High resolution global modelling
Using numerical ocean models, we simulate the global ocean circulation from the surface to the abyss. Despite the increasing wealth of observations from ships, satellites, floats and moorings (e.g. Argo, Aviso, TAO, RAPID) observational coverage is still sparse for vast parts of the deep ocean. However, with the increasing exploitation of the deep seas it is important that we understand ocean currents and their variability not just at the surface, but at all depths. To achieve this goal numerical ocean models are a valuable tool.
The spatial resolution of current climate models is typically 100 km or more for both the ocean and the atmosphere. To predict climatic impacts in smaller regions this is not sufficient. High resolution ocean models such as the ones run and developed at NOC will form the ocean component of the next generation of UK coupled climate models that are likely to provide improved regional forecasts.
In order to be able to reproduce the observed position, strength and variability of ocean currents such as the Gulf Stream or the Kuroshio, ocean models need a high enough spatial resolution (Figure 1 below).
Our global ocean simulations are currently run at resolutions of 1/4° and 1/12° in longitude and latitude, i.e. the smallest scales that are resolved are about 25 by 25 km and 8 by 8 km, respectively. At such resolutions oceanic mesoscale eddies can develop in our model simulations (Figure 2 below).
Satellite observations have shown that mesocale eddies are ubiquitous in the World's oceans. Regions of particularly high eddy activity are the Southern Ocean, especially at the northern border of the Antarctic Circumpolar Current (ACC) or the extensions of western boundary currents after their separation from the continent (e.g. Gulf Stream, Kuroshio). Oceanic eddies are the equivalent of weather systems in the atmosphere: the time and location of their occurrence depend on initial conditions, i.e. in high resolution simulations (eddy-permitting or better) small changes in e.g. temperature and salinity at a given time eventually lead to a different evolution of the ocean eddy field (just as numerical weather forecasts with perturbed initial conditions typically start to diverge after a few days).
The impact of mesocale eddies on the large-scale ocean circulation is far from being fully understood and is the subject of ongoing research in the high resolution global ocean modelling group at NOC.
Model development and improvement
Physical oceanography is an active field and new knowledge from observations and theoretical studies requires that we constantly improve and update our models to stay at the forefront of ocean science. Of particular importance is the validation of our model against ocean observations (link to ocean observation group). Differences between simulations and observations often point to model weaknesses that we need to address.
The high resolution ocean modelling group at NOC is part of a core team of institutions in France, the UK, and Italy, who are responsible for the development of the NEMO (Nucleus for European Modelling of the Ocean) model, which is widely used in European research centres as well as in operational oceanography (e.g. MetOffice - FOAM, MERCATOR). The development of NEMO at NOC benefits from the extensive experience that the modelling team has acquired during the development of the OCCAM model.
Currently, we are undertaking our first steps with a 1/12° version of NEMO. At this resolution the model simulates mesoscale eddies and the narrow boundary currents well. Animations from the first few years of integration can be found here (link to animations: spinup of eddies and boundary currents).
The 1/4° version of NEMO forms the ocean component of the coupled ocean-atmosphere model HadGEM3-H. HadGEM3-H is the latest generation of climate models in the UK, developed jointly between NERC and the MetOffice in the framework of the Joint Weather and Climate Research Programme (JWRCP) .
The 1/4° NEMO model also forms a physical framework into which biogeochemical models are incorporated (Ocean Biogeochemical Modelling group). However, even a resolution of 1/12° is not sufficient for a good representation of coastal environments. Nonetheless, model output from our global ocean simulations can be used as an offshore boundary condition for shelf models that are able to resolve coastal processes (link to webpage of coastal processes/impacts group at NOCL).
Running global ocean models such as NEMO at a high resolution requires high performance computing (HPC) facilities. An in house cluster at NOC allows us to perform development runs locally. However, for longer simulations we rely on HECToR, the UK National supercomputing Service (http://www.hector.ac.uk).