamental differences between Holocene and Eem climates. The main results can be summarized as follows:

• a stable Holocene climate being characterized by an active biological mechanism leading to homeostasis and showing almost negligible climate sensitivity.
• for a small and linear increase rate of the anthropogenic climate forcing (e.g. 0.015 Wm^-2 per year), the model gives a constant temperature increase (0.3 K) over a limited time interval, being correlated only with the rate of forcing increase and not with the absolute forcing. This might give an explanation for the plateau with more or less constant temperatures between about 1940 and 1980.
• as soon as D M reaches its maximum value, stabilization is not longer possible and temperature increases parallel to forcing, as observed since 1980.
• for a temperature increase D T of 0.5 K, the model gives an increase rate of the average normalized biomass of 0.026 per year being equivalent to a carbon sink of 1.5 Gt per year (based on an estimated affected plant biomass of approximately 200 GtC equivalent to M = 1.7/0.5 = 3.4). This result is compatible with the "missing carbon sink" of 1.6±1.4 Gt per year (explained here as the control signal for stabilization).
• as soon as a critical forcing is exceeded, the system responds with an abrupt temperature increase and a collapse of a part of the biosphere (D M displays a sudden transition from +1 to -1).
• for the Eem interglacial (e = 4%), the model shows an interesting structure of different stability regimes: for perihelion in January a stable Holocene type climate, for perihelion in July a stable Eem optimum or Interstadial climate, for intermediate perihelion dates stability of all 3 climates. As a result of forcing fluctuations of the order of a few Wm^-2, abrupt transitions between the different stable climates can occur. Within the above mentioned hysteresis range, the "activation energy" needed to trigger a transition is sensitive to variations of the amplitude of the yearly temperature cycle and so, transition probabilities due to a constant stochastic forcing considerably vary within relatively short time intervals (centuries) in Eem. Further, in many situations, transitions occur directly between the E and I state, and the intermediate H state, though stable, is difficult to reach. These properties seem compatible with the general behavior of the Eem climate as observed with the GRIP ice core [3].
• because transpiration takes place mainly during daylight and during the warm season, day-night and summer-winter asymmetries of climate change as observed in Central Europe [17] are a natural outcome of the proposed conceptual model.

Our analyses should not be construed as proof that global climate is actively controlled by the northern biosphere. The quantity M might allow other interpretations leading to a similar analytical form of equation (7) and therefore similar results. It would not be surprising, however, if more detailed analyses would show that the active biological part of the earth’s surface significantly contributes to climate control, so supporting the idea of the "Gaia hypothesis" [18]. In fact, a refinement of the above described model by Füssler [19, 20] taking into account more detailed physiological processes, especially a feedback-loop involving root-dynamics under drought, leads to a similar hysteresis behavior as described above. In addition, other theoretical investigations with simplified models [21] underline the importance of atmosphere-vegetation feedbacks. Simulations with GCMs taking also increasing atmospheric CO₂ concentrations into account clearly show significant physiological and structural vegetation feedbacks in the climate system [22]. Finally, it is expected that a CO₂ doubling would induce major changes in broad vegetation types of about two-thirds of the forests in medium and high northern latitudes [23, p.5-7].

References

back