The Earth system is dynamically stable but with strong feedbacks. Its behaviour resembles more the physiology of a living organism than that of the equilibrium box models of the last century (Lovelock 1986). Broecker (1991) has shown by observation and models that even the wholly physical models of the Earth system are non-linear, often because the properties of water set critical points during warming and cooling. These include the heat-driven circulation of the oceans. The phase change from ice to water is accompanied by an albedo change from 0.8 to 0.2 and this strongly affects climate (Budyko 1969). There are other purely physical feedbacks in the system: the ocean surface stratifies at 12-14 °C, the rate of water evaporation from land surfaces becomes a problem for plants at temperatures above 22-25 °C and atmospheric relative humidity has a large direct effect on the size and effective albedo of aerosols. In a simple energy balance model, Henderson-Sellers & McGuffie (2005) show the large climate discontinuity between the ice-free and icy worlds and marked hysteresis.
Model systems that include, in addition to geophysics, an active and evolving biota self-regulate at physiologically favourable temperatures. Lovelock & Kump (1994) described a zero-dimensional model of a planet that self-regulated its climate; it had land surfaces occupied by plants and the ocean was a habitat for algae. This model system was normally in negative feedback with respect to temperature or CO2 increase, but when subjected to a progressive increase of CO2 or heat flux, regulation continued at first, but as the critical CO2 abundance of 450 ppm, or heat input of 1450 Wm-2, was approached, the sign of the feedback changed to positive and the system began to amplify and did not resist change. At the critical point, amplification rose steeply and precipitated a 6 °C rise in temperature. Afterwards the system returned to negative feedback and continued to self-regulate at the higher temperature. As with the ice albedo feedback, there was marked hysteresis and reducing CO2 abundance or heat flux did not immediately restore the state prior to the discontinuity.
The justifications for using this tiny zero-dimensional model to argue against the powerful forecasts of the giant global climate models are these. First, it is a model in which the biota and the geosphere play an active dynamic role, as in the model Daisyworld (Watson & Lovelock 1983) from which it has descended. Second, it makes predictions that are more in accord with the Earth's history. It suggests that attempts at amelioration should take place before the critical point is reached. Unfortunately, when the large effect of unintentional cooling by shortlived pollution aerosols is taken into account, we may already be past this point and it would be unwise to assume that climate change can simply be reversed by reducing emissions or by geo-engineering.
An engineer or physiologist looking at the IPCC forecasts for this century would find unconvincing their smooth and uninterrupted temperature rise until 2100, something expected of the equilibrium behaviour of a dead planet such as Mars. A glance at the Earth's recent history reveals a climate and atmospheric composition that fluctuates suddenly as would be expected of a dynamic system with positive feedback. The long-term history of the Earth suggests the existence of hot and cold stable states that geologists refer to as the greenhouses and the icehouses. In between are metastable periods such as the present interglacial. The best-known hot house happened 55 Myr ago at the beginning of the Eocene period (Tripati & Elderfield 2005; Higgins & Schrag 2006). In that event, between 1 and 2 teratons of CO2 were released into the air by a geological accident. Putting so much CO2 in the air caused the temperature of the temperate and Arctic regions to rise by 8 °C and of the tropics by 5 °C and it took about 200 000 years for conditions to return to their previous states. Soon we will have injected a comparable quantity of CO2 and the Earth itself may release as much again when the ecosystems of the land and ocean are adversely affected by heat.
The rise in CO2 55 Myr ago is thought to have occurred more slowly than now; the injection of gaseous carbon compounds into the atmosphere might have taken place over a period of about 10 000 years, instead of about 200 years as we are now doing. The great rapidity with which we add carbon gases to the air could be as damaging as is the quantity. The rapidity of the pollution gives the Earth system little time to adjust and this is particularly important for the ocean ecosystems; the rapid accumulation of CO2 in the surface water is making them too acidic for shell-forming organisms (Royal Society 2005). This did not appear to happen during the Eocene event, perhaps because there was time for the more alkaline deep waters to mix in and neutralize the surface ocean. Despite the large difference in the injection times of CO2, the change in the temperature of approximately 5°C globally may have occurred as rapidly 55 Myr ago as it may soon do now. The time it takes to move between the two system states is likely to be set by the properties of the system more than by the rate of addition of radiant heat or CO2.
There are differences between the Earth 55 Myr ago and now. The Sun was 0.5 per cent cooler and there was no agriculture anywhere so that natural vegetation was free to regulate the climate. Another difference was that the world was not then experiencing global dimming - the 2-3 °C of global cooling caused by the atmospheric aerosol of man-made pollution (Ramanathan et al. 2007). This haze covers much of the northern hemisphere and offsets global heating by reflecting sunlight and more importantly by nucleating clouds that reflect even more sunlight. The aerosol particles of the haze persist in the air for only a few weeks, whereas carbon dioxide persists for between 50 and 100 years. Any economic downturn that reduced fossil-fuel use would reduce the aerosol density and intensify the heating and so would the rapid implementation of the Bali recommendation for cutting back fossil-fuel use.
It is sometimes assumed that the temperature of the sunlit surface of a planet is directly related to the albedo of the illuminated area. This assumption is not true for forested areas. The physiological temperature regulation of a tree normally keeps leaf temperature below ambient air temperature by évapotranspiration, the active process by which ground water is pumped to the leaves; the trees absorb the solar radiation but disperse the heat insensibly as the latent heat of water vapour. I have observed in the southern English summer that dark conifer tree leaves maintain a surface temperature more than 20 °C cooler than an inert surface of the same colour.
Was this article helpful?