But determining the precise changes which we can expect remains an ongoing challenge – a challenge which can only be addressed via the utilisation of highly sophisticated and well developed climate models.
The science of climate change prediction is based on a number of underlying sets of knowledge. Climate scientists use their understanding of the climate to describe the global system both physically and mathematically. Using this knowledge, climate models are built with computational code, and global simulations are performed on supercomputers.
The Earth’s climate system is composed of many interacting components. Climate models are constructed out of these individual system components, which include the atmosphere, the oceans, sea ice, ice sheets, and land vegetation systems including forests, farmland and steppes. Teams of researchers are constantly refining the physical and mathematical descriptions of these component parts, always with the aim of increasing the predictive accuracy of the models.
Before performing future simulations, all climate models are first assessed and validated for quality. This is conducted using a validation method whereby the simulated climate data is compared with modern-day observations from satellites and ground based measurements. Such system parameters as land temperature, sea surface temperature, rainfall rates and snow accumulation are checked between the model and reality. Generally there is a good agreement, and the models are capable of simulating our world quite realistically.
Once the present-day climate is validated by comparison with real world observations, the models can be used to predict the future, running simulations over decades and centuries under varying CO2 emission scenarios, evaluating the best and worst-case outcomes.
The very same climate models can also be used to simulate past climate environments (palaeoclimates), for example the last ice age (ending about 10,000 years ago), or the Eocene warm period (lasting from 56 to 34 million years ago). Again, the output model parameters – sea surface temperature for example – can be checked with temperature reconstructions from geological evidence. Among different models there is a wide spread in the accuracy of simulating particular past climate features, with some models performing better than others in predicting regional temperatures, usually reflecting particular model biases in present day simulations.
Ongoing climate model development
Like all fields of future prediction, climate change modelling and future Earth prediction is in a state of constant progression. Over the past two to three years for example, developments in ice sheet modelling are allowing the simulation of the melting ice caps on Greenland and Antarctica.
A 2015 research article in the journal Geophysical Research Letters documents one of the first studies to utilise a so-called dynamic ice sheet model, which has the capability of simulating melting ice caps and the subsequent runoff of fresh water into the oceans. The article, entitled ‘Response of Atlantic overturning to future warming in a coupled atmosphere-ocean-ice sheet model’, shows that melting ice has an effect of freshening the oceans, which in turn has the potential to trigger larger climate changes.
Lead author Paul Gierz, of the Alfred Wegener Institute in Germany, says, ‘in certain cases this melting can slow down or even shut off the ocean circulation, meaning large reductions in the strength of the Gulf Stream – a realistic projection for the future with potentially massive implications, especially for the European climate’.
Gierz says, ‘essentially what this implies is that by reducing the heat transport via the Gulf Stream, global warming over Europe is reduced, suggesting that climate models which do not adopt a dynamic ice sheet model have thus far overestimated the warming pattern across Europe and other parts of the Northern Hemisphere.’
Interestingly, the article also outlines the results of simulations extending as far as the year 2500. Two patterns appear to emerge: one high CO2 emission warming scenario where the the Gulf Stream is permanently weakened, and another weaker CO2 emission scenario where the stream weakens but eventually recovers after year 2300.
Gierz says, ‘in the more extreme scenario, with a permanently weakened ocean circulation, the global heat budget is seriously modified in the long-term future, with large implications for future weather and climate trends, affecting phenomena such as the frequency of storms and flood events across Europe’.
‘It appears that performing climate simulations using dynamic ice sheet components allows for increased model realism, advancing our forecasts for this coming century and beyond.’ he says.
But ice sheets are only part of the story. The development and innovation of other model components, from ocean processes to cloud parameterisations to forest changes, will also lead to improved models and future forecasts; forecasts capable of generating information which can help policy makers in their assessments of future risk and, ultimately, enable effective planning.
Dr. Conor Purcell is a postdoctoral researcher at Professor Jennifer McElwain’s Program for Experimental Atmospheres and Climate at University College Dublin. He specialises in future climate change prediction.