Predicting the oceanic mesoscale dynamics in the Australian region

Download files
Access & Terms of Use
open access
Copyright: Woodham, Robert
Altmetric
Abstract
The skill and error growth characteristics of oceanographic forecasts produced by the Australian 'BLUElink' system are investigated. The principal variable investigated is Sea Surface Height (SSH), because it is the primary indicator of the mesoscale dynamics; however Sea Surface Temperature (SST) and sub-surface temperatures are also investigated. Forecast performance is assessed using simple statistical metrics, and model skill is measured relative to persistence and climatology. Verifying data is a behind real time BLUElink analysis, constrained by observations unavailable to the forecast. The East Australian Current separation and eddy-shedding region is identified as the most challenging to forecast in the Australian region, and is analysed in detail. The performance along the continental shelf is also described. In the study region, the forecast baseline for SSH (SST) is established as persistence for the first 20 days (12 days), then climatology thereafter. Relative to this baseline, BLUElink forecasts in this region show skill throughout the 6-day forecast period. The distribution of variability between different space and timescales is investigated. Complex Empirical Orthogonal Function (CEOF) analyses are used to investigate eddy properties and their relationships with baroclinic processes. The temporal and spatial distribution of forecast error shows that a skilful forecast is much more likely than one with no skill. A particular failure mode related to eddy mergers is identified, with two distinct events leading to poor performance during the study period. The initial error at real time, attributed to the initialisation procedure and the latency of the observational data, is shown to contribute a significant portion of the total forecast error for the system studied. These findings justify the research and development applied to initialisation. Observations unavailable to the BLUElink system show that there is skill in forecasting SSH along the continental shelf. BLUElink forecasts capture Coastally-Trapped Waves (CTWs) particularly well. Contrary to earlier studies, it is shown, using a CEOF analysis, that the majority of the CTW variance propagates as continuous features between the south-west and north-east. Although modulated by the shelf width, CTWs can be coherently tracked through sharply changing coastline orientations, the shallow Bass Strait and wind forcing regions.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Author(s)
Woodham, Robert
Supervisor(s)
Robertson, Robin
Brassington, Gary
Alves, Oscar
Kiss, Andrew
Creator(s)
Editor(s)
Translator(s)
Curator(s)
Designer(s)
Arranger(s)
Composer(s)
Recordist(s)
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
2013
Resource Type
Thesis
Degree Type
PhD Doctorate
UNSW Faculty
Files
download whole.pdf 19.66 MB Adobe Portable Document Format
Related dataset(s)