Modelling South Africa’s most capricious current system - the Agulhas

Stennis Space Center in Mississippi, USA

By Dr Bjorn Backeberg, Co-Director, Nansen-Tutu Centre for Marine Environmental Research

Thanks to support from SAEON and a grant from the US Office of Naval Research, I was fortunate to visit the Navy Research Laboratory (NRL) at the NASA Space Centre in Stennis, Mississippi in the USA to work with fellow marine environmental researchers James Richman and Jeff Book.

The visit was undertaken in my capacity as a co-investigator on the project "Key performance indicators of global operational models around southern Africa, in particular improving the understanding of model numerical sensitivity in the Agulhas System: a choke point affecting global climate". The project is led by Dr Juliet Hermes, Manager of SAEON’s Egagasini Node.

The purpose of the visit was twofold - to identify and scope the work that Kyle Cooper, an MSc student from the University of Cape Town (UCT) will be doing during his visit in early 2014 and to identify potential avenues of collaboration with NRL.

I was also able to use this opportunity to attend the GODAE (Global Ocean Data Assimilation Experiment) OceanView 2013 Symposium to present work being undertaken on modelling, data assimilation and re-analyses evaluations in the greater Agulhas Current system, including the work of the student on the project, Kyle Cooper.

Understanding southern Africa's unique oceans

The background to this project lies in the fact that the oceans surrounding southern Africa are unique in that an eastern and western boundary current system interact with each other. The region is characterised by intense mesoscale variability including anticyclonic and cyclonic eddies, the retroflection of the Agulhas Current and its interaction with the Benguela Upwelling system via eddies and filaments. These processes are not well understood and the ability to model this complex area is limited. Forecasting the system accurately remains a significant challenge.

It is essential that there should be a better understanding of the Agulhas System given its significant impact on South African weather, climate and resources. In particular, the Agulhas acts as a choke point affecting global climate, as highlighted in a recent paper published in Nature and in the Chapman Conference, held in South Africa in October 2012 (Beal, L., et al., 2011: Nature; the American Geophysical Union Chapman Conference, Spier, Stellenbosch 2012: The Agulhas System and Its Role in Changing Ocean Circulation, Climate, and Marine Ecosystems).

The role of ocean models

In the absence of a coherent observing system, ocean models play an important role in advancing our understanding of this complicated ocean system. However, due to the non-linear nature of the ocean dynamics in the Agulhas system, providing realistic simulations of the region is a challenging task. In fact, very few models accurately simulate the mesoscale dynamics associated with the Agulhas Current, and in particular its retroflection.

A commonality among ocean models of the Agulhas is a tendency toward an early retroflection with too many and regular anticyclonic eddies drifting into the South Atlantic, along a narrowly constrained trajectory. Global models have, in the past, been too coarse to resolve this dynamic system. However, new state-of-the-art forecast models are of high enough resolution, and their ability to accurately resolve regional dynamics needs to be evaluated.

Addressing the Global Change Grand Challenge

Operational oceanography, in particular ocean modelling and forecasting, has recently been highlighted as an area of importance in South Africa (see the South African Department of Science and Technology's Global Change Grand Challenge), but there is very limited capacity for this in South Africa. There are currently a number of global forecasting ocean models available to the South African research community but, as yet, none of these models have been evaluated in the oceans around South Africa. This analysis will help develop the modelling skills within South Africa as well assisting in global numerical modelling programmes by identifying how well these models work around South Africa.

Work being carried out in South Africa on the Hybrid Coordinate Ocean Model (HYCOM) involves understanding the sensitivity of the model to different numerics in the Agulhas region. A regional model of HYCOM for the Agulhas has been developed and research with the regional model has shown that the fourth order momentum advection scheme significantly improves the simulation compared with the "standard" second order scheme more commonly used (http://www.ocean-sci.net/5/173/2009/os-5-173-2009.html). Work is currently underway in South Africa to implement data assimilation in the regional model.

A key aspect of the reason for this trip has been to build collaborations with NRL in order to share both assimilated and non-assimilated HYCOM experiences and build our understanding of the ability for the NRL operational model to resolve the Agulhas. The model output will be made available to the South African team who are developing methods of intermodal and observational comparisons for the region.

Topics for future collaboration

While in the United States, I gave two presentations at NRL. The first of these was entitled "Numerical sensitivity and data assimilation experiment(s) in the Agulhas system", and the second was a wrap-up presentation suggesting the following topics for future collaboration: How to reduce upstream eddies in model simulations of the Agulhas; At what resolution do the deep currents impact the upper ocean stearing; and What impact do tides have on the mesoscale variability in the region?

In addition, I presented two posters at the GODAE OceanView 2013 Symposium entitled "Assimilating along-track SLA data using the EnOI in an eddy resolving model of the Agulhas system" and "Developing key performance indicators for global operational models around Southern Africa". The latter was on behalf of Kyle Cooper.

Appreciation and thanks are given to Augustus Vogel and the US Office of Naval Research for the grant, and to SAEON for the support and opportunity. I would also like to thank Jim Richman for hosting me for my stay in Mississippi and the other NRL colleagues for their support and discussions.