Register      Login
Marine and Freshwater Research Marine and Freshwater Research Society
Advances in the aquatic sciences
Shopping Cart: ( empty )
You are here: Home > Journals > MF > MF04218
RESEARCH ARTICLE
Contents Vol 56(5)

Development of a robust classifier of freshwater residence in barramundi (Lates calcarifer) life histories using elemental ratios in scales and boosted regression trees

Mike Cappo A C , Glenn De’ath A , Steve Boyle A , James Aumend A , Roland Olbrich A , Frank Hoedt A , Colton Perna B and Gregg Brunskill A
+ Author Affiliations
- Author Affiliations

A Australian Institute of Marine Science, PMB 3, Townsville MC, Qld 4810, Australia.

B Australian Centre for Tropical Freshwater Research, James Cook University, Townsville, Qld 4811, Australia.

C Corresponding author. Email: mcappo@aims.gov.au

Marine and Freshwater Research 56(5) 713-723 https://doi.org/10.1071/MF04218
Submitted: 13 August 2004  Accepted: 7 March 2005   Published: 24 July 2005

Abstract

Field and experimental studies showed that solution-based analysis of scales could be used to discriminate the long-term freshwater residents in the coastal fishery for catadromous barramundi. A new, robust classification technique was developed using boosted regression trees (MART) and its performance was compared with traditional linear discriminant analysis (LDA). The non-parametric MART had errors 33–81% less than LDA, and could account for non-linear relationships and interactions among elemental ratios. The best model used Sr : Ca, Ba : Ca, Fe : Ca and Mn : Ca in scales as predictors of salinity regime. Analysis of scales collected repeatedly from sub-adult fish of known environmental history showed the MART classifier could identify fish of freshwater origin until at least 10 months residence in seawater, and possibly several years, but scale growth rate could affect the temporal stability of the classifier after that time. The experiment indicated an approximate fourfold rise in Sr : Ba ratios in new scale margins, which were strongly classified by the MART as coming from saltwater fish, but inner scale sections of the same scales were still correctly classified as coming from freshwater fish. We conclude that solution-based elemental analyses of whole scales, and also annuli within scales, could offer a cost-effective, non-destructive technique to help understand the mechanisms causing enhanced year-class strength following high freshwater outflows.

Extra keywords: environmental flows, fish scales, geochemical markers, statistical discrimination.


Acknowledgments

We wish to thank the numerous anglers and researchers who provided scales and fish carcasses and assisted with electrofishing operations. Staff of Argus security and the AIMS Mariculture Facility helped feed and maintain captive fish, and J. Williams, M. Ireland and C. Syms assisted greatly with laboratory work and ANOVA. We are especially grateful to the generosity of Burdekin residents for access to private property.


References

Belanger, S. E. , Cherry, D. S. , Ney, T. J. , and Whitehurst, D. K. (1987). Differentiation of freshwater versus saltwater striped bass by elemental scale analysis. Transactions of the American Fisheries Society 116, 594–600.
| Crossref | GoogleScholarGoogle Scholar | Breiman L., Friedman J. H., Olshen R. A., and Stone C. G. (1984). ‘Classification and Regression Trees.’ (Wadsworth International Group: Belmont, CA.)

Coffey, M. , Dehairs, F. , Collette, O. , Luther, G. , Church, T. , and Jickells, T. (1997). The behaviour of dissolved Barium in estuaries. Estuarine, Coastal and Shelf Science 45, 113–121.
| Crossref | GoogleScholarGoogle Scholar | Freund Y., and Schapire R. (1996). Experiments with a new boosting algorithm. In ‘Machine Learning: Proceedings of the Thirteenth International Conference’. pp. 148–156. (Morgan Kauffman: San Francisco, CA.)

Friedman, J. H. (2001). Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics 29, 1189–1232.
| Crossref | GoogleScholarGoogle Scholar | Hastie T. J., Tibshirani R. J., and Friedman J. H. (2001). ‘The Elements of Statistical Learning.’ (Springer-Verlag: New York.)

Li, Y. , and Chan, L. (1979). Desorption of Ba and 226Ra from river-borne sediments in the Hudson Estuary. Earth and Planetary Science Letters 43, 343–350.
| Crossref | GoogleScholarGoogle Scholar | R Development Core Team (2004). ‘R: A Language and Environment for Statistical Computing.’ (R Foundation for Statistical Computing: Vienna, Austria.) Available at http://www.R-project.org, verified June 2005.

Ridd, P. V. , and Stieglitz, T. (2002). Dry season salinity changes in tropical mangrove and saltflat fringed estuaries. Estuarine Coastal and Shelf Science 54, 1039–1049.
| Crossref | GoogleScholarGoogle Scholar | Ripley B. D. (1996). ‘Pattern Recognition and Neural Networks.’ (Cambridge University Press: Cambridge.)

Russell D. J., and Rimmer M. A. (1997). Assessment of stock enhancement of barramundi Lates calcarifer (Bloch) in a coastal river system in far northern Queensland, Australia. In ‘Developing and Sustaining World Fisheries Resources. The State of Science and Management. Second World Fisheries Congress’. (Eds D. A. Hancock, D. C. Smith, A. Grant and J. P. Beumer.) pp. 498–503. (CSIRO Publishing: Melbourne.)

Sire, J. Y. , and Akimenko, M. A. (2004). Scale development in fish: a review, with description of sonic hedgehog (shh) expression in the zebrafish (Danio rerio). The International Journal of Developmental Biology 48, 233–248.
| Crossref | GoogleScholarGoogle Scholar | PubMed |

Staunton-Smith, J. , Robins, J. B. , Mayer, D. G. , Sellin, M. , and Halliday, I. A. (2004). Does the quantity and timing of fresh water flowing into a dry tropical estuary affect year-class strength of barramundi (Lates calcarifer)? Marine and Freshwater Research 55, 787–797.
| Crossref | GoogleScholarGoogle Scholar |

Stone, M. (1974). Cross-validatory choice and assessment of statistical predictions (with discussions). Biometrika 64, 29–35.


Wells, B. K. , Bath, G. E. , Thorrold, S. R. , and Jones, C. M. (2000a). Incorporation of strontium, cadmium, and barium in juvenile spot (Leiostomus xanthurus) scales reflects water chemistry. Canadian Journal of Fisheries and Aquatic Sciences 57, 2122–2129.
| Crossref | GoogleScholarGoogle Scholar |

Wells, B. K. , Thorrold, S. R. , and Jones, C. M. (2000b). Geographic variation in trace element composition of juvenile weakfish scales. Transactions of the American Fisheries Society 129, 889–900.
| Crossref | GoogleScholarGoogle Scholar |

Wells, B. K. , Rieman, B. E. , Clayton, J. L. , Horan, D. L. , and Jones, C. M. (2003a). Relationships between water, otolith, and scale chemistries of westslope cutthroat trout from the Coeur d'Alene River, Idaho: the potential application of hard-part chemistry to describe movements in freshwater. Transactions of the American Fisheries Society 132, 409–424.
| Crossref | GoogleScholarGoogle Scholar |

Wells, B. K. , Friedland, K. D. , and Clarke, L. M. (2003b). Increment patterns in otoliths and scales from mature Atlantic salmon Salmo salar. Marine Ecology Progress Series 262, 293–298.


Wells, B. K. , Thorrold, S. R. , and Jones, C. M. (2003c). Stability of elemental signatures in the scales of spawning weakfish, Cynoscion regalis. Canadian Journal of Fisheries and Aquatic Sciences 60, 361–369.
| Crossref | GoogleScholarGoogle Scholar |