Find Out More


ROADMAP aims to expand current understanding of how the Northern Hemisphere ocean surface state and ocean dynamics influence the extratropical atmospheric circulation, as well as associated impact-relevant weather and climate extremes, across space and time scales, short synoptic to decadal-planetary, under both present day and future climate conditions.


Project Portfolio

Specifically, ROADMAP will address:

  1. the impact of the ocean circulation, especially the Atlantic Meridional Overturning Circulation, on large-scale sea surface temperature (SST) patterns [WP1];
  2. the changing modes of variability of the Northern Hemisphere western boundary current extensions and what novel ocean-eddy resolving climate models can say about their future evolution [WP1];
  3. how and on which time scales extratropical ocean-atmosphere interactions control the tropospheric eddy-driven jets, cyclone variability (storm track), blocking events and the associated dynamical link to extreme conditions; and how such controls can be modified by global warming [WP2];
  4. the impact of tropical El Niño Southern Oscillation and Madden Julian Oscillation SST anomalies on the mid-latitude and polar atmospheric circulation [WP3];
  5. the multidecadal links between tropical and subtropical North Atlantic, and inter-basin connections between the Atlantic and the Pacific Oceans, as well as modifications of linkages under climate change conditions [WP3];
  6. the role of the Northern Hemisphere ocean surface state (SST and sea ice) for driving impact relevant atmospheric extremes, such as atmospheric and marine heat waves and droughts, including compound weather extremes and Mediterranean mesoscale cyclones; both large-scale natural variability modes and climate-change induced anomalies will be considered [WP4];
  7. the identification of key spatial-temporal variability patterns as well as cross-scale causal coupling between different variability modes of ocean and atmosphere [WP5 jointly with WP2 and WP3].

ROADMAP will exploit the wealth of simulations recently produced in other (international) research activities, such as CMIP6 and single-model large ensemble (~100 members) of simulations as well as frontier ocean-eddy resolving simulations from H2020 project PRIMAVERA. Existing simulations will be complemented with dedicated sensitivity experiments, encompassing cutting-edge numerical modelling techniques, such as pacemaker experiments based on data assimilation and interactive ensemble modelling. The sensitivity experiments will partly employ very high-resolution atmospheric grid configurations. Novel observational indices will be developed to investigate yet poorly understood historical ocean circulation variability. Analysis will be based on new advanced dynamical and statistical methods as well as novel approaches from the field of machine learning designed to infer complex, non-linear relationships. Analysis will also be performed in a multi-model framework, crucial for assessing the robustness of the results. Key results achieved will be disseminated to the scientific, stakeholder and climate service community as well as the general public.

The ROADMAP consortium encompasses leading climate research institutions from seven European countries, including universities as well as institutions providing (national) meteorological and climate services. ROADMAP will continue a long-standing history of international collaboration between its partners within the framework of previous joined projects, making significant contributions to climate variability, predictability and response, as well as climate extremes, particularly in the North Atlantic/European sector.

ROADMAP will address the following four key open questions:

Q1: How will future changes in northern hemisphere ocean circulation—western boundary currents and the Atlantic meridional overturning circulation—influence SST fronts and large-scale SST variability patterns?

While the ocean is warming globally, this warming has not been uniform in the last decades. This warming “hole” has been associated with the weakening of the AMOC. The processes behind these changes are still unknown, and whether they are driven by changes in the atmosphere, cryosphere or a result of ocean circulation changes is a topic of ongoing scientific research. These broad-scale changes in ocean warming and salinity have knock on consequences for the ocean –atmosphere interface.

Understanding the origin of these ocean surface changes is critical, for knowledge purposed and from the perspective of predicting and constraining future projections.

Western boundary currents (WBCs) conduit heat and available moisture polewards in all ocean basins and are intimately linked with the existence of storm tracks in certain regions. The frontal regions in the zonal extensions of these currents exhibit both broad meridional shifts and periods of intense eddying, with associated strong, mesoscale SST fronts. Increased eddying and broadening of WBCs have a distinct impact on the atmospheric storm track. Changes in the Gulf Stream ongoing in the Atlantic since 2005 are coincident with a weakening of the AMOC. Whether this timing is coincidental is an important question in understanding the future fate of WBCs.

Q2: How does mid-latitude ocean–atmosphere interaction influence the northern hemisphere tropospheric eddy-driven jet, the stratospheric polar night jet and atmospheric blocking on seasonal and longer timescales, and how the interaction is affected by global warming?

Variability of the tropospheric eddy-driven jet (TEDJ), the stratospheric polar night jet (SPNJ) and blocking are important for densely populated areas in Eurasia and North America as they affect the variability of storm tracks (cyclones), weather regimes and extremes.

The dynamics of TEDJ, SPNJ, and blocking involve atmospheric eddies, which are in turn strongly influenced by the lower atmospheric boundary conditions. The TEDJ, SPNJ and blocking provide a dynamical framework to understand the interactions between the ocean and different aspects of the atmospheric circulation including extremes. However, the driving role of the ocean versus atmospheric dynamics is not fully understood, and the interplay between the influence of external radiative forcing and oceanic forcing complicates the matter.

Predictive skill has been recently demonstrated in large-ensemble seasonal and interannual predictions of the North Atlantic Oscillation (NAO), which variability is linked to the TEDJ, the SPNJ, and blocking, but the ocean-induced dynamical processes behind the significant NAO predictability present a fundamental, open scientific question. It is still unclear to what extend the tropical and extratropical SST-forcing, stratospheric, Arctic sea-ice and external natural and anthropogenic radiative forcing may play a role. It is also unclear, which other extratropical atmospheric modes of variability and weather regimes linked to the TEDJ can be predicted outside the NAO.

Q3: How do tropical-extratropical atmosphere-ocean interactions and inter-basin teleconnections impact the Northern hemisphere tropospheric eddy-driven jet, stratospheric polar night jet and atmospheric blocking on seasonal and longer-timescales? Will the interactions and teleconnections change in the future?

Tropical-extratropical teleconnections and inter-basin linkages at seasonal and longer time scales are important issues to be tackled from a climate point of view because long-term variability often results from a change in frequency of specific events occurring at shorter time scales. There is evidence that the Madden-Julian Oscillation (MJO) has an impact on the atmospheric circulation in mid- and high-latitude, influencing the position of the Pacific and Atlantic eddy-driven jets, the North Atlantic Oscillation and the occurrence of blockings. Tropospheric and stratospheric pathways have been discussed to explain the influence of El Niño Southern Oscillation (ENSO) and MJO onto the midlatitudes. In the Atlantic Ocean, the main mode of variability, known as the Atlantic Multidecadal Variability (AMV) shows alternating basin-wide anomalies at the multi-decadal time scale. Although the oceanic advection associated to the overturning circulation and subpolar gyre explains the presence of the mid-latitude subpolar anomalies, the underlying processes associated to the tropical part of the AMV are however not well known. Not known is indeed if they result from the atmospheric forcing or from oceanic processes.

Q4: How will changes in the major oceanic and atmospheric patterns impact atmospheric and marine extremes in the tropics and extra-tropics, including extra-tropical cyclones?

Recent decades have seen a strongly non-uniform warming of the ocean, which caused changes to both atmospheric extremes (AE, e.g. heat waves or floods) and marine extremes (ME, e.g. marine heat waves). There is rising evidence that some extremes are synergetic, but substantial uncertainties remain in understanding the drivers of compound events. Understanding how single/compound AE or ME respond to a change in surface climate could reveal potential for enhancing sub-seasonal to decadal predictions of extreme events as well as assessing the risks and potential damages of anthropogenic climate change.

ROADMAP will investigate observations, atmospheric and oceanic reanalysis data together with already available or planned model simulations from CMIP5, CMIP6, EU-Project PRIMAVERA/BlueAction, MPIM Grand Ensemble [73] and DCPP experiments.

Compared to observations, models provide longer time series and physically consistent sets of atmospheric and oceanic parameters without temporal and spatial gaps.

However, the caveat of analysing the observations and the existing climate model simulations listed above is that it is impossible to explicitly extract the climate response to a specific perturbation in a particular region and to a specific process. Therefore, in order to ‘strengthen our understanding of climate variability and extremes resulting from the interactions with the oceans’ and deconstruct the responsible physical processes, ROADMAP will perform dedicated sensitivity experiments with a hierarchy of state-of-art climate models.

ROADMAP novel experimental investigation methods:

- Global coupled climate simulation with an eddy-resolving nest ocean model (1/10°) over the North Atlantic to study the propagation of subsurface heat and salt anomalies by lagrangian methods (WP1).

- Frontier HighResMIP/PRIMAVERA eddy-resolving simulations to assess the changes in WBC position and eddying (WP1). This class of models exhibit much reduced bias in circulation and surface properties.

- Novel advanced statistical-dynamical analysis (such causal effect networks and finite-amplitude local wave activity (FALW) theory) and cutting-edge numerical modelling techniques (such as interactive ensemble modelling, pacemaker experiments in a 20th reanalysis frame where SST is assimilated) to better understand the role of oceanic processes within the Atlantic and Pacific basins, including sea-ice, in driving low-frequency atmospheric variability (WP2, WP3).

- Pacemaker experiment (wind stress perturbation), analysis with advanced blind source separation (BSS) methods and cross-scale causality inference to assess the role of tropical-extratropical and inter-basin large-scale linkages on key atmospheric circulation patterns which are closely related to climate extremes (WP3, WP5).

- Atmosphere-only simulations with nudged stratosphere to assess the relative importance of the tropospheric vs. stratospheric pathways in the tropical-extratropical linkages (WP3).

The current national ROADMAP proposal will contribute to the above-mentioned study of the propagation of subsurface heat and salt anomalies by lagrangian methods over the North Atlantic (WP1). Moreover, ROADMAP will rely on:

- Pattern selection techniques
Both state-of-the-art climate observations and model outputs constitute examples for big data, within which the spatio-temporal patterns of essential climate variables like SST, SLP, surface temperature and precipitation that reflect key processes and mechanisms are not always apparent from the very beginning. In order to identify such patterns, advanced blind source separation (BSS) methods will be used (e.g., Projection Pursuit, Independent Component Analysis (ICA), conditional ICA, Independent Subspace Analysis (ISA), all of which are based on information-theoretic measures like multivariate mutual information (MI), joint negentropy and cross high-order cumulants), which constitute examples for the growing number of novel approaches from the field of machine learning and artificial intelligence.

- Multivariate interactions
Complex, nonlinear, multivariate interactions like triadic wave resonances allow efficient energy transfer across time and space scales in geophysical fluid dynamics and result in asymmetric, non-Gaussian and fat-tailed pdfs of climate variability modes leading to enhanced extremes, both in frequency and intensity. Such interactions will be quantified in terms of cumulants (e.g., skewness and kurtosis) and their spectral decomposition (bispectrum) as well as their generalizations. Multivariate interactions among the full set of variability modes (beyond wave resonances), will by studied using the interaction information (IT), along with robust tests to discard false artificial dependencies. IT will be diagnosed under different climate forcings and for different formats of the output provided by the other WPs, thus identifying sources of synergetic predictability. In addition, we will decompose relevant time series into scale-specific amplitude and phase components, the statistical association and causality among which will be quantified by information theoretic approaches.


Kick-Off Meeting


Scheduled Events

Soon available

Media Outreach

Soon available


Ana Russo