LONG-TERM CURRENT PREDICTION IN THE CENTRAL GREAT-BARRIER-REEF

Citation
Dm. Burrage et al., LONG-TERM CURRENT PREDICTION IN THE CENTRAL GREAT-BARRIER-REEF, Continental shelf research, 14(7-8), 1994, pp. 803-829
Citations number
21
Categorie Soggetti
Oceanografhy
Journal title
ISSN journal
02784343
Volume
14
Issue
7-8
Year of publication
1994
Pages
803 - 829
Database
ISI
SICI code
0278-4343(1994)14:7-8<803:LCPITC>2.0.ZU;2-E
Abstract
Long-term current records in the Great Barrier Reef region are needed to address major ecological problems such as periodic outbreaks of the coral predator, Crown of Thorns Starfish (COTS, Acanthaster sp.). In situ current meter data are sparsely distributed and typically availab le for deployments shorter than 1 year. A suite of linear systems mode ls has therefore been developed which allow low-frequency along shelf currents to be specified using readily available meteorological and oc eanographic forcing data. The models which are essentially statistical , nevertheless reflect our understanding of regional hydrodynamics. Us ing optimally lagged multi linear regression they allow predictions to be made quickly and economically from input time series and a few spe cified parameters. Models were calibrated using current meter mooring data obtained from a transect across the central Great Barrier Reef in 1985 and validated using data from similar deployments in 1987 and 19 90. The models which are simplest to implement, perform well and are s tatistically efficient are those based on the geostrophic across-shelf momentum balance. Using as inputs coastal sea levels or, when availab le, offshore sea level differences, they can precisely predict current s over time spans of up to 25 years. They accurately respond to fluctu ations at weather time scales and, when offshore differences are used, at seasonal and inter-annual scales. The predicted currents are being used to drive advective models of COTS larvae dispersal and to set bo undary conditions for more complex numerical hydrodynamic current simu lation models.