Climate change forecasts used to set policy and billions of pounds in investment are flawed, according to new research from Lancaster University Management School (LUMS).
Complex climate models have been used by scientists to reach a consensus (through the International Panel on Climate Change, the IPCC) of global warming of 0.2 °C per decade. But this fundamental finding for governments and the global population continues to be fiercely contested by sceptics of the role of human activity in climate change.
The competing interest groups involved have led to a decline in confidence generally in the wake of claims of manipulated data from the University of East Anglia, and incorrect projections – such as Himalayan glaciers disappearing by 2035 .
The new study by Robert Fildes and Nikolaos Kourentzes at the Lancaster Centre for Forecasting applies the latest thinking on forecasting to the work of climate change scientists, in a bid to make 10 and 20 year ahead climate predictions more accurate and trustworthy for policy-makers, and help address growing doubts over the realities of climate change.
Such decadal forecasts have the most relevance to current thinking and policy plans and if they are to be credible and useful, they need to demonstrate their accuracy. But the forecasts produced by the current models do not achieve this.
The authors set out a new basis for ‘decadal’ forecasting which is to be a major component of the next IPCC assessment report. Using a combination of models, with statistical benchmarking as checks, current forecasts prove almost certainly less accurate than they could be. Inaccurate climate forecasts costs the world considerable money.
The implication is that the climate modelling community needs to open up its research agenda.
As yet it has not demonstrated that it can produce better forecasts than simpler statistical methods.
A consequence of this, explored by Fildes and Kourentzes, is that the overwhelming focus on limiting green house gases alone may well be mis-guided.
The hydrologist Keith Beven’s work on modelling carried out in the Lancaster Environment Centre leads to the same conclusion.
In short, eclectic forecasting methods and a wide range of policy responses are what is needed if we are to overcome the problems of emerging warming.