eCite Digital Repository

Evaluation of hydrodynamic ocean models as a first step in larval dispersal modelling


Vasile, R and Hartmann, K and Hobday, AJ and Oliver, E and Tracey, S, Evaluation of hydrodynamic ocean models as a first step in larval dispersal modelling, Continental Shelf Research, 152 pp. 38-49. ISSN 0278-4343 (2018) [Refereed Article]

Copyright Statement

Copyright 2017 Elsevier Ltd.

DOI: doi:10.1016/j.csr.2017.11.001


Larval dispersal modelling, a powerful tool in studying population connectivity and species distribution, requires accurate estimates of the ocean state, on a high-resolution grid in both space (e.g. 0.5–1 km horizontal grid) and time (e.g. hourly outputs), particularly of current velocities and water temperature. These estimates are usually provided by hydrodynamic models based on which larval trajectories and survival are computed. In this study we assessed the accuracy of two hydrodynamic models around Australia – Bluelink ReANalysis (BRAN) and Hybrid Coordinate Ocean Model (HYCOM) – through comparison with empirical data from the Australian National Moorings Network (ANMN). We evaluated the models’ predictions of seawater parameters most relevant to larval dispersal – temperature, u and v velocities and current speed and direction - on the continental shelf where spawning and nursery areas for major fishery species are located. The performance of each model in estimating ocean parameters was found to depend on the parameter investigated and to vary from one geographical region to another. Both BRAN and HYCOM models systematically overestimated the mean water temperature, particularly in the top 140 m of water column, with over 2 °C bias at some of the mooring stations. HYCOM model was more accurate than BRAN for water temperature predictions in the Great Australian Bight and along the east coast of Australia. Skill scores between each model and the in situ observations showed lower accuracy in the models’ predictions of u and v ocean current velocities compared to water temperature predictions. For both models, the lowest accuracy in predicting ocean current velocities, speed and direction was observed at 200 m depth. Low accuracy of both model predictions was also observed in the top 10 m of the water column. BRAN had more accurate predictions of both u and v velocities in the upper 50 m of water column at all mooring station locations. While HYCOM predictions of ocean current speed were generally more accurate than BRAN, BRAN predictions of both ocean current speed and direction were more accurate than HYCOM along the southeast coast of Australia and Tasmania. This study identified important inaccuracies in the hydrodynamic models’ estimations of the real ocean parameters and on time scales relevant to larval dispersal studies. These findings highlight the importance of the choice and validation of hydrodynamic models, and calls for estimates of such bias to be incorporated in dispersal studies.

Item Details

Item Type:Refereed Article
Keywords:hydrodynamic model, larval dispersal, BRAN, HYCOM, IMOS, ANMN, Australia, continental shelf
Research Division:Earth Sciences
Research Group:Oceanography
Research Field:Biological oceanography
Objective Division:Environmental Management
Objective Group:Coastal and estuarine systems and management
Objective Field:Assessment and management of coastal and estuarine ecosystems
UTAS Author:Vasile, R (Ms Roxana Vasile)
UTAS Author:Hartmann, K (Dr Klaas Hartmann)
UTAS Author:Oliver, E (Dr Eric Oliver)
UTAS Author:Tracey, S (Associate Professor Sean Tracey)
ID Code:122736
Year Published:2018
Funding Support:Australian Research Council (LP120200164)
Web of Science® Times Cited:1
Deposited By:Sustainable Marine Research Collaboration
Deposited On:2017-11-24
Last Modified:2019-02-21

Repository Staff Only: item control page