Written by Marian Wiltshire, Emlyn Jones (CSIRO) and Gary Brassington (BoM) for publication in Marine Matters.
Ocean conditions can be unpredictable. This unpredictability creates risk and uncertainty for industries and sectors that rely on the ocean, especially our maritime and naval industries.
Bluelink addresses this challenge by using a variety of observational data streams from numerous sources to create a comprehensive suite of ocean forecasts that predict all types of marine weather scenarios, from local beach conditions to oceanic interactions on a global scale.
Over the past two years Bluelink’s regional forecasting system (ROAM-Ocean), has been applied to more than 100 domains in the Australian region and performed over 3,900 individual model simulations. ROAM-Ocean is nested within the Ocean Model, Analysis and Prediction System (OceanMAPS) an operational global ocean model and assimilation system.
OceanMAPS is the global ocean forecasting component of Bluelink and is forced with atmospheric fields from the Australian Community Climate and Earth-System Simulator (ACCESS). To ensure accurate predictions of the physical state of the ocean, observations taken from the sea (in-situ) and the and space (via satellite) are processed and ingested into the prediction system via a process called data assimilation, to create accurate, near-real time estimates of circulation. More specifically, OceanMAPS assimilates multiple satellite platforms for sea surface temperature and altimetry as well as in situ platform (e.g., Argo, XBT, CTD). OceanMAPS is a Category 1 system operated by BoM, forming the backbone of the national ocean forecasting system producing daily forecasts out to seven days.
Both OceanMAPS and ROAM-Ocean are routinely assessed against withheld in-situ observations (e.g. gliders, moorings and radars). OceanMAPS forecasts are assessed against unassimilated reference observations from the global ocean observing system [e.g., Argo, drifting buoys (SST and currents) and Jason-altimetry] and compared with other international systems from the UK, France, Canada and the USA. The results are accessible via an internet service (http://18.104.22.168/monitoring/index.php?pg=class4_stats) in near real-time and summary charts are published annually. Each of the ROAM-Ocean simulations is automatically assessed against IMOS observations (including data from Satellite Sea Surface Temperature, Argo floats and Ocean Gliders) using software (EnKF-C, https://github.com/sakov/enkf-c) developed by Pavel Sakov (BoM).
A majority of the ROAM-Ocean assessments take place along the shelf off southern Western Australia and northern New South Wales, using the repeat transects from the IMOS Ocean Gliders (both the deep water Sea Gliders and the more coastal Slocum Gliders).
Of key interest to the Bluelink partners are upper ocean dynamics and our ability to predict their evolution in time. Such dynamics include (but are not limited to) the surface mixed layer (mixed layer depth and temperature gradients), surface currents and the position of fronts and eddys.
More recently, Bluelink started to use data collected from the IMOS Ocean Radar sites (Newcastle, Coffs and Perth), whilst this assessment is still in a preliminary phase, the data will be used to assess the modelled current velocities.
The IMOS Ocean Glider data is extremely important for Bluelink because there are very few other observations available to assess the 3D structure and temporal frequency of models in shelf areas. The vertical resolution provided by the Ocean Glider data is especially useful in this respect.
Whilst the IMOS Ocean Glider data is proving important, it is somewhat difficult to use as it contains a lot of sub-meso-scale features which are notoriously challenging to predict in a model. A model can often generate these sub-mesoscale features, but they may not be in the right place at the right time. The prediction of sub-meso scale features is an area of active research, and as observing systems generate larger quantities of high resolution data, and modelling systems increase in resolution, such features may become skillfully forecast.
The Bluelink models and forecasts are shared with Defence, marine planners, and maritime industry and safety authorities, to help guide their activities at sea and near shore.
With a better understanding of Australia’s oceans and beaches, innovative approaches can be created to understand the marine environment and support its sustainability, as well as boosting the efficiency and profitability of ocean-based industries and services.