South African operational oceanography ::

capabilities & aspirations

Björn Backeberg (NTC, UCT), Mike Roberts (DEA) and many others…

GOVST-IV :: 5-9 November 2012, Rio de Janeiro

Outline • Are there plans in SA for a national centre responsible for operational

marine forecasting?

• How tight are (or would be) the links between research institutes and the above operational centre?

• Is there investment into highly educated personnel (oceanographers, mathematicians), supercomputing facilities and data service technology

• What are the local and international partnerships and the local user needs?

• Background :: OceanSAfrica initiative

• Regional in-situ observations network

• Numerical modeling capabilities

• Data assimilation plan (aspirations)

• Achievements & answers to the questions.

DEA Lead + UCT Lead +

Conceptualised during a workshop in 2009.

Multi-institutional initiative.

Develop operational oceanography activities in South Africa.

Facilitate communication between institutions & activities.

OceanSAfrica :: Aims

• Provide a strategically-directed platform amongst African marine research institutes for the identification, development and showcasing of new capabilities in the fields of ocean modelling, earth observation, in situ observation and dissemination technologies.

• Develop vehicles to transfer these capabilities to operationally mandated marine agencies.

• Develop new national and regional human capacity and infrastructure in the marine science, technological and technical domains.

• Create mechanisms for focused and sustainable development of this expertise and infrastructure.

• Provide regional components for integration into international operational and scientific programmes and commitments, such as JCOMM, GMES and GMES-Africa, and appropriate GEO, GOOS and GODAE tasks.

Marine Remote Sensing Unit Lead – Council for Scientific and Industrial Research, NRE Earth Observation

• Focus :: Algorithm & product development, calibration/validation, curation and dissemination.

• Consolidating, expanding and adding value to existing national and regional facilities, such as FP7 DevCoCast/EAMNet and AMESD/MESA, to provide routine earth observation data and user training for marine and freshwater domains across Africa.

• Developing new algorithms and products for African shelf-sea, oceanic and freshwater users and contributing to sensor and algorithm validation programmes for ESA and NASA.

• Maximising calibration/validation contributions and uptake readiness for forthcoming earth observation missions, strongly focusing on the Sentinel series for marine & aquatic physics & biogeochemistry.

Existing operational EO capabilities ::

disseminating to users in Africa through EAMNet

NRE-EO already processes and provides > 10 daily products to all sub-Saharan Africa through web and GeoNetCast and has archive >70 TB multi-sensor data based at CHPC

www.csir.co.za

In situ observations Lead – Department of Environmental Affairs, Branch Oceans & Coasts

• Integrate existing coastal and ocean observations from various institutions.

• Consolidating, expanding and adding value to existing autonomous and ship based observational facilities, particularly through improved communication links and near real-time data access.

• Developing new scientific and technical capabilities to use currently available autonomous platforms and sensors, such as gliders, profiling floats and buoys.

• Developing new low-cost platforms and sensors to improve African in situ observational capabilities.

Regional in situ observations network

Example Existing DEA Capability 34 UTR sites throughout the SWIO 3 x coastal data buoys deployed at the moment in SA 1 x super buoy in 1000 m in Agulhas Current

SA developed buoys, ranging from deep oceanic through mid-size coastal to lightweight bio-optical, all providing real time multi-sensor capabilities More info @ www.cfoo.co.za

Delayed-time observations network

International research partners

Motivation:

Climatic importance of the Agulhas system

Natural laboratory for testing and developing new methodologies.

Ocean modeling Lead – Dept. Oceanography

• Developing South African and regional expertise and capacity in regional ocean modelling and data assimilation to monitor and predict the ocean state. Assessing regional ocean current and wave forecasting models for critical southern African shelf sea regions using the ROMS, HYCOM and Deltares platforms.

• First stage demonstration of data assimilation capabilities using HYCOM with the aim of generating long-term regional ocean hind- and forecasts with direct application to both operational forecasting and reanalyses, for a wide variety of users.

• Building capacity and regional expertise in numerical ocean modelling and operational forecasting through postgraduate/short course training and provision of modelling tools and code modules.

Regional modeling capabilities

Regional Ocean Modelling System (ROMS)

Hybrid Coordinate Ocean Model (HYCOM)

Focus :: Develop two different types of ocean models (ROMS & HYCOM) high resolution regional modelling.

Facilitate a multi-model approach to simulating and studying the regional oceanography, including the impact of different numerics on the solution.

More info @ www.simocean.org.za

CPU resources in South Africa

Centre for High Performance Computing National facility

Harpertown: 48 nodes: Sun Blade X6250 (2 x Intel Xeon quad core 3.0 Ghz/16GB RAM). Compute blade (48 nodes) running SLES 10 64 bit. Nehalem: 288 nodes Sun Blade X6275 (2 x Intel Xeon quad core 2.93 Ghz/12GB RAM). Compute blades (288 nodes) running SLES 10 64 bit.

University of Cape Town Facility

CORE: 26 computation nodes (64 bit Opteron 2x6, 2.86 GHz), managed by the queue submission system Sun Grid Engine (SGE)

Sensitivity of numerical models

Barnier et al., 2006 Backeberg et al., 2009

Model simulations are sensitive to numerical choices.

In the Agulhas this is due to the high levels of oceanic turbulence.

Models are not able to accurately represent the retroflection.

Comparison of model vorticities with two different momentum schemes: 2nd vs. 4th order approximations

Eddy kinetic energy in different models

:: Proposed 3 year plan :: HYCOM & data assimilation

• Sensitivity simulation experiments and validation

▫ Adjusting the reference densities in HYCOM to a range more suitable for the Agulhas region.

▫ Experimenting with the ratio of σ : ρ (sigma : isopycnal) coordinates.

▫ Test the new 4th order momentum advection switch.

• Produce a hindcast simulation for the static ensemble to be used in the data assimilation experiments.

• Implement data assimilation (EnOI) in the Agulhas

▫ Assimilate along-track SLA & Ostia SSTs

• Produce a reanalysis simulation for the altimeter era (1993 − present)

• Produce assimilated forecasts for the region.

Develop an assimilation system for

Observing System Experiments &

Observing System Simulation Experiments

:: Long-term aspiration ::

Ch

all

en

ge

s:

min

imse

wea

ther

rel

ate

d i

nci

den

ts

Im

pro

ve

sea

rch

an

d r

escu

e

South African Marine Safety Agency

9 centres incl. Centre for Sea Watch

and Response

Ch

all

en

ge

s:

min

imse

wea

ther

rel

ate

d i

nci

den

ts

Im

pro

ve

sea

rch

an

d r

escu

e

South African Marine Safety Agency

9 centres incl. Centre for Sea Watch

and Response

Ch

all

en

ge

s:

min

imse

wea

ther

rel

ate

d i

nci

den

ts

Im

pro

ve

sea

rch

an

d r

escu

e

South African Marine Safety Agency

9 centres incl. Centre for Sea Watch

and Response

Forecasts

1. Waves (including freak waves) 2. Currents 3. Busters (coastal lows) 4. Tides 5. Storm surge/wave setup 6. Specific location info: Oil

&Gas, harbours etc. 7. Lagrangian products on

demand

A potential user of South African operational oceanography

Achievements since 2009 • Developed channel of communication between institutes (pillars) involved

in the various aspects of ocean monitoring and modelling.

Answering the questions

• Are there plans in SA for a national centre responsible for operational marine forecasting?

➞Yes : South African Weather Service / Dept Environmental Affairs

• How tight are (or would be) the links between research institutes and the above operational centre?

→ Very. R&D capabilities located at academic institutes

• Is there investment into highly educated personnel (oceanographers, mathematicians), supercomputing facilities and data service technology

➞Yes, but not explicitly for operational activities.

• What are the local and international partnerships and the local user needs?

➞Established some international partnerships in terms of capacity building

➞ Local users / needs: SAMSA, … others?