Short- to Medium-Range Ocean Forecasts: Delivery and Observational Requirements
Gary B. Brassington(1), Adrian Hines(2), Eric Dombrowsky(3), Shiro Ishizaki(4), Frank Bub(5), Mark Ignaszewsk(6)
Download this paper »
The Global Ocean Data Assimilation Experiment (GODAE) proposed in OceanObs'99 has proved so successful at demonstrating the practical utility of ocean forecasting that many research systems have been implemented as "operational" systems. GODAE has demonstrated that the global ocean observing system over the past decade can constrain an eddy-resolving ocean model to achieve skilful nowcasts and forecasts. In turn, these services are supporting a rapidly growing user community and a wide range of applications.
Mesoscale ocean circulation is a frontier science. Ocean eddies have been found to be ubiquitous over the entire ocean, accounting for more than 50% of the dynamic height variability. Evidence has recently emerged that geostrophic turbulence can also be systematic and contribute to the mean circulation. Further, model simulations indicate that eddies can extend into the abyssal depths supporting a complex range of stratified and un-stratified interactions that are difficult to substantiate with the present observing system. Intensive observational campaigns are required to resolve hypotheses posed by both theory and models.
At present, there is no redundancy in the observing system for ocean forecasting, all observations available in real-time can be demonstrated to have a positive impact. However, the impact is determined by the quality, quantity, independence and the statistical covariance of the observations to the model ocean state and circulation variables. Satellite altimetry has a dominant impact on global performance metrics due to its coverage and statistical covariance with the sub-surface ocean state. Satellite altimetry is therefore the highest priority platform for ocean forecasting. Improving the real-time coverage would improve both performance and robustness. Satellite SST (Sea Surface Temperature) has a high impact on the upper ocean state estimation but less direct impact at depth. Satellite SST provides an important constraint to a multi-variate analysis with altimetry. Maintenance of present multi-sensor, multi-platform satellite SST is a priority. The direct impact of in situ observations is high local to the observation but the overall impact is limited by the short time/space covariances of the ocean model state and the coverage relative to that from remote sensing. The likelihood that in situ observations are available local to a practical application is low at present coverage.
The optimality of ocean forecasting systems is an important property to keep in mind when specifying observational requirements. Less optimal systems by definition require more observations to achieve the same performance. Development in ocean forecasting is none the less constrained by the pace of scientific innovation and the computational resources available for this high-end application. Achieving the performance required by the user will require advances in all aspects of the system. It is however reasonable to claim that the present observing system under-samples the mesoscale variability.
Many important applications will be able to make progress through an incremental improvement to GOOS (Global Ocean Observing System) as it exists in 2009 such as fisheries management, marine park management and coupled weather forecasting. These applications either have more time to evaluate the forecasts, to improve decision making or have less sensitivity to the forecast errors. However, there are a set of applications for which there is an immediate need for a higher level of forecast performance (e.g., search and rescue, marine accident and emergency and defence). These applications are ad hoc and typically highly localised in time and in space. A practical alternative strategy to expanding GOOS by an order of magnitude to achieve this performance would be to locally and temporarily (over the period of an event) enhance the observing system. Developing practical observation deployment strategies that maximise performance while minimising cost as well as developing national capabilities represent a significant challenge over the next decade. Some suggested components to a rapid and dense observing system are outlined.
There is a range of applications that will have high national and commercial value, which will be supported by the expansion of national observing networks predominantly focused within exclusive economic zones. Maximising progress across the range of applications will require both national and international planning and coordination between GOOS and these national and commercial efforts.
Robustness and consistency of the forecast performance is essential to the delivery of quality services and achieving the maximum impact to applications. At present, the performance is acutely sensitive to changes to the observing system. Improving the integrity of the observing system and related downstream product delivery infrastructure will improve the robustness of existing ocean forecast systems. The global observing system, open data access policies, real-time reporting have all played roles in achieving rapid innovation and capacity building. Maintaining these ingredients and making strategic and targeted investments over the next decade will support both the next generation of systems as well as the current and future development of systems within the Asia, Africa and South America.
Correspondence should be addressed to E-mail: email@example.com
This paper shall be cited as:
Brassington, G. & Co-Authors (2010). "Short- to Medium-Range Ocean Forecasts: Delivery and Observational Requirements" in Proceedings of OceanObs’09: Sustained Ocean Observations and Information for Society (Vol. 1), Venice, Italy, 21-25 September 2009, Hall, J., Harrison, D.E. & Stammer, D., Eds., ESA Publication WPP-306, doi:10.5270/OceanObs09.pp.08
Rights to reproduction of individual articles are held by the authors. The source of the article (these proceedings) shall be cited.