Another way that predictions can be made more accurate is by ensemble computing. Computer models can be very temperamental, and slight changes can result in wildly different predicted outcomes. Ensembles allow lots of simultaneous scenarios to be calculated, which can be statistically assessed to see how likely on outcome is. That way, weather scientists will have an idea whether any given prediction equates to a freak occurrence or if it’s more likely to happen.
Climate science on the other hand can require a lot of computer time to simulate changes that happen over millennia – it might take months or years. Faster computers obviously shorten this time. In 2012, scientists announced that their petascale model of the Earth’s climate, going back tens of thousands of years, indicated that the Earth warmed on the last occasion when lots of CO2 filled the atmosphere. This is important, because the argument over whether increased atmospheric CO2 causes global warming, or global warming has historically led to more CO2 being leaked into the atmosphere, influences the position governments should take with regard to CO2-producing industries.