L.F. Borchert, V. Koul, M.B. Menary, D.J. Befort, D. Swingedouw, G. Sgubin, J. Mignot
Environmental Research Letters
Work Package 1
Work Package 5
Climate prediction on a decadal scale is very important for climate policy and planning, but chaotic variability within the climate system often reduces the skill of these predictions. This study assesses the decadal skill of the latest CMIP6 climate models at predicting the chaotic variability portion of European summer temperatures, as well as whether a link to North Atlantic temperatures, described here for the first time, can improve the predictions. The team found these models’ predictions are an improvement on those of older models, but they still rely on predicting external temperature forcing, with less skill for internal variability. The more-predictable variation in North Atlantic Ocean temperature can, however, be used to improve the European temperature predictions beyond the forcing. Improving our decadal climate predictions is a key part of informing effective climate policy, helping limit the impact of future climate change.
Predicting how the climate may change up to ten years in the future is of great importance to society, as changes on this timescale can have a high impact and may require relatively urgent action. However, decadal climate prediction is complicated by the influence of chaotic, internal climate variability in the climate system. This is the natural differences in the climate from year to year, as opposed to changes driven by external forcings such as climate change, which are more predictable. Some internal variability however, such as changes in the North Atlantic Ocean, can be predicted with some skill. Links between North Atlantic variability and European climate could help improve decadal European climate predictions. This study uses the latest generation of climate models, dubbed CMIP6, to examine whether ocean variations can be used to improve decadal-scale summer temperature predictions over Europe. Such a technique may help improve planning for near-future climate change and allow us to reduce its impact.
The team found that the skill of their CMIP6 models was improved compared to their predecessors. They were good at predicting European summer temperature thanks to predicting the temperature response to external forcing accurately. They were less good, however at predicting internal variability. This was addressed using simulations of North Atlantic temperature, which has more easily modelled internal variability. The team observed a correlation between temperature changes in the North Atlantic and Southern Europe. By using a dynamical-statistical modelling approach this correlation can be exploited, improving predictions of European temperature over the base model. These results highlight the importance of external forcing in the skill of decadal climate predictions, but also show how predictable internal variability can be used to improve our decadal predictions. Better representations of the links between different parts of the climate system could further improve our prediction of variables of interest. Such improvements in decadal predictions are a key part of informing effective climate policies, mitigating and adapting to the impacts of future climate change.
This paper used an ensemble of 8 CMIP6 model systems, each with 10 ensemble members. These produced initialised predictions, using contemporary observations to begin the simulations. These began every year after 1960 and then ran for 10 years, producing retrospective forecasts (called hindcasts). This study focussed on simulations in the period 1970-2015, so as to maximise the variety of simulations available by ensuring simulations of all lengths are included. Another model ensemble subjected to a standardised external forcing was used to isolate the internal variability in the simulations. Observational data from the HadCRUT5 and HadISST datasets, and the ERA5 reanalysis, are used to assess the skill of the model predictions. The dynamical-statistical approach relies on predictions of North Atlantic Ocean temperature in the CMIP6 models, from which changes in Southern European summer temperature are statistically inferred.
Understanding how to improve our predictions of climate change a decade ahead is of great importance to the plans being made today. Communities and policymakers can use improved climate predictions to formulate the best-informed policies to reduce the impact of future climate change, including adaptation and mitigation measures.
We assess the capability of decadal prediction simulations from the Coupled Model Intercomparison Project phase 6 (CMIP6) archive to predict European summer temperature during the period 1970-2014. Using a multi-model ensemble average, we show that Southern European (SEU) summer temperatures are highly predictable for up to 10 years in CMIP6. Much of this predictive skill, is related to the externally forced response: historical simulations explain about 90% of observed SEU summer temperature variance. Prediction skill for the unforced signal of SEU summer temperature is low: initialized model simulations explain less than 10% of observed variance after removing the externally forced response. An observed link between unforced SEU summer temperature and preceding spring Eastern North Atlantic – Mediterranean sea surface temperature (SST) motivates the application of a dynamical-statistical model to overcome the low summer temperature skill over Europe. This dynamical-statistical model uses dynamical spring SST predictions to predict European summer temperature, and significantly increases decadal prediction skill of unforced European summer temperature variations, showing significant prediction skill for unforced Southern European summer temperature 2-9 years ahead. As a result, dynamical-statistical models can benefit the decadal prediction of variables with initially limited skill beyond the forcing, such as summer temperature over Europe.