Home | Contact ST  

Feature Article

Using Ocean Forecast Data To Improve Sonar Range Prediction
Harnessing Advances in Operational Oceanography For Anti-Submarine Warfare and Other Naval Applications

By Cmdr. Robert Woodham
Director of Oceanography and Meteorology
Royal Australian Navy
Sydney, Australia

and
Jarrad Exelby
Science Team Leader, Anti-Submarine Warfare Modelling and Simulation
Maritime Operations Division
Defence Science and Technology Organisation
Edinburgh, Australia



Operational oceanography is going through a period of very rapid development around the world. Supported by an ever-increasing volume of observations, numerical models have begun to routinely forecast the ocean’s physical properties, such as temperature, salinity, current and sea surface height. These models are capable of resolving such features as mesoscale eddies, surface mixed layers, upwelling events, oceanic fronts and coastally trapped waves. As resolutions increase, dynamic processes in the complex inshore environment are starting to be effectively modeled.

The main driver for this ongoing revolution seems to be an increasing awareness of the role of the global ocean in climate change and the health of the planet in general. But naval forces are also keenly aware of the importance of understanding and predicting the environment in which they might have to fight. This fits into the broader emphasis in the military on information superiority, or network-centric warfare (NCW), in which actionable information is rapidly collected and disseminated throughout the battlespace using the latest technologies. Environmental information, including meteorological and oceanographic (METOC) information, is regarded by modern navies as a vital component of information superiority and NCW, allowing naval forces to optimize their weapons and sensors and maneuver to take advantage of both the prevailing and the forecast environmental conditions.

The complexity and variability of the underwater environment means that predicting the performance of anti-submarine warfare (ASW) sonar systems is a challenging problem. Despite this complexity, maritime forces have traditionally relied on climatologies or point observations as the basis for sonar range prediction.

Climatologies are based on time-averaged (often monthly) observed data, and they are therefore unable to represent dynamics on shorter timescales. In the worst case, bimodal dynamics (such as inside or outside an eddy, or north or south of a front) can be represented as a mean of the two, a state which never occurs in the real ocean. Mixed-layer climatologies can under-represent the mixed-layer depth, since the mean of two temperature profiles has a mixed-layer depth equal to the shallower of the two mixed layers, not their mean. Temperature profile point observations, usually made using expendable bathythermographs, are valid only for the time and place at which they were made. These can quickly become irrelevant as the ocean evolves, the ship moves away from the location of the observation or the operator becomes interested in sonar contacts at larger ranges in potentially different ocean conditions.

ROAM forecast for Australia’s Tasman Sea on September 27, showing sea surface temperature (in Celsius) and surface currents (in meters per second), displayed in the Fleet METOC viewer.

But these shortcomings are often poorly understood by naval personnel, as evidenced by the common use of the term “range of the day” and the doctrinal concepts of predicted sonar range and tactical sonar range, each represented by a single value. This reveals an ignorance of the reality that sonar ranges vary with azimuth and can change markedly on much shorter timescales, even from ping to ping.

Furthermore, the complexity of the undersea environment can result in a rapidly changing and cluttered active sonar display, which complicates interpretation and hinders the classification of contacts.

However, the recent advent of ocean forecasting models offers a source of high-resolution oceanic data that is able to capture the spatial and temporal variability of the ocean. This data can then be used as input to improve the performance of sonar range prediction systems and assist sonar teams in understanding the undersea environment.

For the Royal Australian Navy (RAN), this is particularly timely, given the Australian government’s emphasis on ASW in its 2009 Defence White Paper. As part of this effort, the RAN has developed new tools to help harness the possibilities that new ocean forecasting models offer in ASW operations.

Example TESS 2 sonar range prediction for the Tasman Sea on September 27, computed using 3D ROAM data. The sonar parameters used are fictional. (Figure courtesy of Thales Australia)

BLUElink Ocean Forecasting System
The Australian BLUElink ocean forecasting system went live in August 2007. As currently implemented at the Australian Bureau of Meteorology (BOM), it runs twice per week, producing forecasts out to six days. Although the system has global coverage, it uses a telescoping grid geometry, which provides its maximum horizontal resolution of 10 kilometers in the region bounded by 90° east and 180° east longitude and 16° north to 75° south latitude. The vertical resolution is 10 meters in the top 200 meters of the ocean and is somewhat coarser at greater depths.

The BLUElink system’s main components are: a data assimilation system, which uses an ensemble optimal interpolation method to generate a snapshot of the ocean’s state from a range of observations; the forecasting model itself, which is based on the U.S. Geophysical Fluid Dynamics Laboratory’s Modular Ocean Model Version 4; and various data management systems, which provide observations and surface forcings from the BOM’s global atmospheric model and also manage model output. This output consists of daily means of temperature, salinity, sea surface height and current.

It is worth noting that the physical oceanography around Australia is particularly complex. The East Australian Current and Leeuwin Current affect the eastern and western seaboards respectively, while the Pacific-Indonesian Throughflow influences the Timor and Arafura Seas and the Antarctic Circumpolar Current dominates waters to the south of the Australian continent.

Other oceanographic phenomena in the region include upwelling events, internal waves, solitons, extreme tidal ranges and abundant freshwater inflows. The BLUElink system has made insights into all of these phenomena, not only through its routine forecasts, but also as a result of various “reanalysis” model runs. These reanalyses are generated using the data assimilation system to ingest archived, historical ocean observations, while being forced with reanalyzed surface fluxes from an atmospheric model.

Example concept demonstrator screenshot, showing PPI of probability of detection overlaid on bathymetry.

A Limited-Area Ocean Model
The BLUElink project has also delivered a limited-area model, the Relocatable Ocean-Atmosphere Model (ROAM), to the RAN. This model is capable of being set up anywhere in the high-resolution domain of the BLUElink model, and it can generate forecasts at resolutions down to one kilometer. Although the atmospheric and oceanic model components of the system are not coupled, this is a likely future direction.

One important feature of ROAM is the ease with which model runs can be set up and executed. This is achieved using a relatively simple user interface, which allows RAN forecasters, who do not have an in-depth understanding of the component models, to set up model runs quickly and easily.

BLUElink and ROAM data is visualized by the RAN in a geographic information system using an extension called Fleet METOC Viewer. This software can work with a range of atmospheric and oceanographic data in Network Common Data Form (netCDF).

Not only is netCDF the output data format used by the BLUElink and ROAM models, but it is also widely used for environmental data by agencies around the world.


Sonar Range Prediction
The horizontal resolutions of BLUElink and ROAM provide ample detail over spatial scales typical of ASW areas of operation. Where conditions and the characteristics of the target allow, resolutions of this order are also sufficient to provide a number of grid points over typical sonar ranges. For these reasons, the RAN has recently developed a version of its Tactical Environmental Support System Version 2 (TESS 2) sonar range prediction software that is able to ingest BLUElink and ROAM data as the basis for spatially dependent predictions. This means that the full complexity of the physical oceanographic environment is used as the basis for sonar performance predictions, greatly improving decision-making during the planning and execution of ASW operations.


Improved Sonar Display
A concept demonstrator has been developed at the Defence Science and Technology Organisation (DSTO) that focuses on assisting the operator in interpreting the sonar picture through distinguishing echoes and potential targets from environmental clutter. It does this by providing the operator with more accurate and timely modeled sonar performance and environmental data.

This demonstrator has been successfully trialed using sonar data from both an active towed array system and hull-mounted sonar systems, and it has shown improvement compared with climatology or point-observed in-situ measurements.

The demonstrator currently operates on a stand-alone laptop and can be used during at-sea exercises or in the laboratory for post-trial analysis. Plans are under way to incorporate the environmental and modeled information into the real-time displays that sonar operators view as an overlay on the conventional sonar data display picture. This would fuse information that would otherwise be presented in disparate ways in order to help operators interpret the sonar picture. For example, by overlaying a sonar plan position indicator (PPI) display with modeled probability of detection graphics and bathymetry, active sonar returns can easily be correlated with bottom features such as canyon walls, seamounts, slopes, wrecks or pipelines. The operator can thus assess the likelihood of a sonar return being a target of interest and can use the range prediction modeling to identify areas of higher probability of detection, which may lead to more alerted detection opportunities. Alternatively, areas of lower probability of detection would warrant a more careful and thorough search.

This approach avoids the complication of trying to reconcile sonar information from a dedicated sonar display (which might present echoes on a Cartesian azimuth/range display) with topographic information from a separate nautical chart (either in paper copy or displayed on its own dedicated computer system) and also with sonar performance predictions from a third system.

In addition, the concept demonstrator displays other useful information, such as propagation-loss graphs, modeled probability of detection in range and depth, sediment type, system settings and ship data (such as ship’s heading, speed and latitude/longitude).

These displays can be automatically updated in real time using live feeds of position, speed, heading, sonar depth and transmit mode information, thus avoiding the need to manually run the sonar performance model offline.


Conclusions
The recent availability of high-resolution forecasts of the physical oceanic environment opens up exciting new opportunities for improved sonar range prediction. It also offers vastly increased spatial awareness in the complex undersea operating environment.

Like other navies around the world, the RAN has been a driving force behind operational oceanography in Australia. This approach is necessary in order to achieve the Australian government’s vision of improved ASW capability and also because the physical oceanography of the Australian region is exceptionally complex.

As well as using oceanic data from the global BLUElink forecasting system, the RAN has implemented its own limited-area, high-resolution ocean-atmosphere model, which can be set up over areas of interest quickly and easily. It has also taken steps to incorporate this newly available, high-resolution oceanographic data into its sonar range prediction models in order to improve the planning and execution of ASW operations—and undersea warfare in general.

Furthermore, DSTO has developed a concept demonstrator designed to assist in the classification of sonar contacts, enabling submarines to be prosecuted—and false targets to be disregarded—much more quickly, ensuring the success of ASW operations.


Acknowledgments
The authors wish to thank Lt. Amy Bulters of the RAN for her kind assistance with Figures 1 and 2.


References
For a full list of references, please contact Robert Woodham at robert.woodham@defence.gov.au.



Cmdr. Robert Woodham has 20 years of naval experience, including 15 as a meteorology and oceanography specialist. His experiences include anti-submarine warfare operations in the North Atlantic and am­phibious exercises in the Mediterranean and Coral seas. He holds an M.Sc. in oceanography and is the Royal Australian Navy’s director of oceanography and meteorology.

Jarrad Exelby completed a B.Sc. (Hons.) in meteorology and oceanography at Flinders University, South Australia, in 1995. He then joined the Maritime Operations Division of the Defence Science and Technology Organisation, where he works on scientific and operational aspects of anti-submarine warfare sonar performance modeling, simulation and capability analysis.


-back to top-

-back to to Features Index-

Sea Technology is read worldwide in more than 110 countries by management, engineers, scientists and technical personnel working in industry, government and educational research institutions. Readers are involved with oceanographic research, fisheries management, offshore oil and gas exploration and production, undersea defense including antisubmarine warfare, ocean mining and commercial diving.