For 20 years, Professor Johns was a firm believer in the view that Carbon Dioxide is the most likely cause of the current global warming. He even required applicants to his department to explain the "simple physics" basis on which many global warming arguments rest. His view was consistent with the position that the technology already exists for us to reduce the use of carbon-based fuels by 80% with negligible impact on living standards. Indeed, if we do not take this view, it follows that in a few hundred years our descendants will return to the Stone Age when carbon-based fuels are exhausted. His view began to change in the mid 1990's when he read the Club of Rome book "Factor Four". The premise upon which the book is based is that we can halve our energy use and simultaneously double our standard of living. This view was entirely consistent with his view that an 80% reduction in use of carbon-based fuels is readily achievable and, in the longer term, a 100% reduction can, and must, be achieved. However, the science in "Factor Four" is weak.

In particular, "Factor Four" puts great emphasis on the significance of the temperature/CO2 correlation discovered in the Vostok ice-core record. For anyone who has designed systems to remove acid gases (including carbon dioxide) from exhaust streams it is immediately obvious that the correlation is simply a solubility effect. Many of these "scrubbing" systems work by absorbing acid gases in low temperature solutions, and releasing the gases in relatively pure form by heating the solutions. The solutions are then cooled and recycled. Thus, we know that (in common with other solutions) the higher the temperature, the lower the solubility of carbon dioxide in ocean water. Thus, as temperature increases the pressure of carbon dioxide gas over the oceans increases. It follows that the Vostok correlation can thus be explained by a simple well-established correlation that is easily confirmed by laboratory measurements. The cause and effect is verified by the observation that, at a given temperature, lower carbon dioxide concentrations precede periods of rising temperature and higher carbon dioxide concentrations precede periods of falling temperature. This effect is expected if the oceans exhibit thermal inertia so that carbon dioxide concentrations (as a measure of ocean temperature) lag atmospheric temperatures. This observation is consistent with hypothesis that the forces that drive natural climate change drive temperatures upwards, despite relatively low carbon dioxide concentrations and pull them downwards despite relatively high carbon dioxide concentrations. Alternatively, carbon dioxide has not played a significant role in climate change over the past 400,000 years. This finding does not prove that carbon dioxide is not currently playing a significant role in climate change. The atmospheric concentrations are now relatively much higher than they have been for 400,000 years. Thus, the Vostok data might be irrelevant to the current period of global warming. Whether or not that is the case, the coupling between carbon dioxide concentration and temperature found in the Vostok ice-core record does not support the hypothesis that increased carbon dioxide concentrations cause rising temperature. On the contrary, they support the conclusion that high carbon dioxide concentrations are caused by high temperatures that drive carbon dioxide from the oceans.

Having found that one of the major props to the carbon theory did not support the theory at all, we turned to the next strong indicator that the current global warming is caused by high atmospheric carbon dioxide concentrations. The assertion is frequently made that we are currently experiencing the highest temperatures and most sustained period of temperature rise for over 10,000 years. Simply examining the Vostok record for the last 10,000 years shows this assertion to be false. There have been higher temperatures in the last 10,000 years, and the current rate of temperature is less than the average found during similarly sustained periods of global warming in the last 10,000 years. Thus, two of the major props to the carbon hypothesis turn out to be false.

The third prop to the carbon theory of global warming is the "simple science" argument. The argument goes that carbon dioxide is almost transparent to radiation from the sun, but absorbs the infra-red radiation from the surface of the Earth. Thus, it "traps" heat and causes global warming. This argument certainly convinced me for many years; it just seems to be common sense. However, we have found that the temperature/CO2 concentration coupling of the Vostok record does not support the carbon hypothesis. We have also found that it is entirely false to claim that the current warming is exceptional in its duration and temperature rise. We should, therefore, re-examine the "simple science" proposition. There are two major flaws in the proposition. The first is that, at a given temperature, fundamental science says that absorptivity and emissivity must be identically equal. (Otherwise, we could trivially create a perpetual motion machine that would deliver infinite energy without any source). The second flaw is that there is a strong interaction between water vapour and carbon dioxide. Thus, in the presence of excess water vapour the absorptivity of carbon dioxide is strongly suppressed. Indeed, at sea level, the contribution of carbon dioxide to infra-red absorption is negligible compared to the natural variability of water vapour concentration in the atmosphere. Consequently, we would expect any contribution of carbon dioxide to global warming to arise from carbon dioxide concentrations at high altitude, above the levels at which water vapour concentrations are significant. In the absence of greenhouse gases, radiation from the surface of the Earth would go straight through this high-altitude layer and dissipate into space. At the same time, heat would rise from the surface of the Earth through convection currents, and 100% of this convected heat would return to the surface through convection. (These currents rise at the equator and in two bands that lie to the equator side of the polar circles. They descend again over the desert regions North and South of the Equator and onto the poles). This heat-flow pattern may be altered by the presence of greenhouse gases. These gases are emissive as well as absorptive. Thus, the layer now absorbs some of the radiant heat from the surface of the Earth and re-radiates it upwards and downwards. That which is re-radiated upwards is lost into space. That which is re-radiated downwards increases the heat flux to the surface of the Earth. If this were the only factor, it would certainly cause a measure of global warming. However, this emissive layer will also radiate some of the heat brought up to it by convection. In the absence of the layer, 100% of the convected heat would be returned to the surface of the Earth. However, the greenhouse gas now radiates a proportion of that heat into outer space (as a further complication, some of the heat absorbed by carbon dioxide will be returned to Earth by convection). Thus, the science is not so simple. If the convected heat is negligible, the layer could cause a measure of global warming. If the convected heat is large compared to the radiant heat absorbed, the greenhouse gases could cause significant global cooling. Indeed, it is possible that the additional carbon dioxide that we are releasing is cooling the Earth and hence reducing the effect of a natural period of global warming. In summary, the "simple science" argument fails. The conclusion may well be correct, and carbon dioxide might be warming the planet. However, the effect is not simple and obvious. The net effect of carbon dioxide results from a subtle balance.

Finally, we come to the argument that the IPCC has a number of reliable climate models from which it concludes that there is a 90% probability that carbon dioxide is causing global warming (and a 10% probability that it arises from natural climate variability). How can all of those scientists be wrong? First of all, none of them are saying that they are 100% certain that carbon dioxide is causing global warming. Thus, if it turns out not be causing the warming, none of them are wrong. You can state that, in throwing dice, there is a low probability of getting a "6". If it then comes up "6", you are not wrong; there was a low probability. Another observation to make is that few of these climate scientists are main-stream scientists. Most come from a weather-forecasting (meteorology) background. They are not used to formulating and testing models like other scientists because they cannot go back to the laboratory to test the various elements of the model; they have to wait for the climatic conditions to arise that test the models. That is why it takes decades for them to progress from predicting one day ahead to making rather uncertain predictions 7 days ahead. The weather is chaotic. Thus, in principle, it is impossible for us to predict very far ahead. Small scale disturbances grow to become large effects. These small disturbances may be at a molecular level that is too small for us to measure, or may result from external effects coming from outer space (for example, dust). Chaotic systems typically develop cycles (which when there are several superimposed cycles of different frequencies look random). The statistics of the variability are a characteristic of the chaotic system. There are two approaches to modelling chaotic systems. The first approach is to determine accurately one point in the cycle, and then predict as far ahead as you can before your predictions are swamped by chaos. The second is to allow the model to run for a long time until sufficient results are obtained to determine the statistical properties of the chaotic system. In the latter approach, the initial conditions for the simulation may be relatively unimportant. The typically chaotic behaviour develops from almost any starting point. For weather forecasting, the first approach is adopted. No emphasis is placed on ensuring that the chaotic statistics are properly reproduced. The emphasis is only on the accuracy of the model for the relatively short period until it is swamped by chaos. When meteorologists approach climate modelling, they adapt their familiar weather-forecasting approach. Thus, there is no emphasis on the chaotic statistics. The emphasis is entirely on forecasting for a few decades ahead during which time they hope that the model is not swamped by chaos. Typical of the models is the well-regarded model produced by the Hadley Centre in Exeter (part of the UK Meteorological Office). It has a detailed model of the geography of the planet with all the mountains, oceans and continents in their correct places. It also places great emphasis on setting the correct initial conditions. Thus, it ensures that parameters are adjusted so that it matches the current temperature and temperature rise rate, and fits the climate change experienced over the last few decades. It ignores the chaotic statistics. Thus, if you remove carbon dioxide from the simulation, it settles to a "pre-industrial climate" (that is the climate experienced in the late 1700s and early 1800s). It does not reproduce the natural variability of climate found in the ice-core records. These records clearly reveal that there is no such thing as a "pre-industrial climate". The climate has always varied. Indeed, there is a 50/50 chance that we would experience the current global warming with no excess carbon dioxide releases. The Hadley model is probably valuable in showing the likely effect of changing climate on weather patterns. However, it is useless in determining the incremental effect of carbon dioxide because it is incapable of reproducing the natural variability of climate. The IPCC is dominated by meteorologists and the IPCC criteria for accepting a model are exactly the criteria met by the Hadley model. There is no requirement that the models reproduce the natural variability of climate in the absence of releases of carbon dioxide. We conclude that current models are of no value in determining whether carbon dioxide influences climate. Their value is in determining the effect of climate on weather if carbon dioxide drives climate change.