Reducing the solar radiation that reaches Earth will have potentially significant consequences beyond limiting the mean temperature of the planet; it may reduce annual rainfall, especially in the Americas and northern Eurasia.
Harvard geoengineers are set to spray sun-reflecting chemical particles into the atmosphere to cool the planet from a balloon at 80 000 feet over Fort Sumner, New Mexico . Chief investigator David Keith manages a multimillion dollar research fund awarded by Microsoft founder Bill Gates, and has already commissioned a study by a US aerospace company that made the case for large-scale deployment of solar radiation management technologies. The experiment, to be conducted with James Anderson within a year, will release tens to hundreds of kilograms of particles to measure the impacts on ozone chemistry and test ways of making sulphate aerosols the appropriate size.
Many scientists are opposed to geoengineering experiments, preferring to study the impacts of sulphuric dust emitted by volcanoes, and to use modelling to identify the risks. A British field test involving a balloon and hose-pipe to pump water into the sky, which was part of the government-funded Stratospheric Particle Injection for Climate Engineering (Spice) project (see  Skyhook to Save the Climate?) was cancelled after public outcry.
But there are good reasons why geoengineering should not be considered.
The obvious way to combat climate change is to cool the planet by reducing emissions of greenhouse gases and removing them from the atmosphere. That means using less energy, replacing fossil fuels by renewables, halting deforestation, and adopting sustainable farming practices. As documented in two major reports published by ISIS [3, 4] Food Futures Now: *Organic *Sustainable *Fossil Fuel Free , Green Energies – 100% Renewable by 2050, all the necessary technologies are available and getting better and cheaper every day, only the political will is missing.
Geoengineering offers an alternative quick fix, which is to reduce the amount of radiation reaching Earth’s surface in the first place. There are several suggestions about how this might be done, for example, by putting mirrors into space or sulphur particles into the stratosphere, or by increasing the brightness of clouds by spraying sea salt into them (see  GeoEngineering A Measure of Desperation, SiS 41). Geoengineering involves making changes on a planetary scale [5, 6].
A major drawback of geoengineering, whether it involves solar radiation management (SRM) or other measures such as fertilizing the oceans in the hope of increasing the absorption of carbon by phytoplankton, is that it is likely to be very difficult to reverse. If we sent small particles into the stratosphere what could happen if they drifted out of position, or coalesced, or came back down sooner than we anticipated? Not only could the result be a massive waste of valuable resources, we could end up having done irreparable harm to the planet.
But if we could manage to put into the sky something that actually accomplishes what it was designed to do, i.e., reduce the amount of radiation reaching Earth by just the amount necessary to compensate for the reduction in outgoing radiation caused by the greenhouse effect; wouldn’t that solve the problem of climate change without having to deal with any of the political and economic obstacles to more conventional measures?
Unfortunately, Earth’s climate is a very complex system, and it responds to far more than just the amount of energy in and energy out averaged over the four seasons and the entire surface of the planet. The precise spatial and temporal variations in energy distribution can have very different effects on global climate, and different SRM measures will lead to different effects. Furthermore, SRM measures are in no way equivalent to reducing greenhouse gas emissions. And should SRM measures fail, we are still left with too much greenhouse gases in our atmosphere.
It is not hard to see why the differences should matter. For example, an important factor in driving the weather is the differences in temperature and pressure between neighbouring areas. The onshore breezes that are so common in coastal areas in the summer arise because the air over the land is warmer, and therefore at a lower pressure than the air over the sea. The temperature differential can also lead to the formation of clouds near the shore where the two air masses meet. Neither the breezes nor the clouds would be there if the temperature were the same on the sea as on the land, even if the mean temperature for the area was the same.
Thus, two climate strategies that produced consistently different patterns of heating and cooling on the surface of the Earth would have different effects on the climate. What we need to know is whether the differences would be large enough to matter, and only detailed modelling can tell us that. Work has begun, and there is a long way to go before we can predict with confidence what will happen; but it is already becoming clear that SRM would have serious unintended consequences for the climate.
Earth’s climate is a highly complex system and consequently very difficult to model. Judgements over which effects to include and what approximations to make will differ from one research group to another. That’s why it is important to have several climate models rather than just one consensus simulation. When the different models make similar predictions, we can be far more confident of the result. Given the complexity of the climate and also of the models, it is not at all surprising that the different models disagree on how much the temperature will rise as the greenhouse gas concentration increases. On the other hand, that they all agree Earth will get warmer and — under a reasonably optimistic estimate of future carbon emissions — by no less than 2 °C, is a very robust result that we would be very ill advised to ignore.
Comparing the effects of limiting greenhouse gases on the one hand and reducing the incoming radiation on the other is even more challenging than modelling the effects of increasing CO2 level. Work has begun and because it is important to be able to compare the results from different models, much of it is being devoted to a project to find out how consistent and therefore how trustworthy the different models are .
Recently, an international team led by H Schmidt at Max Planck Institute for Meteorology in Hamburg compared four different climate models, one from the Institute itself the others from Hadley Centre in the UK, Institut Pierre Simon Laplace in France, and the Norwegian Meteorological Institute in Oslo respectively . To start each model, the CO2 level is set to four times what it was in the preindustrial era, and the solar constant – the amount of radiation reaching the surface of Earth – adjusted so that Earth’s mean temperature remains what it was in the preindustrial era. They then ran the models for 50 years.
As you would expect, the mean temperature averaged over the entire surface of the Earth remains roughly the same in all the models. On the other hand, the variation in temperature as we move north or south is reduced from the preindustrial. Given that, and bearing in mind the importance of temperature gradients in determining the weather, it is not surprising that the patterns of precipitation change. Rainfall is reduced on average over the entire planet, with strong effects over the Americas and northern Eurasia.
The total global cloud cover is also reduced in all the models. This contributes to the change in albedo (reflectivity) of the planet, which drops by about 2% in all four models. The models all predict a stronger effect in Europe but disagree on what would happen in large parts of the tropics or subtropics. Note that the reduction in albedo means that less of the sun’s radiation is reflected from Earth, so more particles or mirrors would be required to reduce the incoming radiation sufficiently to maintain the preindustrial mean global temperature.
There are of course many uncertainties in the calculations. The sudden quadrupling of CO2 is not realistic, though the rise to four times the preindustrial level is within the bounds of the current climate change models, albeit at the high end of the range of predictions. On the other hand, holding the mean global temperature constant is very much a best case scenario, and secondary effects such as decreased precipitation may well be underestimated.
An Earth with a high level of greenhouse gases and a geoengineering scheme that compensates by reducing the amount of incoming solar radiation is not the same as an Earth with lower level of greenhouse gases and no shield. The mean annual temperature averaged over the whole planet may be the same, but within that there will be changes, some quite marked. It is highly likely that total precipitation will be significantly (and unevenly) reduced, as will the total cloud cover. Beyond that, it is too early to say what will happen, which is all the more reason for being very cautious indeed about geoengineering.
It is obvious that if a geoengineering project goes wrong, the planet could be badly damaged. We are now discovering that there could be very harmful consequences even if it goes right.