Community Research and Development Information Service - CORDIS

Final Report Summary - STEPS (Strategies toward enhancing prediction of climate and its impacts)

Climate affects our daily lives and changes in climate can have large impacts on society and economy. It is thus of great interest to predict climate. However, our ability to foresee changes in climate beyond a season and in most regions around the globe remains limited. Model error and poor understanding of the mechanisms for climate variability are two key limiting factors. STEPS addressed both.

STEPS contributed to the construction of a climate super model. This is a novel approach to reduce model error and improve climate prediction. A super model consists of a set of models that are combined so that they synchronise with each other and individual model errors compensate. In this way the super model is superior to any of the individual models. As a first ever demonstration of the potential of super modelling climate, we constructed a super model from two versions of one climate model by combining them only at the ocean-atmosphere interface. The super model showed marked improvement in the simulation of Tropical Pacific climate that was superior to the individual models. In collaboration with scientists from Germany, we further demonstrated the power of the approach by developing a super model with higher model resolution and an even better simulation of the Tropical Pacific. Greater improvements in super climate modelling may be achieved by combining models with larger differences and by increasing synchronization. Scientific exchange facilitated by STEPS between researchers in the USA and Norway on super climate modelling led to new ideas on enhancing synchronization of climate models, and contributed to generic software for combining different model components. This work is now being extended with a Marie Currie Fellowship awarded to Gregory Duane (a scientist from the USA visiting the University of Bergen), and an ERC Consolidator grant awarded to Noel Keenlyside that aims to create a super model from three state-of-the-art European Climate models.

STEPS has advanced understanding of Northern Hemisphere climate variability and of uncertainties in its simulation and prediction. STEPS research work extended from the Atlantic to the Pacific, covered timescales from intra-seasonal to decadal, and included climate models and analysis of paleo-proxy data. There were five main achievements. (1) A multi-proxy index for Atlantic multidecadal variability (AMV) was constructed using five corral records from the Tropical Atlantic. The index is the first marine based multi-proxy reconstruction for AMV. It extends 90 years prior to the instrumental record and demonstrates the persistence of AMV. (2) Coordinated multi-model analysis identified consistent features of simulated AMV, such as the impact of ocean circulation on sea surface temperature (SST), and identified where major uncertainties remain, such as the role of salinity. Further analysis of individual models increased understanding of the role of salinity and ocean-atmosphere interaction in AMV. (3) Additional modelling studies showed how random atmospheric forcing can generate multiple timescales of oceanic variability, and how ocean-atmosphere variability can be enhanced by proper representation of stratosphere-troposphere interaction. (4) Through idealised coupled model experiments we have demonstrated how Atlantic SST may influence Pacific variability, and how Tropical Atlantic SST variations could enhance ENSO prediction skill. Furthermore, we showed how model error can impact the simulation of Tropical Atlantic climate. (5) An innovative modelling approach was developed to better represent ocean-atmosphere interaction, leading to a major improvement in the simulation of intra-seasonal climate variability in the Indo-Pacific region.

STEPS also investigated climate impacts of socio-economic importance in three different areas. Firstly, the predictability of the Kiremt (boreal summer) rainfall in Ethiopia was investigated. While El Niño is known to impact Kiremt rainfall the mechanisms are not fully understood. Through detailed analysis of observations and targeted atmospheric modelling experiments, we showed how warm SST anomalies in the central Pacific impact the large-scale atmospheric circulation and reduce rainfall over Ethiopia. Results showed that up to 50% of seasonal rainfall variability over parts of the Ethiopian highlands is driven by SST variations over the Tropical Pacific. The link between the representation of this mechanism and the prediction skill of state-of-the-art climate models is now being investigated. Secondly, STEPS contributed to a better understanding of the factors controlling future changes in tropical cyclone (TC) activity on regional scales. Analysis of TC changes simulated by the atmospheric model ECHAM5 showed a dramatic 22% reduction in the southern hemisphere, compared to a 6 % reduction over the northern hemisphere. Large-scale circulation changes associated with a weakening of the tropical Walker Circulation were responsible for these regional changes. While the occurrence frequency of intermediate and weak storms decreased, the strongest storms were found to increase in frequency in both hemispheres. Third, we showed that global warming may cause the Madden-Jullian Oscillation – the dominant pattern of intra-seasonal climate variability – to increase in amplitude (by ~30%) and frequency. The coupling of dynamical and thermodynamic responses to global warming drove this intensification.

During the STEPS project, the PI helped organise two summer schools and two workshops where project results were disseminated. One summer school focused on the super modelling approach and the other on climate variability in the Tropical Atlantic. One workshop addressed climate predictability in the North Atlantic sector, and the other the predictability of climate extremes. Furthermore, a number of conference sessions were organised, and many presentations on project were given.

STEPS research has contributed to the development of a new and promising strategy – super modelling – to reduce model systematic errors. Furthermore, our work contributes to a better understanding of the mechanism for climate variability in the Atlantic and Pacific and how SST variations around the globe influence rainfall. This research may lead to better predictions of climate and its socio-economic impacts.

Related information

Reported by

UNIVERSITETET I BERGEN
Norway
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top