Unfortunately scientist can slip into using cliches. This is despite the fact that imagination and creativity are probably the key skills to successful science. A quick Google Scholar search on the term ‘Goldilocks’ reveals over 24,000 scientific papers and books that use the term and only a small number of them involve three bears! It has become shorthand for ‘just right’. Understanding planetary temperatures is no exception. Planet Earth is said to be located in the Goldilocks Zone of the solar system – not too far from the sun to be too cold, not to close to be too hot. In practice, ‘just right’ conditions have prevailed throughout much of the 4.6 billion years of Earth history, the big question is why?
Just right: not too hot, not to cold
The probability of the Earth’s temperature remaining at suitable level for life to survive during much of the 4.6 billion years of Earth history has been described as a kin ‘to crossing a motorway blindfold during rush hour’. When the Earth was first formed the solar constant, that is the amount of radiation received by the Earth from the sun, was lower than it is today. The sun was dimmer and has been slowly warming ever since. The solar constant has risen throughout Earth history but the average planetary temperature has remained broadly the same give or take a few wobbles. The climate and the chemical properties of the Earth now and throughout its history seem always to have been optimal for life The question is why and how?
This observation led Professor James Lovelock to the radical conclusion that the Earth could regulate its own temperature just like any warm blooded animal. He postulating thermo-regulation and coined the term Gaia – Goddess Earth – and suggested that the Earth was living. In this context a living being is something able to regulate and optimise its environmental conditions. He knew how to sell a story and the Gaia hypothesis was born.
Whether the Earth can truly be described as ‘living’ is largely irrelevant, the crux of the issue is that positive and negative feedback loops within planetary systems may operate against extremes just like the human body. To hot and you begin to sweat, to cold and your body develops gooseflesh and a shiver. The search was on to find self-regulating planetary systems and a number of such systems have emerged over the years.
One of the first of these was the role of dimethyl sulphide in cloud formation. For condensation to occur you need condensation nuclei, tiny particles of dust or particulate matter around which water vapour can condense. It is where ideas of cloud seeding in drought stricken areas come from. Lovelock and his colleagues were able to demonstrate that a key condensation-nuclei was sulphur dioxide derived from dimethyl sulphide from plankton on the ocean surface. As a result they found a large planetary feedback mechanism (Fig. 1). The so called CLAW hypothesis was named after the authors involved in its discovery Charlson, Lovelock, Andrea and Warren (1987).
Figure 1: The CLAW hypothesis. (Source: Plumbago-Own work, CC BY 2.5, en.wikipedia.org/w/index.php?curid=12987086)
The sun acts to increase the growth rates of phytoplankton due either to increased surface temperature and/or increased availability of sunlight. Certain phytoplankton, such as coccolithophorids, synthesise dimethylsulphoniopropionate (DMSP), and their enhanced growth increases production. As this breaks down dimethyl sulphide (DMS) is produced first in seawater, and then in the atmosphere. DMS is oxidised in the atmosphere to form sulphur dioxide an important condensation nuclei. The water content of clouds increases as a result increasing their reflectivity to incoming solar radiation (enhanced albedo). The result is less solar heating of the ocean surface and a decrease in phytoplankton production of DMS resulting in a self-regulatory system.
Scientific modelling
Geographical research is not all about direct observations and fieldwork; there is more to geography than standing in the rain measuring things. Physical and laboratory based experiments all play a part as does modelling. Modelling involves the conceptual or numerical simulation of a series of processes that give rise to something like a landform or ecological species distribution. Modelling is a central to our understanding of many geographical phenomena; it lies at the heart of the weather forecast on your smart phone or TV.
There is a range of different opinions as to what actually constitutes a ‘model’ but the simplest way of looking at it as any construct which generates a prediction. It follows that modelling, like experimentation and observation, is simply an activity that enables theories to be tested and examined critically. Here are a just a few different types of geographical model:
- Conceptual models. These are theoretic expressions, often in cartoon form, of the key variables involved in a particular process. They are designed to scope, inform, communicate and help the reader frame ideas for testing. Figure 2 is a conceptual model.
- Physical models. Flumes and wave machines are used sometimes to test coastal or flood defences. Here reality is scaled down and observed in microcosm.
- Empirical/statistical models. These are based on observed data and attempt to look for causal relationships between several variables. Lots of data is collected and statistics are used to pull out salient points or relationships.
- Numerical models. We can use physics and maths to create a set of equations which describe elements of a process. For example, we can describe soil strength or the rate of glacier flow via simple equations. These types of models can vary along a continuum from those that are specifically grounded in a time and place to those that are more abstract. If we ground and parameterise these equations with data from a specific place or time we can attempt to model and thereby reconstruct (or predict) phenomena at a different time. For example, we might try to create a model of an ice sheet that once covered Britain and vary the input parameters to see how it behaves. The experiment here is to find the input parameters that best recreate what we think happened in the past. Alternatively we may create a predictive model such as a general circulation model (GCM) of the Earth’s atmospheric systems and their interaction with land and sea in order to forecast the weather. The predictions made by the model become the test of its accuracy. At the other end of the continuum are more abstract models that are not grounded in a specific place and time but allow us to explore more theoretical relationships between the variables involved.
- Agent based models. This type of model is a form of numerical model which empower individual elements with the power to make decisions. Imaging a large crowded stadium and someone needlessly shouts ‘bomb’. Each member of the crowd is capable of make its own decision; to you run left, right or duck. Perhaps less dramatic is to think of a herd of wildebeest migrating on the Serengeti; in theory each animal can make a choice as to the route it will follow, although it may be highly influenced by other members of the herd. We can simulate this numerically by giving each agent in the model an individual decision making capacity according to a set of pre-set rules and step back and observe. There are lots of ecological and geographical systems that lend themselves to this type of modelling.
Modelling has played an important role in the debate about Gaia.
Lovelock and colleagues developed a conceptual model of a self-regulating world, known as Daisy world. They were not modelling the Earth specifically simply illustrating how self-regulation might work.
Figure 2: Conceptual illustration of daisy world. Think about the distribution of daises and why they occur where they do.
Daisy world is a very elegant model (Fig. 2). To start with they imagine a world inhabited by just two types of daisy – black and white ones – warmed by a sun similar to our own that starts cold and gradually warms up. Now at first the planet is cold and black daisies thrive since their black pigmentation allows them to absorb more radiation than their white counter parts which tends to reflect solar radiation (i.e. white daisies have a higher surface albedo or reflectance). The planet is warmer than one might expect, but as the solar constant rises the black daisies begin to be scorched by the sun. Progressively white daisies thrive, since they can reflect some of the excess heat and keep the planet cooler. Planetary temperature is moderated in this way as the solar constant changes over time. It is an elegant illustration of a thermal regulation without any teleological need (i.e. need for sentient thought). The model can be made more complex by the introduction of grey daisies which thrive in intermediate conditions and by introducing grazing animals and then predators (Fig. 3).
Figure 3: Output from Daisy World model showing how temperature is regulated.
The model does not recreate reality or simulate a process that actually took place on Earth, but provides a conceptual illustration or demonstration of the principles of self-regulation and as such it is a supremely elegant piece of science communication.
Interestingly you can run a daisy world simulation for yourself using an agent based model (http://ccl.northwestern.edu/netlogo/models/Daisyworld). While the original daisy world model was not an ABM the idea lends itself to individual based modelling and is a nice way of demonstrating the potential of such modelling approaches.
Composing computer programs to solve scientific problems is like writing poetry. You must choose every word with care and link it with the other words in perfect syntax. There is no place for verbosity or carelessness. To become fluent in a computer language demands almost the antithesis of modern loose thinking.
James Lovelock
Temperature regulation
So we know – or at least I hope you do – that daisies are not the answer just the illustration. This still leaves the question of how the earth has moderated its temperature. The simple answer is through plate tectonics.
One of the most important volcanic product is carbon dioxide; in terms of volume and impact it is far greater than anything else produced by an eruption. The longest chains of volcanoes are those found at divergent plate boundaries, such as the mid-Atlantic ridge. Large volumes of carbon dioxide are produced by sea floor spreading and at subduction zones. So periods of geological time like the Mesozoic which are known to have been particularly warm are often associated with periods of active plate tectonics in this case the opening of the South and North Atlantic.
So if volcanism increases the carbon content and drives greenhouse conditions, what removes carbon dioxide? Oceans absorb carbon dioxide but by far the most significant process is weathering. Most rock forming minerals are silicate based and are easily attached by weak acid rain. As rain fall it combines with carbon dioxide in the atmosphere and critically in the soil to form a weak carbonic acid (CO2 + H2O – H2CO3). This weak acid attacks the silicates to give a carbonate which is removed in solution and is ultimately deposited in the oceans. In the equation below X stands for any cation like sodium, potassium, or magnesium which are common rock forming elements.
So weathering removes, or scrubs, carbon dioxide from the atmosphere. The more weathering the more carbon dioxide is removed. No weathering is favoured by warm and damp conditions and by surface area. Think of a flat hill top, exposed rock is weathered and a weathered regolith is produced (add organic matter and you would have a soil). As the regolith builds up it begins to slow the weathering rate since the fresh rock is increasingly at depth. Now contrast this with a steep slope under similar conditions. Here the regolith as it forms is removed by gravity and fresh bedrock remains at the surface to be weathered. More weathering will result than on the fat surface. So uplift and mountain building, both a consequence of plate tectonics, favour weathering. So if we have lots of plate convergence creating fold mountains and uplift we tend to draw down atmospheric carbon dioxide.
Figure 4: The feedback between carbon dioxide, global temperature and weathering.
We are not sure what came before plate tectonics when the earth was very molten and very young, perhaps the first 0.5 billion years or so. But with the advent of plate tectonics came movement and with it the creation of continental crust. This was probably formed by the accretion of volcanoes along continental edges and their erosion to create sediments. Slowly by the processes of subduction at convergent boundaries continental rocks were formed. We call these ancient pieces of crust cratons or shield areas. The effect was to create land surfaces for weathering drawing down carbon dioxide rapidly. Not only was the process creating land for future life but also terra forming the atmosphere.
The faint sun in the early days of the Earth was countered by a thick greenhouse atmosphere, helping to retain heat. As the crust and weathering kicked in the carbon dioxide levels began to fall rapidly. Terra forming by bacteria and algae boosted oxygen levels at the expense of carbon dioxide. As the solar constant increased the earth needed less greenhouse gases. The complex interaction between volcanoes, plate tectonics, weathering and the evolution of life is responsible for maintaining global temperature in a equable state. It is perhaps odd to think that life of earth was dependent on greenhouses gases.
Final word: to live or not?
Lovelock and his colleagues were brilliant agents of ‘spin’. The choice of Gaia and the implication that the Earth was something to be revered as a living system was inspired. Unfortunately it also caused many more staid scientists to shake their head and shrouded the scientific value of their core ideas in a lot of mystic metaphysical nonsense. It all depends on how you define whether something is living.
They never set out to create a teleological model or to imply that Earth processes were driven by a higher ‘purpose’ simply to demonstrate how a complex system could self-organise. Complex systems whether living or not can result in emergent behaviours in which the sum is greater than the component parts.
What distinguishes a Complex Adaptive System (CAS) from a pure multi-agent system (MAS) is the focus on system survival and characteristics such as self-similarity, complexity, emergence and self-organization. A MAS is defined as a system composed of multiple interacting agents but in a CAS, the agents as well as the system are adaptive. Complex Adaptive Systems are therefore characterised by a high degree of adaptive capacity, giving them resilience in the face of perturbation. The Earth is an example of a CAS, resilient in the face of perturbation, a collection of self-similar agents adapting in ways that the system behaviour emerges beyond the behaviours of the individual agents. The behaviour of the collective is not predicted by the behaviour of the components.
References
Charlson, R. J., Lovelock, J. E., Andreae, M. O. and Warren, S. G. (1987). “Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate”. Nature. 326 (6114): 655–661. Bibcode:1987Natur.326..655C. doi:10.1038/326655a0.
Lovelock, J.E. (2000) [1979]. Gaia: A New Look at Life on Earth (3rd ed.). Oxford University Press. ISBN 0-19-286218-9.
Lovelock, James (2007). The Revenge of Gaia. Penguin. ISBN 0-14-102597-2.
Professor Sandcastle 