#03 2023

Learning operational coastal modelling from the best

By Giles Fearon, Senior Operational Ocean Modeller, Egagasini Node, NRF-SAEON

Claude-Louis Navier and George Gabriel Stokes would be amazed to see how their Navier-Stokes equations, penned almost 200 years ago, have come to life in the ocean models of today. They can produce truly beautiful patterns of flow, made even more beautiful by the fact that they depict the reality of the ocean. Well, almost. 

Ocean models will never be truly perfect representations of reality. As the old saying goes, “all models are wrong, but some are useful”. In the spirit of making ocean models as useful as possible, Jennifer Veitch (numerical ocean modeller at the Egagasini Node) and I were lucky enough to attend the ninth meeting of the Coastal Ocean and Shelf-Seas Task Team (COSS-TT) of OceanPredict, held in Montreal from 2 to 4 May 2023.

OceanPredict is a network of international scientists concerned with the development of operational ocean forecast systems, which can be thought of as weather forecast systems, but for the ocean. COSS-TT is a task team within OceanPredict with a specific focus on coastal and regional seas. For me, this is the centre of the bullseye in terms of my personal interests.

The majority of human marine activities take place in coastal areas, so this is where the societal benefits of ocean models are most directly realised.

The Navier-Stokes equations in action around Southern Africa, as revealed by a snapshot of sea surface temperature from a global ocean model (https://marine.copernicus.eu/access-data/myocean-viewer). The model shows the warm and energetic Agulhas Current flowing down our east coast, and the cold Benguela Upwelling System on our west coast.

Operational ocean modelling within SAEON

Jenny and I both presented at the meeting, shedding light on some of SAEON’s operational ocean modelling developments currently underway in support of the Department of Forestry, Fisheries and the Environment’s Ocean and Coastal Information Management System (OCIMS).

The COSS-TT meeting was for me the first time engaging with international scientists working in the field of operational ocean modelling. The meeting attendees boasted experts from across the globe, however Jenny and I were the only representatives from Africa, making our perspective somewhat anomalous. Our unique perspective lies in the fact that our efforts are still in their infancy, and are relatively under-resourced, while many of our international partners have had large teams of scientists, engineers and developers working on their systems for 25 years. While this may make us a small fish in a big ocean, starting late comes with certain advantages.

The philosophy of OCIMS has been to build systems which address specific stakeholder needs, rather than building a system as imagined by scientists but which no one wants to use. This is a true strength of our strategy which was noted by the meeting attendees.

Another advantage of starting late is that we can build on the myriad tools and data products which are now freely available, which just were not around 25 years ago. So, although we are relatively under-resourced, we have been able to build tailor-made and optimised models which can really make a tangible impact.

An example of this is our Algoa Bay forecast model – a light-weight system designed to be easily redeployed on any infrastructure (https://somisana.ac.za/explore/1). The model provides high-resolution ocean forecasts for five days into the future at hourly intervals, in a region where it is needed most.

A potential oil spill from the ongoing offshore oil bunkering (ship-to-ship refuelling) operations in the bay is a constant risk to this ecologically sensitive area, and we are currently developing the functionality to use our model to predict the trajectories of potential oil spills to mitigate against the ecological impacts.

Take-home message 

Although the COSS-TT meeting was truly international, the in-person attendees numbered only about 30, making it small enough to facilitate open and interesting discussion within the group. This allowed us to make some important and extremely helpful contacts and stimulated many ideas on how best to proceed with the development of our ocean forecast systems.

The take-home message for me was that we should continue to focus our limited resources on what would be most beneficial to our stakeholders. This can be easier said than done, when as scientists we are naturally drawn to what we find most interesting. These conversations were therefore important in sharpening my thinking on some aspects of potential development. For example, experts in the assimilation of observations into ocean models provided useful insight into what we should consider before embarking on such a (long and apparently arduous!) journey.

Other interactions showed that the accuracy of our forecast information could be more easily enhanced by developing weighted ensemble means from a handful of carefully designed forecast simulations. Another potential focus area for maximising our impact involves downscaling our models further into port and estuary scales, possibly requiring flexible mesh models.

As an operational ocean modeller at SAEON, this is truly an exciting time to be plying one’s trade. There are so many interesting areas on which to focus, with the added benefit that one’s efforts will be of some value to society.

I wonder if Claude-Louis Navier and George Gabriel Stokes ever thought that their equations would end up becoming so useful, in addition to being so beautiful?

COSS-TT members and participants in the 9th COSS-TT meeting in Montreal, May 2023.