An ecological approach to re-establishing Australian freshwater populations: an application to in the Murrumbidgee catchment

Edited by

Brendan Ebner1, 2, Jason Thiem1, 2, Mark Lintermans1, 2 and Dean Gilligan3

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, .

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

3NSW Department of Primary Industries, Narrandera Centre, PO Box 182, Narrandera NSW 2700, Australia.

October 2006

2003/034 FRDC Report 2003/034

©Copyright Fisheries Research and Development Corporation and Parks, Conservation & Lands, 2006.

This work is copyright. Except as permitted under the Copyright Act 1968 (Cwlth), no part of this publication may be reproduced by any process, electronic or otherwise, without the specific written permission of the copyright owners. Neither may information be stored electronically in any form whatsoever without such permission.

ISBN 0 9775019 2 2

Preferred citation: Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D. (eds.) (2006). An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. Final Report to Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

Published by Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

DISCLAIMER The authors do not warrant that the information in this book is free from errors or omissions. The authors do not accept any form of liability, be it contractual, tortious or otherwise, for the contents of this book or for any consequences arising from its use or any reliance placed upon it. The information, opinions and advice contained in this book may not relate to, or be relevant to, a reader's particular circumstances. Opinions expressed by the authors are the individual opinions of those persons and are not necessarily those of the publisher or research provider.

An ecological approach to re-establishing Australian freshwater cod populations ii

FRDC Report 2003/034

TABLE OF CONTENTS

NON-TECHNICAL SUMMARY...... 1 ACKNOWLEDGMENTS ...... 4 CHAPTER 1 . BACKGROUND TO THE AUSTRALIAN FRESHWATER COD ...... 5 Distribution and declines in Australian freshwater cod...... 5 Current approaches to the conservation and recovery of Australian freshwater cod...... 8 Fate of fingerling stockings of trout cod...... 10 Radio-tracking as a tool to monitor dispersal ...... 12 Need ...... 13 OBJECTIVES...... 14 CHAPTER 2 . ESTIMATING RICHNESS AND CATCH PER UNIT EFFORT FROM BOAT ELECTRO- IN A LOWLAND RIVER IN TEMPERATE AUSTRALIA...... 15 INTRODUCTION ...... 15 METHODS ...... 16 Study site ...... 16 Fish sampling...... 17 Data analysis ...... 18 RESULTS ...... 19 Detection of species ...... 20 Catch per unit effort...... 21 Correlation of species within operations...... 22 Analysis of fish community structure amongst grouped samples ...... 23 Estimating species richness ...... 23 DISCUSSION...... 24 Estimating species richness at a site...... 24 Fish community distribution ...... 26 Catch per unit effort...... 26 Limitations of this study...... 27 Recommendations ...... 28 CHAPTER 3 . MOVEMENT OF HATCHERY-REARED AND WILD- REARED MACQUARIENSIS STOCKED IN A LARGE LOWLAND RIVER...... 29 INTRODUCTION ...... 29 METHODS ...... 31 Site description...... 31 Source of fish...... 32 Surgery and release ...... 32 Radio-telemetry...... 33 Data analysis ...... 34 RESULTS ...... 35 Dispersal...... 35 Homing in wild fish...... 39 iii An ecological approach to re-establishing Australian freshwater cod populations

FRDC Report 2003/034

Total range and home-range of individual fish ...... 40 DISCUSSION...... 43 Possible rejection of radio-tags...... 43 Survivorship ...... 44 Direction and magnitude of dispersal...... 44 Homing...... 45 Home-range use and dispersal ...... 45 Home-range establishment model...... 47 The dispersal model ...... 48 CONCLUSION...... 48 CHAPTER 4 . METHODS FOR RECORDING ACTIVITY OF AN ENDANGERED PERCICHTHYID: EFFECTS OF TEMPORAL RESOLUTION...... 50 INTRODUCTION ...... 50 METHODS ...... 51 Study site ...... 51 Fish collection and surgery ...... 51 Radio-telemetry...... 52 Data analysis ...... 53 RESULTS ...... 53 DISCUSSION...... 57 CHAPTER 5 . EXPERIENCED INDIVIDUALS: DESCRIBING FISH MOVEMENT AND IDENTIFYING UNDERLYING MECHANISMS AT THE AMONG HOME-RANGE SCALE...... 60 INTRODUCTION ...... 60 The Australian Percichthyidae as a case study ...... 61 DESCRIPTION OF MAHRS ...... 63 Empirical studies ...... 63 a) Migration...... 63 b) Home-range shift ...... 64 c) Exploration ...... 64 DESCRIPTIVE MODELS OF HOME-RANGE SHIFT ...... 65 Existing conceptual models...... 65 A new conceptual model ...... 66 TEMPORAL RESOLUTION...... 67 MECHANISMS UNDERPINNING MOVEMENT...... 69 Capability and causal factors ...... 69 EMPIRICAL STUDIES...... 69 Navigation...... 69 Decisions...... 71 MECHANISTIC MODELS...... 72 KNOWLEDGE GAPS...... 73 MOVEMENT, MECHANISMS AND MODELS...... 74 APPLICATION TO AUSTRALIAN PERCICHTHYID CONSERVATION...... 74 Conservation and MAHRS...... 74 Priorities ...... 74 Species protection and recovery ...... 75 Comprehensive description of movement ...... 75 An ecological approach to re-establishing Australian freshwater cod populations iv

FRDC Report 2003/034

Mechanisms for movement...... 76 FUTURE RESEARCH DIRECTIONS ...... 77 CHAPTER 6 . FATE OF TWO-YEAR OLD, HATCHERY-REARED MACCULLOCHELLA MACQUARIENSIS, STOCKED INTO TWO UPLAND RIVERS ...... 79 INTRODUCTION ...... 79 METHODS ...... 81 Site description...... 81 Surgery and implantation of radio-tags...... 81 Radio-tracking ...... 83 Data analysis ...... 83 Mortality, radio-tag rejection and missing fish ...... 84 RESULTS ...... 84 Movement and dispersal ...... 84 Mortality and radio-tag rejection ...... 85 DISCUSSION...... 90 Possible rejection of radio-tags...... 90 Dispersal...... 91 Mortality ...... 92 CHAPTER 7 . VIDEO MONITORING IN UPLAND STREAMS...... 95 INTRODUCTION ...... 95 METHODS ...... 96 Study site ...... 96 Study design ...... 96 Camera placement ...... 97 Processing of video footage...... 97 Presence of released fish ...... 99 RESULTS ...... 99 Released fish and radio-tracking data...... 99 Water quality...... 100 Camera Footage: Burkes Creek Crossing...... 100 Camera Footage: Spur Hole...... 101 DISCUSSION...... 104 CHAPTER 8 . FINAL DISCUSSION AND RECOMMENDATIONS...... 105 THE THREE MODELS ...... 105 Detection...... 105 Dispersal...... 106 Mortality ...... 106 Patterns of survivorship...... 107 CONSERVATION STOCKING OF MACCULLOCHELLA...... 109 Pre-release phase...... 109 Post-release phase ...... 109 RECOMMENDATIONS...... 111 BENEFITS AND ADOPTION...... 112 FURTHER DEVELOPMENT...... 113 PLANNED OUTCOMES...... 114 CONCLUSION...... 115 v An ecological approach to re-establishing Australian freshwater cod populations

FRDC Report 2003/034

REFERENCES...... 116 APPENDICES...... 140 APPENDIX 1 . INTELLECTUAL PROPERTY...... 140 APPENDIX 2 . STAFF...... 141 APPENDIX 3 . COMMUNICATION OUTPUTS...... 142 APPENDIX 4 . STANDARD SURVEY EFFORT INVOLVING BOAT ELECTRO-FISHING BY STATE FISHERIES AGENCIES IN THE MURRAY– DARLING BASIN...... 166 APPENDIX 5 . AERIAL TELEMETRY METHODOLOGY – A PILOT STUDY OVER THE COTTER CATCHMENT ...... 167 APPENDIX 6 . MOVEMENTS AND HABITAT USE OF WILD M. MACQUARIENSIS IN THE MURRUMBIDGEE RIVER (EXPERIMENT 1)....170 APPENDIX 7 . MOVEMENTS OF INDIVIDUAL C27 RELEASED IN THE COTTER RIVER...... 179 APPENDIX 8 . A PRELIMINARY PROTOCOL FOR USING UNDERWATER FILMING AS A TECHNIQUE FOR STUDYING RELEASES OF ...... 180

An ecological approach to re-establishing Australian freshwater cod populations vi

FRDC Report 2003/034

LIST OF FIGURES

Figure 1.1. Conceptual model of the fate of sub-adult M. macquariensis stocked as fingerlings...... 11 Figure 2.1. Study reach of the Murrumbidgee River near Narrandera, New South Wales, Australia...... 17 Figure 2.2. The presence of each species (■) detected in the study reach from each electro-fishing operation. Operations occurred from downstream (operation 1) to upstream (operation 157) in an almost continuous reach of river, except for three breaks that could not be navigated by boat involving 1, 0.5 and 1.5 km of river, respectively...... 21 Figure 2.3. The effect of sample size and re-sampling strategy on 95% confidence intervals of mean catch for (a) C. carpio, (b) M. macquariensis, (c) M. ambigua, (d) G. marmoratus, (e) R. semoni, (f) M. peelii peelii, and (g) B. . The upper and lower confidence intervals are plotted as consecutive (black) and random (grey) re-sampling strategies...... 22 Figure 2.4. The predicted species richness for the study reach based on a consecutive re-sampling strategy using three species richness estimators. The means (± SD) are calculated based on 1000 iterations (from n samples) and are represented by black and grey lines, respectively. The dashed line represents the total species richness detected in this study from 157 operations...... 23 Figure 2.5. The predicted species richness for the study reach based on a random re- sampling strategy using three species richness estimators. The means (± SD) are calculated based on 1000 iterations (from n samples) and are represented by black and grey lines, respectively. The dashed line represents the total species richness detected in this study from 157 operations...... 24 Figure 3.1. Location of the study reach along the Murrumbidgee River, New South Wales, Australia...... 31 Figure 3.2. Length-weight comparisons of hatchery reared (●) and wild (□) M. macquariensis implanted with radio-tags...... 34 Figure 3.3. Dispersal of hatchery (■) and wild (■) M. macquariensis groups in the Murrumbidgee River. Time after release: (a) 1 month, (b) 3 months, (c) 6 months, (d) 9 months, and (e) 12 months. Distance moved is expressed as upstream and downstream direction in 5 km categories...... 36 Figure 3.4. Minimum distances moved between radio-tracking intervals (as detected by manual radio-tracking) by all (a) wild and (b) hatchery M. macquariensis and, (c) the water discharge ( ) and temperature (--) in the Murrumbidgee River at Narrandera. Positive and negative values indicate a direction of movement upstream and downstream, respectively...... 37 Figure 3.5. Fate of radio-tagged (a) wild and, (b) hatchery M. macquariensis released into the Murrumbidgee River near Narrandera...... 39 Figure 3.6. Maccullochella macquariensis movements from October 2003 representing: (a) sedentary home-range behaviour for the entire study, following dispersal, (b) home-range relocations that occurred once or multiple times, (c) large-scale movements outside of home-ranges that did not result in a home- range relocation. The first point in each figure (zero) represents the release location (all fish were released at the same location)...... 42 Figure 3.7. A conceptual model categorising the movement of 29 M. macquariensis over the 13-month study. The number of individuals exhibiting specific emigration and home-range behaviour is shown. The occurrence of each vii An ecological approach to re-establishing Australian freshwater cod populations

FRDC Report 2003/034

behaviour is scored only once per individual...... 47 Figure 4.1. Location of the study reach in southern New South Wales, Australia...... 51 Figure 4.2. The effect of decreasing the temporal resolution of sampling on estimates of diel range (▲) and diel area (■), based on radio-tracking of Individual 5...... 55 Figure 4.3. The diel activity of nine individuals in the Murrumbidgee River based on the minimum distance moved between four hourly radio-tracking locations (mean ± S.E.). White and grey sections denote day and night, respectively. Data is grouped over two consecutive diel periods...... 56 Figure 4.4. Activity of Individual 5, based on continuous logging of signal-strength from its radio-transmitter, 14–19 November 2004. Data is grouped at 10 min intervals and plotted as the variation in radio-transmitter signal-strength. Open circles at the top of the graph denote boat activity (predominantly investigator related) within the reach. Day and night periods are represented in white and grey respectively...... 57 Figure 5.1. Home-range shift models expressed as flow diagrams: a) the conceptual model of Crook 2004a; b) the revised model of Crook 2004b...... 66 Figure 5.2. A conceptual decision model of fish movement...... 67 Figure 5.3. Hypothetical movement of an individual in: a) the first two years of life; b) the third year of life; and as observed by c) monthly radio-tracking during the third year of life; and d) by daily radio-tracking in a short-term study commencing at 2 years of age. Semi-permanent locations or home-ranges are labelled L, M, N and O, with each light grey oval representing a home-range, dark grey ovals represent core areas, and arrows indicate direction of movement. Temporary locations are represented by P, Q and R, and apparently at N in c). However, the latter is an artefact of radio-tracking on a coarse temporal resolution...... 70 Figure 6.1. Location of the release sites (▲) on the Cotter and Murrumbidgee rivers in the Australian Capital Territory, Australia...... 82 Figure 6.2. Dispersal of individuals released in the Cotter (black) and Murrumbidgee (grey) rivers at (a) 1 month, (b) 3 months, and (c) 6 months after release. Sample size at each time period is indicated...... 85 Figure 6.3. Fate of individuals released in (a) the Cotter River and, (b) the Murrumbidgee River...... 87 Figure 6.4. The terminal locations and number of individuals released in the Cotter River, represented as (a) the time from release until mortality or tag rejection and (b) the fate of individuals. Note: large distances moved probably resulted from transport by avian predators...... 89 Figure 6.5. The terminal locations and number of individuals released in the Murrumbidgee River, represented as (a) the time from release until mortality or tag rejection and (b) the fate of individuals. Note: distances moved may have resulted from transport by avian predation. The final location of a radio-tag recovered 84 km downstream from the release site is not shown...... 90 Figure 7.1. Location of the study sites (▲) on the Cotter River in the Australian Capital Territory, Australia...... 96 Figure 7.2. Camera mounted on a large tripod at the Spur Hole site on the Cotter River. (Photograph: Parks, Conservation and Lands)...... 98 Figure 7.3. The position of underwater cameras three and four (centre and bottom right of photo), recording gear under tarpaulin (top left of photo) and above water camera (top left) at the Spur Hole site on the Cotter River. (Photograph: An ecological approach to re-establishing Australian freshwater cod populations viii

FRDC Report 2003/034

Parks, Conservation and Lands) ...... 98 Figure 7.4. Schematic representation of the locations where cameras were placed (with corresponding camera number) and pool dimensions at a) Burkes Creek Crossing and b) Spur Hole sites (not to scale)...... 99 Figure 7.5. Discharge of the Cotter River at Vanitys Crossing (approximately 3 km downstream of Spur Hole) prior to, during (grey shading) and post underwater video monitoring...... 101 Figure 8.1. Monthly survivorship of M. macquariensis based on radio-tracking following release of: (a) wild fish (n=31); and (b) hatchery fish (n=27) in the Murrumbidgee River near Narrandera; (c) hatchery fish (n=36) in the Murrumbidgee River, ACT; (d) hatchery fish (n=36) in the Cotter River, ACT, in the current study; and (e) hatchery fish (n=9) in the Cotter River, ACT (from Ebner, Johnston & Lintermans, unpublished data)...... 108

LIST OF TABLES

Table 1.1. Summary of distribution and of Australian freshwater cod (from Lintermans et al., 2005b)...... 6 Table 1.2. General distribution, period of decline and current distribution of Australian freshwater cod (from Lintermans et al., 2005b)...... 8 Table 1.3. Stocking programs for freshwater cod species. (Modified and updated from Lintermans et al., 2005b)...... 9 Table 2.1. Summary of the fishes surveyed in the study (the number of caught and observed individuals was combined)...... 20 Table 2.2. Species-specific non-detection over 157 boat electro-fishing operations in a 13 km reach of the Murrumbidgee River...... 20 Table 3.1. Summary statistics for 29 wild M. macquariensis exhibiting different types of movements following the establishment of a home-range. TL and weight are expressed as mean ± SE...... 43 Table 4.1. The simulated reduction in accuracy associated with decreasing the temporal resolution of sampling on the range and area of river used (combined over two consecutive diel periods) by 10 individuals (based on four hourly radio- tracking)...... 54 Table 4.2. The estimated proportion of available river used by 10 individuals over two consecutive diel periods...... 56 Table 5.1. Australian members of the Percichthyidae, maximum age and longevity (from McDowall, 1996; Allen et al., 2002) and conservation status nationally..62 Table 6.1. The mean (± SE), total length (TL), weight and subsequent fate of radio- tagged individuals following release at two sites. Numbers of individuals are in parentheses...... 85 Table 6.2. The number of individuals suffering mortality, predation/scavenging or radio-tag rejection seven months after release. Numbers of individuals are given in parentheses...... 88 Table 7.1. Dates of filming for both Burkes Creek Crossing (control site) and Spur Hole (treatment site)...... 97 Table 7.2. Sub-samples of continuous footage viewed, including site, treatment date and time of day...... 99 Table 7.3. Number of M. macquariensis present at the Spur Hole during filming, confirmed by radio-tracking...... 100 ix An ecological approach to re-establishing Australian freshwater cod populations

FRDC Report 2003/034

Table 7.4. Evaluation of footage captured by each camera across all sampling periods at Burkes Creek Crossing on the Cotter River...... 102 Table 7.5. Evaluation of footage captured by each camera across all sampling periods at Spur Hole on the Cotter River...... 103

An ecological approach to re-establishing Australian freshwater cod populations x

FRDC Report 2003/034

NON-TECHNICAL SUMMARY

2003/034 Title: An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment

Principal Investigator: Mr Mark Lintermans Address: Parks, Conservation & Lands Department of Territory & Municipal Services ACT Government GPO Box 158 Canberra ACT 2601 Tel (02) 6207 2117 Fax (02) 6207 2122 Email: [email protected]

Co-Investigator: Mr Brendan Ebner Address: Parks, Conservation & Lands Department of Territory & Municipal Services ACT Government GPO Box 158 Canberra ACT 2601 Tel (02) 6207 2117 Fax (02) 6207 2122 Email: [email protected]

Dr Dean Gilligan NSW Department of Primary Industries Narrandera Fisheries Centre PO Box 182 Narrandera NSW 2700 Tel (02) 6959 9031 Fax (02) 6959 2935 Email: [email protected]

OBJECTIVES:

1. Compare population responses of sub-adult trout cod in large versus small river habitats.

2. Compare population responses of trout cod that are stocked as sub-adults with those originally stocked into the wild as fingerlings.

3. Apply and further develop innovative underwater video camera technology as a tool for investigating habitat use in freshwater environments.

1 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

NON-TECHNICAL SUMMARY

OUTCOMES ACHIEVED TO DATE

This study shows that the re-establishment of cod populations based on release of on- grown fish is not an immediate and straightforward alternative to the existing practice of stocking fingerlings into rivers. The pattern of trout cod functioning/existing/ in small river reaches, for extended periods punctuated by occasional large-scale movement, has implications for fisheries managers engaged in protecting trout cod populations through improved site selection for re-establishing species and rehabilitating habitat. This project has demonstrated that the re-establishment of ecologically sustainable fisheries of Australian freshwater cod populations is unlikely to meet with broadscale success without commitment to more informed recovery programs than those based primarily on stocking of fingerlings.

This project consisted of two field experiments primarily designed to determine if dispersal of post-juvenile trout cod, Maccullochella macquariensis, is responsible for the apparent lack of success following stocking of this species into numerous riverine sites. The study also served to trial the release of a species of Australian freshwater cod as on-grown individuals rather than as fingerlings.

The first experiment involved the use of radio-tracking, to compare the dispersal of trout cod stocked as sub-adults (two-year old, hatchery fish) with those originally stocked as fingerlings (unknown age, wild fish) in the Murrumbidgee River at Narrandera, New South Wales. The collection of wild fish for radio-tagging was achieved by boat electro-fishing in a river reach 13 km in length (study reach). The survey revealed a high relative abundance of resident trout cod in contrast to . Analysis of the survey information, and notably the spatial pattern of Murray cod catches, was used to discuss requirements for detecting the presence of cod populations when abundance is low.

After successfully implanting 27 hatchery and 31 wild fish with radio-tags, both groups were released at a site positioned at the midpoint along the reach that had previously been electro-fished. Radio-tracking at fortnightly intervals for 3 months post-release, revealed 52% of the wild group homed to the original capture location (within one home-range) whilst the remainder established home-ranges elsewhere within the reach. The hatchery group primarily (61%) dispersed substantial distances, 5–55 km, downstream. Thirty-three percent of the group remained close to the release site (<5 km). Radio-tracking was subsequently conducted at monthly intervals up until 13 months post-release. During this period rejection of radio-tags or mortality occurred for the majority of the hatchery individuals. In contrast, only a single wild individual was confirmed as a radio-tag rejection or mortality in the 13 months after release. Following initial establishment of a home-range, wild fish demonstrated a high level of fidelity to a home-range, spanning a maximum of about 300 m. A subset of seven wild individuals were observed to undertake unsynchronised large-scale movements, in the order of 10–60 km, resulting in return to their home-range in all but one case. The remaining individual was found to occupy three separate home- ranges, including eventual return to the home-range that it had occupied at the beginning of the study. Consequently, a conceptual model was developed to distinguish dispersal from other types of movement.

An ecological approach to re-establishing Australian freshwater cod populations 2 FRDC Report 2003/034

The strong pattern of home-range occupation that emerged from 13 months of radio- tracking, also lead us to undertake a 5 day empirical study of the small-scale diel movement of ten individuals within their home-ranges. This revealed use of small home-ranges (<300 m in length) positioned on outside bends.

Finally, the poor survivorship of the hatchery-reared sample demonstrated the inadequacy of using hatchery-reared individuals as surrogates for wild fish. This indicated a need to refocus the second experiment toward the issues of trial stocking of on-grown fish, quantifying survivorship and minimising the possibility of radio-tag rejection.

The second experiment involved radio-tracking, following release, of 36 sub-adult trout cod (two-year old, hatchery fish) into both the Murrumbidgee and Cotter rivers in the upper Murrumbidgee River catchment (Australian Capital Territory). The schedule of radio-tracking used in the first experiment was replicated in this experiment. Additionally, underwater video cameras were used to monitor the release pool and a control pool in the Cotter River. The underwater surveillance was unsuccessful as a result of poor water clarity in the Cotter River catchment at the time of release (as a function of sediment run-off following a major bushfire event).

Mortality was rapid in both rivers. Within 3 months, 6% and 19% survivorship occurred in the Cotter and Murrumbidgee rivers, respectively. By 7 months post- release 100% mortality or radio-tag rejection had occurred in both groups. It is likely that these patterns represent real patterns of survivorship, as rejection of radio-tags was unlikely, based on available evidence. Mortality was linked to predation or scavenging by cormorants and water rats in the Cotter River and to a lesser extent in the Murrumbidgee River, although, this probably reflects a superior ability to retrieve radio-tags and fish remains in the former.

Collectively these experiments recorded poor survivorship of on-grown two-year old trout cod at a lowland and two upland sites, in the process demonstrating the effectiveness of radio-tracking for monitoring such releases. Whilst the majority of wild fish did not disperse from their initial home-range in the first experiment, indications were that large individuals can move on scales of tens of kilometres. Further experimental research is required to determine if dispersal explains the disappearance of sub-adult trout cod from fingerling-stocking sites.

KEYWORDS: dispersal, survivorship, stocking, on-growing, radio-tracking, Maccullochella, trout cod, Murray cod, telemetry, home-range, species recovery

3 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

ACKNOWLEDGMENTS This research project has been funded primarily through the Fisheries Research and Development Corporation; Parks, Conservation & Lands (ACT Government) and New South Wales Department of Primary Industries with additional contributions from a number of organisations including the Victorian Department of Primary Industries, Cooperative Research Centre for Freshwater Ecology, Murray–Darling Basin Commission, Freshwater Ecology at the Arthur Rylah Institute, RecFish Australia, State Water and the Department of Infrastructure, Planning and Natural Resources (NSW).

Thankyou to FRDC staff in particular project managers: Jane Graham, Crispian Ashby and Matt Barwick; and steering committee members: Dean Ansell, Anthony Chariton, Mark Dunford, John Koehn, Peter Liston, Simon Nicol and Sue Vink.

Staff listed in Appendix 2 provided support and field assistance throughout the project.

An ecological approach to re-establishing Australian freshwater cod populations 4 FRDC Report 2003/034

CHAPTER 1. BACKGROUND TO THE AUSTRALIAN FRESHWATER COD

Mark Lintermans1, 2 and Brendan Ebner1, 2

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

Lintermans, M. & Ebner, B. (2006). Background to the Australian freshwater cod. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.5–14. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

There are four percichthyid freshwater cod species or sub-species in the Maccullochella in Australia; Murray cod (Maccullochella peelii peelii), (M. peelii mariensis), trout cod (M. macquariensis) and (M. ikei). Prior to the 1970s only one cod species (Murray cod) had been recognised, with trout cod only formally described in 1972 (Berra & Weatherley, 1972). The two coastal cod species (Mary River cod and Eastern freshwater cod) were later identified and described by Rowland (1985, 1993). The following summary of the conservation status and recovery approaches for Australian freshwater cod is from Lintermans et al. (2005b).

All four cod species are or have been popular target species for recreational anglers as they are large freshwater fish, with maximum sizes in excess of 15 kg. However, all four species have declined in range and abundance with all declared threatened under the Commonwealth Environment Protection and Biodiversity Conservation Act 1999 or various State/Territory legislation. All cod species except Murray cod are listed as endangered nationally, with Murray cod considered the least threatened of the four, having only been listed as vulnerable nationally in 2003 (Table 1.1).

Distribution and declines in Australian freshwater cod Two of the four cod species in Australian freshwaters (Mary River cod and Eastern freshwater cod) have relatively small historical and current distributions, however, the two Murray–Darling Basin species (Murray cod and trout cod) had significantly larger historical distributions (Table 1.2). The historical distribution and decline of each of the cod species is summarised below.

Mary River cod Freshwater cod were present in a number of south-east Queensland (Qld) coastal catchments at the time of European settlement, including the Mary, Brisbane–Stanley, Albert–Logan and Coomera basins. It has never been confirmed that the cod in these river systems all belonged to the same species. By the 1970s, cod had apparently disappeared from all but the Mary River.

5 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Table 1.1. Summary of distribution and conservation status of Australian freshwater cod (from Lintermans et al., 2005b). Species Natural State/Territory 1EPBC 2ASFB Distribution conservation status Status (excluding status translocations) trout cod Murray–Darling ACT: Endangered Endangered Critically Basin in Vic., NSW: Endangered Endangered NSW, ACT Vic.: Threatened3

Eastern Clarence and NSW: Endangered Endangered Endangered freshwater Richmond Rivers in cod NSW

Mary Mary and Brisbane Qld: Protected Endangered Critically River cod Rivers in Qld Endangered

Murray Murray–Darling Vic.: Threatened3 Vulnerable not assessed cod Basin in Qld, NSW, ACT: not listed Vic., SA, ACT NSW: not listed Qld: not listed SA: not listed 1 Environment Protection and Biodiversity Conservation Act 1999 2 Australian Society for Fish Biology 2004 3 Trout cod assessed as Critically Endangered and Murray cod assessed as Endangered (DSE 2003)

Trout cod As trout cod were only formally recognized as a separate species in 1972 (Berra & Weatherley, 1972), there is some confusion regarding the historical distribution of trout cod. The type specimen was originally collected in the in the early 1800s, from where it has not been subsequently recorded other than a single unconfirmed record from the Turon River in the 1970s (Morris et al., 2001). investigations of the Murray in 1949–50 by J. O. Langtry (Cadwallader, 1977) recorded trout cod as present in the between about Mildura and the junction with the Ovens River, and as relatively abundant between Yarrawonga Weir and the Barmah–Millewa Forest (near Echuca). Further, the South Australian Museum (Adelaide) houses two specimens of trout cod from the Lower Murray in South Australia (SA) (collected in 1928 and 1932 (Berra, 1974)), indicating that the original distribution of the species extended at least that far downstream in the Murray River. The present distribution of the remnant population in the Murray River extends from the Yarrawonga Weir downstream to Torrumbarry Weir, a river distance of approximately 200 km.

Trout cod were formerly present in the Murrumbidgee River, with records from Hay, Narrandera, Wagga Wagga and Canberra (Berra, 1974; Douglas et al., 1994). The last reported specimen in the lower Murrumbidgee was collected near Narrandera in 1969 and in the upper Murrumbidgee near Tharwa in 1976 (Gilligan, 2005b). Subsequently, the species was considered extinct throughout the Murrumbidgee catchment (Berra, 1974; Lintermans et al., 1988).

An ecological approach to re-establishing Australian freshwater cod populations 6 FRDC Report 2003/034

The species was also formerly abundant in the Ovens River in (Cadwallader, 1977), as well as the Goulburn, Campaspe, Broken, King, Buffalo and Mitta Mitta rivers (Cadwallader & Gooley, 1984; Douglas et al., 1994). Natural populations of trout cod no longer occur in any of these rivers (Douglas et al., 1994; Koehn et al., 1995).

Prior to the initiation of the captive-breeding and reintroduction programs in New South Wales (NSW), Victoria and the Australian Capital Territory (ACT), two self- sustaining populations existed in addition to the remnant Murray River population, both were the result of translocations in the early 1900s. The first was a population translocated from an unknown source population into Cataract Dam, NSW in 1914. The second was a population translocated from the Goulburn River into Seven Creeks, Victoria in the 1920s. Since 1986, captive-breeding programs in NSW (Narrandera Fisheries Centre) and Victoria (Snobs Creek Freshwater Fisheries Hatchery) have reintroduced trout cod fingerlings into a number of waterways within their historical distribution (e.g. Douglas et al., 1994). There is a lack of information regarding the establishment of self-sustaining populations in any of these locations, however, fisheries agencies in NSW, Victoria and the ACT are in the process of collecting data on the status of these reintroductions.

Eastern Freshwater cod Historical accounts from the 1800s and early 1900s indicate that large populations of Eastern freshwater cod were once present in the Clarence and Richmond systems of northern NSW (Bawden, 1888; Rowland, 1985, 1993). However, despite the apparent abundance of cod, a dramatic decline in the population occurred in both catchments throughout the late 1800s and early 1900s. Several large fish-kills in the late 1930s, coincided with dynamiting and railway construction on the Richmond River and ultimately brought about almost total extinction of the species in the Richmond River catchment. By the early 1970s only one small self-sustaining population of the species was known to exist, occurring in the Mann–Nymboida River, a sub-catchment of the Clarence, where isolated pristine habitats persisted (Merrick & Schmida, 1984; Rowland, 1985, 1993, 1996; Faragher & Harris, 1993; Harris & Rowland, 1996). However recent surveys have detected small remnant populations in a few other headwater streams within the Clarence River catchment (Ian Wooden, NSW DPI, unpublished data).

Murray cod Murray cod was an abundant species, with explorers’ journals and early commercial fishery reports noting the size and numbers of cod present (Rowland, 1989, 2005; Scott, 2005). Murray cod formed a major part of the diet of Aboriginal tribes living adjacent to inland waters.

There has been a significant decrease in the abundance of Murray cod from historical levels. The commercial fishery for cod (which developed in the mid to late 1800s) peaked in the early 1900s and then declined, although, there was a subsequent small peak in the 1950s (Rowland, 1989). However, the catch had decreased significantly by the mid 1960s and whilst it has fluctuated in more recent decades (e.g. Gilligan 2005a, b) it has remained at low levels relative to that in the 1800s and early 1900s (Rowland, 1985, 1989; Reid et al., 1997; Ye et al., 2000; Kearney & Kildea, 2001;

7 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Lintermans & Phillips, 2005). The commercial fishery for Murray cod ceased across the Murray–Darling Basin in 2003.

Whereas Murray cod was once common over much of its range, it is now patchily distributed, with few locations containing abundant cod populations. The NSW Rivers Survey failed to catch a single cod across 20 randomly selected sites in the Murray, Murrumbidgee and Lachlan catchments (Harris & Gehrke, 1997). This contrasted with the results from the Darling drainage, where Murray cod were collected at 7 of 20 sites (Harris & Gehrke, 1997). There is concern for the conservation status of Murray cod in Victoria, where it is now considered endangered (DSE, 2003), particularly with the loss of the Mitta Mitta River population following the construction of Lake Dartmouth (Koehn et al., 1995) and recent fish kills in the Ovens River, Goulburn River and Broken Creek (Koehn, 2005).

Table 1.2. General distribution, period of decline and current distribution of Australian freshwater cod (from Lintermans et al., 2005b). Species Distribution Historical Period of Approximate Abundance Decline Current Distribution (River kms) trout cod Southern Murray– Abundant in 1950–70 ~ 200 km Darling Basin southern range

Eastern Clarence and Abundant 1920s and ~300 km freshwater cod Richmond rivers 1930s

Mary River cod Brisbane–Stanley, Abundant Prior to 1930s 300–500 km Albert–Logan, Coomera and Mary river basins Murray cod Murray–Darling Abundant 1920s and 1000s of km Basin (commercial 1930s; then fishery) 1950s

Current approaches to the conservation and recovery of Australian freshwater cod Trout cod was the first cod species (and the first Australian fish species) to have a national recovery plan (Douglas et al., 1994), with hatchery production and restocking commencing in 1986–87. This recovery plan has since been updated twice, by Brown et al., 1998) and the Trout Cod Recovery Team (2004). The major activities of trout cod recovery are the implementation of a total fishing closure during the spawning season in part of the Murray River inhabited by the remnant population (from Yarrawonga Weir downstream to Tocumwal), restocking with hatchery-bred fingerlings, and habitat enhancement or management activities. The conservation stocking program for trout cod has been the most extensive of the three endangered cod species (excluding Murray cod), having been in operation for 20 years and

An ecological approach to re-establishing Australian freshwater cod populations 8 FRDC Report 2003/034 releasing more than 1 million fish across three jurisdictions in the Murray–Darling Basin (Table 1.3).

Hatchery production and restocking of the Mary River cod commenced in the early 1980s, however, a detailed recovery plan was not implemented until 1996 (Simpson & Jackson, 1996). While the ultimate aim of the recovery plan is to restore cod populations in their historical habitats, the plan also allows for limited take by anglers in a small number of water storage impoundments. A targeted riverine restocking program in the Mary River system commenced in 1997, in conjunction with habitat rehabilitation works. Recovery efforts are now focusing on restoring populations in the Brisbane–Stanley, Albert–Logan and Coomera rivers (Simpson & Jackson, 1996).

Table 1.3. Stocking programs for freshwater cod species. (Modified and updated from Lintermans et al., 2005b). Species Declared Stocking Program Total Number Threatened/ Commenced Stocked protected trout cod Pre-1984 1986 (Vic.) **447 350 (Lake, 1971) (NSW) 900 000

Eastern 1984 1988 (one-off release) 220 000 Freshwater cod 1997 (main program)

Mary River cod 1994 1983 *416,300

Murray cod 2003 late 1970s (Vic.) **3.134 million (NSW) *2.85 million (ACT) 442 741 (Qld) ***320 000 * number stocked prior to 2004 ** number stocked prior to 2005 ***number stocked prior to April 2001

Eastern freshwater cod were first bred and stocked in 1988 as a by-product of a research project, with the main stocking program commencing in 1997 (Table 1.3). A recovery plan for the species was finalized in 2004 (NSW Fisheries, 2004); however, a draft recovery plan had guided recovery actions since 1999. As with the other recovery programs for endangered cod species, the emphasis of the recovery plan is a combination of establishing new populations through stocking, protecting existing populations, and habitat enhancement (NSW Fisheries, 2004).

Murray cod are widely sought by anglers (Henry & Lyle, 2003; Park et al., 2005) and are now the only freshwater cod species that can be legally kept (except for Mary River cod stocked for recreational purposes in a small number of impoundments in Qld). Captive-breeding techniques for percichthyids were developed in the late 1970s and 1980s (Rowland, 1983a, b, 1988; Cadwallader & Gooley, 1985; Ingram & Rimmer, 1992). Murray cod stocking began in NSW in 1977 and increasing numbers of Murray cod have been stocked in all Murray–Darling Basin jurisdictions since

9 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

1983. Approximately one million Murray cod are now stocked annually (Ingram & De Silva, 2004; Lintermans et al., 2005b). However, these fish are stocked for recreational purposes rather than for conservation reasons (Lintermans, 2006). Only one stocking of Murray cod has been undertaken for conservation purposes, with >30 000 fingerlings released in the lower in 2005 in response to a fish kill that occurred in 2004 (Gilligan, personal communication). A national recovery plan for Murray cod is currently being drafted.

Current stock enhancement activities for Australian freshwater cod rely almost exclusively on the release of hatchery-bred fingerlings of 3–5 cm total length (NSW Fisheries, 2003; Lintermans, 2005). There have been occasions where yearling or on- grown fish have been released, but these represent a minuscule proportion of stock enhancement activities (Simpson et al., 2002; NSW Fisheries, 2003; Todd et al., 2004; Lintermans, 2005). The perception of stocking as the panacea to fisheries problems is widespread but inaccurate (Harris, 2003), and there has been relatively little effort devoted to ascertaining whether stocking is successful in establishing new populations. Historically (pre-captive husbandry), translocation of adult and sub-adult fish occurred, but this was for recreational fishery purposes, not conservation of the species (Cadwallader & Gooley, 1984; Douglas et al., 1994). The lack of alternative stock enhancement approaches (to stocking fingerlings) and the lack of evidence for self-sustaining populations resulting from fingerling stocking in establishing populations of threatened percichthyids is a cause for serious concern (Lintermans, 2006). The potential for stocking on-grown fish to minimise mortality of newly released individuals has long been recognized (see Brown et al., 1998), but economic (high cost of on-growing fish) or hatchery (lack of available grow-out ponds) constraints have militated against this. The high mortality of newly released captive- bred individuals is a major problem with stock enhancement programs worldwide, with highest mortality occurring during or immediately after release (Olla et al., 1994, 1998; Brown & Day, 2002). A pilot project to the current project demonstrated the suitability of implanting M. macquariensis with radio-tags and examined movement of a small number of on-grown trout cod after release into a stream (Ebner et al., 2005). That study revealed individuals either remained close to the release site or exhibited net downstream movement. Essentially, complete mortality of all individuals comprising the sample occurred 4–12 months after release (Ebner, Johnston & Lintermans, unpublished data).

Fate of fingerling stockings of trout cod Since 1986, hatchery-produced fingerlings have been stocked into a number of catchments in NSW, Victoria and the ACT, in an effort to re-establish populations of this species (Douglas et al., 1994; Brown et al., 1998; ACT Government, 1999; Gilligan, 2005b; NSW DPI, 2005). In recent years, large rivers have been the focus of stocking efforts particularly in NSW and the ACT. There has been infrequent, sporadic or no monitoring at many stocking sites, and knowledge of successful long- term survivorship has more commonly arisen from angler reports. Survival and growth of fingerlings through to an age of 2–3 years (juveniles or sub-adults) has been documented at a number of sites (Faragher et al., 1993; Lintermans, 1995; Brown, 1998; Brown & Nicol, 1998; Douglas & Brown, 2000; Gilligan, 2005b; NSW DPI, 2005). Whereas, the number of 3–5 year old individuals (at which age they are sub- adult or young adults: Harris & Rowland, 1996) being detected is much lower. Reports of large adult individuals have been uncommon. Possible explanations for the

An ecological approach to re-establishing Australian freshwater cod populations 10 FRDC Report 2003/034 scarcity of adults in re-stocked populations are: 1) larger fish (adults) are subject to high mortality including pressure and/or natural mortality; 2) larger fish remain at the release site but are undetected by monitoring; and/or 3) juveniles remain at the release site until they reach sub-adulthood at which stage they disperse, becoming difficult to detect (Figure 1.1).

Fingerling stocking

Detection of sub-adults

Mortality Dispersal Cryptic behaviour (e.g. natural or angler or related) low abundance

Non-detection of adults

Figure 1.1. Conceptual model of the fate of sub-adult M. macquariensis stocked as fingerlings.

Generally, fish populations suffer greatest natural mortality in early life history phases (Jones, 1991) and this presumably applies to trout cod (Todd et al., 2004). Widespread concealment of the rapid harvest by anglers, of all M. macquariensis as they approach or soon after they attain adult size is unlikely. Therefore it is unlikely that these larger fish are subject to high mortality, though, our understanding of mortality in this species is a major knowledge gap in recovering the species (Simon Nicol, personal communication). It is also unlikely (at least for riverine sites) that larger fish remain at the site and are not detected by monitoring. It has been suggested for an impoundment stocking (Bendora Reservoir, ACT), that large fish are occupying deep-water habitats (up to 47 m) beyond the reach of conventional sampling gear (Lintermans, 1995). Electro-fishing has been a commonly employed monitoring technique for trout cod and other freshwater cod (Faragher et al., 1993; Growns et al., 2004) and has proven successful in detecting a wide range of size classes at a subset of sites (J. Koehn, personal communication; Ian Wooden, unpublished data; Chapter 2). Certainly this is the case in the Murrumbidgee River at Narrandera, where sub- adult and adult fish are common (Ian Wooden, unpublished data; Chapter 2).

Despite indications that juveniles can be gregarious (Douglas et al., 1994), it is possible that as individuals grow and mature they become territorial, leading to dispersal from the release sites, since under hatchery conditions individuals become particularly aggressive towards one another from a very early age (Brett Ingram, Department of Primary Industries Victoria, personal communication). Therefore, to improve re-establishment of populations of M. macquariensis it is important to determine the movement patterns of sub-adults following stocking in rivers. In this study, stocking of on-grown fish is also tested as an alternative reintroduction strategy to releasing fingerlings.

11 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Radio-tracking has proven an effective method for monitoring the dispersal, habitat use and survivorship of hatchery-reared, on-grown fish, most commonly salmonids (Jepsen et al., 1998; Dieperink et al., 2001; Aaestrup et al., 2005). An initial pilot study based on two-year old M. macquariensis (n=9) released into a small upland stream bound by upstream and downstream barriers, revealed a majority of individuals remained at or near the release site in the first few months post-release (Ebner et al., 2005). Gradually the number of individuals that had moved or dispersed downstream from the release site increased throughout the study and all individuals died within twelve months post-release. On-grown, hatchery-reared individuals have not been radio-tracked in lowland rivers. Radio-tracking of wild individuals from the lowland Murray River population has revealed that M. macquariensis are typically associated with complex structural woody habitat (SWH – also referred to as snags or large woody debris) in higher-flow parts of the Murray River below Yarrawonga (Koehn, 1997). Research conducted in the mid-Murrumbidgee revealed a similar association with SWH based on electro-fishing survey (Growns et al., 2004). These studies provide data on habitat use, and the current study serves to refine this knowledge by exploring home ranges and diel behaviour. Since habitat attributes are an important consideration in the selection of stocking sites (Douglas et al., 1994) and habitat restoration is considered an important element of recovering this species (Brown et al., 1998), developing an understanding of M. macquariensis habitat use in river systems is fundamental to returning this species to a substantial portion of its former range.

Radio-tracking as a tool to monitor dispersal Radio-tracking has been used to monitor several aspects of the biology and ecology of fishes. This includes monitoring of movement (e.g. Keefer et al., 2006), habitat use (e.g. Crook et al., 2001), interactions among individuals (e.g. Hyvarinen & Vehanen, 2004), and survivorship (e.g. Dieperink et al., 2001) including survivorship in fisheries (Nelson et al., 2005). Movement has been considered at large scales in terms of migration (e.g. Keefer et al., 2006) and immigration (e.g. Crook, 2004a), and at smaller scales in terms of diel activity (David & Closs, 2003) and diel mobility (Ovidio et al., 2000) often focusing on individuals within home-ranges. Radio-tracking has also been used as a basis for quantifying home-range (e.g. Fish & Savitz, 1983), and to investigate habitat-use during migration (Økland et al., 2001) and non-migratory movements (Essington & Kitchell, 1999).

The short to medium-term success of hatchery-reared fishes stocked into rivers (Jepsen et al., 1998; Dieperink et al., 2001; Aarestrup et al., 2005) has been investigated by radio-tracking. Radio-tagging is typically restricted to on-grown fishes as a consequence of threshold body-size requirements for tag attachment. This approach enables rates of mortality and dispersal to be quantified, facilitating adaptive management of fisheries (e.g. Aarestrup et al., 2005) or threatened species recovery (e.g. Skalski et al., 2001).

An ecological approach to re-establishing Australian freshwater cod populations 12 FRDC Report 2003/034

Need Fingerling stocking is widely used to establish or augment recreational fisheries or for conservation purposes worldwide (Brown & Day, 2002) including in Australia (NSW Fisheries, 2003; Phillips, 2003; Lintermans, 2006). However, Australian and overseas studies have demonstrated that fingerling stockings succeed in a subset of cases and often the success of these programs is difficult to evaluate with conventional approaches (Minckley, 1995; Brown & Day, 2002; Lintermans, 2006). Factors that can enhance the chances of fingerling and conservation stockings being successful include: pre-release training in predator recognition and avoidance, exposure to habitat complexity prior to release, and selecting for genetic traits that favour such characters; size at release, timing of release, and abiotic and biotic characteristics of the release site (Brown & Day, 2002; Thorfve, 2002; Vilhunen & Hirvonen, 2003; Salvanes & Braihwaite, 2005; Vilhunen et al., 2005; Vilhunen 2006). These factors are only just starting to be investigated in Australia (Brown & Warburton, 1997; Bearlin et al., 2002; Harris, 2003; Hutchison et al., 2006; Lintermans, 2006).

This project takes the opportunity to explore the possibility of re-establishing adult cod populations through seeding with fewer but much larger individuals (not fingerlings) as identified by Todd et al. (2004). The research has particular relevance to Australian freshwater cod as large, long-lived fishes (Anderson et al., 1992; Rowland, 1998a) capable of large clutch sizes (tens of thousands of eggs per female (Rowland, 1998b)). Theoretically, stocked individuals are capable of spawning annually for decades to produce hundreds of thousands of progeny within their lifetime. To this end, a breeding population can be established from a small number of individuals stocked at large body-size (e.g. individuals measured in kilograms rather than grams) negating most natural mortality that occurs at small size (Ebenman & Person, 1988). Furthermore, Australian freshwater cod are apex predators (Ebner, in press) that may serve as indicators of river system health due to their unique position at the top of the food web (Kearney & Kildea, 2001). The logic follows that a self- sustaining cod population in a river represents a measurable endpoint for a healthy, functioning ecosystem.

The study species, Maccullochella macquariensis is Australia’s most imperilled cod species. It was once an important recreational species and a component of the inland commercial fishery of the Murray–Darling Basin. Today extensive efforts to recover the species and establish new populations remain limited by our lack of understanding of what happens to stocked individuals during sub-adulthood (Figure 1.1).

13 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

OBJECTIVES

1. Compare population responses of sub-adult trout cod in large versus small river habitats.

2. Compare population responses of trout cod that are stocked as sub-adults with those originally stocked into the wild as fingerlings.

3. Apply and further develop innovative underwater video camera technology as a tool for investigating habitat use in freshwater environments.

An ecological approach to re-establishing Australian freshwater cod populations 14 FRDC Report 2003/034

CHAPTER 2. ESTIMATING SPECIES RICHNESS AND CATCH PER UNIT EFFORT FROM BOAT ELECTRO-FISHING IN A LOWLAND RIVER IN TEMPERATE AUSTRALIA

Brendan Ebner, 1, 2 Jason Thiem, 1, 2 Dean Gilligan, 3 Mark Lintermans, 1, 2 Ian Wooden3 and Simon Linke2, 4

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

3NSW Department of Primary Industries, Narrandera Fisheries Centre, PO Box 182, Narrandera NSW 2700, Australia.

4The Ecology Centre, School of Integrative Biology, University of Queensland, St. Lucia, QLD 4072, Australia.

Ebner, B., Thiem, J., Gilligan, D., Lintermans, M., Wooden, I. & Linke, S. (2006). Estimating species richness and catch per unit effort from boat electro-fishing in a lowland river in temperate Australia. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.15–28. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

INTRODUCTION

Accurate estimates of the diversity and abundance of fishes can be difficult to obtain in large lowland rivers where habitat characteristics including variable depth, complex woody structure, heterogeneous habitat, flow and turbidity limit the effectiveness of sampling gear (Thévenet & Statzner, 1999). Species-specific traits that affect detection and the spatial distribution of different species compound the problem (Fausch et al., 2002). Currently, boat electro-fishing represents one of the more useful means of surveying fishes in these habitats (e.g. Harvey & Cowx, 1996). Single-pass sampling often forms the basis of routine monitoring with this technique (e.g. Odenkirk & Smith, 2005). Subsequently priorities for monitoring programs include determining the length of river to survey at a site (Lyons, 1992; Patton et al., 2000; Hughes et al., 2002) an issue intimately related to representative sampling (Cao et al., 2002), and calibrating single-pass sampling with information of actual fish density (Ricker, 1975; Pusey et al., 1998; Odenkirk & Smith, 2005). This study is concerned with the first of these issues.

It has been demonstrated that both the length of stream sampled and the sampling effort used at a site influence the accuracy of fish diversity estimates (Lyons, 1992; Hughes et al., 2002; Smith & Jones, 2005). This is considered to be a function of a) the longitudinal accumulation of microhabitats encountered with increasing stream length and b) the distribution of rare species (Cao et al., 2001; Kennard et al., 2006). Consequently the relationship between species accumulation and the length of stream surveyed (and/or sampling effort) can be used to determine the minimum stream length to survey at a site to achieve the desired objectives of particular monitoring applications (Cao et al., 2001). The minimum length of stream required for surveying

15 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 fish species richness at a site has been investigated particularly in temperate North America in the context of boat and backpack electro-fishing (Lyons, 1992; Hughes et al., 2002; Meador, 2005; Smith & Jones, 2005). In many instances, substantial stream length must be surveyed to estimate species richness (Lyons, 1992; Cao et al., 2001; Hughes et al., 2002; Smith & Jones, 2005). Notably, Hughes et al. (2002) concluded that reliable estimates of 95% of the species richness required sampling from a length of river equalling 85 times the mean wetted channel width and that 100% of species were detected within 300 channel widths.

In Australia, studies based on sampling in comparable lengths of stream to that proposed by Hughes et al. (2002) remain to be achieved, although, the effects of sampling effort or stream length have been examined in much shorter lengths of stream (i.e. <1 km: Faragher & Rodgers, 1997; MDBC, 2004a; Kennard et al., 2006). Meador (2005) commented that survey in 500–1000 m of stream might produce better estimates of species richness in low diversity fish communities (<10 species) relative to more species rich communities. In temperate Australia riverine fish communities frequently contain low species richness (Figure 2 & Figure 3 in Gehrke & Harris, 2000). Routine monitoring is frequently conducted at sites with river length measuring hundreds of metres to 1 km (Appendix 4). Building on the approach of Harris & Gehrke (1997), the Sustainable Rivers Audit in the Murray–Darling Basin (MDBC, 2004a; Lintermans et al., 2005a) is starting to develop approaches to monitoring riverine fish communities at the scale of catchments and basins. Consequently, this study aimed to: a) develop an understanding of the relationship between sampling effort and representative samples of fish species richness; b) determine the sampling effort required to obtain precise estimates of the catch per unit effort of individual species; and c) examine the spatial distribution of the fish community at a site, within a lowland river reach in temperate Australia.

METHODS

Study site The study was undertaken in a lowland reach of the Murrumbidgee River, near Narrandera, in southern inland New South Wales, Australia (Figure 2.1). River red- gum ( Dehnh.) were conspicuous along much of the riparian zone of the river and channel widths are usually in the order of 70 m (Growns et al., 2004). River depths of 1–2 m are typical throughout the study area, with depths of 3– 5 m commonly encountered on outside bends. Bare sand/mud interspersed with fallen trees or branches (particularly of E. camaldulensis) comprised the dominant in-stream habitats.

An ecological approach to re-establishing Australian freshwater cod populations 16 FRDC Report 2003/034

Figure 2.1. Study reach of the Murrumbidgee River near Narrandera, New South Wales, Australia.

Fish sampling From 11–24 September 2003, a survey of the fish community encompassed an almost continuous 13 km length of river (apart from three short sections of shallow, sand bar that could not be navigated), involving 157 replicate boat electro-fishing operations. The electro-fishing boat was a 4.5 m aluminium hull fitted with a 7.5 kW Smith- Root© Model GPP 7.5 H/L electro-fishing unit. Two anodes were suspended from booms mounted on the bow of the boat and a cathode was mounted along each side of the hull. The electro-fisher was operated at 1000 volts DC, 60 hz, 35 percent of range with an average current draw of 4.5 amps. Each replicate operation was undertaken for 5 min (elapsed time), equating to a mean (± SE) of 92 ± 0.35 s of applied power. All navigable habitats were surveyed in proportion to availability by criss-crossing the channel and continuing in an overall upstream direction (to prevent re-sampling habitats). An operation covered a mean distance of about 80 m of river. Two operators collected immobilised fish from the bow of the boat by dip net. A support boat followed about 50 m downstream of the electro-fishing boat to ensure that stunned fishes and particularly trout cod Maccullochella macquariensis (Cuvier) that were slow to surface were collected and fully recovered (note that the primary purpose of this survey was to collect M. macquariensis for a radio-tracking study: see Chapter 3). Captured fish were allowed to recover in a re-circulating holding tank aboard the boat. Recovery tanks have proven particularly useful for Australian freshwater Maccullochella spp., as these species typically take quite some time to recover from electro narcosis, due to a characteristic jaw-gaping or ‘lock-jaw’ observed in these fish (Harris, 1995; Koehn, 1995).

17 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

All captured individuals were identified to species and returned to the river following recovery, with the exception of M. macquariensis. Individuals of this species were measured and weighed, and returned to the river if <550 g and retained for subsequent radio-telemetry studies if >550 g (Chapter 3). The number of fish that were observed whilst electro-fishing, but not captured was also recorded. These individuals were identified to species except on six occasions when individuals were clearly an Australian freshwater cod, but could not be distinguished as being either M. macquariensis or Murray cod Maccullochella peelii peelii (Mitchell), in which case the individuals were identified to genus. Unidentified individuals of Maccullochella are included in a summary of the survey (Table 2.1) but were not included in analyses. Data from observed and caught fishes was pooled following Harris & Gehrke (1997).

Water temperature, conductivity and turbidity were recorded at 15 min intervals from a single location in the middle of the study reach using a Hydrolab (Loveland, CO).

Data analysis A number of approaches were used to investigate spatial pattern in species richness and catch per unit effort of each species. This involved a) examining sequences of operations in which species were not detected, b) calculating confidence intervals associated with mean catch per unit effort of each species based on random and consecutive re-sampling of data, c) checking for correlation between species, d) testing for similarity in species richness among sub-reaches, and e) assessing the performance of species richness predictors in relation to sampling effort.

Record was made of all sequences of consecutive operations where a species was not detected. These data were summarised as mean, minimum and maximum sequences in terms of number of operations. Where a sequence included the beginning or end of the 13 km reach (i.e. operation 1 or operation 157) a potential underestimate occurred. Nevertheless these sequences were retained in calculating summary statistics since their removal caused a greater bias in the output.

To determine the minimum sampling effort required to produce catch per unit effort of individual species within narrow confidence intervals, two practical sampling strategies were simulated in this study. In the first strategy, samples were taken by drawing groups of consecutive operations – for example a five-operation sample could consist of operation 83, 84, 85, 86, and 87. This mimicked a common method of survey (see references in Appendix 4). The second strategy was based on drawing random samples of operations – for example a five-operation sample comprising operation 3, 9, 41, 83 and 129. A macro written in Microsoft Excel 2000® sub- sampled data at one-operation increments either based on the consecutive field sampling order or based on a random order, repeating this process 1000 times for sample size equals 2 to 75 (i.e. up to about half the sample). Ninety-five percent confidence intervals were generated for each sample size by taking the 2.5 and 97.5 percentiles. These confidence intervals were not spaced even distances from the mean, as data did not follow a normal distribution due to the large number of zero values in the data set.

We investigated possible interrelationships between species by correlating abundance within operations, based on data standardised to catch-per-minute of electro-fishing

An ecological approach to re-establishing Australian freshwater cod populations 18 FRDC Report 2003/034

(power-on time). Due to the non-normal distribution of data, Spearman Rank correlations (Sokal & Rohlf, 1995) were required and the abundance of each species was ranked across operations. Given that multiple comparisons were undertaken, a Bonferroni correction (Sokal & Rohlf, 1995) was applied to reduce the experiment- wise error rate. With k = 21 comparisons, the significance level equivalent to α = 0.05 is p = 0.002.

Analysis of fish community structure was undertaken to determine whether fish communities within neighbouring parts of the reach were more similar than those from more distant parts of the 13 km reach (equivalent to an auto-correlation). Two analyses were undertaken. The first using the average of five consecutive operations (i.e. operations 1–5, 6–10, 11–15 etc.) and the second using the average of ten consecutive operations (i.e. operations 1–10, 11–20, 21–30 etc.). Analysis based at a resolution of individual operations could not be undertaken owing to the substantial variability among operation-specific data, and the predominance of zeros in the data set. Analyses were performed on standardised catch-per-minute of electro-fishing on time. Spatial analysis of fish community structure was undertaken using a hierarchical agglomerative cluster analysis in Primer 5.1.2 (Plymouth Marine Laboratory). Data was square root transformed to equalise the influence of common and rare species within the analysis. Similarities between fish assemblages within each group of five or ten consecutive operations were calculated using the Bray-Curtis similarity measure (Bray & Curtis, 1957).

A second matrix was created which represented the spatial proximity of the 31 grouped samples (each the average of five consecutive operations) and 15 grouped samples (ten consecutive operations) (adjacent grouped samples having a proximity value of 1, grouped samples separated by one reach having a proximity value of 2, grouped samples separated by two reaches having a proximity of 3 etc.). The ‘proximity’ matrix was correlated with the Bray-Curtis similarity matrix of sampling data using the RELATE function in Primer, with 999 permutations. This function uses the Spearman Rank correlation to relate corresponding values in the two matrices.

To determine an appropriate level of sampling for estimating species richness, we calculated three estimators: Jackknife 1 and 2 (Burnham & Overton, 1979) and Chao 2 (Chao, 1987). These estimators were calculated based on simulated consecutive and random sampling strategies. Data was re-sampled with replacement at five-sample increments, from five to 75 samples, with this process repeated 1000 times for each sample size. An additional comparison, the average species richness from re-sampled data, is used.

RESULTS

During the sampling period mean (± SE) daily water temperature was 14.4 ± 0.05oC, mean conductivity was 180 ± 0.4 μS.cm-1 and mean turbidity was 15.0 ± 0.3 NTU. Seven fish species were captured or observed in this study (Table 2.1). The alien species carp Cyprinus carpio (204 individuals) and the native trout cod, M. macquariensis (131 individuals) were frequently encountered. The former were schooling or solitary, whereas, M. macquariensis was usually solitary and only occasionally found with more than a single individual associated with a single wood structure. Macquaria ambigua (45 individuals) were moderately

19 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 common, whereas Murray cod M. peelii peelii (15 individuals) were infrequently encountered. Six Maccullochella spp. were observed though not identified to species level. The small native Australian smelt Retropinna semoni was moderately common (41 individuals) and two species ( marmoratus and silver perch Bidyanus bidyanus) were rarely encountered (Table 2.1).

Table 2.1. Summary of the fishes surveyed in the study (the number of caught and observed individuals was combined). Fish species Catch Cyprinus carpio Linnaeus 204 Maccullochella macquariensis (Cuvier) 131 Macquaria ambigua (Richardson) 45 Gadopsis marmoratus (Richardson) 4 Retropinna semoni (Weber) 41 Maccullochella peelii peelii (Mitchell) 15 Bidyanus bidyanus (Mitchell) 4 Maccullochella spp. 6

Detection of species The sampling effort required to detect different species was variable, species-specific and generally a function of the frequency of capture (Table 2.2, Figure 2.2). In regard to the least common species, the mean number of consecutive operations (±1 SE) where G. marmoratus and B. bidyanus was not detected was 39 ± 17 and 31 ± 7, respectively. The maximum number of operations between detection was substantially greater than that used in standard surveys (44–74 operations) in regard to four of seven species (Table 2.2, Figure 2.2, Appendix 4).

Table 2.2. Species-specific non-detection over 157 boat electro-fishing operations in a 13 km reach of the Murrumbidgee River.

Species Sequence Number of (consecutive operations of non-detection) sequences

Mean (± SE) Min Max C. carpio 2 ± 0.2 1 7 39 M. macquariensis 2 ± 0.3 1 8 33 M. ambigua 4 ± 0.9 1 18 28 G. marmoratus 39 ± 17 7 74 4 R. semoni 15 ± 5 1 44 10 M. peelii peelii 10 ± 4 2 49 14 B. bidyanus 31 ± 7 9 53 5

An ecological approach to re-establishing Australian freshwater cod populations 20 FRDC Report 2003/034

B. bidyanus G. marmoratus R. semoni M. peelii peelii M. ambigua M. macquariensis C. carpio

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 break (1 km) break break (0.5 km) break (1.5 km) (1.5 break town boat ramp

Number of operations upstream boatramp

Figure 2.2. The presence of each species (■) detected in the study reach from each electro-fishing operation. Operations occurred from downstream (operation 1) to upstream (operation 157) in an almost continuous reach of river, except for three breaks that could not be navigated by boat involving 1, 0.5 and 1.5 km of river, respectively.

Catch per unit effort Data re-sampling revealed that the confidence intervals of the mean catch per operation of each species differed according to whether a consecutive or random strategy was used. Convergence of confidence intervals began occurring for each of the seven species within the first 10 operations for both consecutive and random re- sampling (Figure 2.3). In the case of C. carpio and M. macquariensis, confidence intervals converged more quickly at smaller sample sizes for the random compared to the consecutive re-sampling strategy (Figure 2.3a and b, respectively). Beyond 30 consecutive operations little convergence in confidence intervals occurred, whereas, confidence intervals continued to converge after 70 operations based on random re- sampling. For M. ambigua and M. peelii peelii (Figure 2.3c and f, respectively) convergence of confidence intervals generally occurred at comparable sample sizes for both consecutive and random re-sampling strategies. However, in the case of G. marmoratus, R. semoni and B. bidyanus (Figure 2.3d, e and g, respectively), the convergence of confidence intervals occurred at smaller sample sizes when data was re-sampled using the consecutive strategy.

21 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

4 (a) C. carpio 3

2

1

0 3 (b) M. macquariensis

2

1

0 2 (c) M. ambigua

1

0 0.4 (d) G. marmoratus

0.2

0.0

3 (e) R. semoni

Catch per operation per Catch 2

1

0 1.0 (f) M. peelii peelii

0.5

0.0 (g) B. bidyanus 0.4

0.2

0.0

0 10203040506070 Number of operations Figure 2.3. The effect of sample size and re-sampling strategy on 95% confidence intervals of mean catch for (a) C. carpio, (b) M. macquariensis, (c) M. ambigua, (d) G. marmoratus, (e) R. semoni, (f) M. peelii peelii, and (g) B. bidyanus. The upper and lower confidence intervals are plotted as consecutive (black) and random (grey) re- sampling strategies.

Correlation of species within operations Of the 21 possible inter-species correlations, none identified significant relationships between any two species. Only one pair of species, C. carpio and M. ambigua had a significant association (r = 0.21, p = 0.008), however, this relationship was not significant when Bonferroni correction was applied.

An ecological approach to re-establishing Australian freshwater cod populations 22 FRDC Report 2003/034

Analysis of fish community structure amongst grouped samples There were significant relationships between fish community composition and proximity of grouped samples when analysed at the scale of five consecutive operations (Global R = 0.-196, p = 0.001) and ten consecutive operations (Global R = -0.279, p = 0.010). This suggests that there was significant clustering amongst grouped samples (i.e. the fish assemblage in neighbouring parts of the reach was more similar than that across distant parts of the reach).

Estimating species richness Based on re-sampling of the entire dataset, 25–35 consecutive operations were required to predict the total species richness (i.e. to approach the asymptotic value of that achieved from 157 operations) (Figure 2.4). Sampling based on consecutive operations beyond this level of effort led to overestimation of total species richness with all estimators (>35 Jackknife 1, 30–60 Jackknife 2, 40–50 Chao 2), although, use of Jackknife 2 and Chao 2 result in comparable estimates of the true species richness by 65 and 55 operations, respectively (see asymptote of species richness curves in Figure 2.4). In comparison, the random re-sampling strategy predicted total species richness from 20 operations and thereafter consistently tracked the total species richness (Figure 2.5).

12 a) Jackknife 1 b) Jackknife 2 10

8

6

4

2

0 12 c) Chao 2 d) Average richness 10 Species richness Species 8

6

4

2

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Number of operations

Figure 2.4. The predicted species richness for the study reach based on a consecutive re-sampling strategy using three species richness estimators. The means (± SD) are calculated based on 1000 iterations (from n samples) and are represented by black and grey lines, respectively. The dashed line represents the total species richness detected in this study from 157 operations.

23 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

12 a) Jackknife 1 b) Jackknife 2 10

8

6

4

2

0 12 c) Chao 2 d) Average richness 10 Species richness Species 8

6

4

2

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Number of operations

Figure 2.5. The predicted species richness for the study reach based on a random re- sampling strategy using three species richness estimators. The means (± SD) are calculated based on 1000 iterations (from n samples) and are represented by black and grey lines, respectively. The dashed line represents the total species richness detected in this study from 157 operations.

DISCUSSION

Estimating species richness at a site Length of stream surveyed and survey effort was shown to affect the detection of rare species and therefore the estimates of total species richness in this study. The finding parallels that of North American studies in revealing that a substantial level of sampling effort must be applied within a threshold extent of river to obtain representative estimates of fish diversity (Cao et al., 2001; Hughes et al., 2002; Smith & Jones, 2005). The current study investigates the issue under Australian conditions. Conversely, boat electro-fishing has typically been applied at river sites for watershed-level sampling (Smith & Jones, 2005) with the purpose of making spatial comparisons or comparing representative samples of Australian fish communities through time (Harris & Gehrke, 1997; MDBC, 2004a). In these programs, effort in the order of 8–15 replicate boat electro-fishing operations represent standard practice at a site, with operations ranging from 2–5 min of elapsed-time (Appendix 4). This level of effort is usually applied to about one kilometre of river (Appendix 4) and species richness predictors have not been used. While this approach is effective for reporting on the health of fish communities over a large extent through time or for identifying the possible effects of large scale impact (e.g. river regulation), results from the current study indicate that this level of effort is unlikely to provide an accurate estimate of species richness at the smaller scale of a river reach (comparable to a site).

An ecological approach to re-establishing Australian freshwater cod populations 24 FRDC Report 2003/034

In the current study an accurate assessment of species richness at a site was achieved by a) increasing the standard electro-fishing effort to at least 20–25 operations positioned randomly and comprising operations in the order of 90 s on-time (comparable to 5 min elapsed-time), and b) applying a species richness estimator. Data sub-sampling demonstrated that an accurate estimate of species richness could be obtained from a moderate increase in sampling effort compared to that more commonly used in watershed-level sampling in the Murray–Darling Basin (Appendix 4). A secondary issue (besides increasing sampling effort) is the spatial application of operations at a site. Comparison of consecutive (current practice by state agencies) and random sampling strategies revealed that the latter provides a superior basis for accurately predicting total species richness (compare Figure 2.4 and Figure 2.5). Specifically, estimates levelled more rapidly with random as opposed to consecutive sampling (Figure 2.4, Figure 2.5). This reflects a level of auto-correlation in the data. Specifically, sampling in adjacent areas of a study reach (represented by analysis of consecutive operations) was found to increase the likelihood of pseudo- replicating the same subset of the fish community. By positioning the location of each replicate operation within a study area randomly, the likelihood of sampling alternative representations of the fish community present within the study area is increased. As a result, randomly distributing the sampling locations of individual replicate operations within a study area is more likely to collect a broader representation of the structure of a fish assemblage present, and consequently, is more likely to optimise the number of taxa collected.

While Jackknife 1 and Chao 2 required slightly more operations to provide an accurate species richness estimate (about 25), the Jackknife 2 estimator attained this earlier, but showed consistently higher variation in relation to increasing sampling effort (Figure 2.5). The optimal estimator is likely to depend on the river, as the number or proportion of species with a total abundance of 1 or 2 or that occur in only one or two operations leads to estimator specific bias (Cao et al., 2004). The transferability of these findings to other sites, rivers and times is a necessary step in developing an appreciation of the effect of sampling effort on estimates of fish species richness in Australian lowland rivers.

Survey effort in excess of that recommended here for predicting representative species richness is required to sample total species richness at a site. This is typical in community survey (Cao et al., 2001; Cao et al., 2002; Hughes et al. 2002). In this regard, the detection of two locally rare species, G. marmoratus and B. bidyanus, required on average 39 and 31 operations, respectively, in the current study. Gadopsis marmoratus is a benthic species that frequently shelters in structure (Koehn & O’Connor, 1990; Jackson et al., 1996) and is probably poorly sampled by boat electro-fishing in deep turbid rivers. It is also at the downstream edge of its distribution at this site (Jackson et al., 1996). Bidyanus bidyanus is now patchily distributed and rare in much of the Murray–Darling Basin (Clunie & Koehn, 2001). Perhaps of more concern is the low detection of M. peelii peelii and R. semoni, both of which would not be generally considered rare in the Murray–Darling Basin (MDBC, 2004b; Lintermans & Phillips, 2005; Lintermans, in press)(also see section on catch per unit effort below). In fact, the maximum number of consecutive operations required to detect each of these four species, ranged from 44 to 74 (Table 2.2). This highlights the risk of not detecting species when boat electro-fishing with pre-existing levels of survey effort in Australian rivers (Appendix 4). The amount of effort

25 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 required to obtain an estimate of true species richness at a site in Australian systems is likely to be river specific, habitat specific and dependent on temporal scale if North American studies serve as a guide (e.g. Hughes et al., 2002). Species richness estimates from a number of riverine sites are required before the transferability of this finding can be appreciated in the Australian context. In turn this has the capacity to influence the design of watershed-level fish community sampling programs (e.g. MDBC, 2004a) in addition to improving localised studies that occur within river reaches (e.g. where a threatened species exists as a small population fragment).

Fish community distribution Inter-specific patterns in distribution were generally not evident at the within operation scale in the current study. This possibly reflects a low diversity fauna comprising generalists that do not require strong delineation in habitat use, or a lack of heterogeneous habitat at the site. Alternatively, species associations or disassociations exist but were not detected. Indeed, lowland Murray–Darling Basin fishes are known to exhibit inter-specific differences in habitat-use at finer resolution than that examined in the current study (Koehn, 1997; Koehn & Nicol, 1998; Crook et al., 2001; Boys et al., 2005).

The highly sought after recreational species M. ambigua and M. peelii peelii were encountered primarily away from close proximity to the township of Narrandera (Figure 2.2),whereas, the nationally endangered (and fully protected) M. macquariensis was distributed throughout the study reach and was the second most common species in the survey. The distribution and abundance of the latter species was not unexpected since the reach represents a successful re-stocking site at least in the first generation for the species (Gilligan, 2005b). Habitat-use by these three species and in particular the Maccullochella spp. is generally similar in that they have an affinity for structure including structural woody habitat (Koehn & Nicol, 1998) and does not explain the difference in distribution between M. macquariensis and both M. ambigua and M. peelii peelii. A plausible explanation is that recreational angling effort is impacting on stocks centred close to Narrandera. The stocking success of M. macquariensis and the fact that catch and release is practised due to its protected status probably account for their distribution in this area, whereas, the other two species are being retained by anglers. Similar patterns have been observed from study of the endangered eastern freshwater cod M. ikei Rowland (Wooden unpublished data), the giant Australian crayfish Astacopsis gouldi Clark (Walsh & Nash, 2002) and Murray River crayfish Euastacus armatus (von Martens) (Lintermans & Rutzou, 1991).

Catch per unit effort Findings of this study indicate that the amount of sampling effort required to accurately estimate catch per unit effort is species-specific. Mean catch per unit effort of species that were infrequently caught (B. bidyanus, G. marmoratus, M. peelii peelii) could not be distinguished from nil catch based on lower confidence intervals (Figure 2.3). In contrast, the lower confidence intervals were above zero for the three most abundant species (Figure 2.3). That variation in estimates of catch per unit effort was a function of the total count of a species is consistent with the findings of others (Cao et al., 2002). Patterns of aggregation can also affect estimates of catch per unit effort (Pennington & Volstad, 1994). For instance, despite similar total catch of M. ambigua (n=45) and R. semoni (n=41), the catch per unit effort of the latter was

An ecological approach to re-establishing Australian freshwater cod populations 26 FRDC Report 2003/034 less precise (Figure 2.3). For example, R. semoni was only detected in 11 operations in the current survey and was abundant when found, leading to high variation in estimates of catch per unit effort (Figure 2.2). In contrast, M. ambigua was found in 36 operations and occurred in low abundance within operations, leading to reduced variation in estimates of catch per unit effort (Figure 2.2).

Of the three species for which catch per unit effort was estimated with a high level of precision, two, C. carpio and M. ambigua are widespread and abundant in the Murray–Darling Basin. These species represent possible indicator species for use in adaptive management exercises or to monitor shifts in the fish fauna. In contrast, catch per unit effort of the recreational angling species M. peelii peelii reflects a) low density of the species in this part of the Murrumbidgee River, and/or b) that boat electro-fishing is inefficient for detecting this species based on even large amounts of sampling effort. The former is most likely assuming that the abundant congener M. macquariensis is similar to M. peelii peelii in its morphology and behaviour, and therefore in detection by electro-fishing. Either way, the efficacy of techniques for monitoring M. peelii peelii stocks requires scrutiny if this key recreational species is to be effectively managed.

Substantial sampling effort is required to estimate catch per unit effort of rare species with precision let alone to estimate population size, and represents a major obstacle in managing threatened fishes. The B. bidyanus population in the Murrumbidgee River provides an example (Gehrke et al., 1995; Gerhke & Harris, 2000; Gilligan, 2005b; this study). Conversely, the endangered M. macquariensis is at sufficient density in this reach of the Murrumbidgee River for catch per unit effort to be measured with precision, therefore affording the opportunity to conduct adaptive management (Walters & Holling, 1990) at the within-reach scale. In river systems comprising numerous threatened fishes, one of the major functions of surveys is detecting remnant populations. A major challenge in the Murray–Darling Basin is then conducting focussed research including adaptive management on these remnants where success and failure can be measured at the population level. Furthermore, single-pass electro-fishing is likely to be a particularly important method for surveying these remnant populations since the increased potential for injury and mortality as a function of multiple-pass electro-fishing becomes unacceptable from a conservation perspective (Kennard et al., 2006).

Limitations of this study It must be acknowledged that the current survey design had a number of limitations. Shallow water posed the greatest problem in this regard and certain large and complex woody habitat could only be accessed at the exposed edges. Reduced electro-fishing effectiveness probably also occurred in association with deep holes (~3–6 metres). Nevertheless boat electro-fishing proved a useful technique that could be repeatedly applied along a substantial length of river. Furthermore boat electro-fishing has been demonstrated to be an especially useful technique for sampling fish communities in large lowland rivers including in the Murray–Darling Basin (Faragher & Rodgers, 1997; MDBC, 2004a). Instead of using this single gear type, use of multiple gear types would likely have sampled particular species more effectively and may have increased the number of species detected. For instance, methods other than electro-fishing have detected rare species, when multiple gear types have been applied simultaneously (Faragher & Rodgers, 1997; MDBC, 2004a; Lintermans et al., 2005a;

27 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Kennard et al., 2006). Boat electro-fishing is also biased towards catching large- bodied fish (Zalewski & Cowx, 1990). Furthermore, small species, including Hypseleotris spp., can exist at low density in the river channel and remain inactive for substantial periods of time, making them difficult to detect (Ebner, personal observation).

Fishes can also be absent from the river channel in time and space. Native species including Murray River rainbowfish Melanotaenia fluviatilis (Castelnau), Gadopsis marmoratus, bony herring Nematolosa erebi (Günther), and Hypseleotris spp. and the alien species redfin perch Perca fluviatilis Linnaeus and goldfish Carassius auratus Linnaeus are detected periodically in low abundances (Gilligan, 2005b; Baumgartner, in press; Wooden, unpublished data) in the vicinity of this study reach. Baumgartner (in press) recorded 12 species from a section of river immediately downstream of our study area. However in that study, M. fluviatilis, Hypseleotris spp., and N. erebi were only captured within the Yanco Weir pool (see location in Figure 2.1) in lentic waters. These species are common in the lower reaches but are typically rare in the middle reaches of the Murrumbidgee River possibly as a function of human impacts, including barriers to migration and thermal pollution (Gilligan, 2005b; Baumgartner, in press).

Recommendations We recommend that in Australia a) species richness predictors be incorporated into surveys of fish communities in large lowland rivers, to guide effort in the field and to improve data analysis (Hughes et al., 2002; Chao et al., 2005), and b) that the empirical approach used in this study be replicated at a small number of sites and times. In regard to the latter, boat electro-fishing at lowland river sites exceeding 1 km in length, is worthy of attention. In this way an appreciation of the trade-offs in allocating sampling effort within and among sites can be obtained, and the costly methods of surveying fish species richness further refined (cf. Pennington & Volstad, 1994). This study challenges the use of relatively uniform amounts of boat electro- fishing effort at lowland sites in the Murray–Darling Basin (Harris & Gehrke, 1997; MDBC, 2004a) and directs attention to the issue of river-specific sampling-effort based on indications from studies of fish communities in temperate North American rivers (Cao et al., 2001; Cao et al., 2002; Hughes et al., 2002; Smith & Jones, 2005).

An ecological approach to re-establishing Australian freshwater cod populations 28 FRDC Report 2003/034

CHAPTER 3. MOVEMENT OF HATCHERY-REARED AND WILD-REARED MACCULLOCHELLA MACQUARIENSIS STOCKED IN A LARGE LOWLAND RIVER

Brendan Ebner1, 2 and Jason Thiem1, 2

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

Ebner, B. & Thiem, J. (2006). Movement of hatchery-reared and wild-reared Maccullochella macquariensis stocked in a large lowland river. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.29–49. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

INTRODUCTION

Effective species reintroduction programs (Wallace, 2000) including those for threatened fishes (Brown & Day, 2002) are usually based on a detailed platform of experimental research. Comparison of wild and hatchery-reared fish in the hatchery (Metcalfe et al., 2003; Salvenes & Braithwaite, 2005) and the wild (Thorstad et al., 1998; Dieperink et al., 2001; Bettinger & Bettoli, 2002) has provided useful insights into the shortcomings of stocking hatchery-reared fish especially endemic salmonids in the Northern Hemisphere. There has not been a similar research focus in Australia, despite widespread stocking of freshwater fishes including threatened endemic percichthyids (Lintermans et al., 2005b; Gillanders et al., 2006; Lintermans, 2006).

Maccullochella macquariensis (Cuvier) is a nationally endangered freshwater fish of the family Percichthyidae, endemic to rivers in the southeast of the Murray–Darling Basin (Ingram & Douglas, 1995). The species is listed as endangered in the Australian Capital Territory (ACT) and New South Wales (NSW) and critically endangered in Victoria (Morris et al., 2001; DSE, 2003; ACT Government, 2006). The species is known to have existed historically in the Murray, Murrumbidgee and Macquarie rivers in NSW/ACT and the Goulburn, Broken, Campaspe, Ovens, King, Buffalo and Mitta Mitta rivers in Victoria (Berra & Weatherley, 1972; Berra, 1974; Cadwallader, 1977; Greenham, 1981; Cadwallader & Gooley, 1984). The species is expected to also have existed in the Lachlan catchment in NSW, however there are no historical records of this species existing in that catchment (Berra & Weatherley, 1972). Between 1950 and 1970, there were large-scale reductions in distribution from much of this former range (Cadwallader & Gooley, 1984; Douglas et al., 1994). In the Canberra region, the species became locally extinct in the late 1970s (Lintermans et al., 1988). There are now only three self-sustaining populations: a natural remnant population in the Murray River below Yarrawonga; and two translocated populations, in Seven Creeks above Euroa in Victoria, and in Cataract Dam south of Sydney in NSW. The Seven Creeks population is the result of translocations of ‘cod’ from the Goulburn River in 1921 and 1922 (Douglas et al., 1994). The Cataract Dam

29 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 population is the result of translocations from an unknown source population in 1914 (Gehrke & Harris, 1996).

Efforts to recover the species have involved the largest freshwater conservation- stocking program in the country (Lintermans et al., 2005b; Lintermans, 2006). Since 1986, hatchery produced fingerlings have been stocked into areas where the natural population was locally extinct in NSW, Victoria and the ACT, in an effort to re- establish populations (Douglas et al., 1994; Gilligan, 2005b). Large rivers have been the focus of stocking efforts as they were assumed to be the primary habitats of M. macquariensis. There has generally been minimal or sporadic monitoring at reintroduction sites, and knowledge of successful survivorship has been mainly from angler reports. Survival and growth of fingerlings through to the age of 2–3 years (juveniles) has been documented at a number of sites (Faragher et al., 1993; Lintermans, 1995; Brown, 1998; Brown & Nicol, 1998; Douglas & Brown, 2000; Gilligan, 2005b; NSW DPI, 2005). The number of 3–5 year old individuals (sub-adult or young adults, Harris & Rowland, 1996) captured in monitoring programs is much lower. Reports of large adult individuals have been uncommon. Possible explanations for the scarcity of adults in re-stocked populations are: 1) larger fish (adults) are subject to high mortality including angling pressure and/or natural mortality; 2) larger fish remain at the release site but are undetected by monitoring; and/or 3) juveniles remain at the release site until they reach sub-adulthood at which stage they disperse, becoming difficult to detect.

Generally, fish populations suffer greatest natural mortality in early life history phases (Jones, 1991) and this presumably applies to M. macquariensis (Todd et al., 2004). Widespread concealment of the rapid harvest by anglers, of all M. macquariensis as they approach or soon after they attain adult size is unlikely. Therefore it is unlikely that these larger fish are subject to high mortality, though, an understanding of mortality in this species is a major knowledge gap in recovery of the species (Simon Nicol, Department of Sustainability and Environment Victoria, personal communication). It is also unlikely that larger fish remain at the site and are not detected by monitoring. Electro-fishing has been a commonly employed monitoring technique for M. macquariensis and other freshwater cod (Faragher et al., 1993; Growns et al., 2004) and has proven successful in detecting a wide range of size classes at a subset of sites (Brown & Nicol, 1998; John Koehn, Department of Sustainability and Environment Victoria, personal communication; Ian Wooden, unpublished data; Chapter 2). Certainly this is the case in the Murrumbidgee River at Narrandera, where sub-adult and adult fish are common (Ian Wooden, unpublished data; Chapter 2).

This study tests the possibility that post-juvenile dispersal is the reason for the scarcity of sub-adult and adult records from stocking locations. Despite indications that juveniles can be gregarious (Douglas et al., 1994), it is possible that as individuals grow and mature they become territorial, leading to dispersal from the release sites, since under hatchery conditions individuals become particularly aggressive towards one another from a very early age (Brett Ingram, Department of Primary Industries Victoria, personal communication). Therefore, to improve re-establishment of populations of M. macquariensis it is important to determine the movement patterns of sub-adults following stocking in rivers. In this study, stocking of on-grown fish is also tested as an alternative reintroduction strategy to releasing fingerlings.

An ecological approach to re-establishing Australian freshwater cod populations 30 FRDC Report 2003/034

METHODS

Site description The study was located in a lowland reach of the Murrumbidgee River, between Berembed and Gogeldrie weirs (105 river km) in southern NSW, Australia (Figure 3.1). This is one of twelve M. macquariensis stocking sites in the Murrumbidgee catchment, with 85 000 fingerlings stocked at this location (Narrandera) between 1996 and 2000 (Gilligan, 2005b). Good numbers of wild fish captured by subsequent surveys indicate the reach to be the most successful stocking site for the species in NSW (Growns et al., 2004; Gilligan, 2005b). In this reach, water is diverted at Berembed, Yanco and Gogeldrie weirs to meet irrigation requirements between spring and autumn. In-stream flow remains a function of rainfall during winter (Ebsary, 1992). The geomorphology includes a transition from lower confined floodplains to open floodplains (Young et al., 2001). River red-gum Eucalyptus camaldulensis is common along much of the riparian edge of the river and channel widths are in the order of 70 m (Growns et al., 2004). River depths of 3–5 m are commonly encountered on outside bends. The dominant in-stream habitat comprises structural woody habitat consisting of fallen trees or branches, particularly of river red-gum.

Figure 3.1. Location of the study reach along the Murrumbidgee River, New South Wales, Australia.

31 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Source of fish Two experimental release groups were used for this study. The first (hatchery) group, comprised 29 M. macquariensis (two years of age, 310–429 mm total length (TL); 0.513–1.567 kg) sourced from Snob's Creek Research Station in Victoria and transported to the Narrandera Fisheries Centre in NSW. These individuals were bred from wild broodstock (Murray River natural population). Initially, post-larvae were reared in a fertilised earthen pond under semi-natural conditions. After harvesting the pond, fingerlings were transferred to hatchery facilities where they were weaned onto an artificial extruded diet, then on-grown in 500 L circular fibreglass tanks that were part of an intensive recirculating system. Methods used for fry rearing, weaning and on-growing were similar to those employed for Murray cod (M. peelii peelii (Mitchell)), which are described in more detail by Ingram (2004). The second (wild) group, comprised 32 M. macquariensis (age unknown, 370–635 mm TL; 0.599–3.704 kg) originally stocked as fingerlings and subsequently collected by boat electro-fishing from the Murrumbidgee River at Narrandera prior to surgery (Chapter 2). These individuals were collected between 10.3 km downstream and 4.7 km upstream of the release site, and a GPS record was obtained at each capture location (Chapter 2). Hatchery M. macquariensis were characterised by heavier individuals than the wild group of the same length (Figure 3.2), due to large fat deposits in the peritoneal cavity.

Surgery and release Radio-tags (F1830, 35, 40 and 50, Advanced Telemetry Systems (ATS), Isanti, USA, ≤ 2% of fish weight, two-stage radio-transmitters, 150–152 MHz, pulse coded, duty cycle of five seconds on/ seven seconds off) were surgically implanted into hatchery and wild fish from 12–16 September and 16–24 September 2003, respectively. The process involved anaesthetising individuals using 5 ml Alfaxan (Jurox, Rutherford, Australia) per 10 L of water, and inserting the radio-tag through a 3 cm incision to the peritoneal cavity (along the ventral line anterior of the anus and posterior to the pelvic fin base). A modified cannular was used to exit the aerial of each radio-tag from the anterior of the caudal peduncle. The incision was closed with two or three sutures, and the temporary skin adhesive Vetbond (3M, St. Paul, USA) was applied and given approximately 20–30 sec to dry. An intra-muscular injection of the antibiotic enrofloxacin (Baytril) was administered at 0.1 ml kg-1, immediate-posterior of the nape. For external identification, individuals were tagged with a dart tag between the second and third dorsal spines. Individuals initially recovered in a darkened enclosure holding 200 L of aerated water at 5 parts per thousand NaCl. Upon regaining swimming ability, individuals were transferred to large circular concrete enclosures that held between 500 and 1000 L at 5 ppt NaCl. Hatchery and wild individuals were held in separate enclosures, with each of the two largest wild individuals held in separate enclosures to prevent aggressive interactions. One wild individual died immediately post-surgery and autopsy revealed internal bleeding from a severed artery within the peritoneal cavity.

On 25 September 2003, 21 hatchery individuals implanted with radio-tags were released into the Murrumbidgee River, five river kilometres upstream of Narrandera (Figure 3.1). On 26 September, 29 wild individuals implanted with radio-tags were released at the same location. Surgery was repeated on a further 10 individuals (two wild and eight hatchery individuals) that showed signs of loose sutures and/or open incisions, with three hatchery individuals rejecting their first radio-tags. On 1 October,

An ecological approach to re-establishing Australian freshwater cod populations 32 FRDC Report 2003/034 two hatchery individuals were euthanased since they had developed severe infection around the incision and showed no sign of healing externally (despite swimming and behaving similarly to other healthy individuals). The remaining six hatchery and two wild individuals had healed incisions and were released on 1 October.

Radio-telemetry A two-person crew, aboard a 3 m aluminium punt powered by an 8 hp, two-stroke outboard motor, undertook manual radio-tracking approximately fortnightly for the first 3 months and then monthly up until 13 months post-release. Individuals were located during daylight hours using a scanning receiver (Australis 26k, Titley Electronics, Ballina, Australia or R4100, ATS) and a three-element Yagi antenna (Titley Electronics or ATS). Spatial locations of individuals were recorded by taking three waypoints with a hand-held GPS (Garmin GPSII Plus or Garmin GPS 76 Marine Navigator; Figure of merit (F.O.M.) ≤ 5.0 at about 95% of locations, F.O.M. ≤ 7.0 at 100% of locations).

A record was made of the habitat occupied by each individual, including positioning within either a straight, the inside bend or the outside bend of a river channel. Water depth (± 0.1 m) was recorded from a depth sounder (Eagle™ Strata 128, Catoosa, USA) mounted to the stern of the boat. Distance from the bank was estimated by both crew members and averaged. Structural habitat that each individual was contacting was characterised as structural woody habitat, clay bank, macrophytes, or an absence of structure referred to as open-water. Habitat was determined visually and from depth sounder images obtained on repeated passes.

From 12–13 November 2003, remote radio-telemetry data loggers (DCCII Model D5041, ATS) connected to a receiver (R4100, ATS) and powered by 12 V power supply were positioned at Berembed and Yanco weirs, 75 km apart (see Figure 3.1). These loggers scanned 60 frequencies sequentially (58 implanted radio-transmitters, and two reference transmitters that were retained in the vehicle and used during field exercises) over approximately 30 min. These telemetry loggers were disconnected on 14 November 2004. The approximate detection range of these loggers was 1 km.

33 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

4000

3000

2000 Weight (g) 1000

0 0 100 200 300 400 500 600 700

Total length (mm) Figure 3.2. Length-weight comparisons of hatchery reared (●) and wild (□) M. macquariensis implanted with radio-tags.

Data analysis The average of three replicate spatial points recorded at the location of an individual was used to minimise GPS error (except where one of these replicates was clearly an outlier i.e. >10 m from either of the other replicate points in which case the two similar replicates were averaged). Large-scale movement of individuals was quantified from records of the remote telemetry logger stations combined with those arising from manual radio-tracking. Movement distances were calculated using ArcView 3.2™ based on maps digitised at the 1:25 000 scale. For large-scale movement and dispersal, each spatial location was shifted onto the river mid-line. A polyline was generated based on the sequential locations of each individual using the Movement Extension in ArcView (Hooge & Eichenlaub, 1997). This was used to construct a time series of a) the proximity of individuals to the release point, and b) distance moved between consecutive radio-tracking events. Dummy values were not used if individuals were not located during manual radio-tracking.

Homing was categorised as return to a location within one home-range, in either an upstream or downstream direction, from where an individual was originally captured (relevant to wild fish only), based on a maximum home-range size of 272 m for a subset of this group (Chapter 4). Total range was calculated as the distance between the furthest upstream and downstream location of an individual in the study. This did not include the capture locations of wild individuals but did incorporate records from the remote telemetry loggers.

An ecological approach to re-establishing Australian freshwater cod populations 34 FRDC Report 2003/034

Home-range was estimated from monthly locations of individuals beginning 8 weeks after release, as at this stage >90% of wild individuals had established fidelity to a site, and so any potential biases from the use of a single release site are removed. Site fidelity was classified as repeated monthly locations (i.e. a minimum of two locations) <272 m upstream or downstream from a previous location. In general, movements were either contained within this length of river or vastly exceeded this distance. Linear home-range estimates were taken as the minimum direct distance between upstream and downstream extremities, where site fidelity was clearly established (inaccuracy of the base-map prevented calculation of linear home-range along the river channel midline). Minimum convex polygons (MCP) were used to estimate home-range (Hooge & Eichenlaub, 1997), where a minimum of 10, monthly fixes was obtained unless otherwise stated.

Water temperature was collected at half-hourly intervals using a Hydrolab (Sue Vink, CSIRO Land and Water, unpublished data) and averaged to provide a daily record. River discharge was collected at daily intervals (NSW DNR, 2004). Statistical analyses were conducted in Statistix for Windows (version 2.0) and transformed, where necessary, to achieve the assumption of normality (Tabachnick & Fidell, 1989).

RESULTS

Dispersal Dispersal from the release site was characterised by a different scale of movement between the two M. macquariensis groups (Figure 3.3). Individuals from the wild group remained within 5 km of the release site (upstream or downstream) for the majority of the study with the exception of a maximum of four individuals on any fortnightly or monthly radio-tracking exercise (Figure 3.3). Typically these exceptions involved movements up to 15 km upstream or downstream from the release site, although, one individual was recorded 40 km upstream of the release site over a four and a half month period (Figure 3.3). Conversely, the hatchery group exhibited pronounced modal classes 0–5 km and 5–10 km downstream of the release site, plus large-scale downstream movement of individuals within one month of being released (Figure 3.3). Two individuals that had passed through Yanco Weir and were 40–50 km downstream of the release site represented the largest of these downstream movements. At 7–8 weeks post-release the location of individuals from the hatchery group was similar to the previous fortnightly fix and generally remained stable until the completion of the study (Figure 3.3). Additionally, a hatchery individual that was undetected at 4 weeks post-release was located 10–15 km upstream of the release site 7 weeks post-release where it remained for the entire study (Figure 3.3).

Up until 5–6 weeks after release both the hatchery and wild groups remained close to the release site, with the hatchery group also dispersing downstream whilst the wild group moved in either direction (Figure 3.4a and b). After this time, there was minimal movement by hatchery individuals (Figure 3.4b). Similarly, there was minimal movement by the majority of wild individuals, however, a subset of individuals displayed pronounced upstream and/or downstream movements (Figure 3.4a). These movements were not synchronised, occurring in November and December 2003, and April, May and September 2004 based on monthly radio- tracking exercises (Figure 3.4a). These movements showed no discernible relationship

35 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 with changes in discharge, temperature (Figure 3.4c) or expected spawning period (Koehn & O’Connor, 1990).

25 20 (a) 1 month n= 26 15 n= 31 10 5 0 25 20 (b) 3 months n= 26 15 n= 30 10 5 0 25 20 (c) 6 months n= 26 15 n= 30 10 5 0 25 20 (d) 9 months n= 22 15 n= 24 10 5 Number of individuals 0 25 20 (e) 12 months n= 19 15 n= 20 10 5 0 5-0 0-5 10-5 5-10 65-60 60-55 55-50 50-45 45-40 40-35 35-30 30-25 25-20 20-15 15-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45

(downstream) (upstream) site Yanco Weir Berembed Weir Release Gogeldrie Gogeldrie Weir Distance (km)

Figure 3.3. Dispersal of hatchery (■) and wild (■) M. macquariensis groups in the Murrumbidgee River. Time after release: (a) 1 month, (b) 3 months, (c) 6 months, (d) 9 months, and (e) 12 months. Distance moved is expressed as upstream and downstream direction in 5 km categories.

An ecological approach to re-establishing Australian freshwater cod populations 36 FRDC Report 2003/034

40 (a) 20

0

-20

-40 40 (b)

20

Distance moved (km) 0

-20

-40 15000 30 (c) )

-1 25 C) o 10000 20

15 erature ( erature

5000 10 p

5 Tem Discharge (Ml day (Ml Discharge 0 0 Oct Dec Feb Apr Jun Aug Oct 2003 2004

Figure 3.4. Minimum distances moved between radio-tracking intervals (as detected by manual radio-tracking) by all (a) wild and (b) hatchery M. macquariensis and, (c) the water discharge ( ) and temperature (--) in the Murrumbidgee River at Narrandera. Positive and negative values indicate a direction of movement upstream and downstream, respectively.

37 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

In addition to movements detected by monthly and fortnightly radio-tracking, the downstream (Yanco Weir) telemetry logger (set-up after two hatchery individuals had passed through the weir (Figure 3.3)) detected the presence of five individuals (one hatchery and four wild M. macquariensis) encountering the weir on which it was stationed (Appendix 6). Interestingly all five individuals did not stay within detection range of the logger for more than a single logger cycle (approximately 30 mins). The hatchery individual was the first record on the logger (November 2003), travelling approximately 11 km downstream from its normal home-range and then returning to its previous location. This individual displayed the fastest recorded swimming speed, covering 10.9 km in a maximum time of four hours. In all four instances of wild individuals encountering Yanco Weir, individuals vacated an established home-range to undertake downstream movements of between 27 and 33 km, followed by return journeys to the vacated home-range. One of these movements occurred in December 2003, two in February 2004 and another in April 2004. All journeys were circular, with fish subsequently located at the previously occupied structural woody habitat. Minimum swimming speeds for these individuals varied from 0.04 to 0.32 km h-1.

Confirmed survival of individuals from the two groups, wild and hatchery, differed greatly throughout the course of this study (Figure 3.5). In the case of the wild group, 29 individuals were alive six-months after release (Figure 3.5a). One individual had a failed radio-transmitter (heard failing during radio-tracking) and one individual was suspected (and later confirmed) as a mortality or radio-tag rejection. Twelve months after release 19 individuals were alive and being radio-tracked, with 11 radio- transmitter failures and no further radio-tag rejections or mortality (Figure 3.5a). A light aircraft was used to search for a number of these failed transmitters and verified they were not operational in the study area or surrounding areas. One individual was reported as an angler capture and subsequent release. This fish was radio-tracked following capture for three months until its radio-transmitter failed.

For the hatchery group, survival of individuals declined rapidly, with only 14 individuals alive after one-month (Figure 3.5b). This number decreased to six individuals two-months after release, with three individuals alive after 12 months. Three radio-tags were recovered from shallow water during low-flow events in the river at three, six and nine-months after release. Further, a large number of radio-fixes for hatchery individuals were taken in repeated locations for the duration of the study from ‘suspect’ locations in either relatively shallow water, or open water devoid of suitable structure, or associated with access points along the river. Three of these radio-tags were verified as radio-tag rejections or mortality through 24 hr radio- tracking (i.e. no movement detected), as was one radio-tag from the wild group. In these cases it could not be determined if rejection of radio-tags or mortality had occurred. In contrast, the majority of wild individuals regularly moved short but detectable distances. As a consequence, further analyses were solely conducted on individuals from the wild group.

An ecological approach to re-establishing Australian freshwater cod populations 38 FRDC Report 2003/034

Alive Confirmed mortality or rejection Suspected mortality or rejection Failed radio-transmitter 35 a) 30

25

20

15

10

5

0 30 b)

25

Number of individuals Number 20

15

10

5

0 12345678910111213

Time elapsed (months)

Figure 3.5. Fate of radio-tagged (a) wild and, (b) hatchery M. macquariensis released into the Murrumbidgee River near Narrandera.

Homing in wild fish Sixteen of 31 wild individuals homed to their original capture location in this study (Appendix 6). The majority of individuals originally captured within 3 km either side of the release site (91%) homed to their capture location at some stage during the study. In contrast individuals captured between 3 and 5 km either side of the release site only homed on 40% of occasions. At capture distances greater than 5 km from the release site fish did not home back to the original capture site. Homing success was a

39 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 function of an individual having been captured in close proximity to the release site (T = 3.39; d.f. = 29, P <0.005). The direction (upstream or downstream of release site) did not affect homing ability (T = 1.96; d.f. = 14; P = 0.07).

Typically, homing to a previous capture location was rapid. Ten of sixteen individuals had homed by the second radio-track at 12 days post-release. Five of these individuals homed in less than seven days to locations within 2 km of the upstream side of the release site. The time required to home was greater for the remaining six individuals, taking between 21 and 67 days.

Individuals exhibiting homing behaviour typically remained at their home-site once they had returned, with the proportion of subsequent radio-locations at the home-site ranging between 31 and 100%. However, on one occasion an individual took 12 days to home, stayed in this location for approximately 8 weeks and then relocated to a different home-site.

Total range and home-range of individual fish The total range (mean ± standard error (SE)) of wild M. macquariensis was 8840 ± 2151 m for the entire study (Appendix 6). This was significantly different (T = 4.16; d.f. = 43.8; P <0.005) when compared to total range following establishment of home- ranges for 90% of the group (6660 ± 2202 m). This finding is consistent with localised home-range movements following initial dispersal and/or homing.

Twenty-nine of the 31 wild M. macquariensis either returned to their pre-capture home-range or established new home-ranges after initial dispersal movements. One individual was not recorded establishing a home-range as its radio-transmitter failed early in the project. A second individual was excluded from home-range analysis as it was verified as a radio-tag rejection or mortality following 24 hr radio-tracking. Based on the remaining 29 individuals, the mean linear home-range was 78 ± 13 m and ranged between 8 and 270 m. In the case of three individuals that had undertaken home-range shifts, more than one linear home-range was estimated where at least two radio-tracking fixes were obtained. Each of these estimates was of comparable size to individuals that did not undertake home-range relocations. The home-range area of 19 individuals (with a minimum of 10, monthly radio-tracks) based on minimum convex polygons was between 53 and 4073 m2, with a mean of 1070 ± 302 m2. Home-range overlap occurred for a small number of individuals immediately upstream of the release site including two individuals which inhabited the same log upon establishment of home-ranges, for the duration of the study.

A number of different types of movement over the 13-month study period were observed following the establishment of home-ranges for 29 wild M. macquariensis. These were grouped into three categories of movement that are discussed in turn. Examples of each are provided in Figure 3.6.

(1) Sedentary home-range movements Following home-range establishment, through either homing or dispersal, 18 of 29 M. macquariensis exhibited restricted movements for the entire study (Figure 3.6a). For example, nine of these individuals exhibited homing behaviour to their capture location post-dispersal, established tight home-ranges and were not located outside of these for the remainder of the study (or until transmitters failed in some cases). This

An ecological approach to re-establishing Australian freshwater cod populations 40 FRDC Report 2003/034 occurred for homing in upstream or downstream directions. The remaining nine individuals also exhibited strong site fidelity and only localised movements were observed following dispersal, however, these individuals set up homes in different locations to their capture location, in both upstream and downstream directions from the release site. It should be noted that one of these individuals either took much longer to establish a clear home-range or had a larger home-range than was calculated as it did not exhibit fidelity to a restricted area for consecutive periods of time until six months after release. Collectively this group of 18 individuals comprised 10 females, two males and six individuals of unknown sex (Table 3.1). TL (mean ± SE) of individuals was 443 ± 10 mm and weight was 1147 ± 95 g (Table 3.1).

(2) Home-range relocations Five individuals exhibited home-range relocations in this study (Figure 3.6b). This occurred when an established home-range (based on consecutive fixes at repeated locations) was vacated and a new home-range was established. Two female and three unknown sex M. macquariensis exhibited home-range relocation behaviour and mean (± SE) TL and weight for this group was 407 ± 11 mm and 838 ± 71 g, respectively (Table 3.1). In the case of four of these individuals, single home-range shifts were recorded, varying in distance from 863 m to 2.4 km (see example W12, Figure 3.6b). The fifth individual in this group (W6) exhibited four home-range shifts, at least one of which was to a previous home-range (Figure 3.6b). These shifts ranged in distance from 1 to 36 km, with the last recorded movement of 36 km to within five metres of a previous location five months earlier (W6, Figure 3.6b).

(3) Circular journeys Six individuals exhibited movements not associated with home-range movements or relocations (Figure 3.6c). Four of these individuals undertook single large-scale return movements and were recorded on the downstream Yanco Weir logger. These movements were characterised by an initial downstream movement of 27 to 33 km, a single record on the downstream logger and a subsequent return of the same distance to a home-range (see W26, Figure 3.6c). Another individual (W31, Figure 3.6c) undertook four large-scale movements following the establishment of a home-range. These movements ranged between 2.3 and 20 km and always culminated in a return to a discernible home-range. It should be noted that this individual was detected for 9.5 of the 13 months in a distinct home-range and exhibited fidelity to this location. The remaining individual in this group was located within its home-range on all except for two consecutive radio-tracking occasions. The scale of movement undertaken during these two months is unknown. This group of six M. macquariensis comprised of one female, two males and three individuals of unknown sex (Table 3.1). Mean (± SE) TL and weight for this group was 494 ± 20 mm and 1633 ± 223 g, respectively (Table 3.1).

Habitat-use Wild M. macquariensis displayed a preference for outside bend habitats (53.3% of locations) within their home-range (Appendix 6). The locations of home-range movements were less frequently encountered on straight sections of river channel (23.7%), inside bends (17.3%) and mid-channel (5.7%). M. macquariensis also displayed distinct patterns of habitat use within home-ranges. Structural woody habitat was the predominant habitat type used, comprising 83% of fixes. Open water

41 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

(13%), rock (2%), indeterminate habitat (2%) and undercut banks (<1%) were infrequently used and M. macquariensis were not found to use emergent macrophytes.

400 1600 (a) W1 W27 0 1200

-400 800

-800 400

-1200 0

-1600 -400 3000 W12 W6 (b) 40000 2500

2000 30000

1500 20000 1000 10000 500

0 0 20000 Proximity to release site (m) to release Proximity (c) W26 W31 0 15000

-10000 10000

5000 -20000 0 -30000 -5000 Oct Dec Feb Apr Jun Aug Oct Oct Dec Feb Apr Jun Aug Oct

2003 2004 2003 2004 Figure 3.6. Maccullochella macquariensis movements from October 2003 representing: (a) sedentary home-range behaviour for the entire study, following dispersal, (b) home-range relocations that occurred once or multiple times, (c) large- scale movements outside of home-ranges that did not result in a home-range relocation. The first point in each figure (zero) represents the release location (all fish were released at the same location).

An ecological approach to re-establishing Australian freshwater cod populations 42 FRDC Report 2003/034

Table 3.1. Summary statistics for 29 wild M. macquariensis exhibiting different types of movements following the establishment of a home-range. TL and weight are expressed as mean ± SE. Sedentary Relocations Circular journeys

Number 18 5 6

Sex 10♀, 2♂, 6? 2♀, 3? 1♀, 2♂, 3?

TL (mm) 443 ± 10 407 ± 11 494 ± 20

Weight (g) 1147 ± 95 838 ± 71 1633 ± 223

DISCUSSION

Low survival of hatchery-reared M. macquariensis and dissimilar dispersal of hatchery and wild groups recorded in this study, demonstrate the unsuitability of using on-grown fish to examine the dispersal patterns of those originally stocked as fingerlings. Instead, stocking of large on-grown individuals should be considered as trial re-establishment exercises based on stocking at a more advanced stage than as fingerlings. These trials are not trivial since stocking of on-grown M. macquariensis has been suggested as an alternative to stocking fingerlings. This suggestion is in response to the current low levels of fingerling production and the relatively poor success rate to date of stocking fingerlings in re-establishing populations (Todd et al., 2004).

Initially the issue of possible radio-tag rejection is discussed to facilitate dealing with specific aspects of survivorship and dispersal.

Possible rejection of radio-tags It is possible that hatchery fish (in this study) rejected radio-tags following release into the Murrumbidgee River. This is indicated by a small number of cases of rejection or signs of unsatisfactory recovery from surgery (e.g. incomplete closing of the incision, tearing of tissue at suture entry points) that occurred within days of implantation prior to release. This contrasts with rapid healing of the incision and complete retention of radio-tags by wild fish in this study (although two individuals showed minor signs of reduced incision closure). Furthermore, two-year old hatchery M. macquariensis showed complete recovery from implantation with radio-tags based on identical methods in an aquarium trial (Ebner et al., 2005). The unsatisfactory post-surgery recovery of about one third of hatchery individuals initially in the current study was probably a function of their notably rotund and fat laden abdomens relative to the wild fish (Figure 3.2). However it is unlikely that radio-tag rejection was widespread following release of the hatchery group since two-thirds of the sample were healing well following initial surgery and the remainder showed signs of healing following re-surgery (with the exception of the two individuals euthanased). Radio- tagging of excessively fat hatchery fish should be avoided in future studies. It would also be useful to minimize incision length and radio-tag size, and provide an increased holding period for observation post-surgery.

43 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Survivorship Survivorship of wild and hatchery fish at 13 months post-release was 95% and 9%, respectively. Comparable studies have centred on salmonids (Thorstad et al., 1998; Dieperink et al., 2001; Bettinger & Bettoli, 2002). Of these, both Dieperink et al. (2001) and Bettinger & Bettoli (2002) report higher survival of wild than hatchery fish. Predation mediated by morphological and behavioural differences was suggested as the explanation for the outcome in both studies. Similarly, differences in the body shape (Figure 3.2) and dispersal of hatchery and wild groups (Figure 3.3) represent possible explanations for higher survival of wild than hatchery groups in the current study. In contrast, Thorstad et al. (1998) recorded 77% and 9% survival of hatchery and wild salmon, respectively. It was argued that greater energy stores of hatchery fish resulted in their higher survival over winter (Thorstad et al., 1998). Clearly, better condition of hatchery fish did not confer such a benefit in the current study. More likely, the familiarity of wild fish with the river environment was critical to their success (Brown & Day, 2002).

The causes of mortality were unclear in the current study. A number of factors including low visibility, structural woody habitat and moderately deep water prevented retrieval of radio-tags or location of fish remains in most instances. In the Cotter River (ACT), positions of radio-tags provided an indication of avian and predation or scavenging following a previous release of M. macquariensis (Ebner, Johnston & Lintermans, unpublished data) with similar results in releases of salmonids overseas (Dieperink et al., 2001). Based on the position of radio-tags in the current study, (primarily associated with structural woody habitat), a distinction could not be made between radio-tag rejection and predator effects, let alone among broad predator groups (i.e. avian, riparian mammal, fish, anglers) that might have been involved. This highlights a need for improved methods for detecting the timing and cause of mortality of M. macquariensis (e.g. Dieperink et al., 2001) especially following releases in turbid lowland rivers.

Direction and magnitude of dispersal There was a difference in the dispersal patterns of the hatchery and the wild group in this study. The wild group returned to the reach, 13 km in length, from which they had originally been collected. Whereas the hatchery group remained near the release site or dispersed in a downstream direction from the release site. This is the first attempt to investigate movement of hatchery and wild Maccullochella species simultaneously, although, remarkably similar findings arose from a study of rainbow trout (Oncorhynchus mykiss (Walbaum)) in the Clinch River, USA (Bettinger & Bettoli, 2002). Fish were released immediately below a dam in that study, whereas, there was opportunity for upstream dispersal in the current study.

Comparison can also be made with the dispersal of wild M. macquariensis from a study in the Murray River (Koehn, 1997) where the species was found to exhibit little movement except during a major flooding period. Conversely, Ebner et al. (2005) released and radio-tracked nine hatchery-reared M. macquariensis in the Cotter River (within the upper Murrumbidgee River catchment). In that study, individuals remained within or near to the pool of release or exhibited net downstream dispersal. Collectively these studies indicate that downstream dispersal of on-grown individuals is likely and that individuals stocked as fingerlings or that naturally occur are capable of fidelity to a river reach.

An ecological approach to re-establishing Australian freshwater cod populations 44 FRDC Report 2003/034

Downstream dispersal of hatchery fish and the downstream explorations of four wild fish and one hatchery fish (involving encounter with Yanco Weir before returning home), poses a paradox for the management of this . It is widely accepted by river and fisheries managers that weirs pose a major ecological threat to riverine ecosystems and specifically migratory fishes within the Murray–Darling Basin (Harris & Mallen-Cooper, 1994; Lintermans & Phillips, 2004). Conversely, weirs may prove useful initially for concentrating M. macquariensis within small reaches in order to establish minimum viable populations (Gilpin & Soulé, 1986; Ebner et al., 2005). This approach to stocking would facilitate concentrated research, monitoring and compliance efforts (Ebner et al., 2005) and is especially pertinent in view of the low annual hatchery production of this species (Gilligan, 2005b) relative to that predicted as necessary for establishing wild populations (Todd et al., 2004).

Homing Half of the wild group were found to return to reoccupy the location from which they were originally captured in this study. It is considered that this represents an ability to home. Further, the rapid homing of a substantial number of individuals, including 30% of the group within a fortnight (i.e. by the second radio-tracking occasion) indicates effective searching capability and/or navigational capability of this species. It also represents the first evidence of experimental displacement and subsequent homing in M. macquariensis or any of the Maccullochella. Large-scale (10–100s km) return migrations have also been recorded by radio-tracking M. peelii peelii (Koehn, 1997), M. peelii mariensis (Rowland) (Simpson & Mapleston, 2002) and M. ikei (Rowland) (Butler, 2001), following release at capture locations. Collectively these findings indicate that all extant species of Maccullochella are capable of homing.

Homing success was a function of an individual having been captured in close proximity to the release site. Similarly, more than half of a sample of Macquaria ambigua (Richardson) were found to home following displacements of about 2 km (Crook 2004a; n=10) though not following displacement of ~25 km (Crook, 2004b; n=15). Crook (2004a) suggested that a number of non-homing individuals encountered high quality habitat not long after release and probably elected not to home. If in the future, threatened species of Maccullochella were to be translocated, it would be informative to have some understanding of the maximum scale at which they home and whether or not unsuccessful homing behaviour will affect the success of translocation.

Home-range use and dispersal Four types of behaviour were distinguished in this study following the establishment of a home-range by 29 wild individuals. These were: (a) occupation of a home-range, (b) home-range shift, (c) return home-range shift, and d) return exploration. The observed frequency of these behaviours is summarised in Figure 3.7. Not included are the individual that experienced radio-transmitter failure and another that either rejected a radio-tag or died early in the study. Each of these behaviours is outlined before placing these findings into the newly developed home-range establishment model of lowland fishes (Crook, 2004a, b).

(a) Occupation of a home-range Various definitions of home-range including those with a statistical basis are available, though, in general the term refers to the area in which an animal conducts

45 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 its everyday activity (Blundell et al., 2001). In the current study, home-ranges were small with mean length of river used 78 ± 13 m, ranging from 8 to 270 m and centred on structural woody habitat on outer bends. Further, these results were confirmed by intensive radio-tracking of a subset of individuals (n= 9, Chapter 4). Similarly, large hatchery-reared M. macquariensis have been recorded moving on average 47 to 292 m in the Cotter River (n=9, Ebner et al., 2005) and individuals in the remnant population of the Murray River traversed 100 to 500 m (n=4, Brown & Nicol, 1998).

Individuals were found to have overlapping home-ranges (also see Chapter 2, Chapter 4) and in some cases individuals exhibited long-term (1 year) co-occupation of structural woody habitat, in the current study. There is little comparable information of intraspecific interaction in percichthyids. Butler (2001) reported that M. ikei individuals (n=6) forage in close proximity to one another at the head of a pool in the Nymboida River. Collectively, these observations challenge the idea that Australian freshwater cod are wholly territorial and solitary. Understanding intraspecific interactions that occur in the home-range occupation phase is a necessary step toward re-establishing self-sustaining populations of Maccullochella.

(b) Home-range shifts A home-range shift occurs when an individual vacates a home-range and establishes a new home-range (Crook, 2004a). In the current study, five individuals undertook home-range shifts and one of these, Individual W6, undertook multiple shifts (Figure 3.6). Home-range shifts have been recorded in a number of radio-tracking studies of freshwater fishes (Armstrong et al., 1998; Armstrong et al., 1999) including studies of members of the Percichthyidae (Simpson & Mapleston, 2002; Crook, 2004a, b; Ebner et al., 2006). There have also been reports of percichthyids undertaking multiple home-range shifts (Simpson & Mapleston, 2002; Crook, 2004b; Ebner et al., 2006).

(c) Return home-range shifts A return home-range shift occurs when an individual returns to occupy a home-range that it has occupied previously though not in its most recent home-range occupation phase. The behaviour occurs in a subset of home-range shift cases. In the current study, only one individual exhibited a return home-range shift (Figure 3.7; also see W6 in Figure 3.6). Return home-range shifts have been recorded in studies of other percichthyids including M. peelii peelii (Koehn, 1997), M. p. mariensis (Simpson & Mapleston, 2002) and M. ambigua (Crook, 2004b).

(d) Return explorations A return exploration occurs when an individual leaves its home-range and returns to that home-range upon re-entering the home-range occupation phase (Crook, 2004b). Four wild individuals and a single hatchery-reared individual were recorded undertaking return explorations downstream to Yanco Weir and back to a home- range, in the current study (Figure 3.7). Return explorations occurring on smaller spatial scales may have occurred but remained undetected, since remote telemetry loggers were only positioned at the upstream and downstream extremities of the study reach and monthly manual radio-tracking was used. For instance radio-tracking Cyprinus carpio (L.) every 5 days, Crook (2004a) detected return explorations on scales of hundreds of metres to a few kilometres.

An ecological approach to re-establishing Australian freshwater cod populations 46 FRDC Report 2003/034

Release of fish

Initial dispersal

29 Home -range

18 No Leave home ?

Yes 11

Establish new 5 home-range? Yes

No

Return to a 1 previous Yes home-range?

No

6 Return to same Yes home-range?

Figure 3.7. A conceptual model categorising the movement of 29 M. macquariensis over the 13-month study. The number of individuals exhibiting specific emigration and home-range behaviour is shown. The occurrence of each behaviour is scored only once per individual.

Home-range establishment model Much of the movement recorded in this study is comparable with that of Crook (2004a, b) and is largely in agreement with the revised home-range shift model proposed in the second of these publications. Individuals occupy small home-ranges

47 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 most of the time with a minority of individuals undertaking explorations away from a home-range resulting in home-range shifts or an immediate return to the original home-range. The current study and Crook (2004b) also recorded an individual(s) undertaking return home-range shifts. Chapter 5 incorporates this behaviour into a conceptual model and discusses its significance. Suffice to say, this behaviour is indicative of fish using a mental map that encompasses multiple potential home- ranges.

The cause of the return explorations on a scale of tens of kilometres in the current study is unknown. Crook (2004b) considered explorations to be searches for more profitable habitat. Large-scale adult movement has often been ascribed to reproductive activity in percichthyids (Koehn, 1997; O’Connor, et al., 2005). There is also an indication that high flow may be an important factor or a contributing stimulus for percichthyid movement (Koehn, 1997; O’Connor et al., 2005). The return explorations observed in the current study were asynchronous among individuals and occurred outside the known spawning period of the species. In this regard our findings are similar to that for M. p. mariensis (Simpson & Mapleston, 2002).

Only large individuals performed return explorations in this study. Focussing on a range of size classes and in particular large individuals (e.g. >2 kg) may prove useful in future investigations of the movement of this species. Clearly, there is a need for better understanding of the ecological mechanisms underpinning these return explorations (Chapter 4). Presumably this will be achieved by complementing the radio-tracking approach with other techniques including confirming reproductive status, manipulating the dominance hierarchy or investigating resource availability.

The dispersal model In revisiting the model of dispersal as an explanation for the apparent disappearance of M. macquariensis from fingerling-stocking sites, it should be noted that low levels of emigration occurred in the Murrumbidgee River at Narrandera in the current study. This at least superficially indicates that the dispersal model is not the likely explanation. However, four important points temper this: (a) the situation in the Murrumbidgee River between Yanco and Berembed weirs is likely to be non- representative, since it is the premier M. macquariensis re-introduction site in NSW (Gilligan, 2005b); (b) observation of five individuals encountering and not passing Yanco Weir in 13 months is not insignificant especially considering the possibility that this behaviour may be more common in the population when scaled to the life- span of M. macquariensis; (c) the wild group may also have comprised individuals that are unlikely to disperse over large distances, since they were collected near their original stocking site (cf. the Restricted Movement Paradigm, Gowan et al., 1994); and (d) dispersal ability of juvenile M. macquariensis is unknown and needs to be addressed. Therefore considerable research is required before the merit of the dispersal model can be fully assessed.

CONCLUSION

Hatchery-reared, sub-adult M. macquariensis stocked at this developmental stage, exhibited low survivorship and different dispersal patterns, to those originally stocked as fingerlings. Consequently, hatchery fish should not be used as a surrogate for examining the dispersal of wild fish. Wild sub-adults and adults can home following

An ecological approach to re-establishing Australian freshwater cod populations 48 FRDC Report 2003/034 displacement or exploration, occupy small home-ranges for much or all of the year and are capable of emigrating to home-ranges used in their past.

49 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

CHAPTER 4. METHODS FOR RECORDING ACTIVITY OF AN ENDANGERED PERCICHTHYID: EFFECTS OF TEMPORAL RESOLUTION

Jason Thiem,1, 2, Brendan Ebner,1, 2 and Ben Broadhurst1

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

Thiem, J., Ebner, B. & Broadhurst, B. (2006). Methods for recording activity of an endangered percichthyid: effects of temporal resolution. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.50–59. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

INTRODUCTION

Radio-tracking is an effective tool for quantifying the movement and habitat-use of freshwater fishes (Priede & Swift, 1992; Winter, 1996; Lucas & Baras, 2000). The technique has been useful in the study of home-range and diel activity, since individuals can be located frequently without the need for recapture (Winter, 1996). Frequent manual radio-tracking increases the likelihood of disturbing the study animal and generating auto-correlated data (White & Garrott, 1990; Winter, 1996). Conversely, there can be significant information loss associated with less frequent tracking (Baras, 1998; Ovidio et al., 2000; LØkkeborg et al., 2002). Thus the temporal resolution of radio-tracking is important in the study of home-range (Baras, 1998; Ovidio et al., 2000).

Recently in Australia, radio-tracking has been used to study the localised movement of large freshwater fishes, primarily percichthyids (Butler, 2001; Crook et al., 2001; Simpson & Mapleston, 2002; Crook, 2004a, b; Ebner et al., 2005; Ebner et al., 2006). Large percichthyids are site-attached and have a small home-range as adults over most or all of the year (Koehn, 1997; Simpson & Mapleston, 2002; Crook, 2004a, b). Diel activity patterns include nocturnal activity (Butler, 2001; Ebner et al., 2005; Ebner et al., 2006) and matinine activity (Simpson & Mapleston, 2002). However, the extent to which the temporal resolution of radio-tracking affects these estimates is unknown and limits the management of particular species including the nationally endangered trout cod Maccullochella macquariensis (Cuvier) (Percichthyidae).

Maccullochella macquariensis is endemic to rivers in the southeast of the Murray– Darling Basin (Ingram & Douglas, 1995) and has undergone large-scale reductions in distribution from much of this former range (Cadwallader & Gooley, 1984; Douglas et al., 1994). As a result, conservation stockings of hatchery-produced fingerlings have occurred at selected rivers within the Murray–Darling Basin since the late 1980s in an effort to re-establish the species, with mixed success (Douglas et al., 1994; Gilligan, 2005b). In the current study, movement of M. macquariensis was investigated, at a conservation-stocking site where fingerlings reach adulthood. The primary aim of this study was to identify an appropriate temporal resolution for radio-

An ecological approach to re-establishing Australian freshwater cod populations 50 FRDC Report 2003/034 tracking M. macquariensis within diel periods. Additionally, the potential for observer effects resulting from manual radio-tracking was investigated with the aid of remote telemetry.

METHODS

Study site The study area was located in a lowland reach of the Murrumbidgee River, 5 km upstream of the township of Narrandera (173 m ASL) in southern New South Wales (NSW), Australia (Figure 4.1). Narrandera is one of twelve M. macquariensis stocking sites in the Murrumbidgee catchment, with 85 000 fingerlings stocked at this location between 1996 and 2000 (Gilligan, 2005b). Subsequent surveys have identified successful survival and growth of individuals (Growns et al., 2004; Gilligan, 2005b). The river channel has widths of 60–70 m and maximum water depths of 3–5 m occur on outside bends of the river. River red-gum Eucalyptus camaldulensis is common along both banks and fallen trees or branches of this species comprise the dominant in-stream structure and habitat for M. macquariensis (Growns et al., 2004).

Figure 4.1. Location of the study reach in southern New South Wales, Australia.

Fish collection and surgery Thirty–one fingerling stocked (termed ‘wild’) M. macquariensis were collected from the study area over 2 weeks in September 2003 by boat electro-fishing. Individuals greater than 550 g were retained for surgery. An additional 27 on-grown, 2 year-old

51 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

(termed ‘hatchery’) M. macquariensis were sourced from a state government hatchery. Total length (TL) was measured to the nearest 1 mm and weight was recorded to the nearest 1 g. Radio-tags were surgically implanted into the peritoneal cavity under anaesthesia (internal body implants with a 30 cm trailing whip antenna, models F1830, 35, 40 and 50, Advanced Telemetry Systems (ATS), Isanti, USA). Weight of radio-tags in air was between 11 and 25 g to accommodate a range of fish sizes and was kept to ≤ 2% of fish mass. Pulse coded, two–stage radio-transmitters were used on a frequency of 150–152 MHz and programmed on a pulse rate of 5 s on and 7 s off to increase battery life (warranted for between 230 and 504 d). For external identification individuals were also tagged with a dart tag between the second and third dorsal spines. Further details of tagging procedures are outlined in (Chapter 3). Individuals were held for 2–15 d under hatchery conditions following surgery and all individuals were released at the same location in late September 2003. Three radio- telemetry exercises occurred from 14–19 November 2004.

Radio-telemetry Continuous remote telemetry A modified technique of David & Closs (2001) was used to remotely monitor the activity of a single M. macquariensis (Individual 5, Table 4.1) continuously from 2000, 14 November until 0900, 19 November 2004 (Australian Eastern Standard Time). The individual was selected since an overhanging branch above its home site (as determined by 13 months of radio-tracking, Chapter 3) was ideal for mounting an antenna above the water surface. Two remote data loggers (ATS, DCII Model D5041) each connected to a receiver (ATS, R4100) were set-up on shore. One receiver was connected by a 20 m length of co-axial cable to a modified gap-loop antenna (Titley Electronics, Ballina, Australia) approximately 0.5 m above the water surface, and was termed ‘Logger A’. A 20 m length of co-axial cable was connected from the second receiver to a horizontally placed three-element Yagi antenna (Titley Electronics) and then fixed to a tree on shore in the middle of the home-range based on estimates from previous radio-tracking of this individual. This arrangement was termed ‘Logger B’. Signal-strength received from a radio-transmitter was recorded at 5 s intervals. The standard deviation of signal-strength was plotted in 10 min grouped intervals to examine signal variability as a measure of activity (David & Closs, 2001).

The detection distance of Logger A and B was tested at fixed distances prior to logging using a radio-transmitter in a weighted plastic bag positioned immediately below the surface of the water and on the river bottom. The detection distance of Logger A was approximately 25 m and Logger B was approximately 80 m (incorporating any potential movements to the opposite bank). Boat movements into and out of the study reach were noted, including activities by researchers and local angler traffic.

Scan sampling (following Martin & Bateson, 1993) Following a 13 month study, ten radio-tagged individuals (nine wild individuals including Individual 5; one hatchery individual, Individual 9, Table 4.1) that had functioning radio-transmitters and were in close proximity to one another, were used for intensive radio-tracking within diel periods. A two–person crew manually radio- tracked from a 3 m aluminium punt, powered by a Yamaha 8 hp petrol motor. Each individual was located at four hourly intervals from 1400, 15 November until 1400, 17 November 2004 using a three-element Yagi antenna connected to a receiver unit

An ecological approach to re-establishing Australian freshwater cod populations 52 FRDC Report 2003/034

(Australis 26k, Titley Electronics). Each location was recorded using a handheld Global Positioning System unit (GPS) (Garmin® Pro GPSII Plus or GPS76 Marine Navigator) by taking three waypoints (Figure of merit (FOM) ≤ 5.0 at about 95% of locations, FOM ≤ 7.0 at 100% of locations).

Focal-animal sampling (following Martin & Bateson, 1993) A two–person crew, aboard a 3 m aluminium punt powered by an electric motor (Shakespeare®), located Individual 5 (Table 4.1) at one hourly intervals from 0600, 18 November until 0700, 19 November 2004. The location of the individual was determined using a 3–element Yagi antenna connected to a receiver unit (Australis 26k, Titley Electronics). Each location was recorded using a handheld GPS as per scan sampling.

Data analysis GPS records were averaged to provide a single spatial location for each individual per radio-tracking fix. Spatial data were plotted in GDA94 format in ArcView 3.2™ (ESRI, USA) over a base-map, generated by walking both banks of the river with handheld GPS units. A polyline was generated based on the sequential locations of each individual using the Animal Movement Extension in ArcView (Hooge & Eichenlaub, 1997) and used to construct a time series of the distance moved between consecutive radio-tracking intervals (activity). Linear range (the distance between the most upstream and downstream points) and area used (Minimum Convex Polygons (MCP)) were also calculated with this software. For comparison with MCP an available area was calculated by constructing a line perpendicular to the river channel at the most upstream and downstream fix of each individual and estimating the wetted area within this. Linear range and area estimates were sub-sampled to simulate decreasing the temporal resolution of radio-tracking. Spatial radio-tracking data collected at four hourly intervals was sub-sampled at eight and 12 hourly intervals by using every second and third location respectively. The same approach was used for spatial radio-tracking data collected at hourly intervals, whereby sub-sampling at two hourly intervals involved using every second location, up until 12 hourly sub- sampling. Statistical analysis was conducted using Statistix For Windows (version 2.0) with data transformed, when necessary, to achieve the assumption of normality (Tabachnick & Fidell, 1989).

RESULTS

The total length of river (linear range) used by individuals over two consecutive diel periods ranged from 6–272 m, with a mean (± S.E.) of 83.1 ± 30.0 m (Table 4.1). There was a significant positive correlation between fish length (mm) and range (m) over the two diel periods (Pearson Correlation, P<0.05). The area of river used (MCP) over the two consecutive diel periods ranged between 18.5–4603.5 m2 and averaged 1284.1 ± 621.1 m2 (Table 4.1). There was also a significant positive correlation between fish length (mm) and area (m2) over the two diel periods (Pearson Correlation, P<0.05).

Simulations of sub-sampled range of individuals (from data collected four hourly) resulted in a mean percentage reduction (± S.E.) of 23.8 ± 10.2 and 23.9 ± 8.1 for locating individuals at eight and 12 hourly intervals, respectively (Table 4.1). A more pronounced effect occurred with sub-sampling of area based estimates, with a mean

53 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

percentage reduction of 36.7 ± 9.6 and 48.4 ± 10.7 when locating individuals at eight and 12 hourly intervals, respectively (Table 4.1). Additionally, there was a small increase in the diel range estimate (75 to 84 m (12%)) corresponding to four as opposed to one hourly radio-tracking (Figure 4.2). In sharp contrast, increasing the temporal resolution of sampling from four to one hourly increased the diel area estimate from 307 to 1010 m2 (329%), respectively (Figure 4.2).

Table 4.1. The simulated reduction in accuracy associated with decreasing the temporal resolution of sampling on the range and area of river used (combined over two consecutive diel periods) by 10 individuals (based on four hourly radio-tracking). Range reduction (%) Area reduction (%) if locating every n hours if locating every n hours

Fish ID TL Weight Sex Range Eight 12 hourly Area Eight 12 hourly (mm) (g) (m) hourly (%) (m2) hourly (%) (%) (%) 1 371 672 ? 25 60 32 292 61 47 2 370 599 F 26 73 0 280 77 57 3 412 834 F 15 53 0 191 71 42 4 405 827 ? 6 0 0 19 0 0 5 430 1066 ? 90 0 49 658 42 84 6* 466 1249 ? N/A N/A N/A N/A N/A N/A 7 461 1292 M 109 28 28 925 36 51 8 481 1384 ? 272 0 59 4604 32 94 9† 407 1247 ? 31 0 0 128 0 0 10 575 2587 M 174 1 48 4463 12 59 Mean 83.1 23.8 23.9 1284.1 36.7 48.4 (S.E.) (30.0) (10.2) (8.1) (621.1) (9.6) (10.7)

*Individual 6 had a weak radio signal and as such was excluded from range and area calculations due to difficulties in locating the individual whilst active. †Denotes on-grown hatchery individual.

An ecological approach to re-establishing Australian freshwater cod populations 54 FRDC Report 2003/034

1200 100

1000 80 )

2 800 60 600 40 400 Diel area (m Diel range (m) 20 200

0 0 02468101214 Sampling resolution (hours)

Figure 4.2. The effect of decreasing the temporal resolution of sampling on estimates of diel range (▲) and diel area (■), based on radio-tracking of Individual 5.

The mean (± S.E.) area of river used by nine individuals over two consecutive diel periods, 1284.1 ± 621.1 m2, represented only a small proportion (14.7 ± 3.2%) of the river that was available within the upstream and downstream limits of total range (Table 4.2), indicating that individuals selected for specific habitats. All ten individuals (including Individual 6 that had a weak radio-signal enabling occasional detection) were always located in the thalweg during the day. For nine of ten individuals this corresponded to a location in close proximity (<20 m) to the outer river bend. An exception was the position of one individual on the inside of a bend at the upstream end of a braided channel section. Nine of the ten individuals did not cross from one side of the river to the other during the study. One individual (Individual 10, Table 4.1) was recorded crossing the channel. This occurred in a straight section of river that was without shallow water associated with either bank. Additionally, four individuals had overlapping ranges during this study. Two of these individuals were found to co-inhabit the same hollow log during the day for the duration of the study (and indeed throughout much of the 13–month period of a larger-scale experiment, Chapter 3).

The movements of M. macquariensis were generally greater at night, dusk and dawn relative to daytime (Figure 4.3). The period of greatest movement (mean ± S.E.) was between 1800 and 2200, 37.27 ± 12.6 m, with the period of least movement between 1000 and 1400, 8.98 ± 4.34 m (Figure 4.3). Differences in distance moved between time periods were non-significant (One–way ANOVA, d.f.=107, P>0.05).

55 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Table 4.2. The estimated proportion of available river used by 10 individuals over two consecutive diel periods.

Fish ID MCP estimate Available river Proportion of available (m2) (m2) river used (%)

1 292 2293 13 2 280 1926 15 3 191 1128 17 4 19 585 3 5 658 6160 11 6 N/A N/A N/A 7 925 8374 11 8 4604 20516 22 9 128 2332 5 10 4463 12672 35 Mean 1284.1 6220.8 14.7 (S.E.) (621.1) (2227.7) (3.2)

60

50

40

30

20 Distance moved (m) 10

0 14:00- 18:00- 22:00- 02:00- 06:00- 10:00- 18:00 22:00 02:00 06:00 10:00 14:00

Time interval Figure 4.3. The diel activity of nine individuals in the Murrumbidgee River based on the minimum distance moved between four hourly radio-tracking locations (mean ± S.E.). White and grey sections denote day and night, respectively. Data is grouped over two consecutive diel periods.

An ecological approach to re-establishing Australian freshwater cod populations 56 FRDC Report 2003/034

Data from Logger A was excluded from analysis, due to a combination of a slight shift in the resting-site of Individual 5 and the small detection area of this type of antenna resulting in very few records on the logger. However, data from Logger B recorded sporadic variations in signal-strength for the first three daylight periods (Figure 4.4). In comparison, variations in signal-strength were consistently higher and sustained for longer periods of time during the night or during dusk or dawn periods, indicating periods of heightened activity. Additionally, radio-transmitter signal strength varied repeatedly throughout the fourth daylight period of remote telemetry logging (Figure 4.4). This period coincided with a decrease in water discharge (4376– 3336 ML/day) and river height (0.21 m decrease) (NSW DNR, 2004). The fourth daylight period also coincided with more intensive radio-tracking of Individual 5, changing from four to one hourly (Figure 4.4).

Logger Logger 4-hourly Focal Logger test test scan sampling sampling test

120

100

80

60

40

Signal variabilitySignal (SD) 20

0 24 48 72 96 120

Time (hours)

Figure 4.4. Activity of Individual 5, based on continuous logging of signal-strength from its radio-transmitter, 14–19 November 2004. Data is grouped at 10 min intervals and plotted as the variation in radio-transmitter signal-strength. Open circles at the top of the graph denote boat activity (predominantly investigator related) within the reach. Day and night periods are represented in white and grey respectively.

DISCUSSION

Maccullochella macquariensis occupied small (<300 m) lengths of river over consecutive diel periods in this study and the size of home-range was positively correlated with fish length. These individuals had previously demonstrated fidelity to the same diurnal locations throughout much of the previous year (Chapter 3). Similarly this species has been shown to be relatively sedentary in the Murray River (Koehn, 1997; Koehn & Nicol, 1998). The small home-range of M. macquariensis is consistent with that of other Australian percichthyids (Butler, 2001; Simpson & Mapleston, 2002; Crook, 2004a, b).

57 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Individual M. macquariensis demonstrated a preference for movements along outer river bends within diel periods in this study. These outer bends are associated with deeper sections of river channel and contain higher quantities of structural woody habitat in lowland rivers (Hughes & Thoms, 2002; Koehn et al., 2004). Efforts to recover M. macquariensis and other threatened percichthyids should be aided by an improved understanding of their lateral movements in large lowland rivers. This information provides the basis for strategic placement of structural woody habitat in stream restoration programs (e.g. Nicol et al., 2002).

Differences between estimates of available river area and the area used by each individual reflected the fidelity of M. macquariensis to outer bends. Where linear movements of a species predominate, standard methodologies to calculate home- range often produce considerable over-estimates (Blundell et al., 2001). Therefore equating home-range size to the area of river within the upstream and downstream limit of a radio-tracked individual (e.g. Gust & Handasyde, 1995) is an overestimate when applied to M. macquariensis in a large lowland river. The lateral distribution of M. macquariensis within a large lowland river is likely to be a direct response to in- stream habitat differentiation (Hughes & Thoms, 2002; Koehn et al., 2004).

The logger method detected distinct nocturnal activity of an individual whereas coarse-scale manual radio-tracking did not. There was an indication of greater movement during crepuscular and nocturnal periods, based on manual radio-tracking of ten individuals. Simpson & Mapleston (2002) found that the activity of Mary River cod Maccullochella peelii mariensis (Rowland) was matinine. However that study made use of real-time radio-tracking within one–hour periods. Our findings indicate that application of the continuous remote telemetry method of David & Closs (2001) based on increased sample sizes (e.g. David & Closs, 2003) is an effective means of elucidating the diel activity patterns of M. macquariensis.

The cause of the shift from nocturnal to both nocturnal and diurnal activity of an individual in this study, is unknown. The shift corresponded to both the use of one hourly boat-based manual radio-tracking and a change in discharge and river height. This demonstrates the capacity to use variation in signal-strength from remote loggers to monitor disturbance (e.g. by the researcher, releases from dams) in experiments. To date the application of variation in signal-strength has only been used to record diel activity (see Baras et al., 1998; David & Closs, 2001; Hiscock et al., 2002; David & Closs, 2003).

Reducing the temporal resolution of radio-tracking, within diel periods, decreased estimates of home-range in this study. Comparable findings have been reported from the use of radio or acoustic telemetry at coarse (Baras, 1998; Ovidio et al., 2000) or fine (LØkkeborg et al., 2002) temporal resolution relative to the current study. To produce reasonable estimates of the extent of diel-range, four hourly radio-tracking (of about five to ten individuals) appears to represent the most pragmatic solution when a maximum of two researchers are available. At a minimum, observations should be replicated among seasons to produce meaningful estimates of home-range size and shape. At this stage preliminary findings indicate that M. macquariensis inhabit small reaches of river on a scale of tens to hundreds of metres, within the deeper outer bank of the Murrumbidgee River. Consequently, the recovery of this

An ecological approach to re-establishing Australian freshwater cod populations 58 FRDC Report 2003/034 species can probably be conducted within small reaches of river and specific instream habitats prioritised for rehabilitation.

59 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

CHAPTER 5. EXPERIENCED INDIVIDUALS: DESCRIBING FISH MOVEMENT AND IDENTIFYING UNDERLYING MECHANISMS AT THE AMONG HOME-RANGE SCALE

Brendan Ebner1, 2 and Jason Thiem1, 2

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

Ebner, B. & Thiem, J. (2006). Experienced individuals: describing fish movement and identifying underlying mechanisms at the among home-range scale. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.60–78. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

INTRODUCTION

The home-range concept was first applied to fish by Gerking (1950, 1953) in studying brook char (Salvelinus fontinalis) in streams. Subsequently, it has been found that many fishes occupy a restricted area, termed a home-range, for a substantial part of their existence in the post-larvae phase (Rodríguez, 2002). However whilst movement beyond the home-range often occupies a short period of time in an individual’s life, it can have important consequences at the individual and population level (Brooker & Brooker, 2002). Movement beyond the home-range can take many different forms including migration, emigration, home-range shift or exploration (Kynard, 1997; Weisser, 2001; Crook, 2004a, b). These types of movement underlay important population processes including dispersal, recolonisation and spawning migration. Increasingly, conservation efforts rely on knowledge of these processes as humans continue to fragment the landscape (Clobert et al., 2001; Brooker & Brooker, 2002; Gruber & Henle, 2004).

This review will focus on fish movement at the among home-range scale (MAHRS). It is opportune to conduct this review for three reasons. First, the proliferation of automated technologies for collecting known-status-data is rapidly enhancing capabilities for observing and describing individual fish movements (Lucas & Baras, 2000; Cooke, Hinch et al., 2004; Heupel et al., 2006). Known-status-data arises when individuals can be located at will without capture, which enables the calculation of probabilities associated with processes including dispersal and habitat use (Bennetts et al., 2001). Devices facilitating this type of data collection include: radio and acoustic telemetry, pit tagging, underwater video, acoustic surveillance (Lucas & Baras, 2000), and video-acoustics (Mueller et al., 2006). Second, conceptual models of individual fish movement have recently been developed (Crook, 2004a, b) and remain to be refined through clarification of their logic structure and by accumulating further empirical data. Third, there has been significant progress in understanding the mechanisms that underpin fish movement (see Braithwaite & Burt de Perera, 2006). Much of this research demonstrates mechanisms by which fish orientate themselves in captive environments and its relevance to MAHRS remains to be ascertained.

An ecological approach to re-establishing Australian freshwater cod populations 60 FRDC Report 2003/034

Furthermore, descriptive and mechanistic aspects of MAHRS have predominantly been investigated in isolation, and an integrated discussion of progress in both fields is warranted.

Here, current understanding of how and why fish move at the among home-range scale, is reviewed based on four objectives. First, to identify examples of improvements that can and have been made to the descriptive elements of MAHRS, including the incorporation of the home-range shift model(s) (Crook, 2004a, b). In particular, this is achieved by including return home-range shifts and migrations into that conceptual model(s), and by simplifying one of its existing constructs. Second, to provide a more comprehensive mechanistic basis to the resulting conceptual model. This is achieved by reviewing studies that specifically investigate mechanisms for movement in fishes and by invoking population and ecological explanations for such movement from the wider fish population biology and ecology literature. Third, to examine the role that an individual’s experience has in determining its movement. This largely arises in the context that technologies enabling collection of known- status-data are making collection of comprehensive information on the history of individual movement feasible. Therefore guidance on the types of questions that can be answered with these data is of current interest. Fourth, to explore applications of the new conceptual model through discussion of current and potential conservation based research of Australian members of the freshwater family Percichthyidae. Examples from the study of percichthyids are used where possible throughout the document in preparation for achieving this fourth objective in the final section prior to summarising the review.

The Australian Percichthyidae as a case study Restoring or maintaining large-scale passage (tens to hundreds of kilometres) for fish has become a management priority in rivers of eastern and especially south-eastern Australia for about two decades (Beumer, 1980; Beumer & Harrington, 1982; Eicher, 1982; Harris, 1984a, b; Mallen-Cooper & Harris, 1990; Mallen-Cooper, 1992; Harris & Mallen-Cooper, 1994; Swales, 1994a; Mallen-Cooper et al., 1995; Lintermans & Phillips, 2004; MDBC, 2004B). Movement of large percichthyids, particularly those in the Murray–Darling Basin, has been central to these developments. This includes the seminal mark-recapture study by Reynolds (1983) that recorded large movements of M. ambigua (among other species). Later, Mallen-Cooper and colleagues (Mallen- Cooper et al., 1995; Mallen-Cooper, 1996) investigated performance of the Torrumbarry fishway on the Murray River and Koehn (1997) radio-tracked M. peelii peelii undertaking moderate to large-scale spawning migrations. Subsequently, there has been increased research attention given to the issue of large-scale movement and migration, and widespread installation of fishways suitable for passage of native species in the Murray–Darling Basin and along other Australian rivers (Harris & Mallen-Cooper, 1994; Stuart & Mallen-Cooper, 1999; Stuart & Berghuis, 2001; Lintermans & Phillips, 2004).

The Percichthyidae is a southern hemisphere family of freshwater fishes comprising extant species on the South American and Australian continents (Nelson, 1994). In Australia this group is under threat from human impact particularly in the south-east and central east of the continent where the majority of percichthyid species are distributed and human development is greatest (Harris & Rowland, 1996; Unmack, 2001; Unmack et al., submitted). This impact is reflected in current

61 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 conservation listings of several percichthyids (Table 5.1). All members of the freshwater cod genus Maccullochella, are listed nationally or internationally with M. ikei, M. macquariensis and M. peelii mariensis listed as endangered and M. peelii peelii as vulnerable (Table 5.1). The perches include the endangered Macquaria australasica (Table 5.1), and the closely related Guyu wujalwujalensis that has a highly restricted distribution and is subsequently likely to become a listed species (Pusey & Kennard, 2001; Mark Kennard, Griffith University, personal communication). Three pygmy perch are listed, N. oxleyana as endangered and N. obscura and N. variegata as vulnerable (Table 5.1).

In particular the review will draw from percichthyid examples where possible for three reasons. First, due to the urgent conservation needs of these fishes particularly regarding human impacts on MAHRS and the persistence of some remnant percichthyid populations or species (e.g. Ebner et al., 2006). Second, the group includes a diversity of species differing in their MAHRS (Koehn & Nicol, 1998; Crook, 2001, 2004a, b; O’Connor et al., 2005; Jamie Knight, Southern Cross University, unpublished data). Third, the group has been more comprehensively studied in terms of movement than any other Australian freshwater fish group.

Table 5.1. Australian members of the Percichthyidae, maximum age and longevity (from McDowall, 1996; Allen et al., 2002) and conservation status nationally.

Genus Species Sub Maximum Longevity Conservation species body size status Maccullochella ikei 66cm;41kg 1,2endangered macquariensis 85cm;16kg 1,2endangered peelii peelii 180cm;113.5kg 60yrs 1vulnerable peelii mariensis 23.5kg 1endangered 2critically endangered

Macquaria ambigua 76cm;23kg 26yrs australasica 46cm;3.5kg 1endangered 2data deficient colonorum 75cm;10kg novemaculeata 3.8kg;60cm

Nannoperca australis 8.5cm 5yrs obscura 7.5cm 1,2vulnerable oxleyana 7.5cm 1,2endangered variegata 6.5cm 1,2vulnerable

Edelia vittate 6.8cm 5yrs

Nannatherina balstoni 9cm

Bostockia porosa 15cm 6yrs

Guyu wujalwujalensis 10cm 1EPBC 1999. Environment Protection and Biodiversity Conservation Act 1999. (Cwlth) . Accessed 8 August 2006. 2IUCN 2006. 2006 IUCN Red List of Threatened Species. . Accessed 8 August 2006.

An ecological approach to re-establishing Australian freshwater cod populations 62 FRDC Report 2003/034

DESCRIPTION OF MAHRS

Empirical studies There are serious inconsistencies in the terminology used to distinguish among different types of MAHRS in the ecological literature (Swingland & Greenwood, 1983). This has resulted in different descriptions of the same type of movement or alternatively, different definitions of the same term (e.g. migration: see Baker, 1978; Swingland & Greenwood, 1983; Clobert et al., 2001). With advances in technology leading to rapid increase in the amount of descriptive information of animal movement available (cf. Lucas & Baras, 2000), more than ever there is a need for a common set of terms, a language regarding functional types of MAHRS, used by ecologists.

For the purposes of this review, three specific types of MAHRS are identified and described. These are a) migration; b) home-range shift; and c) exploration. In defining these terms we do not seek to differentiate by spatial and temporal scale of movement. Although, we acknowledge that the scale of movement varies between and within species and habitats. Rather these terms are used to distinguish behaviours in an attempt to contextualise the purpose of those movements. a) Migration Migration has been defined differently depending on the purpose of a study or the taxa being investigated (Baker, 1978; Swingland & Greenwood, 1983; Clobert et al., 2001). In this review, migration is confined to mean the regular seasonal movement away from a home-range to settle periodically elsewhere in response to food availability or to spawn (Sinclair, 1983; Kynard, 1997). This excludes short-term movements, including daily feeding movements, that are commonly reported as migrations (e.g. Comeau & Boisclair, 1998). Prominent examples are provided by the large-scale spawning migrations of members of the Salmonidae. In this group spawning migrations of thousands of kilometres are not uncommon and in several cases involve individuals exhibiting a high degree of fidelity to return and spawn in natal streams (e.g. Dittman & Quinn, 1996).

Known-status-data has been useful in revealing specific types of pre-spawning migration movements. For example, Kynard (1997) broke the pre-spawning movements of shortnose sturgeon (Acipenser brevirostrum, Acipenseridae) into three types: 1) a short one-step migration; 2) a long one-step migration; and 3) a short two- step migration. To date, confirmation that percichthyids migrate, is confined to examples relating to spawning. Examples include radio-tracking of large scale movement of Maccullochella peelii peelii during the known breeding season (Koehn, 1997) and synchronised large-scale movement of subsets of individuals of M. ambigua (O’Connor et al., 2005), and observation of aggregating adult Macquaria australasica during the known breeding season (e.g. Cadwallader, 1979). In this regard spawning is not always definitively shown to have occurred and large-scale movement is deduced to be a spawning migration. Examples include studies of percichthyids (e.g. Koehn, 1997; O’Connor et al., 2005) and other fishes particularly those likely to be annual group spawners (Kieffer & Kynard, 1993; van Ginneken & Maes, 2005). In certain cases direct observations of pre-spawning or spawning behaviour, changes in gonad structure or the detection of eggs and larvae validate that the movement had a reproductive function (Acipenseridae: Kieffer &

63 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Kynard, 1993, 1996; Percichthyidae: Mackay, 1973; Cadwallader, 1979; Battaglene, 1991; Salmonidae: Aarestrup & Jepsen, 1998; Johnsen & Hvidsten, 2002). Fishes have also been reported to undertake feeding migrations with some of the best examples coming from apex predators migrating in relation to prey resources (e.g. Polovina et al., 2001; Le Boeuf, 2004). These feeding migrations can be supported by direct observation of feeding (Martin & Martin, 2006), established from gut content analysis or stable isotope analysis (Hobson, 1999) or based on correlation between predator movement and prey movement or availability (Leggett, 1977; Madsen & Shine, 1996). b) Home-range shift Migration can involve the establishment of a temporary home-range and therefore involve a form of home-range shift, however, for the purpose of this review a home- range shift and a migration are treated as mutually exclusive. This simplification enables for the disentanglement of MAHRS relating to predictable autecological phenomena, including spawning migrations of annual spawners and tracking of regular pulses of prey (see examples in Migration, above), from the seemingly less predictable shifts in home-range that have been observed within populations of riverine fishes (Simpson & Mapleston, 2002; Crook, 2004a, b). Furthermore the latter may occur at a much more localised temporal and spatial scale than migrations. Therefore whilst a home-range shift occurs when an individual moves from occupying one home-range to establishing another home-range, this excludes migrations (cf. Crook, 2004a).

The function of home-range shifts is often unknown, however, in a subset of cases has been shown or considered to represent a response to changes in resource availability, directly in terms of habitat and food or as a function of density (Sinclair, 1983; Sullivan et al., 2003; Crook, 2004b) and can be mediated by competition (e.g. Nakano, 1995a, b). Gowan and Fausch (2002) induced home-range shifts in brook char (Salvelinus fontinalis) by excluding the dominant fish in a pool from the optimal foraging location, whereby this fish left the pool or occupied the next best foraging location, displacing the next largest fish (which left the pool). They proposed two explanations for fish leaving the pool 1) that fish had pre-existing knowledge of habitat conditions beyond the pool, or 2) fish were searching in order to gain such knowledge.

Most species of Percichthyidae for which known-status-data is available have been reported to exhibit home-range shift behaviour (Koehn, 1997; Simpson & Mapleston, 2002; Crook, 2004a, b; Chapter 3). The return home-range shift is a particular subcategory of home-range shift, recently recognised and subsequently incorporated into an updated home-range shift model describing the movement of the percichthyid M. macquariensis in a lowland river (Chapter 3). A return home-range shift occurs when an individual makes use of a home-range that it had occupied at some previous time, although, not as it’s most recent home-range (Chapter 3). c) Exploration An exploration involves movement into a foreign or infrequently visited area, resulting in establishment of a new home-range or return to a previously occupied home-range (Crook, 2004b). Alternatively, where home-range occupation is not the final outcome, nomadism may occur. It is foreseeable that the term ‘exploration’ as

An ecological approach to re-establishing Australian freshwater cod populations 64 FRDC Report 2003/034 used in this context, encompasses a range of types of MAHRS with different purposes. These include: a) assessing resource availability beyond the home-range, comprising the assessment of resource abundance (e.g. habitat, food or mate availability) and its availability or occupancy by others (Brooker & Brooker, 2002; Gowan & Fausch, 2002; Crook, 2004b); b) an individual can be expelled from an area by a competitor, predator or physical-chemical disturbance (Sinclair, 1983; Nakano, 1995b; Gowan & Fausch, 2002; Sullivan et al., 2003; Crook, 2004b); and c) an individual can presumably leave to perform a social behaviour other than spawning. In the case of the latter, exploration may perform a social function whereby individuals meet and interact as is the case for certain (Dawson, 1995; Viddi & Lescrauwaet, 2005). It has long been known that intraspecific and interspecific differences in the degree of mobility and nomadism exist in fish (Rodríguez, 2002). Notably, there is evidence that a number of fish populations (Northcote, 1992; Gowan & Fausch, 2002; Rodríguez, 2002) including that of the percichthyid M. macquariensis (Chapter 3; Chapter 4) have both a highly mobile or nomadic faction and a sedentary faction that engages purely in home-range occupation. Whether this has a genetic or ecological basis remains to be clarified (Rodríguez, 2002).

DESCRIPTIVE MODELS OF HOME-RANGE SHIFT

Existing conceptual models Crook’s original model of home-range shift (2004a) and an advanced version (2004b), were referred to earlier and are depicted in Figure 5.1 as flow diagrams. The original model simply described the process of occupying a home-range followed by establishing a new home-range (Figure 5.1a). It did not formally distinguish where and when fish made decisions (note the absence of a decision box/diamond in Figure 5.1), although, possible mechanisms underpinning movement were discussed briefly (Crook, 2004a). The subsequent version distinguished three movement phases: home- range occupation, mobile and home-range shift (Figure 5.1b). This provided an improved description of the fish movements observed in his and other studies (see Crook, 2004b). The revised model included recognition, that an individual can assess habitat whilst in the mobile phase and decide whether to return to its home-range or establish a new habitat (note the decision box in Figure 5.1b; Crook, 2004b). This was a first attempt to incorporate an explanation for shifts in home-range into the model and represents the forerunner, to a mechanistic conceptual model of home-range shift and movement.

In regard to the non-migratory component of adult percichthyid life, the revised home-range shift model (Crook, 2004b) adequately explained the findings from other radio-tracking studies of percichthyids (Simpson & Mapleston, 2002; Crook, 2004a; Chapter 3). This included individuals typically occupying small home-ranges, and a minority of individuals undertaking explorations away from the home-range resulting in home-range shifts or an immediate return to the original home-range. However, percichthyids have also been observed undertaking return home-range shifts in rivers (e.g. Crook, 2004b; Chapter 3). Chapter 3 suggested that this behaviour indicated knowledge of the river landscape beyond the bounds of a home-range. It was then suggested that the conceptual model of home-range shift be broadened to encompass return home-range shifts and migration (Chapter 3).

65 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

A new conceptual model An updated conceptual model of fish MAHRS is presented in Figure 5.2. Migration has been included in the model. The model retains the home-range occupation and mobile phases of its predecessors (Figure 5.1), though, not the home-range shift phase (Figure 5.2). The latter is superfluous in that a home-range shift occurs when an individual establishes a new home-range by exiting the mobile phase and entering the home-range occupation phase (Chapter 3). Return home-range shifts have also been included in the model.

a) Home – range occupation

Mobility Home -range shift

Establish new home -range

b) Home– range occupation phase

Return to original home -range Establish new Exploration home -range

Home- range Is a more shift phase Yes pro fitable No habitat

encountered? Mobile phase

Figure 5.1. Home-range shift models expressed as flow diagrams: a) the conceptual model of Crook 2004a; b) the revised model of Crook 2004b.

An ecological approach to re-establishing Australian freshwater cod populations 66 FRDC Report 2003/034

Yes Home - range occupation Migrate? No phase

Home –range occupation

Emigrate? No

Yes

Establish a new Yes home -range?

No Nomadism Return to a Yes previous home -range?

No

No Return to same Migration home -range?

Yes Mobile Exploration phase

Figure 5.2. A conceptual decision model of fish movement.

TEMPORAL RESOLUTION

When considering the temporal resolution at which movement data is collected, there are two topics of note. The first is obtaining an accurate record of individual movement. In this regard the primary issue has revolved around subsampling animal movement without excessive loss of information. The issue has received considerable attention and knowledge is rapidly developing, particularly in regard to understanding

67 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 the effects of how frequently an individual is located by manual or remote telemetry (Baras, 1998; Ovidio et al., 2000; Løkkeborg et al., 2002). The second relates to an individual’s knowledge of the spatial environment beyond the home-range (i.e. MAHRS). The history or experience of fish has been shown to be an important determinant of fish learning and navigation ability based on laboratory and field experiments (Brown & Laland, 2003; Braithwaite & Burt de Perera, 2006). This has included investigation of the mechanisms that underpin individual MAHRS in a few cases based on field experiments and/or observation (Nakano, 1995b; Gowan & Fausch, 2002). However, these studies are typically conducted over a fraction of the life span of an individual. The importance of the experience of an individual prior to being studied, for instance before being radio-tagged as an adult fish, is largely unknown and yet has the capacity to influence the decisions made by fish and hence MAHRS. These two topics are now discussed in more detail.

There is a constant trade-off between the accuracy of the information gathered (resolution) and the duration of a study (extent) when describing fish movement. Both can result in non-detection of events or misinterpretation of an event. Theoretical examples of this are presented in Figure 5.3. Where events occur over short time periods (Figure 5.3b), but where the study is lengthy in extent and of coarse-temporal scale (Figure 5.3c) (e.g. Koehn, 1997; Chapter 3, Chapter 6), an event may be missed (P, Q, R) or misidentified (N). Conversely studies of short extent and fine-temporal resolution (Figure 5.3d) (e.g. Chapter 4) will not identify events that occur over long time periods (e.g. Figure 5.3b and to a lesser extent c). Thus the selection of an appropriate temporal resolution at which to radio-track has particular relevance to studying percichthyids that undertake short-term explorations and home-range shifts (Simpson & Mapleston, 2002; Crook, 2004b; Ebner et al., 2005; Chapter 3) since these two behaviours are difficult to distinguish if a coarse temporal resolution is employed. Short-term explorations can be missed completely by weekly or monthly radio-tracking (see ‘P’ in Figure 5.3b and contrast with c). Alternatively, a monthly resolution can indicate a short-term exploration (see ‘N’ in Figure 5.3c) when in fact a short-term shift in home-range occurred (Figure 5.3b). The distinction among home- range shift, exploration and short-term migration, has been clouded by the coarse- resolution of radio-tracking in at least two studies of Maccullochella species (Simpson & Mapleston, 2002; Chapter 3). In certain cases problems with incomplete descriptions of fish movement have been overcome by remote telemetry applications and the provision of high-resolution temporal information over lengthy extents (e.g. Cooke, Bunt & Schreer, 2004; Heupel et al., 2006).

The temporal extent and resolution of movement studies also restrict researcher understanding of fish experience and knowledge of the river. In radio and acoustic tracking studies, researchers generally have no prior knowledge of the home- relocation behaviour of an individual prior to its capture and release, and should not assume that this equates to a lack of knowledge on behalf of the fish. For example where experimental displacement occurs, successful homing of an individual is classified by the experimenter as a return to the original site of capture (Baker, 1978). An individual may in fact be in the mobile phase or ready to depart the home-range occupation phase when captured, or alternatively, may locate suitable habitat in between release site and capture location (e.g. Crook, 2004a). The temporal extent of most studies is often especially limiting in this regard. For instance in the hypothetical studies shown in Figure 5.3b, c and d the researcher is unaware of the experience and

An ecological approach to re-establishing Australian freshwater cod populations 68 FRDC Report 2003/034 familiarity that an individual has of the river based on the first two years of its life (Figure 5.3a). This is an important distinction from the previous model (Crook, 2004a, b) which simply assumed that an individual blindly leaves a home-range on an exploration. Consequently, a decision box has been added into the home-range occupation phase in Figure 5.2. Presumably, experience with the river beyond a home-range can influence an individual’s decision to vacate a home-range in the first place (Gowan & Fausch, 2002). However, increasing the extent of a radio-tracking study is not without practical difficulties (see Baras, 1998; Lucas & Baras, 2000).

MECHANISMS UNDERPINNING MOVEMENT

Capability and causal factors Two types of mechanism for movement are categorised here. The first type comprises the capabilities of the individual. It includes both the movement and sensory capabilities of the animal and in a holistic sense is the navigational ability of the individual. The comprehensive review by Braithwaite and Burt de Perera (2006) forms the basis for the discussion in the following paragraphs. The second type refers to the individual’s motivation for staying or moving. Decisions regarding movement can have a physiological, population level, ecological or evolutionary basis. In this regard, debate over the currency for use in individual-based models is directly relevant. Typically a measure(s) of fitness (e.g. expected survival) is used having been derived in any one of a number of possible ways. For instance, Jager and Tyler (2001) discuss the merit of including terms that cater for state dependence or density dependence. Individual-based models are discussed later in this section following the treatment of empirical studies.

EMPIRICAL STUDIES

Navigation Laboratory-based studies have been useful in determining the orientation capability of fish (Braithwaite & Burt de Perera, 2006). Briefly, this has involved investigating: a) sensory capability; b) spatial memory including use of local or global landmarks, dead-reckoning and spatial maps; and c) integration of spatial information based on more than one type of sense (Braithwaite & Burt de Perera, 2006). The research builds on experiments that are often conducted on a model species, and in particular goldfish, guppies or sticklebacks, in captivity. Intra- and inter-population variability in orientation has been identified in sticklebacks (Girvan & Braithwaite, 1998; Braithwaite & Girvan, 2003). Overall, findings indicate that fish orientation is similar to that of terrestrial vertebrates (Braithwaite & Burt de Perera, 2006).

69 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

a)

N (16- 18 months) Nursery M O L (14 - 16 months) (0- 14 months) (22 - 24 months) (18- 22 months)

(day 27) b) P

NN (days 55-65) (days 1 - 26) M (days 28 - 54) O (days 243- 365) R Q (day 242) (day 241) ( days 66-240)

c) (month 2) N (month 1) M (months 9 - 12) O

(months 3-8)

(day 27) d) P N (days 55-65) (days 1 - 26) M (days 28- 54) O

(days 66- 90)

Figure 5.3. Hypothetical movement of an individual in: a) the first two years of life; b) the third year of life; and as observed by c) monthly radio-tracking during the third year of life; and d) by daily radio-tracking in a short-term study commencing at 2 years of age. Semi-permanent locations or home-ranges are labelled L, M, N and O, with each light grey oval representing a home-range, dark grey ovals represent core areas, and arrows indicate direction of movement. Temporary locations are represented by P, Q and R, and apparently at N in c). However, the latter is an artefact of radio-tracking on a coarse temporal resolution.

The movement of fish has been described in many field studies and used as a basis for inferring navigational ability. However, field-based studies aimed at determining the navigational capabilities of fish are far less common. A level of success has been achieved in studying species that move on a small scale (metres to hundreds of metres) in clear water systems including upland streams and coral reefs. Manipulation of landmarks along predictable daily travel routes on coral reefs has revealed that

An ecological approach to re-establishing Australian freshwater cod populations 70 FRDC Report 2003/034 fishes use landmarks and the order of the sequence of those landmarks to navigate (Reese, 1989; Mazeroll & Montgomery, 1998). However, even on coral reefs this type of study is not likely to be easily achieved at the among home-range scale. Not the least because of the difficulty of observing irregular and infrequent events such as home-range shifts.

Decisions Direct observation of fish in the field provides a means for acquiring information on fish behaviour as it relates to fish movement at the within home-range scale (MWHRS) and MAHRS. Major advantages of the technique include detection of: a) behaviours that occur on short time scales (e.g. seconds); and b) identification of important factors from many possibilities including factors that were unknown in the study design phase. Researching site-attached species in the fish communities of clear water systems been especially productive in this regard. Studies of freshwater cichlids (Perrone, 1978; Kuwamaru, 1986; Kuwamura, et al., 1989), coral reef fishes (Sale, 1991; Noda et al., 1994) and temperate reef fishes (Jones, 1984, 1988; Ebeling & Hixon, 1991) are among the more notable examples. However, direct observation has a number of limitations including observer effects (Martin & Bateson, 1993) since the researcher or remote filming device (both referred to as an observer in this case) must be within close proximity to the fish. The fish must then become accustomed to the presence of the observer (e.g. Nakano, 1995b) and specific life history phases are likely to react differently in this regard. The field of view is also relatively small, even under optimal conditions. Therefore studies based exclusively on a visual observation technique, are confined to a small area of study (e.g. Noda et al., 1994; Nakano, 1995b). This invokes tradeoffs whereby a sample population is observed in detail at a fixed site or less thoroughly at multiple sites in time and space (e.g. Noda et al., 1994; Nakano, 1995b).

An exceptional example of studying biological and ecological mechanisms for fish movement at the among-home-range scale is that of Nakano (1995b). Detailed observation of a salmonid (Oncorhnychus masou ishikawai) in upland streams, revealed that within and among-home-range scale movement, is governed by an intraspecific dominance hierarchy (Nakano, 1995b). A quantitative assessment of the aggressive encounters between individuals and the dimensions of territories was developed. A number of factors contributed to the success of this study. The extent of the study reach was relatively small and encompassed the scale of home-range shift behaviour. Individuals could be visually distinguished from their size and markings. A number of behaviours (feeding, aggressive encounters, subordinate behaviour) providing information regarding the mechanisms for staying or shifting from a home- range, were observable. This occurred as a function of Nakano making observations in real-time. Furthermore, the small size of home-ranges (metres) and the close proximity of individuals within a pool, enabled Nakano to simultaneously observe all pool occupants and identify intra-specific interactions. Finally, a large amount of information was obtained to describe WHRMS and MAHRS and the mechanisms for that movement in unison (also see earlier research that provided a platform for this study: Nakano, 1994; Nakano & Furukawa-Tanaka, 1994; Nakano, 1995a).

Conversely, individual fish behaviour as it relates to MAHRS is rarely observed this comprehensively in the field. Indeed in many circumstances it is impractical. Visibility is intermittently or permanently low in many aquatic habitats as a

71 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 consequence of elevated turbidity and/or reduced light. Observing species that are cryptic, nocturnal, associated with complex structure, or that move unpredictably or on large scales, pose difficulties for observation. Unfavourable conditions for researchers or cost prohibitive environments including extremes in water temperature (Ebeling & Hixon, 1991), inaccessibility to remote localities including montane habitats and deep seas, and dangerous situations such as river flooding, or the presence of predators such as sharks or crocodiles, can discourage the direct observation of individual fish behaviour. Applications in many of these instances require non-visual forms of observation or experiments to determine mechanisms that underpin movement.

In these cases, techniques that produce known-status-data, including radio and acoustic telemetry, are particularly useful for identifying mechanisms for three reasons. First, descriptions of movement can be used to identify locations in space and time that warrant detailed observation of mechanisms. For instance, Kieffer and Kynard (1993, 1996) verified successful spawning of tagged shortnose sturgeons (Acipenser brevirostrum) through both capture of milting males and by sampling for drifting eggs and larvae downstream of a suspected spawning location. Additionally, radio-tracking has proved particularly useful in the estimation of levels of bird and mammal predation on fish (e.g. Dieperink et al., 2001). Second, data describing individual movement can itself be examined for correlations with variables. Synchronised departure from an onshore reef lagoon by sharks was detected by acoustic telemetry, and occurred in relation to changes in barometric pressure in the lead up to a major storm event (Heupel et al., 2003). There are also a number of examples of changes in fish movement correlating with environmental variables including changes in stream discharge and/or temperature (Matheney & Rabeni, 1995; Harvey et al., 1999; David & Closs, 2002; Cooke, Bunt & Schreer, 2004; Naughton et al., 2005). Third, factors can be manipulated to test for an effect on individual movement. Examples of factors that have been manipulated in this type of study include: dominance hierarchy (Gowan & Fausch, 2002), discharge (Thorstad & Heggberget, 1998; Berland et al., 2004; Ebner et al., 2005), catch and release angling (Whoriskey et al., 2000) and rearing method in the case of released hatchery fish (Bettinger & Bettoli, 2002; Skalski et al., 2002; Jokikokko & Mäntyniemi, 2003). The approach has also been used to investigate research methodology (e.g. Makinen et al., 2000).

Finally, biotelemetry techniques represent further applications for investigating the physiological, biological and ecological mechanisms for movement in fish and are summarised by Cooke, Hinch et al. (2004). In brief, biotelemetry provides advances in defining functions such as energetic outputs associated with specific behaviours in fish. Cooke, Hinch et al. (2004) provide examples whereby the energetic requirements and stresses of fish passage, spawning ground defence and the stress and survival of angled fish are monitored by biotelemetry.

MECHANISTIC MODELS

Rules of individual fish movement have been explored by dynamic programming models (Railsback et al., 2001) and by learning algorithms (Jager et al., 1993; Clark & Rose, 1997; Van Winkle et al., 1998). Jager and Tyler (2001) explain two primary differences between the two approaches. Dynamic programming models assume that

An ecological approach to re-establishing Australian freshwater cod populations 72 FRDC Report 2003/034 an individual has knowledge of spatial and temporal dimensions of the environment, whereas, learning algorithms assume that an individual accumulates knowledge of the environment through exploration and visiting places that it has previously been to (Jager & Tyler, 2001). The decision to move or stay is made by an individual at its destination in the dynamic programming model but at departure in the learning algorithm (Jager & Tyler, 2001).

From an ecological perspective it would seem reasonable to expect that an individual is capable of deciding to stay or move at both the point of departure and destination. For instance, an individual could be displaced from its home-range having been confronted by a superior conspecific (Olsson & Shine, 2000). Subsequently, the individual might visit a number of potential home-ranges prior to deciding upon where to resume home-range occupation. In this regard a model that has scope for both departure and destination rules is preferable, and subsequently the timing of these decisions is incorporated into the conceptual model (Figure 5.2).

Jager and Tyler (2001) acknowledged that the two types of model could produce different results, and contended that the learning algorithm was more realistic in its assumptions of individual fish knowledge. They used a number of examples of individuals acting according to imperfect knowledge and implementing deficient strategies for surviving in a sub-optimal manner. Currently, the validity of available models remains to be evaluated with empirical information (Jager & Tyler, 2001). In this regard a promising area for future study is field-based evaluation of the role that fish experience, knowledge and learning has on MAHRS.

KNOWLEDGE GAPS

Though a new conceptual model that distinguishes among migration, exploration, and home-range shifts (including return home-range shifts) has been established, three major uncertainties remain:

(a) To date the frequency of these behaviours has only been established in wild populations over a fraction of the lifespan of an adult fish (e.g. Crook, 2004a, b; Chapter 3), and there is a need for more field based data. In lengthy studies, the focus has frequently been on a single behaviour such as spawning migration (e.g. Koehn, 1997; O’Connor et al., 2005), whereas, the simultaneous detection of other types of MAHRS is overlooked. In this respect the development of remote telemetry arrays should dramatically increase capabilities for collecting information on the long-term movement patterns of individuals and populations of fish (Lucas & Baras, 2000; Cooke, Hinch et al., 2004; Heupel et al., 2006).

(b) The mechanisms that underpin individual fish movement are often unknown. In particular applications involving biotelemetry and/or visual observation of mechanisms are a potential solution.

(c) The conceptual model should be further refined by challenging its logic structure and through testing with empirical data. To this end, the definition of what constitutes MWHRS and MAHRS will likely be challenged, particularly as comprehensive empirical datasets describing individual fish movement become available. In the

73 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 process, the choice of which fitness currency(s) such as to use in individual-based models (Jager & Tyler, 2001) should come under increased scrutiny.

MOVEMENT, MECHANISMS AND MODELS

Comprehensive understanding of fish MAHRS remains elusive. Success is likely to come from a) comprehensive description of movement, b) identification of mechanisms that underpin movement, and c) synthesis of these data into an explanatory or predictive model. In this regard, we reiterate that recent technological development provides opportunity for rapid improvement (Lucas & Baras, 2000; Cooke, Hinch et al., 2004; Heupel et al., 2006). There is also a need to transfer understanding of fish navigation based largely on laboratory studies (Braithwaite & Burt de Perera, 2006) to more complex field based scenarios (e.g. Reese, 1989; Mazeroll & Montgomery, 1998).

The success achieved by studying species that move on a small scale (metres to hundreds of metres) in clear water systems including upland streams and coral reefs should not be overlooked as an avenue for continued success (e.g. Noda et al., 1994; Nakano, 1995b). However, even in clear water ecosystems this type of study is not likely to be easily achieved at the among home-range scale. Not the least because of problems with making observations of irregular and infrequent events such as home- range shifts. Many fishes occupy a home-range on a scale greater than metres and move at the among-home-range scale at orders of magnitude beyond that which is easily observed underwater. Furthermore, many species are not easily observed for lengthy periods. Progress in gaining information regarding these species is likely to come from a variety of methods. Subsequently, the conceptual model presented in this study provides a framework for investigating the application of different methods and relating different information types. Further, comparison across studies should be aided by the formal recognition of different types of MAHRS. This issue is progressed in a discussion of conservation requirements of the Australian Percichthyidae.

APPLICATION TO AUSTRALIAN PERCICHTHYID CONSERVATION

Conservation and MAHRS What is important about understanding percichthyid movement at the among-home- range scale from a conservation perspective? The approach taken here is to consider this question at the species and population level. In this regard an important distinction can be made between a) Maccullochella and Macquaria comprising large bodied mobile species (Koehn, 1997; Simpson & Mapleston, 2002; O’Connor et al., 2005) that inhabit rivers including large rivers (referred to here as ‘large percichthyids’) (Table 5.1), and b) Nannoperca comprising small presumably less mobile species that generally occupy lentic waters (referred to as ‘small percichthyids’ here) (Table 5.1).

Priorities Large-scale movement of large percichthyids is of significance to humans because these taxa are fished, most are listed as threatened, and people can identify with them. In contrast, small percichthyids have no consumptive or recreational values and are

An ecological approach to re-establishing Australian freshwater cod populations 74 FRDC Report 2003/034 essentially only of interest due to their threatened conservation status. It is also relevant that large percichthyids are at risk of over harvesting from widespread recreational angling activity and illegal angling (no longer ) (Swales, 1994b). The effect of catch and release recreational angling is yet to be established for percichthyids. The small percichthyids are of no value to anglers but may be subject to small-scale, localised harvesting for aquarium purposes.

Species protection and recovery Two broad strategies for conserving threatened percichthyids are protecting individual remnant populations and protecting or enhancing the integrity and connectivity of multiple populations. In regard to conservation of small percichthyids, understanding individual MAHRS is not currently an accepted management priority. This is largely due to the small scales of movement anticipated for these species. Rather, protecting against local threatening processes is recommended (Arthington & Marshall, 1996; Kuiter et al., 1996; Hammer, 2001). In these cases, understanding and addressing processes, including introduced species and human impacts, is recommended (Arthington & Marshall, 1996; Kuiter et al., 1996; Hammer, 2001). In this respect, much of what has been presented earlier in this review regarding understanding mechanisms, behaviours or processes that can be investigated through observation or experimentation is applicable, though, in relation to MWHRS more than MAHRS. Conversely, where MAHRS is of importance to conserving small percichthyids, is in regard to human assisted dispersal. Translocation is available as a management tool should a population be at risk of extinction, and whilst it would be useful to develop an understanding of the dispersal ability of these fishes, this is not an immediate priority. Therefore the remaining discussion is focussed on large percichthyids for which MAHRS is a conservation issue.

Comprehensive description of movement As a consequence of applying manual radio-tracking and remote telemetry the description of large Australian percichthyid MAHRS is reasonably well progressed. This is despite the limited number of telemetry studies conducted to date and since radio-tracking has only been applied to Australian freshwater fishes since the 1990s (Koehn, 2000). Nevertheless there are a number of ways for improving description of percichthyid MAHRS. First, at key sites there should be long-term use of remote telemetry logger stations set equidistance or at functional points (e.g. weirs, river junctions) on river systems over a large extent (logger arrays, e.g. O’Connor et al., 2005). Further, the initial cost of increasing the number of loggers and therefore the spatial resolution of movement should be considered carefully in view of rate of information gain. At one extreme, consecutive loggers could be located so that radio- fields overlap slightly thus providing continuous known-status-data throughout the logger array or within a subsection of the array. However, inevitably in studying at a large extent (e.g. hundreds of river kilometres) loggers will be sparsely spaced. In these cases within the logger array, intensive studies of home-range behaviour should be encouraged. These studies would be similar to those of Crook (2004a, b) that involve manual radio-tracking of individuals at approximately daily intervals, or use remote-monitoring to confirm residence (e.g. David & Closs, 2001; Chapter 4). Second, an improved understanding of the habitat requirements of large percichthyids in relation to habitat and in particular in relation to human modification of habitat (including river regulation) is required. Macquaria ambigua was found to be relatively sedentary in the Broken River (Crook et al., 2001; Crook, 2004a, b),

75 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 whereas, the species can be mobile on a scale of tens to hundreds of kilometres in the Murray River (O’Connor et al., 2005). The underlying mechanisms for this different movement pattern are unknown, although, O’Connor et al. (2005) suspected it was merely an artefact of Crook radio-tracking outside of the active spawning period in spring. Further, there is a general lack of information regarding the movement of large percichthyid species in the Murray–Darling Basin in relation to unregulated rivers (partly since there are very few such rivers remaining). Third, there is potential for increasing the temporal extent of radio-tracking individuals since the longevity of Maccullochella and Macquaria is in excess of a decade (Anderson et al., 1992a, b; Gooley, 1992; Rowland, 1998a). In this regard, repeatability of spawning migrations and the possibility that migration routes are learned is of particular interest. Fourth, the conceptual model presented in this study (Figure 5.2) should be used in the design phase of studies as a framework for quantifying different types of movement relative to specific factors. Habitat type, impacts of alien species and reintroduction strategy (e.g. variations on life skills training during hatchery-rearing) are factors of immediate interest.

Mechanisms for movement A number of studies have confirmed or assumed that large-scale movement of large percichthyids was for spawning purposes (Koehn, 1997; O’Connor et al., 2005; Gavin Butler, unpublished data). The upstream migration of M. peelii peelii (Koehn, 1997) coincided with the known spawning period of that species in the Murray River catchment (Koehn & O’Connor, 1990). In contrast, M. ambigua were found to move downstream or upstream to spawn based on the synchronised movement of individuals in the mid-Murray River (O’Connor et al., 2005). Butler (unpublished data) was able to confirm spawning by underwater filming of egg guarding adults. As a secondary alternative in researching threatened percichthyids, non-lethal methods for determining reproductive status including cannulation (Ross, 1984) should be further developed, whereas, sacrificial methods including gonad staging have greater applicability to non-listed species.

Studies of the endangered M. peelii mariensis and M. macquariensis have reported unexplained MAHRS that is unlikely to have represented spawning migration (Simpson & Mapleston, 2002; Chapter 3). The reason for this being that the timing of movement was neither synchronised among individuals or that MAHRS occurred outside of the expected or known spawning period. Spawning could also be definitively ruled out if the approach of underwater filming as employed by Butler (unpublished data) was used. However, it should be noted that this method is restricted to clear water habitats and is not likely to be nearly as effective for studying Macquaria spp. unless the spawning-site is known to the researcher, for representatives of this genus are not egg guarders (Cadwallader, 1979; Harris & Rowland, 1996). In these instances, increased sample sizes can prove useful in identifying spawning aggregations (e.g. O’Connor et al., 2005).

The overarching recommendation for empirical research of percichthyid MAHRS is to concentrate resources for describing movement and observing mechanisms in the same place. Where threatened percichthyids are the focus it would seem most advantageous to select a characteristically clear water site(s). This would facilitate the application of direct observation including underwater filming (an alternative is the use of techniques not restricted by turbidity including DIDSONTM: Baumgartner

An ecological approach to re-establishing Australian freshwater cod populations 76 FRDC Report 2003/034 et al., 2006; Mueller et al., 2006). Generally, these conditions occur in upland habitats in Australia as many lowland habitats experience high turbidity. If individuals are to be observed for extended periods of time then small upland pools are likely to be best for overcoming the limitations of a small field of view. Alternatively, filming under controlled conditions including hatchery ponds or small reservoirs (scale of hectares) would prove informative. To this end experimentation and observation in an artificial system of sufficient size to house a sample population is worthy of consideration. This has particular relevance to producing fingerlings and also further on-growing fish specifically in regard to learning life skills important for survival post-release (e.g. predator avoidance, prey capture) (Brown & Day, 2002; Vilhunen & Hirvonen, 2003; Vilhunen, 2006).

FUTURE RESEARCH DIRECTIONS

This review has revealed that at least three fields of research have the capacity to make significant contribution to understanding of fish MAHRS. These fields are description of fish movement, identification of mechanisms for moving, and models that integrate the two. Promising directions for research in each of these fields are apparent from this review.

There is a need for detailed description of fish movement in the field. This will come from collecting known-status-data (Bennetts et al., 2001) in well-planned studies that cater for the extent, scale and resolution of movement (e.g. Petersson & Fausch, 2003). Current technology including radio and acoustic tracking techniques provide the opportunity to sample movement of large-bodied species. Application of remote- continuous telemetry holds promise for revealing habitual and historical movement of fish at the individual and population level responses to the environment (e.g. Heupel et al., 2003). Collection of data on these scales should be matched with careful consideration of the length of time over which individuals are observed. Subsequently, there will be a need to develop a range of data analysis approaches that investigate patterns of movement in relation to multiple variables at multiple scales.

Understanding the mechanisms underpinning decisions for movement and the means by which fish move and orientate in the aquascape is currently limited by a lack of field-based empirical data. Visual observation provides an effective means for rapidly improving understanding in this regard (e.g. Noda et al., 1994; Nakano, 1995b), however, as has been discussed, is limited in a number of ways. It could potentially be used to monitor ecological interactions in the home-range occupation phase of species that exhibit small home-ranges, and used in concert with a technique that collects known-status data to provide information regarding the mobile phase.

Calls for empirically derived data to refine models of individual fish movement (Railsback et al., 1999; Jager & Tyler, 2001) are reinforced by this review. Study of home-range behaviour and movement among potential and realised home-ranges can be advanced by quantifying the frequency of a number of behaviours: migration, return home-range shifts, shifts to new home-ranges and exploration that results in return to the most current home-range of occupation. The conceptual model presented here, provides a mean of distinguishing among these behaviours, and should prove useful in comparing the movement of populations or species. It also serves to focus research on the mechanisms that underpin these movements.

77 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

The conceptual model provides a framework for unifying two disparate fields of study (i.e. description of fish movement, identification of mechanisms for moving). Advancing the understanding of fish MAHRS is likely to come from use of this model and appreciation of both fields of study. In this regard, modelling is likely to have a role in integrating these fields of research and aiding the priority setting for adaptive management of fish populations including threatened Percichthyids.

An ecological approach to re-establishing Australian freshwater cod populations 78 FRDC Report 2003/034

CHAPTER 6. FATE OF TWO-YEAR OLD, HATCHERY-REARED MACCULLOCHELLA MACQUARIENSIS, STOCKED INTO TWO UPLAND RIVERS

Brendan Ebner,1, 2 Jason Thiem1, 2 and Mark Lintermans1, 2

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

Ebner, B., Thiem, J. & Lintermans, M. (2006). Fate of two-year old, hatchery-reared Maccullochella macquariensis, stocked into two upland rivers. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.79–94. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

INTRODUCTION

Radio-tracking has been used as a method for monitoring the short to medium-term success of hatchery-reared fishes stocked into rivers, in particular salmonids in the Northern Hemisphere (Jepsen et al., 1998; Aarestrup et al., 2005). This approach enables rates of mortality and dispersal to be quantified, facilitating adaptive management of fisheries (e.g. Aarestrup et al., 2005) or threatened species recovery (e.g. Skalski et al., 2001). There is scope for similar application in Australia, namely as part of reintroducing threatened freshwater cod of the genus Maccullochella.

The trout cod Maccullochella macquariensis (Cuvier) is one of four freshwater cod taxa in Australia, all of which are nationally threatened. Prior to the 1970s only one cod species (Murray cod, M. peelii peelii (Mitchell)) had been recognised, with M. macquariensis formally described in 1972 (Berra & Weatherley, 1972). Recovery programs have been in place for several years for three of the cod species, with a recovery plan currently being prepared for the fourth (M. peelii peelii).

Maccullochella macquariensis is a large species (maximum size 16 kg and ~850 mm) that occupies a range of habitats including pools, riffles and runs but is usually associated with deeper water and instream cover such as logs and boulders (Douglas et al., 1994; Ingram & Douglas, 1995; Growns et al., 2004). Maccullochella macquariensis was formerly widespread in the southern Murray–Darling Basin, occurring mainly in the lowlands but recorded at approximately 800 m altitude in the Murrumbidgee River. Sexual maturity is reached at 3–5 years of age with spawning occurring in spring. Like all of the freshwater cod species in Australia, M. macquariensis act as top and even apex predators, probably structuring fish communities historically (cf. Ebner, in press). The range and abundance of the species has declined dramatically in the last 50 years, with the species now only known from three self-sustaining populations, two of them a result of historical translocations. The specific reasons for the decline are unknown, but are likely to be related to increased river regulation, impacts of alien fish species, and habitat modification (Douglas et al., 1994; Ingram & Douglas, 1995).

79 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

For two decades conservation stocking of fingerling M. macquariensis has been conducted in the Murray–Darling Basin, in an effort to establish new self-sustaining populations (Lintermans & Phillips, 2005). There have been few signs of success from these stockings, presumably as a function of stocking insufficient numbers of fingerlings, poor release strategies and high post-release mortality (Brown et al., 1998; Bearlin et al., 2002; Todd et al., 2004), however, stocking success has never been thoroughly assessed. Generally, fish populations suffer greatest natural mortality in early life history phases (Jones, 1991) and this presumably applies to M. macquarienis (Todd et al., 2004). Subsequently, stocking fish at more advanced stages than as fingerlings has been modelled and proposed as an alternative strategy (Todd et al., 2004).

Percichthyids have been radio-tracked in Australia for two decades primarily to determine their movement and habitat-use (Koehn, 1997; Crook et al., 2001; Simpson & Mapleston, 2002; O’Connor et al., 2005). In these studies high survivorship has been recorded or assumed, although, there have been cases of failing radio-tags, mortality and indeterminate status (e.g. Crook et al., 2001; O’Connor et al., 2005). However, only recently has radio-tracking been used to monitor releases of hatchery- reared threatened . This has involved: a) examining the suitability of techniques for radio-tagging M. macquariensis based on a 4-month aquaria trial (n=9); and b) radio-tracking small numbers of on-grown M. macquarienis following release at a single site in the lowlands of the Murrumbidgee River and in the Cotter River (an upland tributary of the Murrumbidgee River, in the southeastern part of the Murray–Darling Basin) (Chapter 3; Ebner et al., 2005). These two studies have revealed retention of individuals to release sites or downstream dispersal, and poor survivorship. The finding of downstream dispersal is comparable with that from the more comprehensive literature based on hatchery-reared salmonid releases (Bettinger & Bettoli, 2002, and references therein).

Large rivers have generally been the targets of M. macquariensis re-stocking programs (Douglas et al., 1994; Brown et al., 1998; TCRT, 2004). However the annual production of M. macquariensis fingerlings in government hatcheries is low (relative to production of recreational species in the Murray–Darling Basin: Gilligan, 2005b; Lintermans et al., 2005b; Lintermans, 2006). This only allows small stockings of M. macquariensis that are unlikely to establish (Todd et al., 2004). One alternative is to stock greater numbers of fingerlings at a site (Bearlin et al., 2002; Todd et al., 2004) and possibly to use a small river where there would be less dilution of M. macquariensis at the site and monitoring of survivorship could be more effective. A second alternative (mentioned earlier) is to stock fewer, larger individuals (Todd et al., 2004). This approach has an empirical basis since the only definitive evidence of a self-reproducing population based on reintroducing M. macquariensis came from translocating an unknown number of individuals to Cataract Dam around 1914 (Harris & Rowland, 1996) and less than 150 large (i.e. >fingerling size) individuals into Seven Creeks in the 1920s (Douglas et al., 1994). Evidence is emerging that fingerling-stocking programs have resulted in successful wild spawning of the species at some sites (Douglas & Brown, 2000; King et al., 2005) but the establishment of self-sustaining populations remains elusive.

The aim of the current project was to compare dispersal and survivorship of two-year old, hatchery-reared M. macquariensis released into a large and a small upland river.

An ecological approach to re-establishing Australian freshwater cod populations 80 FRDC Report 2003/034

METHODS

Site description The study was located in two rivers, the Cotter and Murrumbidgee rivers, in the upper Murrumbidgee River catchment in the Australian Capital Territory (ACT), Australia (Figure 6.1). A single site was selected on each of the rivers for the release of two- year old, hatchery-reared M. macquariensis. The Cotter River is an upland stream dominated by short riffle-run-pool sequences, with the release site, the Spur Hole (148o53’E, 35o22’S, 600 m altitude), situated approximately 10 river kilometres upstream of Cotter Reservoir. Mean annual discharge (1988–2005) is 63 ± 11 gigalitres. The Cotter River previously drained a well-vegetated catchment dominated by native vegetation in the upper reaches and plantations of pine in the lower catchment; however, bushfires in 2003 burnt much of the catchment, resulting in losses of riparian cover and increases in water turbidity (Carey et al., 2003). The release site consisted of a 25 m long and 8–12 m wide pool, with water depth extending to 1.4 m, representing a typical pool for this river. During the study period discharge ranged from 12–1011 ML/day and water temperature was 10.3–29.6oC. The Murrumbidgee River within the ACT comprises long, deep, gorge pools interrupted by short riffles, with in-stream habitats dominated by bedrock and sand. Mean annual discharge (1977–2005) is 328 ± 59.6 gigalitres. The release site, Pine Island (149o04’E, 35o26’S, 540 m altitude), consisted of a 400 m long pool, of 20–60 m width and maximum water depths of 11 m. Discharge ranged from 4–1940 ML/day during the study and water temperature was 14.1–31.7oC.

Surgery and implantation of radio-tags Seventy-two on-grown, hatchery M. macquariensis (2 years of age, 330–424 mm (TL); 0.518–1.178 kg; first generation, bred from Murray River brood stock) were sourced from Snobs Creek Research Station in Victoria. Initially, post-larvae were reared in a fertilised earthen pond under semi-natural conditions. After harvesting the pond, fingerlings were transferred to hatchery facilities were they were weaned onto an artificial pellet diet, then on-grown in 500 L circular fibreglass tanks that were part of an intensive recirculating aquaculture system. Methods used for fry rearing, weaning and on-growing were similar to those employed for Murray cod, which are described in more detail by Ingram (2004).

The radio-tags used in this study were internal body implants with a 30 cm trailing whip antenna, models F1820 (8 g), F1830 (11 g), F1835 (14 g) and F1850 (25 g) (Advanced Telemetry Systems (ATS), Isanti, USA) and were obtained without mortality switches based on projected fish body weights at the time of purchase. Each individual was implanted with a radio-tag of ≤ 2% of its body weight. Two-stage radio-transmitters were used on a frequency of 150–152 MHz and programmed on a duty cycle of 5 sec on / 7 sec off to increase battery life. The minimum warranty for battery life of radio-transmitters ranged from 175–504 days depending on radio-tag size. Frequency drift was not encountered despite checking for its occurrence prior to surgery and following release.

81 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Figure 6.1. Location of the release sites (▲) on the Cotter and Murrumbidgee rivers in the Australian Capital Territory, Australia.

Radio-tags were inserted from 27–28 September 2004 under anaesthesia. Individuals were anaesthetised using 0.5 ml Alfaxan (Jurox, Rutherford, Australia) per litre of water. A 2 cm ventro-longitudinal incision was made into the peritoneal cavity along the ventral line anterior of the anus and posterior to the pelvic fin base. A modified cannular was used to exit the aerial of each radio-tag from the anterior of the caudal peduncle. The incision was closed with two or three sutures, and the temporary skin adhesive Vetbond (3M, St. Paul, USA) was applied and given approximately 20–30 sec to dry. An intra-muscular injection of the antibiotic enrofloxacin (Baytril) was administered at 0.1 ml kg-1, immediate-posterior of the nape. For external identification individuals were tagged with a dart tag between the second and third dorsal spines. Individuals were allowed to recover in a darkened enclosure holding 200 L of well-aerated water at 5 parts per thousand NaCl. The dentary was held to open the mouth, permitting the flow of aerated water from a power head over the gills, until an individual was clearly breathing independently. Time required for surgery (mean ± SE) was 5.2 ± 0.12, 15.2 ± 0.48 and 5.9 ± 0.30 min to achieve total anaesthesia, conduct surgery and perform recovery, respectively. Water temperature at the time of radio-tagging was 18.1–24.0oC.

Fish were held for an observation period of 2–3 days and on 30 September 2004 all 72 individuals were anaesthetised with 2 ml/100 l clove oil, and transferred by bucket into two separate 2000 L water containers (mounted on a truck, and aerated with oxygen). The two samples, each containing 36 individuals, of comparable length (T = -0.26; d.f. = 70; P = 0.79) and weight (T = -0.18; d.f. = 70; P = 0.86) were transported from the Snobs Creek Research Station (Victoria) to the ACT (a 6 hr journey). Upon arrival at each release site, samples were again anaesthetised (approximately

An ecological approach to re-establishing Australian freshwater cod populations 82 FRDC Report 2003/034

2 mL/100 L clove oil) and transferred by bucket to a 250 L container at the river edge. River water was bucketed into the container sporadically over 30–40 min to revive M. macquariensis from anaesthesia and acclimate them to river water. Late afternoon 30 September 2004, 36 individuals were released at both Pine Island on the Murrumbidgee River and the Spur Hole on the lower Cotter River. Monitoring independent of this study demonstrated that sufficient prey resources (primarily small fishes and australiense Holthuis) were available at and within close proximity to both release sites (Lintermans, unpublished data).

Four underwater and two above-stream video cameras were deployed at the release site in the Cotter River (before: 14 September; during: 30 September–4 October; after: 21 October 2004) to investigate both initial intra-specific and inter-specific interactions. To maximise the chances of filming such interactions, all fish were released at a single location in each river, resulting in higher than natural densities of released fish. However, filming during and after release proved largely unsuccessful as a consequence of high turbidity arising from rainfall and sediment runoff. Therefore, the results and discussion of the filming component of this study are not expanded upon here and are provided elsewhere (Chapter 7).

Radio-tracking A two-person crew undertook manual radio-tracking surveys weekly for the first 2 weeks, fortnightly for the first 3 months and monthly thereafter. Radio-tracking was conducted on foot or in a 3 m aluminium punt (powered by an 8 hp outboard motor), by directly positioning over the top of a radio-transmitter or where this was not possible by triangulating. Individuals were located during daylight hours using a scanning receiver (Australis 26k, Titley Electronics, Ballina, Australia or R4100, ATS) and a three-element Yagi antenna (Titley Electronics or ATS). The location of each individual was recorded by obtaining three waypoints with a hand-held GPS (Garmin GPSII Plus or Garmin GPS 76 Marine Navigator; Figure of merit (F.O.M.) ≤ 5.0 at about 95% of locations, F.O.M. ≤ 7.0 at 100% of locations). Where individual transmitter frequencies could not be detected in the study area by ground crews, a light aircraft with wing-mounted yagi antennas, was used to search the river corridor and surrounding areas. This occurred six times throughout the study and involved searching a 100 km stretch of the Murrumbidgee River (all radio-tags were accounted for in the Cotter River). When an individual was located by plane (Appendix 5), a ground crew was sent within 24 hrs to obtain a more accurate location. At the end of the study an additional plane tracking exercise was conducted to exhaustively search for missing radio-tags in the Murrumbidgee River (and tributaries), covering >250 km of river.

Data analysis GPS records were averaged to provide a single spatial location of each individual per radio-tracking fix. Spatial data were plotted in GDA94 format in ArcView 3.2™ (ESRI, USA) over a base-map. A polyline was generated based on the sequential locations of each individual using the Animal Movement Extension in ArcView (Hooge & Eichenlaub, 1997) and used to construct a time series of the distance moved between consecutive radio-tracking intervals. Polyline distances were calculated along a river midline. Where small to medium-scale movements took place between consecutive radiolocations, the line generated by the animal movement path extension was snapped to the river midline vertices in ArcMap™. Where large-scale movement

83 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 of an individual was detected, the segment of river midline between two points was measured by converting this line to a shape file and running an area/perimeter update using the ArcView 3.2™ X-Tools Extension. On occasions when individuals were not located they were assumed to have been stationary at their next known location and this potentially resulted in underestimates of movement.

Mortality, radio-tag rejection and missing fish Where an individual was recorded in the same location on consecutive occasions or was in a dubious location (e.g. extremely shallow water with no cover), attempts were made to startle it. If this produced no movement of the individual, attempts were made to locate the radio-tag using a modified gap-loop antenna (accuracy ≈10 cm) (Titley Electronics) and then retrieve it by snorkelling. Where radio-tags or fish remains were retrieved, signs of predator or scavenger involvement (e.g. common water rat (Hydromys chrysogaster Geoffroy) burrows, cormorant (Phalacrocorax spp.) guano) were recorded. In the event of locating a whole carcass, an autopsy was conducted. An individual was also considered to have undergone mortality in cases where a radio-tag was repeatedly found at a location, the radio-tag could not be retrieved (e.g. in deep water), the individual was not startled (see above) and movement was not detected by 4-hourly radio-tracking over a diel period. These cases were not considered to represent radio-tag rejections, and reasoning for this is provided at the beginning of the discussion section. The number of days at large was calculated based on the last recorded movement of an individual. Where mortality occurred, the distance moved from a previous location may have been a result of active transport by a predator.

RESULTS

Movement and dispersal Maccullochella macquariensis moved small distances in both rivers and remained close to the release sites (<3 km) for the duration of the study (Figure 6.2). At 1 month post-release the majority of the sample in the Murrumbidgee River was located immediately upstream of the release site within the release pool. The majority of the sample in the Cotter River remained immediately downstream of the release site, however, a few individuals had begun to disperse downstream in each river (Figure 6.2(a)). Similar patterns were observed at 3 months post-release, with the exception of an upstream (<1 km) movement in the Cotter River (Figure 6.2(b)) made by an individual. The number of radio-tracked individuals had decreased considerably by the 3-month stage. At 6 months post-release a single survivor was being radio-tracked in the Cotter River and was located immediately downstream of the release site (Figure 6.2(c), Appendix 7). No survivors remained in the Murrumbidgee River 6 months after release (Figure 6.2(c)). Dispersal distance was not related to body size in relation to the sample released in either the Cotter or Murrumbidgee rivers (Pearson correlation, p = 0.68 and p = 0.30, respectively).

An ecological approach to re-establishing Australian freshwater cod populations 84 FRDC Report 2003/034

20 (a) 1 month n=11 15 n=22 10

5

0 20 (b) 3 months n=2 15 n=7

10

5

0 20

Number of individuals (c) 6 months n=1 15 n=0 10

5

0 3.5-3 3-2.5 2.5-2 2-1.5 1.5-1 1-0.5 0.5-0 0-0.5 0.5-1 1-1.5

(downstream) Release site (upstream) Distance from release site (km)

Figure 6.2. Dispersal of individuals released in the Cotter (black) and Murrumbidgee (grey) rivers at (a) 1 month, (b) 3 months, and (c) 6 months after release. Sample size at each time period is indicated. Mortality and radio-tag rejection This study identified 100% mortality or radio-tag rejection of M. macquariensis released into the Cotter River at the end of 7 months of radio-tracking (Table 6.1). In the Murrumbidgee River, 86% of the sample were accounted for by mortality or radio-tag rejection by 7 months post-release (Table 6.1). The remaining 14% of the sample is not expected to have comprised individuals that had actively emigrated from the study area as confirmed by radio-tracking from a plane. The fate of these individuals was attributed to predator assisted emigration (or predators destroying radio-tags) or radio-transmitter failure (Table 6.1).

Table 6.1. The mean (± SE), total length (TL), weight and subsequent fate of radio- tagged individuals following release at two sites. Numbers of individuals are in parentheses. Group n TL Weight Radio-tag Mortality (mm) (g) failure or or predator radio-tag assisted rejection emigration Cotter 36 363±3.6 759±26.8 0 100% (36)

Murrumbidgee 36 365±3.9 767±28.4 14% (5) 86% (31)

85 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Survivorship was significantly prolonged in the Murrumbidgee River relative to the Cotter River (Mann-Whitney U test: p<0.05, n=72) (Figure 6.3). Eleven individuals were alive 1-month after release in the Cotter River (Figure 6.3(a)). Survival continued to decline, with seven individuals alive after 2 months, two individuals alive after 3 months and one individual alive after 4 months. The latter survived for 6– 7 months at which stage its radio-tag was recovered on shore. In the Murrumbidgee River, 22 individuals survived for at least 1 month after release at which stage survivorship continued to decline (Figure 6.3(b)). Two months after release 15 individuals were alive. There were no survivors 6 months after release.

Evidence for cause of mortality was more commonly found in the Cotter River than in the Murrumbidgee River. In the Cotter River three intact specimens were recovered, 19 individuals or radio-tags had evidence relating to direct predation or scavenging and 14 individuals were linked to radio-tag rejection or mortality (however carcasses were not recovered) (Table 6.2). In the Murrumbidgee River, of the 31 individuals without radio-tag failure or predator-assisted emigration, two intact specimens were recovered, five individuals or radio-tags had evidence relating to direct predation or scavenging and 24 individuals were classified as radio-tag rejections or mortality (Table 6.2).

In the Cotter River, individuals surviving for less than 2 weeks were predominantly located or recovered in a large gorge approximately 1 km downstream from the release site (Figure 6.4(a)). At 2–4 weeks after release, the terminal locations of six individuals were in Cotter Reservoir, 10–12 river km downstream from the release site. A specimen was also recovered from Condor Creek, a major tributary of the Cotter River. The majority of survivors were located close to the release site or downstream in the gorge. A radio-tag was also located in Cotter Reservoir at greater than 4 weeks after release.

Three intact specimens were recovered from the Cotter River (at 1, 2 and 4 weeks post-release), less than 2 km downstream of the release site, and autopsy did not reveal the cause of death (Figure 6.4(b)). The specimens showed no signs of predation or scavenging, with healing of surgery wounds underway and no evidence of infection. Two individuals had empty stomachs and one had fish vertebrae in its stomach. All seven terminal locations in Cotter Reservoir were attributed to direct predation or scavenging (Figure 6.4(b)), thought to be by cormorants, probably great cormorants (Phalacrocorax carbo (L.)). The remaining individuals that had been predated or scavenged (Figure 6.4(b)) were located immediately downstream of the release site in a gorge. Fourteen individuals represent unidentified mortality or potential radio-tag rejection and were categorised by stationary radio-tags (Figure 6.4(b)) located within close proximity to the release site.

An ecological approach to re-establishing Australian freshwater cod populations 86 FRDC Report 2003/034

Alive Confirmed mortality or radio-tag rejection Suspected mortality or radio-tag rejection Failed radio-transmitter 40 (a) 35

30

25

20

15

10

5

0 40 (b) 35

Number of individuals 30

25

20

15

10

5

0 1234567

Time elapsed (months)

Figure 6.3. Fate of individuals released in (a) the Cotter River and, (b) the Murrumbidgee River.

87 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Table 6.2. The number of individuals suffering mortality, predation/scavenging or radio-tag rejection seven months after release. Numbers of individuals are given in parentheses. Group Intact Stationary Direct specimen radio-tag predation or scavenging Cotter (n=36) 8% (3) 39% (14) 53% (19)

Murrumbidgee (n=31) 7% (2) 77% (24) 16% (5)

Evidence of predation in this study came from the identification of strike-marks on specimens that were not consumed. Evidence of predation or scavenging included multiple radio-tags found under trees (Casuarina cunninghamiana Mique) where cormorants frequently perched in groups, and chewed radio-tags or retrieved radio- tags in or near common water rat (Hydromys chrysogaster) burrows along the riverbank. Direct observations of predation or scavenging were not made in this study; however, multiple observations of live M. macquariensis were made in the early stages of the study, confirming initial survival of some individuals. At least five individuals were probably consumed by H. chrysogaster in the Cotter River. Of these, two were located in the gorge downstream of the release site, with the remaining three located in close proximity to the release site. Indirect evidence indicates that 12 individuals were killed or scavenged by cormorants in the Cotter River. Four of the seven radio-tags located in Cotter Reservoir were recovered less than 10 m apart under a tree commonly frequented by cormorants. Three radio-tags were recovered from the downstream gorge less than 2 weeks after release under a large tree in a large pile of avian guano with a scattering of fish bones and skin. Additionally, two individuals were recovered as partially intact specimens; however, the type of predator/scavenger could not be identified.

In the Murrumbidgee River the terminal location of most individuals was within the release pool or immediately downstream (Figure 6.5(a)). The last individual survived for 5–6 months, with its radio-tag recovered six months post-release, 84 km downstream from the release site near Burrinjuck Reservoir (not presented in Figure 6.5). This radio-tag was recovered under a cormorant perch in water less than 50 cm deep, in a section of river devoid of in-stream structural habitat.

A single intact specimen was retrieved from the Murrumbidgee River at four and five months post-release, respectively. Both had healed surgical incisions and showed no signs of infection. One of these individuals was recovered on shore, four river kilometres downstream from the release site (and its previous location) (Figure 6.5(b)) following a high flow event. The cause of mortality is unknown in both cases; however, no signs of predation or scavenging were evident. Five individuals were killed or scavenged in the Murrumbidgee River and did not represent radio-tag rejections. Three of these were attributed to cormorants. One radio-tag was recovered from a H. chrysogaster burrow in the release pool. In one case a carcass exhibited signs of predation or scavenging, however, the type of predator/scavenger was not evident. Twenty-three of the remaining 24 individuals were confirmed to either be mortalities or have rejected a radio-tag based on retrieval of radio-tags or a total lack

An ecological approach to re-establishing Australian freshwater cod populations 88 FRDC Report 2003/034 of movement following radio-tracking at four-hourly intervals over a diel period. Additionally, an individual was assigned to the mortality/radio-tag rejection category after the radio-tag had been repeatedly located in a position in shallow water and efforts to retrieve the radio-tag were unsuccessful (presumably the radio-tag was buried).

Our capacity to distinguish mortality from radio-tag rejection was lower in the Murrumbidgee than the Cotter River. This difference was attributed to the Murrumbidgee being a larger, deeper and more turbid river, decreasing the opportunity to recover radio-tags and carcasses, and reducing the possibility for visual confirmation of survivors.

Figure 6.4. The terminal locations and number of individuals released in the Cotter River, represented as (a) the time from release until mortality or tag rejection and (b) the fate of individuals. Note: large distances moved probably resulted from transport by avian predators.

89 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Figure 6.5. The terminal locations and number of individuals released in the Murrumbidgee River, represented as (a) the time from release until mortality or tag rejection and (b) the fate of individuals. Note: distances moved may have resulted from transport by avian predation. The final location of a radio-tag recovered 84 km downstream from the release site is not shown.

DISCUSSION

Possible rejection of radio-tags Rejection of radio-tags is considered unlikely in the current study despite frequent recovery of radio-tags and repeated confirmation of stationary radio-transmitters at locations. A number of lines of evidence support this contention. Five intact specimens, including three from the Cotter River and two from the Murrumbidgee River had retained radio-tags and showed good healing of surgical incisions. An

An ecological approach to re-establishing Australian freshwater cod populations 90 FRDC Report 2003/034 aquaria trial including a 4 month observation period post-surgery demonstrated the suitability of this species to radio-tag implantation by the research group (Ebner et al., 2005) and comparable results have been achieved by others (Simon Nicol, Department of Sustainability and Environment, Victoria, personal communication). It should be acknowledged that whilst these aquaria trials provide an indication that the species is suitable for radio-tagging, potentially important biotic and abiotic factors may have been overlooked under controlled conditions (e.g. flow, competitive interactions). Radio-tag rejection has been reported for some hatchery-reared individuals prior to release and was almost certainly a function of re-opened surgical incisions following sutures tearing through the abdominal wall in response to substantial fat deposits and an enlarged abdomen in the first experiment (Chapter 3). In contrast M. macquariensis were not rounded and laden with fat deposits in this experiment. The length of surgical incisions was also greatly reduced in this study in comparison with the study reported in Chapter 3 (about 1.5–2 cm, reduced from 3–3.5 cm). Inspection prior to release revealed excellent healing and no radio-tag rejection. Furthermore, Maccullochella spp. have been successfully radio-tagged in situ (Koehn, 1997; Simpson & Mapleston, 2002). Based on the assumption that rejection of radio- tags was not occurring in the current study, limited dispersal and high mortality in both small and large upland rivers were the major outcomes of releasing hatchery- bred M. macquariensis in this study.

Dispersal Conclusive evidence of large-scale movement or dispersal by live M. macquariensis was not obtained in the current study (Figure 6.2). Radio-tags found substantial distances from release sites (>5 km) were likely the result of transport by avian predators (see discussion of mortality below and Figure 6.4 and 6.5). Rapid mortality may largely explain the lack of large-scale dispersal in the current study relative to that reported from another release of on-grown M. macquariensis (Chapter 3).

In the current study dispersal was on small-scales, typically occurring at tens to hundreds of metres and less frequently kilometres (Figure 6.2). Patterns of dispersal were characterised by retention to release sites and small-scale downstream dispersal. This retention to release sites has also been demonstrated for M. macquariensis when released as fingerlings (Faragher et al., 1993; Douglas et al., 1994). Retention to the release site is largely in agreement with releases of hatchery-reared salmonids (Bettinger & Bettoli, 2002, and references therein). However, there was more substantial downstream movement and dispersal on the scale of kilometres in a previous study of M. macquariensis (Ebner et al., 2005), and particularly in a larger lowland river (Chapter 3). Rapid mortality considerably reduced the opportunity for dispersal to occur in the current study, however, the role of stream morphology in contributing to this reduced dispersal may also warrant consideration. The only other study of M. macquariensis movement in upland rivers (Ebner et al., 2005) involved larger fish (388–511 mm) than that of the current study. It is unknown whether the different stream morphology of upland rivers (short pool/riffle/run sequences as opposed to long reaches of pool-type habitat in lowland rivers) may influence dispersal behaviour. For example, do mesohabitats such as shallow and fast-flowing rapids or riffles provide some constraint to ordinary ranging behaviour in these large- bodied fishes?

91 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Mortality Mortality principally occurred within 2 months of release. This is comparable with the fate of two-year old hatchery-reared M. macquariensis released into the Murrumbidgee River at a lowland site (Chapter 3), although, difficulties with ascertaining the precise time of death were pronounced in that study. In contrast, high survivorship was reported for wild M. macquariensis over the entire 13 months of that study (Chapter 3) and in a prior study in the Cotter River there was complete survivorship of large hatchery-reared M. macquariensis (n=8) over the first 4 months post-release followed by complete mortality 4–9 months post-release (Ebner, Johnston & Lintermans, unpublished data).

A number of differences among these studies provide potential explanation for the observed differences in the rate of survivorship of hatchery-reared M. macquariensis (excluding the issue of radio-tag rejection which was dealt with earlier). Whilst size at release can be an important determinant of survivorship (Brown & Day, 2002), it is unlikely to be the reason for different rates of survivorship among studies of on-grown M. macquariensis (Chapter 3; Ebner, Johnston & Lintermans, unpublished data; the current study). There is considerable overlap in the size distribution of samples in each of these trials. Additionally, the previous Cotter River study (Ebner et al., 2005) made use of a 4-month holding period following surgery and fed live crayfish (Cherax spp.) prior to release. In contrast, M. macquariensis were released into the wild 3 days after surgery in the current study and 1–2 weeks after surgery in Chapter 3. Prolonged survivorship in a previous study (Ebner, Johnston & Lintermans, unpublished data) may be due to the full recovery from surgery afforded under aquaria conditions (e.g. improved healing of incision, opportunity to overcome any initial infection) or the absence of anaesthesia used in the final step of moving fish from the transport vehicle to the stream. Further, the opportunity to recognise, capture and handle live prey during grow-out in farm dams and in aquaria may have been advantageous (Brown & Day, 2002; Thorpe, 2004), whereas, in the current study and that of Chapter 3 hatchery-reared M. macquariensis were fed a pellet diet.

The seasonal timing of release also differed among these studies of M. macquariensis (Chapter 3; Ebner et al., 2005; the current study), and studies of hatchery-reared salmonids indicate time of release to be an important determinant of survival (Cowx, 1998). The previous liberation of hatchery M. macquariensis involved a release in mid–late summer (Ebner et al., 2005), whereas, releases were conducted in early–mid spring in the current study and that of Chapter 3. It is possible that factors relating to time of year, for instance annual cycles of suitable prey availability, influence rates of survivorship. In terms of prey availability in the Cotter River, the abundance of the resident fish fauna (principally Gadopsis bispinosus Sanger and Oncorynchus mykiss Walbaum) was especially low in this study and the previous Cotter River study (Lintermans, unpublished data) as a result of a catastrophic bushfire event in January 2003 (Carey et al., 2003). However, there was a high abundance of large Macrobrachium australiense, ideal prey for M. macquariensis (Baumgartner, in press) in the first few months following release in the previous study, though, not in this study (Ebner, personal observation). The fish fauna remained relatively intact in the Murrumbidgee River following the bushfire event (Lintermans, unpublished data) and therefore decreased prey availability does not provide an explanation for the death of the released M. macquariensis at that site. It would certainly be useful to develop an understanding of the feeding ecology of M. macquariensis following release

An ecological approach to re-establishing Australian freshwater cod populations 92 FRDC Report 2003/034 through further study. To this end, investigation of survival of fish that have received different levels of training in feeding on live prey prior to release is warranted (Brown & Day, 2002; Thorpe, 2004).

Larger fish were released in a previous study in the Cotter River (Ebner et al., 2005) relative to the current study. Larger size can correspond to increased protection from predators (Hyvärinen & Vehanen, 2004). The habitat conditions in the Cotter River were also somewhat different in the earlier study, with the catchment in the very early stages of recovery from the 2003 bushfires. As such, turbidity was higher for extended periods (weeks) after rainfall events, potentially affording more cover to fish. By the time of the release of fish in the current study, the supply of sediment from the catchment had reduced substantially as vegetation cover re-established.

The Murrumbidgee River in the ACT is historically more turbid than the Cotter River, with a lower proportion of the catchment burnt, and the severity of the fire reduced in comparison to that of the Cotter catchment (Carey et al., 2003). However, the Murrumbidgee River has suffered from excess sedimentation for decades as a result of gully erosion and channel incision in the upper catchment (Eyles, 1977; Erskine, 1997; Prosser et al., 2001). As a consequence, much of this coarse eroded material is stored in the channel, and is reworked with every large rainfall event. This coarse sediment has severely degraded pool habitats in the Murrumbidgee River in the ACT, severely reducing substrate diversity, one of the major sources of cover for fish in upland rivers (Lintermans, 2002, 2004). As a result of this loss of substrate cover in the Murrumbidgee in the ACT, coupled with the general scarcity of in-stream structural woody habitat resulting from riparian degradation, refuges for sub-adult fish are likely to be suboptimal (though still capable of supporting populations of the congener M. peelii peelii). Lack of cover coupled with the potential naivety of hatchery-reared fish may have contributed to high mortality of released fish. Habitat rehabilitation and reduction of densities of the potential competitor M. peelii peelii are factors that warrant attention in future reintroductions of this species.

In this study, mortality and/or rejection of radio-tags accounted for 100% and 86% of M. macquariensis released into the Cotter and Murrumbidgee rivers, respectively. In the Murrumbidgee River it is possible that the remaining 14% of the sample was the result of avian or human-assisted (illegal angler-take) emigration from the study area. Certainly the finding of a radio-tag 84 km downstream of the release site along the Murrumbidgee River is indicative of avian predators acting as a vector on a large- scale (see also Marchant & Higgins, 1990) and at the limits of our detection capabilities. Furthermore, the high dependability of radio-tags manufactured by ATS supports the argument that the 14% of missing radio-tags was related to transport or destruction by avian predators or anglers.

Cormorants and to a lesser extent common water rats, were the primary predators or scavengers of M. macquariensis based on available evidence in this study. Of the 72 individuals released, cormorants (probably P. carbo) and common water rats were associated with at least 15 and six cases, respectively. It is likely that these are underestimates and it is acknowledged that our approach was incapable of detecting instream predation by M. peelii peelii in the Murrumbidgee River (M. peelii peelii do not occur in the Cotter River). A number of radio-tags were recovered long after mortality had taken place based on the condition of specimens and movement data,

93 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 and a large number of radio-tags were never recovered (e.g. from deep water) thus precluding an opportunity to conduct autopsies and reach a conclusion regarding the cause of death. Therefore more frequent radio-tracking (e.g. daily) should be incorporated into future studies, at least in the first couple of months post-release. Ensuring hatchery-reared fish are of a sufficient size to receive radio-tags fitted with mortality switches is also recommended.

Phalacrocorax carbo is an effective piscivore capable of feeding on even larger fishes than the M. macquariensis released in the current study (Marchant & Higgins, 1990). A study of P. carbo in southeastern Australia revealed that fish comprised in excess of 80% of the diet, and that the take by cormorants exceeded the commercial and recreational harvest (Coutin & Reside, 2003). Coutin & Reside (2003) recommended that cormorant predation be taken into account in stock enhancement programs. The results of the current study certainly support that recommendation. Hydromys chrysogaster has also been found to prey upon fish comparable in size to that recorded in the current study (Woollard et al., 1978) and avian and mammalian predators have had significant impacts on other hatchery-reared fishes released into the wild (Derby & Lovvorn, 1997; Dieperink et al., 2001; Aarestrup et al., 2005). There are also a number of additional mechanisms that could account for the mortality of M. macquariensis in this study, including predation or competitive exclusion by M. peelii peelii, capture by anglers or succumbing to disease. A better understanding of predator effects and avenues for soft-release of hatchery-reared M. macquariensis into rivers (Griffin et al., 2000) should be incorporated into recovery of this species. To this end, progress toward successful reintroduction of Maccullochella spp. should be achieved by pre-training hatchery-reared fish, discerning their ecology upon release into the wild and adopting an experimental approach to releases that relates survivorship to management options.

An ecological approach to re-establishing Australian freshwater cod populations 94 FRDC Report 2003/034

CHAPTER 7. VIDEO MONITORING IN UPLAND STREAMS

Ben Broadhurst,1 Jason Thiem1, 2 and Brendan Ebner1, 2

1Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

Broadhurst, B., Thiem, J. & Ebner, B. (2006). Video monitoring in upland streams. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.95–104. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

‘Fish are rarely observable to the dry researcher, and require unique methods and techniques to study their behaviour’ (Priede, 1980).

INTRODUCTION

There is a plethora of methods and techniques used for studying the behaviour of fish species. These include electronic tagging, and direct and remote observations (Lucas & Baras, 2000). Remote methods enable researchers to monitor fish behaviour without disturbance (Lucas & Baras, 2000), however, direct observation (e.g. snorkelling and observation from shore) provide means by which to quantify behaviour in real time.

Underwater video monitoring can be used to observe fish behaviour from a remote location (Collins & Hinch, 1993). The technique is commonly applied in marine systems (e.g. Jury et al., 2001; Yau et al., 2001; Holbrook & Schmidt, 2002; Cappo et al., 2004). In freshwater systems, fish energetics (e.g. Trudel & Boisclair 1996; Hinch & Rand, 2000), counts of individuals passing through fish-ways (e.g. Hierbert et al., 2000), spawning behaviours (Cambray, 1997) and foraging activities (Collins & Hinch, 1993) have been documented using underwater video monitoring. Underwater video has rarely been used to study the behaviour or ecology of freshwater fish in Australia, in part probably owing to the unsuitability of using a visual technique in high turbidity, low visibility environments.

Underwater video monitoring was used in this study to describe ecological interactions involving re-introduced trout cod, Maccullochella macquariensis, in an upland river (Cotter River, Australian Capital Territory (ACT)) typically characterised by low turbidity, high visibility conditions. Maccullochella macquariensis is a top predator, feeding predominantly on and fishes (Harris & Rowland, 1996) and an initial trial demonstrated the suitability of underwater video as a means of monitoring diel activity patterns of potential prey species and quantifying ecological interactions among fish in the Cotter River (Ebner, unpublished data). The aims of the current study were to describe ecological interactions involving M. macquariensis in the days following release into a pool in the Cotter River and in the process evaluate the usefulness of the underwater video method.

95 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

METHODS

Study site The study was conducted in the Cotter River, ACT, a high gradient upland stream with a previously well-vegetated catchment and associated high water clarity. However, water clarity was reduced as a consequence of sedimentation associated with wildfires in the Cotter River catchment in January 2003 (Carey et al., 2003). Two study sites were used, both located in a stretch of the Cotter River between Bendora Dam and Cotter Reservoir (Figure 7.1). The Burkes Creek Crossing site was used as the control site (no stocking of M. macquariensis) and was the larger of the two pools. It consisted of a pool approximately 60 m in length, with a width ranging from 5–8 m and a maximum depth of around 1.5 m. A 30 m section of the pool was used for filming. The Spur Hole site, approximately six kilometres downstream of the Burkes Creek Crossing site was used as the treatment site (stocking of M. macquariensis). The Spur Hole pool is approximately 25 m in length, 8–12 m in width and had a maximum depth of around 1.4 m. The entirety of this pool was used for filming.

Figure 7.1. Location of the study sites (▲) on the Cotter River in the Australian Capital Territory, Australia.

Study design Video monitoring was undertaken across three distinct sampling periods to accommodate a before and after, control and impacted (BACI) design and analysis; pre-release, release and post-release. Pre-release filming occurred at both sites before the release of the 36 radio-tagged Maccullochella macquariensis at the Spur Hole site (Table 7.1). Filming at the time of release was only conducted at the Spur Hole as no M. macquariensis were released into the Burkes Creek Crossing pool. The purpose of filming the release was primarily to observe any initial aggressive interaction among individuals in particular in regard to establishing micro-habitat refuges. Post-release

An ecological approach to re-establishing Australian freshwater cod populations 96 FRDC Report 2003/034 filming occurred at both sites (Table 7.1). The presence-absence and activity of potential prey species were to be quantified as the basis for the BACI analysis. Table 7.1. Dates of filming for both Burkes Creek Crossing (control site) and Spur Hole (treatment site). Sampling period Burkes Creek Crossing Spur Hole Pre-release 16/09/2004 14/09/2004 Release Not filmed 30/09/2004 – 04/10/2004 Post Release 02/11/2004 21/10/2004

Camera placement Six black and white low-lux-cameras (SciElex, Tasmania) were used at each site. A camera attached to a large custom made aluminium tripod was set up at both the head and foot of each pool (see Figures 7.2 and 7.3). These tripods gave the cameras an elevation of about three metres above the water and provided a view of the entire width of the river. These cameras were set-up to view the fish as they entered and exited the pool. The remaining four cameras were placed underwater, and arranged so that the majority of each pool could be monitored (see Figures 7.3 and 7.4). Each underwater camera was attached to a small adjustable tripod with foldable footplates welded onto each leg, and weighted to the bottom of the channel by placing rocks on the tripod feet.

All six cameras were connected to a 9 channel multiplexer (SciElex, Tasmania), which divided the frame rate amongst the six cameras. Records at 4 frames per second from each camera, were obtained with a long play video recorder (Mitsubishi Electronics). A 6” LCD screen (TFT model: LEE – 561) facilitated adjustments to the field of view for each camera. A 12V battery powered all equipment. Video recording equipment (VCR, battery, multiplexer and LCD screen) were kept off the ground using a shelf, and housed under a tarpaulin as protection from precipitation. Filming was conducted during daylight hours, 05:00 to 19:00 hours (04:30-19:30 hours on 21/10/2004), to record from one hour before sunrise to one hour after sunset. No artificial lighting was used.

Processing of video footage Footage was viewed with the long play video recorder on a 64 cm flat-screen television. Footage from all four underwater cameras was viewed simultaneously as a two-by-two view set-up on the television screen. Observations were re-viewed with a single camera view on the screen. The two above-water cameras were viewed separately, as one camera per view. All species observations and the duration of each observation were recorded. A total of 24 hours of footage was viewed based on sub- sampling according to that described in Table 7.2. The focus was on morning and late afternoon periods (Table 7.2), as these times were believed to yield the greatest number of observations of fish based on previous study (Ebner, unpublished data). Additionally, the effectiveness of the cameras under high light conditions was investigated by evaluating footage collected during the middle of the day (Table 7.2).

97 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Figure 7.2. Camera mounted on a large tripod at the Spur Hole site on the Cotter River. (Photograph: Parks, Conservation and Lands)

Figure 7.3. The position of underwater cameras three and four (centre and bottom right of photo), recording gear under tarpaulin (top left of photo) and above water camera (top left) at the Spur Hole site on the Cotter River. (Photograph: Parks, Conservation and Lands)

An ecological approach to re-establishing Australian freshwater cod populations 98 FRDC Report 2003/034

Figure 7.4. Schematic representation of the locations where cameras were placed (with corresponding camera number) and pool dimensions at a) Burkes Creek Crossing and b) Spur Hole sites (not to scale).

Table 7.2. Sub-samples of continuous footage viewed, including site, treatment date and time of day. Sampling period Burkes Ck Crossing Spur Hole Pre-Release 16/09/2004: 0500–0800 & 14/09/2004: 0800–1100 1700–1900 & 1700–1900 Release No footage filmed at this site 30/09/2004: 1700–1900 for this period 03/10/2004: 1400–1700 Post-Release 02/11/2004: 1400–1700 21/10/2004: 0430–0730 & 1630–1930

Presence of released fish The presence of M. macquariensis in the release pool was determined by scan- sampling based on radio-tracking of each individual. Radio-tracking was conducted in the morning (~0500 hours) and evening (~1700 hours) (for dates see Table 7.3) at the upstream and downstream end of the Spur Hole Pool with the aid of a three-element Yagi antenna (Titley Electronics, Ballina) and a scanning receiver (Australis 26k, Titley Electronics).

RESULTS

Released fish and radio-tracking data Of the thirty-six M. macquariensis implanted with radio-tags and released into the head of the Spur Hole pool, radio-tracking confirmed the presence of between 4 and 10 individuals in the Spur Hole pool during filming (Table 7.3). The number of individuals present in the Spur Hole at the commencement and conclusion of filming each day, is shown in Table 7.3.

99 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Water quality The filming period (mid September 2004 to late October 2004) coincided with heavy rainfall and associated increased discharge in the Cotter River catchment (Figure 7.5). During this period turbidity levels were elevated by high levels of fine sediment input from the surrounding un-vegetated (burnt) catchment (Broadhurst, personal observation). Heavy rain for a period of 10 minutes was experienced at the Spur Hole site at 14:05 on 02/10/2004 (Mark Lintermans, personal observation). This was followed by a rapid increase in river water turbidity approximately 30 minutes after the rainfall (Mark Lintermans, personal observation). Although turbidity levels were not measured, similar rainfall events in the Cotter catchment resulted in turbidity readings of hundreds of NTU (Mark Lintermans, unpublished data).

Table 7.3. Number of M. macquariensis present at the Spur Hole during filming, confirmed by radio-tracking. Session Date Time Number of individual M. macquariensis present in the pool Release 30/09/2004 19:00 7 Release 01/10/2004 05:00 10 Release 01/10/2004 17:00 9 Release 02/10/2004 05:00 10 Release 02/10/2004 17:00 7 Release 03/10/2004 05:00 6 Release 03/10/2004 19:00 6 Release 03/10/2004 05:00 5 Release 03/10/2004 18:30 5 Post-release 21/10/2004 07:30 4 Post-release 21/10/2004 19:30 4

Camera Footage: Burkes Creek Crossing Water surface glare impacted on the quality of images filmed by the two above-water cameras (cameras 1 and 6) at Burkes Creek Crossing (Table 7.4). Even in the absence of glare, the deeper water (coupled with overall low water clarity) captured on Camera 6 prevented viewing and identification of fishes. Neither of the above water cameras provided footage suitable for analysis at this site (Table 7.4).

None of the underwater cameras from Burkes Creek Crossing provided any footage suitable for analysis. Low visibility (caused by high water turbidity) affected all underwater cameras. Maximum useful visibility was around 30 cm (see Table 7.4). The pool was too wide and too deep to be filmed under the turbid conditions that were prevalent during filming.

An ecological approach to re-establishing Australian freshwater cod populations 100 FRDC Report 2003/034

400

300

200

Discharge ML/Day 100

0

4 04 004 004 00 004 004 0 /2 /2 2 2 2 2 6 7 8/ 9/ 0/ 1/ /0 /0 /0 /0 /1 /1 1 01 01 01 01 0 01

Figure 7.5. Discharge of the Cotter River at Vanitys Crossing (approximately 3 km downstream of Spur Hole) prior to, during (grey shading) and post underwater video monitoring.

Camera Footage: Spur Hole Water surface glare prevented any penetration of the water column by the two above water cameras (cameras 1 and 6) at the Spur Hole site. The entire width of the channel could be seen, however, the benthos could not be seen owing to the glare from the water surface. A rippling water surface at the head of the pool also compromised the quality of footage obtained from camera 1. Overall, neither of the above water cameras (cameras 1 and 6) provided useful footage of aquatic biota for analysis (Table 7.5). No avian or mammalian predators were observed from processed footage.

The only observable record of an identifiable fish, was a two-spined blackfish Gadopsis bispinosus, filmed on underwater camera 2 at the Spur Hole during the post release recording (Table 7.5). Poor visibility due to high water turbidity rendered the other three underwater cameras unsuitable for analysis over the three sampling trips (Table 7.5). Visibility ranged from about 10 cm to about 50 cm (Table 7.5) and was insufficient for viewing the majority of the pool.

101 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Table 7.4. Evaluation of footage captured by each camera across all sampling periods at Burkes Creek Crossing on the Cotter River. CAMERA PRE-RELEASE RELEASE POST RELEASE CAMERA 1. Morning and afternoon NB. No Release session For 1400–1700: Too Above water glare off the surface of the was filmed for the much glare. Cannot see surface on large water. Unsuitable for Burkes Ck Crossing site. anything under the water tripod at the head analysis. surface. For 0800–1100: of the pool. Still poor contrast. Approx. 3m high Unlikely to be able to distinguish a fish unless it is large and at the surface. CAMERA 2. A small portion of the Some debris can be seen On the left bank benthos can be seen but no going past the camera, about 6m d/s of great detail/contrast. A leaf but there is no detail. No Camera 1. Small that went past the lens at a background (i.e. amount of flow distance of approx 20– substrate) can be seen, past this camera 30cm was in quite good just a grey screen. Not focus. For a fish to be suitable for quantifiable identified it would have to analysis. be in close proximity to the camera and be large. CAMERA 3. As for Camera 5. As for Camera 2. On the right bank about 4m d/s of Camera 2. Slow flow area CAMERA 4. As for Camera 5. As for Camera 2. On the left bank about 7m d/s of Camera 2. Slow flow area, no perceptible flow CAMERA 5. Debris can be seen floating As for Camera 2. On the Right past the lens. Little detail, bank about 10m just appears as light or d/s of Camera 3. dark coloured blurry Almost no flow shapes. Unsuitable for analysis. CAMERA 6. Morning and afternoon For 1400–1700: Water Above waters glare off the surface of the too deep and too much surface on large water. Unsuitable for glare to see anything tripod at the foot analysis. subsurface. Unsuitable of the pool. for analysis. For 0800– Approx. 3m high 1100: A little more visible but fish would not be able to be seen unless at the surface. Species recognition severely compromised.

An ecological approach to re-establishing Australian freshwater cod populations 102 FRDC Report 2003/034

Table 7.5. Evaluation of footage captured by each camera across all sampling periods at Spur Hole on the Cotter River. CAMERA PRE-RELEASE RELEASE POST RELEASE CAMERA 1. A combination of high Too much glare on Too much glare on Above water turbidity, fast flowing surface of water to see surface of water to see surface on large water and glare from anything subsurface. anything under the water. tripod at the head rippling surface water Contrast insufficient for Unsuitable for analysis. of the pool. makes viewing any analysis. Approx. 3m high objects subsurface impossible. Unsuitable for analysis. CAMERA 2. Some rocky benthos can Able to see some rock Fair visibility – approx. Most upstream be seen, although water structure but no detail 1m. By far the best underwater clarity is low. discernable. Unsuitable camera for visual camera. Flow near for analysis. analysis. Low numbers camera still of fish seen in 1hr 45min influenced by of footage (only 1 G. upstream riffle. bispinosus). On right bank CAMERA 3. No benthos can be seen. Poor contrast conditions. Very little physical Second Some debris floating Benthos cannot be seen, structure to be seen. underwater past the lens was visible nor any other substrate. Visibility very poor. camera. About 3– but seemed blurry. Unsuitable for analysis. Appears as a blank grey 4m d/s of camera Unsuitable for analysis. screen. Unsuitable for 2 and on the left analysis. bank CAMERA 4. As for Camera 3. A submerged stick can be A stick can be seen about Third underwater seen about 50cm away. 50cm away. Visibility camera. About 3m Some vague benthos can still quite poor. No fish d/s of camera 3 on be seen but detail is seen upon viewing this the left bank minimal. Contrast is poor camera footage as part of and unsuitable for 2x2. Unsuitable for analysis. analysis. CAMERA 5. Camera is mounted too As for Camera 3. Poor Poor visibility- only Most downstream shallow and the surface contrast, unsuitable. about 10cm max. Viewed underwater of the water can be seen. 1hr 45min and saw no camera. Low flow Surrounding water is fish. Not suitable for area on the right quite disturbed and analysis due to poor bank makes viewing anything visibility. difficult. Unsuitable for analysis. CAMERA 6. As for Camera 1. In the afternoon on the 3, Glare and poor contrast Above water 4th of Oct, streambed can limit the effectiveness of surface on large be seen in some parts. this camera. Although tripod at the foot This probably is some of some substrate on the of the pool. the better out of water streambed is visible, this Approx. 3m high camera footage captured, filming is unsuitable for but still not useful for analysis. analysis.

103 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

DISCUSSION

Underwater and above-water filming was unsuccessful in obtaining information of the feeding ecology and intra-specific interactions of M. macquariensis in the current study. This was a consequence of filming in low visibility conditions that resulted from sedimentation following a major bushfire event and rainfall in the Cotter River catchment. Satisfactory visibility had occurred in the Cotter River post-bushfire (Ebner, personal observation), however, filming was compromised by elevated turbidity associated with rainfall immediately prior to the filming period. Video monitoring was not repeated at a later date, due to all M. macquariensis leaving the Spur Hole before water clarity had reached an acceptable level. This highlights a deficiency of using underwater filming in freshwater systems prone to periodic sedimentation and associated high turbidity.

It would be useful to develop a protocol to determine when to film underwater, based on minimum thresholds of water clarity and visibility (Appendix 8). If observations are to be compared among time periods or sites, there is also a need to correct for reduced visual fields under minimum as opposed to ideal conditions in regard to visibility. In this regard it may be useful to establish a predetermined quadrat or standard volume from which a census of behaviour is recorded. Additionally, recent developments in the use of Dual-frequency Identification Sonar (DIDSONTM) provide means for obtaining images in high turbidity environments including lowland rivers (Baumgartner et al., 2006; Mueller et al., 2006).

Elevated turbidity was not the only factor restricting vision of aquatic fauna based on above-water filming. Glare from surface water is also likely to prevent successful filming in this regard. Polarising lenses may remove the effect of glare to an extent. While potential predators of M. macquariensis (see Chapter 6) were not observed in this study, above water filming may be useful in recording the activity of terrestrial and riparian predators following future re-introductions. For this exclusive purpose, above-water cameras could be positioned to observe a substantially larger spatial extent.

An ecological approach to re-establishing Australian freshwater cod populations 104 FRDC Report 2003/034

CHAPTER 8. FINAL DISCUSSION AND RECOMMENDATIONS

Brendan Ebner1, 2, Mark Lintermans1, 2, Jason Thiem1, 2 and Dean Gilligan 3

1Parks, Conservation and Lands, Department of Territory and Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia.

2Cooperative Research Centre for Freshwater Ecology, University of Canberra, ACT 2601, Australia.

3NSW Department of Primary Industries, Narrandera Fisheries Centre, PO Box 182, Narrandera NSW 2700, Australia.

Ebner, B., Lintermans, M., Thiem, J. & Gilligan, D. (2006). Final Discussion and Recommendations. In An ecological approach to re-establishing Australian freshwater cod populations: an application to trout cod in the Murrumbidgee catchment. (Ebner, B., Thiem, J., Lintermans, M. & Gilligan, D., eds), pp.105–115. Final Report to the Fisheries Research and Development Corporation (Project No. 2003/034). Canberra: Parks, Conservation & Lands.

THE THREE MODELS

An explanation for the apparent disappearance of sub-adult Maccullochella macquariensis following stocking as fingerlings at sites within the Murray–Darling Basin (Brown et al., 1998; Chapter 1), has not been completely resolved in the current study. However, a number of issues have been clarified. This includes new information regarding all three potential explanations for the absence of sub-adults that were distinguished in Chapter 1 (Figure 1.1): detection, dispersal and mortality.

Detection Maccullochella macquariensis were the second most common species obtained in the boat electro-fishing survey (described in Chapter 2). The survey also confirmed the suitability of this technique for detecting a wide size-range of M. macquariensis (Chapter 2, 3) as has been the experience of others (Brown & Nicol, 1998; J. Koehn, personal communication; Ian Wooden, unpublished data). This indicates that the species is likely to be detected by boat electro-fishing with the levels of effort normally applied to routine monitoring of this species in the Murrumbidgee River near Narrandera. In contrast, Brown (1998) found that sampling of M. macquariensis populations by electro-fishing and angling resulted in different size-classes of fish favoured by each method, with angling capturing relatively more of the larger size- classes. However it was also demonstrated in Chapter 2 that a much larger survey effort than is typically applied at a monitoring site, may be required to detect rare species at a site in a lowland river of the Murray–Darling Basin. For instance Maccullochella peelii peelii, considered a moderately common species in this river reach and probably of similar catchability to the congener M. macquariensis, was only recorded in seven of 157 electro-fishing operations. Similarly, there were very few M. p. peelii captured by electro-fishing in the pilot Sustainable Rivers Audit of four river valleys (MDBC, 2004a), indicating that electro-fishing may not be the most reliable sampling method for detecting this species in certain habitat conditions (e.g. waters >4 m depth). For this reason it must be emphasised that the failure to detect M. macquariensis at a site based on typical levels of survey effort does not necessarily represent the absence of this species from a river reach.

105 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Dispersal Maccullochella macquariensis exhibited a predominantly site-attached habit, based on radio-tracking of a substantial sample of adult and sub-adult wild M. macquariensis in the lowlands of the Murrumbidgee River for thirteen months in this study (Chapter 3). Similarly adult M. macquariensis are site-attached in the Murray River (Koehn, 1997; Brown & Nicol, 1998; Koehn & Nicol, 1998). An initial interpretation of this might be that dispersal of sub-adult M. macquariensis is unlikely to underlay their apparent disappearance from release sites. However it is also possible that the radio-tracking of this species to date (Koehn, 1997; Brown & Nicol, 1998; Koehn & Nicol, 1998, the current study) was based primarily on individuals that had already undertaken emigration prior to being radio-tagged. It should also be noted that the studies in the Murray River were based on truly wild individuals rather than those stocked as fingerlings, and do not represent a pure test of this model (i.e. Figure 1.1). Certainly wild M. macquariensis are capable of large-scale movement (the current study).

Larger fish from the wild group generally exhibited the greatest levels of movement recorded following initial homing in the current study. Their large-scale movements were explorations documented in Chapter 3. The only large-scale (>5 km) emigration (home-range shift) was undertaken by a small wild individual (752 g). It would be informative to develop a more comprehensive understanding of the frequency and scale of emigration and exploration of M. macquariensis within and outside the range of body-sizes examined in this study. In particular future radio-tracking of individuals in two developmental phases should be a priority. First, the emigration of late juvenile to early sub-adult phase M. macquariensis, corresponding to weights slightly less than those used in the current study (i.e. 100–500 g), should be the highest priority. This information is necessary to inform the dispersal model (Chapter 1, Figure 1.1). Second, the movement and specifically the exploration behaviour of large adults (e.g. >2 kg) is a secondary research priority, as indicated by the current study.

It would also be informative to examine the movement patterns of: (a) wild fish translocated from another site (thus avoiding the potential for naïve, hatchery-based behaviours to confound movement interpretations); and (b) to examine the movement behaviour of such translocated wild fish in the presence and absence of conspecifics. This would provide further insights into how behaviour of stocked fish may be modified by either previous experience (scenario a) or the presence of conspecifics (scenario b). Additionally, high levels of, or complete mortality, of two-year old, hatchery M. macquariensis at three release sites in the current study and a single site in a pilot study (Ebner, Johnston & Lintermans, unpublished data) contrast with apparently high survivorship based on radio-tracking of wild fish (Koehn, 1997; Brown & Nicol, 1998; Koehn & Nicol, 1998; the current study). This demonstrates the unsuitability of using hatchery-reared individuals as a surrogate for wild fish in field experiments.

Mortality Mortality of sub-adult and adult M. macquariensis is unlikely to be a major factor in the Murrumbidgee River at Narrandera based on high survivorship of the wild group in experiment 1 (Chapter 3). This confirms what was considered most likely at that site prior to beginning the current study, based on angler reports (Stephen Thurstan, New South Wales Department of Primary Industries, personal communication) and survey results (Growns et al., 2004; Gilligan, 2005b). However the river reach along

An ecological approach to re-establishing Australian freshwater cod populations 106 FRDC Report 2003/034 the Murrumbidgee River at Narrandera represents the most successful re-introduction site for M. macquariensis based on stocking of fingerlings in New South Wales (NSW) (Gilligan, 2005b). Subsequently, low adult mortality at this site may be atypical of sites stocked with fingerling M. macquariensis. Therefore this information has limited bearing on the mortality-based model (Figure 1.1). Future monitoring of survivorship based on radio-tracking of samples of 1+ and 2+ fish, across sites apparently experiencing a range of survivorship patterns represents a potential way forward for testing the mortality-based model (Figure 1.1).

This study identified a number of potential predators of sub-adult and small adult sized M. macquariensis (Chapter 6). However this result must be viewed in the light of the demonstration that two-year old hatchery fish are not a sound surrogate for wild fish in tests of the models proposed in Chapter 1. In particular the second experiment demonstrated an impact from avian predators, notably great cormorants. Apparently, wild fish possess means of evading these predators. Whether this advantage comes from experience and/or has a genetic basis (Metcalfe et al., 2003; Vilhunen & Hirvonen, 2003) is yet to be determined. It indicates a need for improved information on the survivorship of M. macquariensis at stocking sites, identifying vulnerable life- history phases on a site-by-site basis (cf. Bar-David et al., 2005).

Patterns of survivorship The results of the current study demonstrate the failure of two-year old, hatchery on- grown M. macquariensis to survive at three riverine sites (Figure 8.1). Similarly, a pilot study based on a smaller sample size (n=9) revealed complete mortality (Figure 8.1). Trends in the rate of survivorship among these releases and their possible causation, was discussed in Chapter 6. The common finding of high mortality in all of these trials provides the impetus for discussing future directions for conservation stocking of Maccullochella.

107 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Alive Confirmed mortality or rejection Suspected mortality or rejection Failed radio-transmitter 35 30 25 a) 20 15 10 5 0 30 25 b) 20 15 10 5 0 40 35 c) 30 25 20 15 10 5 0 40 35 d) 30 Number of individuals 25 20 15 10 5 0 10 8 e) 6 4 2 0 12345678910111213 Time elapsed (months) Figure 8.1. Monthly survivorship of M. macquariensis based on radio-tracking following release of: (a) wild fish (n=31); and (b) hatchery fish (n=27) in the Murrumbidgee River near Narrandera; (c) hatchery fish (n=36) in the Murrumbidgee River, ACT; (d) hatchery fish (n=36) in the Cotter River, ACT, in the current study; and (e) hatchery fish (n=9) in the Cotter River, ACT (from Ebner, Johnston & Lintermans, unpublished data).

An ecological approach to re-establishing Australian freshwater cod populations 108 FRDC Report 2003/034

CONSERVATION STOCKING OF MACCULLOCHELLA

Pre-release phase There is scope for pre-release training of Maccullochella comparable to that implemented in other species recovery programs (Wallace, 2000). Advances have been made in terms of hatchery-based training of fish in terms of feeding on live prey (Brown & Day, 2002; Thorpe, 2004), predator evasion (Vilhunen et al., 2005) and habitat use (Salvenes & Braithwaite, 2005). Demonstrator fish can be used to train greater numbers of individuals (Vilhunen et al., 2005) and genetic screening can be used to select individuals likely to benefit from training (Vilhunen & Hirvonen, 2003). For instance a genetic basis for innate predator evasion has been recorded in a study of Arctic char, Salvinus alpinus (Vilhunen & Hirvonen, 2003).

The current study has explored the potential for releasing on-grown M. macquariensis. It has clearly shown that substantially increasing the size and age of individuals at release does not ensure high survival. However, it does not mark the end for conservation stocking of on-grown Maccullochella. Theoretically, the body size of individuals released in this study was at the upper limit of being vulnerable to avian predators and common water rats, though not M. peelii peelii (Ebner, in press). Further increases in the body size of M. macquariensis at release (e.g. stocking >2 year olds) represent a strategy worthy of testing, especially at upland sites beyond the distribution of M. peelii peelii (a potential predator and competitor) in the future. Alternatively, it may be beneficial to investigate releasing on-grown fish of intermediate in size to fingerlings and adults (e.g. 1 and 2 year olds) in concert with pre-release training.

Whilst in the context of radio-tracking studies, moderately large sample sizes (~30 per site) were used in the current study; the number of individuals released was orders of magnitude less than those used in standard releases of fingerlings (Bearlin et al., 2002; Todd et al., 2004; Gilligan, 2005b). For instance the fact that two hatchery individuals out of twenty-seven survived a year in the Murrumbidgee River at Narrandera may not be insignificant in contrast to the annual survivorship of fingerlings stocked into rivers. Survivorship of fingerlings at stocking sites is not easily achieved and represents a major knowledge gap in the recovery of Maccullochella species.

Post-release phase Low survival rates of raised in captivity following release into the wild, including release of hatchery-reared fishes, has often been attributed to insufficient life-skills (Wallace, 2000). To overcome this deficiency, pre-release training (discussed above) and post-release care have been invoked in species recovery programs (Wallace, 2000). Post-release care has also included familiarising animals to the local environment in to which they are to be released, a process known as soft- release (Wallace, 2000). This approach is warranted in the case of Maccullochella spp. as indicated by the rapid mortality of hatchery-reared M. macquariensis following release at three sites in the current study, and since hard releases have totally dominated re-introductions and stockings of the Maccullochella to date.

Thought has been put into selecting sites for Maccullochella spp. primarily centring on their past and present distribution, habitat requirements, avoiding predators

109 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

(primarily introduced Perca fluviatilis and anglers) and reducing the chances of hybridisation between M. macquariensis and M. peelii peelii (Douglas et al., 1995; Simpson & Jackson, 1996; Brown & Nicol, 1998; Brown et al., 1998). Activities to better prepare release sites are currently absent from strategies for releasing Maccullochella, although, the issue is mentioned in the broad sense of improving habitat for species in some cases (e.g. Simpson & Jackson, 1996). Local conditions including habitat (e.g. Simpson et al., 2002), predators (Molony et al., 2004) and competitors (Levin & Williams, 2002; Jokikokko & Jutila, 2004) have been shown to have a bearing on the stocking success of other taxa. Accordingly habitat has been manipulated as part of the species reintroduction process in a subset of cases (e.g. Hutchison et al., 2006). Specifically in regard to reintroducing Maccullochella spp., the issue of predators warrants investigation based on the pattern of mortality observed in the current study (Figure 8.1). This may include: a) manipulation of predators in the initial period of and following release (as suggested by Thorfve, 2002), b) selection of sites with high habitat complexity or addition of habitat to mediate predation (Hutchison et al., 2006), and/or c) pre-release anti-predator training (Kelley & Magurran, 2003; Vilhunen et al., 2005; Vilhunen, 2006). Future recovery plans for Maccullochella spp. should progress to include development of improved short-term release practices. Certainly there is scope for improving releases of hatchery produced Maccullochella through experimenting with variations in soft release (see Wallace, 2000; Molony et al., 2004; Thorfve, 2002). This includes allowing fingerlings an opportunity to recover from transportation and acclimate to environmental conditions prior to release (Wallace, 2000; Thorfve, 2002) and a protocol for inspecting sites to assess their current suitability in the weeks leading up to release.

Timing of release (Cowx, 1994) is an issue that warrants consideration in Maccullochella re-introductions of fingerlings and on-grown fish. For instance avian piscivores may operate at specific feeding grounds according to seasonal cycles or annual events (Marchant & Higgins, 1990). Ebner et al. (2006) recognised that the most vulnerable period for adult was during spawning migration (October–November) from Cotter Reservoir to the Cotter River because of great cormorant predation. The timing of releases of Maccullochella fingerlings is currently driven by hatchery production schedules. For instance fertilised eggs of M. macquariensis and M. peelii peelii are harvested from earthen ponds in spring and early summer and fingerlings are grown-out for release in late summer (Stephen Thurstan, New South Wales Department of Primary Industries, personal communication). High zooplankton productivity in ponds provides opportunity for accelerated growth. However, upon releasing fingerlings into rivers, prey availability becomes less certain (e.g. Shiel et al., 1982). Management options include releasing fingerlings at optimum periods based on monitoring of zooplankton abundance in the river, or on-growing fingerlings to a size at which they can access alternative prey sources (e.g. macro-invertebrates and small fishes).

Future efforts to determine the optimal size or age of Maccullochella reintroduction would also benefit from addressing the issue of imprinting. Animals develop certain abilities very early in life, for instance, salmonids imprint on their nursery habitat based on olfactory cues prior to or soon after hatching (Candy & Beacham, 2000). Equivalent studies on Maccullochella should inform strategies for conservation stockings.

An ecological approach to re-establishing Australian freshwater cod populations 110 FRDC Report 2003/034

RECOMMENDATIONS

This study has demonstrated that three separate releases of two-year old hatchery- reared M. macquariensis at riverine sites were unsuccessful. However, considerable progress has been made by implementing radio-tracking to quantify survivorship and identify timing and in some case likely causes of mortality. This method is to be recommended for future monitoring of Maccullochella releases.

The study and its pilot project (Ebner et al., 2005) have also demonstrated a niche for conducting research specific to monitoring re-introduction of a species of Maccullochella. This builds on hatchery-based knowledge and ecological studies of Maccullochella (e.g. Harris & Rowland, 1996; Lintermans & Phillips, 2005) including radio-telemetry study of resident populations (Koehn, 1997; Koehn & Nicol, 1998; Brown & Nicol, 1998; Butler, 2001; Simpson & Mapleston, 2002; Stuart & Jones, 2002) and in particular the recent development of models for M. macquariensis recovery (Bearlin et al., 2002; Todd et al., 2004). These models have focussed attention on the issue of how to best allocate fingerling stockings in space and time. The current study champions the need to monitor survivorship and describe the mechanisms underpinning survivorship following reintroductions of Maccullochella whether at the fingerling or more advanced stages. To this end, the conceptual model presented in Chapter 5 should aid future comparisons of wild and hatchery released Maccullochella based on research and monitoring.

In Chapters 3, 5 and 8, a lack of scientific understanding of the experience that hatchery reared Maccullochella have or require in rivers, emerged as a theme. Another theme that has arisen from Chapter 3 and 6 and the pilot study (Ebner et al., 2005) is that the density of released individuals could be improved for monitoring, research, compliance and other management purposes, by stocking into small contained river reaches or reservoirs. In this regard, the recovery of Maccullochella should benefit from two initiatives. First, a small number of sites or a single site could be used as a model system of long-term Maccullochella recovery. This would involve consolidation of resources (outlined above) and enable the long-term monitoring of individuals through juvenile, sub-adult and adult phases. A site-focussed approach would also make a number of adaptive management options more feasible. For instance the reduction of predator numbers and the enhancement of habitat could be conducted at levels that are capable of invoking population level changes. Furthermore these changes have the potential to be detected if the density of Maccullochella sp. is sufficient due to the containment of stockings in a small area. Second, there needs to be better integration between the hatchery science and the field ecology components of Maccullochella recovery. This had become particularly evident by the end of experiment 2 (Chapter 6) and relates to a number of issues discussed previously in the current chapter.

111 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

BENEFITS AND ADOPTION

This study demonstrated that stocking of on-grown cod is not necessarily a straightforward solution or an immediate alternative to fingerling stocking. In this regard the study does not directly advocate change to the status quo of widespread stocking of cod fingerlings as a means of enhancing or recovering cod populations in South Australia (SA), Victoria, NSW, the Australian Capital Territory (ACT) and Queensland (Qld). Evaluation of the effectiveness of stocking cod fingerlings has been based on localised monitoring (Faragher et al., 1993; Lintermans, 1995; Harris, 2003; Todd et al., 2004; Hutchison et al., 2006; Lintermans, 2006) or modelling (Bearlin et al., 2002; Todd et al., 2004) and combined with the current study of on- grown cod, illustrates a major gap in the effective management of Australian freshwater cod populations. The widespread lack of scientific evaluation of the effectiveness of stocking, including stocking of fingerlings and on-grown cod is a significant deficiency. There have been in excess of 50 million fingerlings of a variety of native species stocked into the Murray–Darling Basin (Gillanders et al., 2006), and yet there has been no peer-reviewed publication on the effectiveness or otherwise of such programs.

Additionally, from a pragmatic perspective the sustainable management of Australian freshwater cod populations requires an effective means of monitoring cod populations (at low abundance or with low detectability) in large rivers. Without this the effectiveness of management strategies involving re-stocking, angling or habitat modification cannot be evaluated. The current study identifies a number of problems surrounding the issue of quantifying cod abundance (Chapter 2). Through application of substantial electro-fishing effort at a site and statistical techniques based on North American studies, the current study provides to state government agencies a means for improving existing practices of monitoring cod in large rivers (Chapter 2).

Finally, the utility of radio-tracking was demonstrated as a highly suitable and cost effective technique for monitoring reintroductions of cod. Radio-telemetry provided insights into both fine and large-scale aspects of the movement ecology and mortality of this species that would not have been obtainable through other sampling or tagging techniques.

An ecological approach to re-establishing Australian freshwater cod populations 112 FRDC Report 2003/034

FURTHER DEVELOPMENT

Two ways for progressing the re-establishment or enhancement of sustainable cod populations are apparent at the conclusion of the current project. The first is a call for major improvement in the assessment of cod populations and cod stockings into remnant populations or unpopulated river reaches. This will require substantial investment into development of methods that quantify cod populations. In this regard we support the need for life-history phase specific sampling techniques (e.g. Humphries & King, 2003) and the need to develop or refine techniques for effectively and reliably estimating the abundance of large, adult cod. Currently, boat electro- fishing is the most widespread technique utilised, but limitations imposed by deep, highly turbid waters, and a lack of knowledge of the detection capacity (or performance) of standard electro-fishing procedures on a range of cod life stages severely hampers cod assessment. Perhaps most importantly if we are to have ecologically sustainable cod populations, there is a need to quantify adult cod abundance, estimate population sizes and age-structures, quantify mortality and emigration/immigration, and relate these to the relative abundance estimates obtained by rapid assessment techniques (Lintermans & Phillips, 2005; Bearlin et al., 2002; Todd et al., 2004). Such information will allow fisheries managers to manage cod populations effectively.

The second is the need for a comprehensive reintroduction program for Australian freshwater cod. Key elements of this program include: a) integration of hatchery production and field-based monitoring and research; b) an adaptive experimental management approach which relies on reintroduction strategies that are evaluated by field-based monitoring; and c) cross-fertilisation, synthesis and cooperation across programs relating to the different cod species. To this end there is a need for a scoping study to formulate a planned way forward. This should involve a forum to receive input from international experts in freshwater fish reintroduction and those with local relevant technical expertise including government hatchery staff, fish ecologists, modellers and managers. It is also likely that university researchers and organisations that operate across state boundaries will be instrumental in successful species reintroduction programs (Wallace, 2000).

Finally, the simultaneous development of improved monitoring techniques, capabilities and aspects of reintroduction ecology is necessary in addressing the short- term goal of developing an effective means of re-establishing adult numbers of Australian freshwater cod to detectable levels at sites. To achieve this goal it would be of benefit to concentrate re-introduction efforts at a limited number of sites. Investigation of techniques or approaches that minimise the deficits associated with hatchery rearing of adult fish should also be encouraged. Alternative approaches such as the use of translocations, soft-release or other conditioning techniques prior to release could be usefully explored. Furthermore, the potential dispersal of cod on a medium-large-scale (tens to hundreds of kilometres) (Koehn, 1997; Simpson & Mapleston, 2002; this study) could be minimised, and detection and monitoring made more efficient by conducting stocking at sites that are confined by instream barriers. It also has the advantage of consolidating monitoring, research and compliance resources (Ebner et al., 2005). The establishment of successful self-recruiting populations represents the second step in the recovery process.

113 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

PLANNED OUTCOMES

This study has not demonstrated a successful alternative strategy (to fingerling stocking) for re-establishing M. macquariensis into rivers despite three releases of on- grown fish. Understanding how to re-establish adult Australian freshwater cod in the wild, remains an elusive goal. Nevertheless the study has reinforced the need for a sound scientific understanding of species ecology in reintroducing cod species. Specifically, predators (and the implication of lower fitness for hatchery derived adult individuals) have been identified as an important factor in poor post-release survivorship of hatchery-reared cod.

The project provides an empirically derived evaluation of stocked on-grown cod species. This report informs managers of cod species in SA, Victoria, NSW, the ACT and Qld that on-growing cod for release is not a straightforward alternative to the current strategy of stocking fingerlings. The study also serves to encourage state fisheries managers to rigorously evaluate the conservation outcomes (cf. outcomes) of stocking cod species within their jurisdiction.

This project contributes toward the desired outcome of re-establishing Australian freshwater cod species in eastern Australia. This empirical evaluation of releases of on-grown fish complements recent efforts to model reintroduction of cod (Bearlin et al., 2002; Todd et al., 2004) in providing a basis for recovering cod beyond an approach largely based on stocking fingerlings. In doing so, efforts to recover Australian freshwater cod are becoming aligned with the long-standing fish reintroduction programs conducted in the Northern Hemisphere, such as those centred on sturgeons (e.g. Ireland et al., 2002) and salmonids (e.g. Aarestrup et al., 2000; Thorfve, 2002).

An ecological approach to re-establishing Australian freshwater cod populations 114 FRDC Report 2003/034

CONCLUSION

The re-establishment of Australian freshwater cod populations based on release of on- grown fish is not an immediate and straightforward alternative to the existing practice of stocking fingerlings into rivers. In particular mortality, including that caused by avian predators, represents a potential impediment to re-introductions of hatchery- reared cod. In this regard, radio-tracking provided a means for maximising information pertaining to the dispersal and survivorship of individuals following reintroduction in the current study. It would therefore be useful to monitor further reintroductions in this way and set within an adaptive experimental management framework.

Monitoring of a sample of adult and sub-adult M. macquariensis that had been originally stocked as fingerlings, revealed the ability of this species to exist in small reaches of river on outer bends for extended periods, sometimes punctuated by large- scale movement. The ability of this species to function entirely within small areas for long periods indicates the potential to establish viable populations within small reaches of river. Efforts to rehabilitate habitat for this species, at least in the lowlands of the Murrumbidgee River, may be maximised by focussing specifically on outer bends. This provides an example of the usefulness of targeted research based on radio-tracking at appropriate temporal scales, and collection of known-status-data in relation to threatened Australian freshwater cod.

115 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

REFERENCES

Aarestrup, K. & Jepsen, N. (1998). Spawning migration of sea trout (Salmo trutta (L)) in a Danish river. Hydrobiologia 371/372, 275–281.

Aarestrup, K., Jepsen, N., Koed, A. & Pedersen, S. (2005). Movement and mortality of stocked brown trout in a stream. Journal of Fish Biology 66, 721–728.

Aarestrup, K., Jepsen, N., Rasmussen, G., Økland, F., Thorstad, E. B. & Holdensgaard, G. (2000). Prespawning migratory behaviour and spawning success of sea-ranched Atlantic salmon, Salmo salar L., in the River Gudenaa, Denmark. and Ecology 7, 387–400.

ACT Government (1999). Trout Cod (Maccullochela macquariensis): An endangered species. Action Plan No. 12. Canberra: Environment ACT.

ACT Government (2006). Draft aquatic species and riparian zone conservation strategy. Canberra: Arts, Heritage and Environment.

Allen, G. R., Midgley, S. H. & Allen, M. (2002). Field guide to the freshwater fishes of Australia. Perth: Western Australian Museum.

Anderson, J. R., Morison, A. K. & Ray, D. J. (1992a). Age and growth of Murray Cod, Maccullochella peelii (: Percichthyidae), in the lower Murray– Darling Basin, Australia. Australian Journal of Marine and Freshwater Research 43, 983–1013.

Anderson, J. R., Morison, A. K. & Ray, D. J. (1992b). Validation of the use of thin- sectioned otoliths for determining the age and growth of golden perch, Macquaria ambigua (Perciformes: Percichthyidae), in the lower Murray–Darling Basin, Australia. Australian Journal of Marine and Freshwater Research. 43, 1103–1128.

Armstrong, J. D., Braithwaite, V. A. & Fox, M. (1998). The response of wild Atlantic salmon parr to acute reductions in water flow. Animal Ecology 67, 292–297.

Armstrong, J. D., Huntingford, F. A. & Herbert, N. A. (1999). Individual space use strategies of wild juvenile Atlantic salmon. Journal of Fish Biology 55, 1201–1212.

Arthington, A. H. & Marshall, C. J. (1996). Threatened fishes of the world: Nannoperca oxleyana Whitley, 1940 (Nannopercidae). Environmental Biology of Fishes 46, 150.

Australian Society for Fish Biology (2004). Conservation status of Australian fishes 2004. Australian Society for Fish Biology Newsletter 34, 59–64.

Baker, R. R. (1978). The evolutionary ecology of animal migration. London: Hodder and Stoughton.

Bar-David, S., Saltz, D., Dayan, T., Perelberg, A. & Dolev, A. (2005). Demographic models and reality in reintroductions: Persian fallow deer in Israel. Conservation Biology 19, 131–138.

An ecological approach to re-establishing Australian freshwater cod populations 116 FRDC Report 2003/034

Baras, E. (1998). Selection of optimal positioning intervals in fish tracking: an experimental study on Barbus barbus. Hydrobiologia 371/372, 19–28.

Baras, E., Jeandrain, D., Serouge, B. & Philippart, J. C. (1998). Seasonal variations in time and space utilization by radio-tagged yellow eels Anguilla anguilla (L.) in a small stream. Hydrobiologia 371/372, 187–198.

Battaglene, S. C. (1991). The golden perch, Macquaria ambigua (Pices: Percichthyidae) of Lake Keepit, NSW. M. Sc. Thesis, University of New South Wales.

Baumgartner, L. J. (in press). Diet and feeding habits of predatory fishes upstream and downstream of a low-level weir. Journal of Fish Biology

Baumgartner, L. J., Reynoldson, N., Cameron, L. & Stanger, J. (2006). Assessment of a Dual-frequency Identification Sonar (DIDSON) for application in fish migration studies. Fisheries Final Report Series No. 84. Cronulla: NSW Department of Primary Industries.

Bawden, T. (1888). Personal recollections of early days in the Clarence district the stations and all about them. In The Bawden Lectures. June 1886, July 1886, August 1888. The First Fifty Years of Settlement on the Clarence. pp. 88–182. Canberra: R. C. Law. Australia: National Library of Australia.

Bearlin, A. R., Schreiber, E. S. G., Nicol, S. J., Starfield, A. M. & Todd, C. R. (2002). Identifying the weakest link: simulating adaptive management of the reintroduction of a threatened fish. Canadian Journal of Fisheries and Aquatic Science 559, 1709– 1716.

Bennetts, R. E., Nichols, J. D., Lebreton, J., Pradel, R., Hines, J. E. & Kitchens, W. M. (2001). Methods for estimating dispersal probabilities and related parameters using marked animals. In Dispersal (Clobert, J., Dunchin, E., Dhondt, A. A. & Nichols, J. D., eds.), pp. 3–17. New York: Oxford University Press.

Berland, G., Nickelsen, T., Heggenes, J., Økland, F., Thorstad, E. B. & Halleraker, J. (2004). Movements of wild Atlantic salmon parr in relation to peaking flows below a hydropower station. River Research and Applications 20, 957–966.

Berra, T. M. (1974). The Trout Cod, Maccullochella macquariensis a rare freshwater fish of eastern Australia. Biological Conservation 6, 53–56.

Berra, T. M. & Weatherley, A. H. (1972). A systematic study of the Australian freshwater serranid fish genus Maccullochella. Copeia 1, 53–64.

Bettinger, J. M. & Bettoli, P. W. (2002). Fate, dispersal and persistence of recently stocked and resident Rainbow Trout in a Tennessee tailwater. North American Journal of Fisheries Management 22, 425–432.

Beumer, J. (1980). Fish ladders: steps in the right direction. Wildlife in Australia 17, 38–39.

Beumer, J. P. & Harrington, D. J. (1982). A preliminary study of movement of fishes through a Victorian (Lerderderg River) fish-ladder. Proceedings of the Royal Society

117 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034 of Victoria 94, 121–132.

Blundell, G. M., Maier, J. A. K. & Debevec, E. M. (2001). Linear home ranges: effects of smoothing, sample size, and autocorrelation on Kernel estimates. Ecological monographs 71, 469–489.

Boys, C. A., Esslemont, G. & Thoms, M. C. (2005). Fish habitat assessment and protection in the Barwon–Darling and Paroo Rivers. Fisheries Final Report Series No. 78. Cronulla: NSW Department of Primary Industries.

Braithwaite, V. A. & Burt de Perera, T. (2006). Short-range orientation in fish: How fish map space. Marine and Freshwater Behavior and Physiology 39, 37–47.

Braithwaite, V. A. & Girvan, J. R. (2003). Use of water flow direction in a small-scale orientation task. Journal of Fish Biology. 63, 74–83.

Bray, J. R. & Curtis, J. T. (1957). An ordination of the upland forest communities of Southern Wisconsin. Ecological Monographs 27, 325–349.

Brooker, L. & Brooker, M. (2002). Dispersal and population dynamics of the blue- breasted fairy-wren, Malurus pulcherrimus, in a fragmented habitat in the Western Australian wheatbelt. Wildlife Research 29, 225–233.

Brown, A. & Nicol, S. (1998). Trout Cod recovery plan (draft). Melbourne: Department of Natural Resources and Environment.

Brown, A. M., Nicol, S. & Koehn, J. D. (1998). Trout cod recovery plan- major project review report. Melbourne: Department of Natural Resources and Environment.

Brown, C. & Day, R. (2002). The future of stock enhancements: lessons for hatchery practice from conservation biology. Fish and Fisheries 3, 79–94.

Brown, C. & Laland, K. N. (2003). Social learning in fishes: a review. Fish and Fisheries 4, 280–288.

Brown, C. & Warburton, K. (1997). Predator recognition and anti-predator responses in the rainbowfish Melanotaenia eachamensis. Behavioural Ecology and Sociobiology 41, 61–68.

Brown, P. (1998). Goulburn River Trout Cod Stock Assessment, November 1997: growth and Mortality estimates. MAFRI Freshwater Division Fish Survey 97/15. Snobs Creek, Victoria: Marine and Freshwater Resources Institute.

Burnham, K. P. & Overton, W. S. (1979). Robust estimation of population size when capture probabilities vary among animals. Ecology 60, 927–936.

Butler, G. (2001). Age, growth and telemetric tracking of the eastern freshwater cod, Maccullochella ikei (Pisces: Percichthyidae) within the Mann–Nymboida River System, NSW. BSc (Hons) thesis, Southern Cross University, Australia.

Cadwallader, P. L. (1977). J. O. Langtry's 1949–50 Murray River investigations:

An ecological approach to re-establishing Australian freshwater cod populations 118 FRDC Report 2003/034

Fisheries and Wildlife Paper. Victoria.

Cadwallader, P. L. (1979). Distribution of native and introduced fish in Seven Creeks River system. Australian Journal of Ecology 4, 361–385.

Cadwallader, P. L. & Gooley, G. J. (1984). Past and present distributions and translocations of Murray Cod Maccullochella peeli and Trout Cod M. macquariensis (Pisces: Percicthyidae) in Victoria. Proceedings of the Royal Society of Victoria 96, 33–43.

Cadwallader, P. L. & Gooley, G. J. (1985). Propagation and Rearing of Murray Cod Maccullochella peeli at the Warmwater Fisheries Station Pilot Lake Charlegrark. Melbourne: Government Printer.

Cambray, J. A. (1997). The spawning behaviour of the endangered Eastern Cape rocky, Scandalia bainsii (Anabantidae), in South Africa. Environmental Biology of Fishes 49, 293–306.

Candy, J. R. & Beacham, T. D. (2000). Patterns of homing and straying in southern British Columbia coded-wire tagged chinook salmon (Oncorhynchus tshawytscha) populations. Fisheries Research 47, 41–56.

Cao, Y., Larsen, D. P. & Hughes, R. M. (2001). Evaluating sampling sufficiency in fish assemblage surveys: a similarity based approach. Canadian Journal of Fisheries and Aquatic Science 58, 1782–1793.

Cao, Y., Larsen, D. P. & White, D. (2004). Estimating regional species richness using a limited number of survey units. Ecoscience 11, 23–35.

Cao, Y., Williams, D. D. & Larsen, D. P. (2002). Comparison of ecological communities: the problem of sample representativeness. Ecological Monographs 72, 41–56.

Cappo, M., Speare, P. & De'ath, G. (2004). Comparison of the baited remote underwater video stations (BRUVS) and () trawls for assessments of fish biodiversity in inter-reefal areas of the Great Barrier Reef Marine Park. Journal of Experimental Marine Biology and Ecology 302, 123–152.

Carey, A., Evans, M., Hann, P., Lintermans, M., Macdonald, T., Ormay, P., Sharp, S., Shorthouse, D. & Webb, N. (2003). Wildfires in the ACT 2003: report on the impacts in natural ecosystems. Technical Report 17. Canberra: Environment ACT.

Chao, A. (1987). Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43, 783–791.

Chao, A., Chazdon, R. L., Colwell, R. K. & Shen, T. (2005). A new statistical approach for assessing similarity of species composition with incidence and abundance data. Ecology Letters 8, 148–159.

Chessman, B. (2003). Integrated Monitoring of Environmental Flows: State Summary Report 1998–2000. Sydney: NSW Department of Infrastructure, Planning and Natural Resources.

119 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Clark, M. E. & Rose, K. E. (1997). Individual-based model of stream-resident rainbow trout and brook char: model description, corroboration, and effects of sympatry and spawning season duration. Ecological Modelling 94, 157–175.

Clobert, J., Danchin, E., Dhondt, A. A. & Nichols, J. D. (Eds.) (2001). Dispersal. New York: Oxford University Press.

Clunie, P. & Koehn, J. (2001). Silver perch: a resource document Final report for Natural Resource Management Strategy Project R7002 to the Murray–Darling Basin Commission.

Collins, N. C. & Hinch, S. G. (1993). Diel and seasonal variation in foraging activities of pumpkin seeds in an Ontario pond. Transactions of the American Fisheries Society 122, 357–365.

Comeau, S. & Boisclair, D. (1998). Day-to-day variation in fish horizontal migration and its potential consequence on estimates of trophic interactions in lakes. Fisheries Research 35, 75–81.

Cooke, S. J., Bunt, C. M. & Schreer, J. F. (2004). Understanding fish behavior, distribution, and survival in thermal effluents using fixed telemetry arrays: a case study of Smallmouth Bass in a discharge canal during winter. Environmental Management 33, 140–150.

Cooke, S. J., Crossin, G. T., Patterson, D. A., English, K. K., Hinch, S. G., Young, J. L., Alexander, R. F., Healey, M. C., Van Der Kraak, G. & Farrell, A. P. (2005). Coupling non-invasive physiological assessments with telemetry to understand inter- individual variation in behaviour and survivorship of Sockeye Salmon: development and validation of technique. Journal of Fish Biology 67, 1342–1358.

Cooke, S. J., Hinch, S. G., Wikelski, M., Andrews, R. D., Kuchel, L. J., Wolcott, T. G. & Butler, P. J. (2004). Biotelemetry: a mechanistic approach to ecology. Trends in Ecology and Evolution 19, 334–343.

Coutin, P. C. & Reside, J. (2003). Fish predation by great cormorants, Phalacrocorax carbo carboides (Gould 1838) in the Gippsland Lakes, south-eastern Australia. In Interactions Between Fish and Birds: Implications for Management (Cowx, I. G., ed.), pp. 196–211. Oxford: Blackwell Science.

Cowx, I. G. (1994). Stocking strategies. Fisheries Management and Ecology 1, 15– 130.

Cowx, I. G. (1998). Stocking and Introductions of Fish. Oxford: Fishing News Books, Blackwell Science.

Crook, D. A. (2004a). Is the home range concept compatible with the movements of two species of lowland river fish? Journal of Animal Ecology 73, 353–366.

Crook, D. A. (2004b). Movements associated with home-range establishment by two species of lowland river fish. Canadian Journal of Fisheries and Aquatic Science 61, 2183–2193.

An ecological approach to re-establishing Australian freshwater cod populations 120 FRDC Report 2003/034

Crook, D. A., Robertson, A. I., King, A. J. & Humphires, P. (2001). The influence of spatial scale and habitat arrangement on diel patterns of habitat use by two lowland river fishes. Oecologia 129, 525–533.

David, B. O. & Closs, G. P. (2001). Continuous remote monitoring of fish activity with restricted home ranges using radiotelemetry. Journal of Fish Biology 59, 705– 715.

David, B. O. & Closs, G. P. (2002). Behaviour of a stream-dwelling fish before, during, and after high-discharge events. Transactions of the American Fisheries Society 131, 762–771.

David, B. O. & Closs, G. P. (2003). Seasonal variation in diel activity and microhabitat use of an endemic New Zealand stream-dwelling galaxiid fish. Freshwater Biology 48, 1765–1781.

Dawson, T. J. (1995). Kangaroos: biology of the largest marsupials. Sydney: University of New South Wales Press.

Derby, C. E. & Lovvorn, J. R. (1997). Predation on fish by cormorants and pelicans in a cold-water river: a field and modeling study. Canadian Journal of Fisheries and Aquatic Science 54, 1480–1493.

Dieperink, C., Pedersen, S. & Pedersen, M. I. (2001). Estuarine predation on radiotagged wild and domesticated sea trout (Salmo trutta L.) smolts. Ecology of Freshwater Fish 10, 177–183.

Dittman, A. H. & Quinn T. P. (1996). Homing behaviour in pacific salmon: mechanisms and ecological basis. The Journal of Experimental Biology 199, 83–91.

Douglas, J. W. & Brown, P. (2000). Notes on the successful spawning and recruitment of a stocked population of the endangered Australian freshwater fish, Trout cod, Maccullochella macquariensis (Cuvier) (Percichthyidae). Proceedings of the Linnaean Society of New South Wales 122, 143–147.

Douglas, J. W., Gooley, G. J. & Ingram, B. I. (1994). Trout Cod, Maccullochella macquariensis (Cuvier) (Pisces: Percicthyidae) resource handbook and research and recovery plan. Victoria: Department of Conservation and Natural Resources.

Douglas, J. W., Gooley, G. J., Ingram, B. A., Murray, N. D. & Brown, L. D. (1995). Natural hybridization between Murray cod, Maccullochella peelii peelii (Mitchell), and trout cod, Maccullochella macquariensis (Cuvier) (Percichthyidae), in the Murray River, Australia. Marine and Freshwater Research 46, 729–734.

DSE (Victorian Department of Sustainability and Environment) (2003). Advisory List of Threatened Vertebrate Fauna in Victoria. Department of Sustainability and Environment, Melbourne: .

Ebeling, A. W. & Hixon, M. A. (1991). Tropical and temperate reef fishes: comparison of community structures. In The Ecology of Fishes on Coral Reefs (Sale, P. F., ed.), pp. 509–563. Calafornia: Academic Press Inc.

121 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Ebenman, B. & Persson, L. (eds.) (1988). Size-structured Populations: Ecology and Evolution. Berlin: Springer-Verlag.

Ebner, B. (in press). Murray cod an apex predator in the Murray River, Australia. Ecology of Freshwater Fish

Ebner, B., Johnston, L. & Lintermans, M. (2005). Re-introduction of Trout Cod into the Cotter River Catchment. Final Report to the Natural Heritage Trust. Canberra: ACT Government.

Ebner, B., Lintermans, M., Godschalx, S. & Jekabsons, M. (2006). Fish passage, movement requirements and habitat use for Macquarie perch . Final Report to the Department of Agriculture, Fisheries and Forestry Australia. Canberra: Parks, Conservation and Lands.

Ebsary, R. (1992). Regulation of the Murrumbidgee River. In The Murrumbidgee, Past and Present (Roberts J. & Oliver R., eds.), pp. 49–59. Griffith: CSIRO Division of Water Resources.

Eicher, G. J. (1982). Fish passage in New South Wales. Fisheries 7, 2–4.

Erskine, W. D. (1997). Application of geomorphology to the determination of environmental flows on the upper Murrumbidgee River. In Expert Panel Environmental Flow Assessment of the Upper Murrumbidgee River (Anon, ed.). NSW Environmental Protection Authority.

Essington, T. E. & Kitchell, J. F. (1999). New perspectives in the analysis of fish distributions: a case study on the spatial distribution of largemouth bass (Micropterus salmoides). Canadian Journal of Fisheries and Aquatic Science 56, 52–60.

Eyles, R. J. (1977). Changes in the drainage networks since 1820, Southern Tablelands, NSW. Australian Geographer 13, 377–387.

Faragher, R. A., Brown, P. & Harris, J. H. (1993). Population surveys of the endangered fish species trout cod (Maccullochella macquariensis) and eastern cod (M. ikei). Report for the Australian National Parks and Wildlife Service Endangered Species Program. Cronulla: NSW Fisheries.

Faragher, R. A. & Harris, J. H. (1994). The historical and current status of freshwater fish in New South Wales. Australian Zoologist 29, 166–176.

Faragher, R. A. & Rodgers, M. (1997). Performance of sampling-gear types in the New South Wales rivers survey. In Fish and Rivers in Stress: the NSW Rivers Survey (Harris, J. H. & Gehrke, P. C., eds.), pp. 251–268. Cronulla, NSW: NSW Fisheries Office of Conservation and the Cooperative Research Centre for Freshwater Ecology.

Fausch, K. D., Torgersen, C. E., Baxter, C. V. & Li, H. W. (2002). Landscapes to riverscapes: bridging the gap between research and conservation of stream fishes. BioScience 52, 483–498.

Fish, P. A. & Savitz, J. (1983). Variations in home ranges of large mouth bass, yellow perch, bluegills and pumpkinseeds in an Illinois lake. Transactions of the American

An ecological approach to re-establishing Australian freshwater cod populations 122 FRDC Report 2003/034

Fisheries Society 112, 147–153.

Gehrke, P. & Harris, J. (1996). Fish and fisheries of the Hawkesbury–Nepean River system. Cronulla: NSW Fisheries and the CRC for Freshwater Ecology.

Gehrke, P. C., Brown, P., Schiller, C. B., Moffatt, D. B. & Bruce, A. M. (1995). River regulation and fish communities in the Murray–Darling River System, Australia. Regulated Rivers: Research and Management 11, 363–375.

Gehrke, P. C. & Harris, J. H. (2000). Large-scale patterns in species richness and composition of temperate riverine fish communities, south-eastern Australia. Marine and Freshwater Research 51, 165–182.

Gerking, S. D. (1950). Stability of a stream fish population. Journal of Wildlife Management 14, 193–202.

Gerking, S. D. (1953). Evidence for the concepts of home range and territory in stream fishes. Ecology 34, 347–365.

Gillanders, B., Elsdon, T. & Munro, A. (2006). Impacts of native fish stocking on fish within the Murray–Darling Basin. Final Report for MDBC contract No. MD239. Adelaide: University of Adelaide.

Gilligan, D. M. (2005a). Fish communities of the Lower Murray–Darling catchment: status and trends. Fisheries Final Report Series No. 83. Cronulla: NSW Department of Primary Industries.

Gilligan, D. M. (2005b). Fish communities of the Murrumbidgee catchment: status and trends. Fisheries Final Report Series No. 75. Cronulla: NSW Department of Primary Industries.

Gilpin M E. & Soulé M. E. (1986). Minimum viable populations: processes of species extinction. In Conservation Biology: the Science of Scarcity and Diversity (Soulé M. E., ed.). Massachusetts: Sinauer Associates.

Girvan, J. R. & Braithwaite, A. V. (1998). Population differences in spatial learning in three - spined sticklebacks. Proceedings of the Royal Society B: Biological Sciences 265, 913–918.

Gooley, G. J. (1992). Validation of the use of otoliths to determine the age and growth of Murray Cod, Maccullochella peelii (Mitchell) (Percichthyidae), in Lake Charlegrark, Western Victoria. Australian Journal of Marine and Freshwater Research 43, 1091–1102.

Gowan, C. & Fausch, K. D. (2002). Why do foraging streams salmonids move during summer? Environmental Biology of Fishes 64, 139–153.

Gowan, C., Young, M. K., Fausch, K. D. & Riley, S. C. (1994). Restricted movements in resident stream salmonids: a paradigm lost? Canadian Journal of Fisheries and Aquatic Sciences 51, 2626–2637.

Greenham, P. (1981). Murrumbidgee River aquatic ecology study. Report to the

123 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

National Capital Development Commission and the Department of the Capital Territory. Canberra.

Griffin, A. S., Blumstein, D. T. & Evans, C. S. (2000). Training captive-bred or translocated animals to avoid predators. Conservation Biology 14, 1317–1326.

Growns, I., Astles, K. & Gehrke, P. (2003). Spatial and temporal variation in the composition of riverine fish communities. NSW Fisheries Final Report Series, Report No. 54. NSW Fisheries.

Growns, I., Wooden, I. & Schiller, C. (2004). Use of instream wood habitat by Trout Cod Maccullochella macquariensis (Cuvier) in the Murrumbidgee River. Pacific Conservation Biology 10, 261–265.

Gruber, B. & Henle, K. (2004). Linking habitat structure and orientation in an arboreal species Gehyra variegata (Gekkonidae). Oikos 107, 406–414.

Gust, N. & Handasyde, K. (1995). Seasonal variation in the ranging behaviour of the Platypus (Ornithorhynchus anatinus) on the Goulburn River, Victoria. Australian Journal of Zoology 43, 193–208.

Hammer, M. (2001). Molecular systematics and conservation biology of the southern pygmy perch Nannoperca australia (Günther, 1861) (Teleostei: Percichthyidae) in south-eastern Australia. Honours Thesis, Adelaide University.

Harris, J. H. (1984a). A survey of fishways in streams of coastal south-eastern Australia. Australian Zoologist 21, 219–233.

Harris, J. H. (1984b). Impoundment of coastal drainages of south-eastern Australia, and a review of its relevance to fish migrations. Australian Zoologist 21, 235–250.

Harris, J. H. (1995). Electrofishing injuries experiment. In Electrofishing Developments Workshop. Canberra: Cooperative Research Centre for Freshwater Ecology.

Harris, J. H. (2003). Fish stocking and translocation in the Murray–Darling Basin: issues, benefits and problems. Workshop held in Canberra 25–26 September 2002. Statement, recommendations and supporting papers. In Managing fish translocation and stocking in the Murray–Darling Basin. (Phillips, B., ed.) WWF Australia.

Harris, J. H. & Gehrke, P. C. (1997). Fish and Rivers in Stress: The NSW Rivers Survey New South Wales Fisheries and the Cooperative Research Centre for Freshwater Ecology.

Harris, J. H. & Mallen-Cooper, M. (1994). Fish-passage development in the rehabilitation of fisheries in mainland south-eastern Australia. In Rehabilitation of Freshwater Fisheries (Cowx, I. G., ed.), pp. 185–193. Hull International Fisheries Institute, University of Hulll, UK: Fishing News Books.

Harris, J. H. & Rowland, S. J. (1996). Family Percicthyidae: Australian freshwater cods and basses. In Freshwater Fishes of South-Eastern Australia, 2nd edn (McDowall R. M., ed.), pp. 150–163. Australia: Reed Books.

An ecological approach to re-establishing Australian freshwater cod populations 124 FRDC Report 2003/034

Harvey, B. C., Nakamoto, R. J. & White, J. L. (1999). Influence of large woody debris and a bankfull flood on movement of adult resident coastal cutthroat trout (Oncorhynchus clarki) during fall and winter. Canadian Journal and Fisheries and Aquatic Sciences 56, 2161–2166.

Harvey, J. & Cowx, I. G. (1996). Electric fishing for the assessment of fish stocks in large rivers. In Stock Assessment in Inland Fisheries (Cowx, I. G., ed.). Oxford: Fishing News Books.

Henry, G. W. & Lyle, J. M. (2003). The National Recreational and Indigenous Fishing Survey. In NSW Fisheries Final Report Series No. 48. Sydney: NSW Fisheries.

Heupel, M. R., Simpfendorfer, C. A. & Hueter, R. E. (2003). Running before the storm: blacktip sharks respond to falling barometric pressure associated with Tropical Storm Gabrielle. Journal of Fish Biology 63, 1357–1363.

Heupel, M. R. Semmens, J. M. & Hobday, A. J. (2006). Automated acoustic tracking of aquatic animals: scales, design and deployment of listening station arrays. Marine and Freshwater Research 57, 1–13.

Hiebert, S., Helfich, L. A., Weigmann, D. L. & Liston, C. (2000). Anadromous salmonid passage and video image quality under infrared and visible light at Prosser Dam, Yakima River, Washington. North American Journal of Fisheries Management 20, 827–832.

Hinch, S. G. & Rand, P. S. (2000). Optimal swimming speeds and forward-assisted propulsion: energy-conserving behaviours of upriver-migrating adult salmon. Canadian Journal of Fisheries and Aquatic Sciences 57, 2470–2478.

Hiscock, M. J., Scruton, D. A., Brown, J. A. & Pennell, C. J. (2002). Diel activity pattern of juvenile Atlantic salmon (Salmo salar) in early and late winter. Hydrobiologia 483, 161–165.

Hobson, K. A. (1999). Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia 120, 314–326.

Holbrook, S. J. & Schmitt, R. J. (2002). Competition for shelter space causes density- dependent predation mortality in Damselfishes. Ecology 83, 2855–2868.

Hooge, P. N. & Eichenlaub, B. (1997). Animal Movement extension to arcview 1.1. Anchorage, AK, USA: Alaska Biological Center, U. S. Geological Survey.

Hughes, R. M., Kaufmann, P. R., Herlihy, A. T., Intelmann, S. S., Corbett, S. C., Arbogast, M. C. & Hjort, R. C. (2002). Electrofishing distance needed to estimate fish species richness in raftable Oregon Rivers. North American Journal of Fisheries Management 22, 1229–1240.

Hughes, V. & Thoms, M. C. (2002). Associations between channel morphology and large woody debris in a lowland river. In The Structure, Function and Management Implications of Fluvial Sedimentary Systems (Dyer, F. J., Thoms, M. C. & Olley, J. M., eds.), pp. 11–18. Oxfordshire, UK: International Association of Hydrological

125 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Sciences.

Humphries, P. & King, A. J. (2003). Drifting fish larvae in Murray–Darling Basin rivers: composition, spatial and temporal patterns and distance drifted. In Downstream Movement of Fish in the Murray–Darling Basin Workshop Held in Canberra 2003: Statement, Recommendations and Supporting Papers (Lintermans, M. & Phillips, B., eds.), pp. 51–58. Canberra: Murray–Darling Basin Commission.

Hutchison, M., Gallagher, T., Chilcott, K., Simpson, R., Aland, G. & Sellin, M. (2006). Impoundment stocking strategies for Australian native fishes in eastern and northern Australia: with an assessment of the value of scales as tags for stocked barramundi. Queensland: Queensland Department of Primary Industries and Fisheries.

Hyvärinen, P. & Vehanen, T. (2004). Effect of brown trout body size on post-stocking survival and pike predation. Ecology of Freshwater Fish 13, 77–84.

Ingram, B. A. (2004). Murray cod fingerling production and growout. In Development of intensive commercial aquaculture production technology for Murray cod. Final report to the Fisheries Research and development Commission, project No. 1999/328. (Ingram B. A. & De Silva S. S., eds.), pp. 25–65. Victoria: Primary Industries Research.

Ingram, B. A. & De Silva, S. S. (eds.) (2004). In Development of Intensive Commercial Aquaculture Production Technology for Murray cod. Final Report to the Fisheries Research and Development Corporation. (Project No. 1999/328). Alexandra: Primary Industries Research Victoria, DPI.

Ingram, B. A. & Douglas, J. W. (1995). Threatened fishes of the world: Maccullochella maquariensis (Cuvier, 1829) (Pericichthyidae). Environmental Biology of Fishes 43, 38.

Ingram, B. A. & Rimmer, M. A. (1992). Induced breeding and larval rearing of the endangered Australian freshwater fish Trout Cod Maccullochella macquariensis (Cuvier) (Percicthyidae). Aquaculture and Fisheries Management 24, 7–17.

Ireland, S. C., Beamesderfer, R. C. P., Paragamian, V. L., Wakkinen, V. D. & Siple, J. T. (2002). Success of hatchery-reared juvenile white sturgeon (Acipenser transmontanus) following release in the Kootenai River, Idaho, USA. Journal of Applied Ichthyology 18, 642–650.

Jackson, P. D., Koehn, J. D., Lintermans, M. & Sanger, A. C. (1996). Family Gadopsidae. In Freshwater Fishes of South-Eastern Australia, 2nd edn (McDowall, R. M., ed.), pp. 186–190. Australia: Reed Books.

Jager, H. I., DeAngelis, D. L., Sale, M. J., Van Winkle, W., Schmoyer, D. D., Sabo, M. J., Orth, D. J. & Lukas, J. A. (1993). An individual-based model for smallmouth bass reproduction and young-of-year dynamics in streams. Rivers 4, 91–113.

Jager, H. I. & Tyler, J. A. (2001). Discussion on 'Movement rules for individual-based models of stream fish' [S.F. Railsback et al. Ecol. Model. 123 (1999) 73–89]. Ecological Modelling 143, 245–248.

An ecological approach to re-establishing Australian freshwater cod populations 126 FRDC Report 2003/034

Jepsen, N., Aarestrup, K., Økland, F. & Rasmussen, G. (1998). Survival of radio- tagged Atlantic salmon (Salmo salar L.) smolts passing a reservoir during seaward migration. Hydrobiologia 371/372, 347–353.

Johnsen, B. O. & Hvidsten, N. A. (2002). Use of radio telemetry and electrofishing to assess spawning by transplanted Atlantic salmon. Hydrobiologia 483, 13–21.

Jokikokko, E. & Jutila, E. (2004). Divergence in smolt production from the stocking of 1-summer-old and 1-year-old Atlantic salmon parr in a northern Baltic river. Journal of Applied Ichthyology 20, 511–516.

Jokikokko, E. & Mäntyniemi, S. (2003). The survival of stocked Atlantic salmon smolts during sea run and the timing of migration in the river Simojoki, northern Finland. Aquaculture 219, 431–444.

Jones, G. P. (1984). The influence of habitat and behavioural interactions on the local distribution of the wrasse, Pseudolabrus celidotus. Environmental Biology of Fishes 10, 43–58.

Jones, G. P. (1988). Ecology of rocky reef fish of north-eastern New Zealand: a review. New Zealand Journal of Marine and Freshwater Research 22, 445–462.

Jones, G. P. (1991). Post recruitment processes in the ecology of coral reef fish populations: a multifactorial perspective. In The Ecology of Fishes on Coral Reefs (Sale, P. F., ed.), pp. 294–398. California: Academic Press Inc.

Jury, S. H., Howell, H., O'Grady, D. F. & Watson, W. H. (2001). Lobster Trap video: in situ video surveillance of the behaviour of Homarus americanus in and around traps. Marine and Freshwater Research 52, 1125–1132.

Kearney, R. E. & Kildea, M. A. (2001). The status of Murray cod in the Murray– Darling Basin. Report to Environment Australia. Canberra: Applied Ecology Research Group.

Keefer, M. L., Peery, C. A. & Caudill, C. C. (2006). Long-distance downstream movements by homing adult chinook salmon. Journal of Fish Biology 68, 944–950.

Kelley, J. L. & Magurran, A. E. (2003). Learned predator recognition and antipredator responses in fishes. Fish and Fisheries 4, 216–226.

Kennard, M. J., Pusey, B. J., Harch, B. D., Dore, E. & Arthington, A. H. (2006). Estimating local stream fish assemblage attributes: sampling effort and efficiency at two spatial scales. Marine and Freshwater Research 57, 635–653.

Kieffer, M. C. & Kynard, B. (1993). Annual movements of shortnose and Atlantic sturgeons in the Merrimack River, Massachusetts. Transactions of the American Fisheries Society 122, 1088–1103.

Kieffer, M. C. & Kynard, B. (1996). Spawning of the shortnose sturgeon in the Merrimack River, Massachusetts. Transactions of the American Fisheries Society 125, 179–186.

127 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

King, A. J., Crook, D. A., Koster, W. M., Mahoney, J. & Tonkin, Z. (2005). Comparison of larval fish drift in the Lower Goulburn and mid-Murray Rivers. Ecological Management and Restoration 6, 136–138.

Koehn, J. (1995). Boat electrofishing in inland waters. In Electrofishing Developments Workshop. Canberra: Cooperative Research Centre for Freshwater Ecology.

Koehn, J. (1997). Habitats and movements of freshwater fish in the Murray–Darling Basin. In Proceedings of the 1995 Riverine Environment Research Forum (Banens, R. J. & Lehane, R., eds.), pp. 27–32. Canberra, Australia: Murray Darling Basin Commission.

Koehn, J. (2005). The loss of valuable Murray cod in fish kills: a science and management perspective. In Management of Murray Cod in the Murray–Darling Basin: Statement, Recommendations and Supporting Papers. Proceedings of a workshop held in Canberra, ACT 3–4 June 2004 (Lintermans, M. & Phillips, B., eds.). Canberra: Murray–Darling Basin Commission.

Koehn, J. & Nicol, S. (1998). Habitat and movement requirements of fish. In Proceedings of the 1996 Riverine Environment Forum (Banens, R. J. & Lehane, R., eds.), pp. 1–6. Brisbane: Murray–Darling Basin Commission.

Koehn, J. D. (2000). Why use radio tags to study freshwater fish? In Fish movement and migration. Australian Society for Fish Biology Workshop Proceedings. (Hancock, D. A., Smith, D. C. & Koehn, J. D., eds.). Sydney: Australian Society for Fish Biology.

Koehn, J. D., Doeg, T. J., Harrington, D. J. & Milledge, G. A. (1995). The Effects of Dartmouth Dam on the aquatic fauna of the Mitta Mitta River. Report to Murray– Darling Basin Commission. Victoria: Department of Natural Resources.

Koehn, J. D., Nicol, S. J. & Fairbrother, P. S. (2004). Spatial arrangement and physical characteristics of structural woody habitat in a lowland river in south-eastern Australia. Aquatic Conservation: Marine and Freshwater Ecosystems 14, 457–464.

Koehn, J. D. & O'Connor, W. G. (1990). Biological information for management of native freshwater fish in Victoria. Victoria: Department of Conservation and Environment Victoria.

Kuiter, R. H., Humphries, P. A. & Arthington, A. H. (1996). Family Nannopercidae: pygmy perches. In Freshwater Fishes of South-Eastern Australia, 2nd edn (McDowall, R. M., ed.), pp. 168–175. Sydney: Reed Books.

Kuwamura, T. (1986). Parental care and mating systems of cichlid fishes in Lake Tanganyika: a preliminary field survey. Journal of Ethology 4, 129–146.

Kuwamura, T., Nagoshi, M. & Sato, T. (1989). Female-to-male shift of mouthbrooding in a cichlid fish, Tanganicodus irsacae, with notes on breeding habits of two related species in Lake Tanganyika. Enviromental Biology of Fishes 24, 187– 198.

An ecological approach to re-establishing Australian freshwater cod populations 128 FRDC Report 2003/034

Kynard, B. (1997). Life history, latitudinal patterns, and status of the shortnose sturgeon Acipenser brevirostrum. Enviromental Biology of Fishes 48, 319–334.

Lake, J. S. (1971). Freshwater fishes and rivers of Australia. Sydney: Thomas Nelson and Sons.

Le Boeuf, B. J. (2004). Hunting and migratory movements of white sharks in the eastern North Pacific. Memoirs of the National Institute of Polar Research 58, 89– 100.

Leggett, W. C. (1977). The ecology of fish migrations. Annual Reviews in Ecology and Systematics 8, 285–308.

Levin, P. S. & Williams, J. G. (2002). Interspecific effects of artificially propagated fish: an additional conservation risk for salmon. Conservation Biology 16, 1581–1587.

Lintermans, M. (1995). The status of trout cod Maccullochella macquariensis in Bendora Reservoir, ACT: results of 1992/93 and 1995 monitoring. Internal Report 95/4. Canberra: ACT Parks and Conservation.

Lintermans, M. (2002). Fish in the Upper Murrumbidgee Catchment: A Review of Current Knowledge. Canberra: Environment ACT.

Lintermans, M. (2004). Rehabilitation of fish habitats in the Murrumbidgee River, ACT. Final Report to the MD 2001 FishRehab Program. Canberra: Environment ACT.

Lintermans M. (2005). Summary of management policies and fisheries regulations for Murray cod in the Murray–Darling Basin. In Management of Murray Cod in the Murray–Darling Basin: Statement, Recommendations and Supporting Papers. Proceedings of a workshop held in Canberra, ACT 3–4 June 2004 (Lintermans, M. & Phillips, B., eds.), pp 64–69. Canberra: Murray–Darling Basin Commission.

Lintermans, M. (2006). A review of stock enhancement activities for threatened fish species in Australia. In Research, development and extension priorities for stock enhancement, fish stocking and stock recovery. (Sawynok, B, ed.), pp. 30–43. Brisbane: Recfish Australia.

Lintermans, M. (in press). Fishes of the Murray–Darling Basin: an introduction. Canberra: Murray–Darling Basin Commission.

Lintermans, M., Kukolic, K. & Rutzou, T. (1988). The status of Trout Cod, Maccullochella macquariensis in the Australian Capital Territory. Victorian Naturalist 105, 205–207.

Lintermans, M. & Phillips, B. (eds.) (2004). Downstream movement of fish in the Murray–Darling Basin workshop held in Canberra 2003: Statement, recommendations and supporting papers. Canberra: Murray–Darling Basin Commission.

Lintermans, M. & Phillips, B. (eds.) (2005). Management of Murray Cod in the Murray–Darling Basin: Statement, Recommendations and Supporting Papers. Proceedings of a workshop held in Canberra, ACT 3–4 June 2004. Canberra: Murray

129 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Darling Basin Commission.

Lintermans, M., Robinson, W. & Harris, J. (2005a). Identifying the health of fish communities in the Murray–Darling Basin: the Sustainable Rivers Audit. In Proceedings of the 4th Australian Stream Management Conference: Linking Rivers to Landscapes (Rutherfurd, I. D., Wiszniewski, I., Askey-Doran, M. & Glazik, R., eds.), pp. 363–368. Hobart, Tasmania: Department of Primary Industries, Water and Environment.

Lintermans, M., Rowland, S. J., Koehn, J., Butler, G., Simpson, R. & Wooden, I. (2005b). The status, threats and management of freshwater cod species in Australia. In Management of Murray Cod in the Murray–Darling Basin: Statement, Recommendations and Supporting Papers. Proceedings of a Workshop Held in Canberra, ACT 3–4 June 2004 (Lintermans, M. & Phillips, B., eds.), pp. 15–29. Canberra: Murray–Darling Basin Commission.

Lintermans, M. & Rutzou, T. (1991). The status, distribution and management of the Murray crayfish Euastacus armatus in the Australian Capital Territory. Canberra: ACT Parks and Conservation Service.

Løkkeborg, S., Ferno, A. & Jorgensen, T. (2002). Effect of position-fixing interval on estimated swimming speed and movement pattern of fish tracked with a stationary positioning system. Hydrobiologia 483, 259–264.

Lucas, M. C. & Baras, E. (2000). Methods for studying spatial behaviour of freshwater fishes in the natural environment. Fish and Fisheries 1, 283–316.

Lyons, J. (1992). The length of stream to sample with a towed electrofishing unit when fish species richness is estimated. North American Journal of Fisheries Management 12, 198–203.

Mackay, N. J. (1973). Histological changes in the ovaries of the golden perch, Plectroplites ambiguus, associated with the reproductive cycle. Australian Journal of Marine and Freshwater Research 24, 95–101.

Madsen, T. & Shine, R. (1996). Seasonal migration of predators and prey - a study of pythons and rats in tropical Australia. Ecology 77, 149–156.

Mäkinen, T. S., Niemelä, E., Moen, K. & Lindström, R. (2000). Behaviour of gill-net and rod-captured Atlantic salmon (Salmo solar L.) during upstream migration and following radio tagging. Fisheries Research 45, 117–127.

Mallen-Cooper, M. (1992). Swimming ability of juvenile , Macquaria novemaculeata (Steindachner), and juvenile barramundi, Lates calcarifer (Bloch), in an experimental vertical-slot fishway. Australian Journal of Marine and Freshwater Research 43, 823–834.

Mallen-Cooper, M. (1996). Fishways and freshwater fish migration in south-eastern Australia. PhD Thesis, University of Technology.

Mallen-Cooper, M. & Harris, J. (1990). Fishways in mainland south-eastern Australia. In Proceedings of the International Symposium on Fishways.

An ecological approach to re-establishing Australian freshwater cod populations 130 FRDC Report 2003/034

Mallen-Cooper, M., Stuart, I. G., Hides-Pearson, F. & Harris, J. (1995). Fish migration in the Murray River and assessment of the Torrumbarry fishway. Cronulla: NSW Fisheries.

Marchant, S. & Higgins, P. J. (1990). Handbook of Australian, New Zealand and Antarctic Birds. Volume 1. Ratites to Ducks. Melbourne: Oxford University Press.

Martin, P. & Bateson, P. (1993). Measuring behaviour: an introductory guide. Cambridge: Cambridge University Press.

Martin, R. A. & Martin, A. (2006). Sociable killers: new studies of the white shark (a k a great white) show that its social life and hunting strategies are surprisingly complex. Natural History October, 42–48.

Matheney, M. P. & Rabeni, C. F. (1995). Patterns of movement and habitat use by northern hog suckers in an Ozark stream. Transactions of the American Fisheries Society 124, 886–897.

Mazeroll, A. I. & Montgomery, W. L. (1998). Daily migrations of a coral reef fish in the Red Sea (Gulf of Aqaba, Israel): initiation and orientation. Copeia 4, 893–905.

McDowall, R. M. (ed.) (1996). Freshwater fishes of south-eastern Australia, 2nd edn. Sydney: Reed Books.

MDBC (2004a). Fish Theme Pilot Audit Technical Report - Sustainable Rivers Audit. Canberra: Murray–Darling Basin Commission.

MDBC (2004b). Native Fish Strategy for the Murray–Darling Basin 2003–2013. Canberra: Murray–Darling Basin Commission.

Meador, M R. (2005). Single-pass versus two-pass boat electrofishing for characterizing river fish assemblages: species richness estimates and sampling distance. Transactions of the American Fisheries Society 134, 59–67.

Merrick, J. R. & Schmida, G. E. (1984). Australian freshwater fishes: Biology and management. North Ryde: Macquarie University.

Metcalfe, N. B., Valdimarsson, S. K. & Morgan I. J. (2003). The relative roles of domestication, rearing environment, prior residence and body size in deciding territorial contests between hatchery and wild juvenile salmon. Journal of Applied Ecology 40, 535–544.

Minckley, W. L. (1995). Translocation as a tool for conserving imperiled fishes: experiences in western United States. Biological Conservation 72, 297–309.

Molony, B. W., Bird, C. & Nguyen, V. P. (2004). The relative efficacy of stocking fry or yearling rainbow trout (Oncorhynchus mykiss) into a large impoundment dominated by redfin perch (Perca fluviatilis) in south-western Australia. Marine and Freshwater Research 55, 781–785.

Morris, S. A., Pollard, D. A., Gehrke, P. C. & Pogonoski, J. J. (2001). Threatened and potentially threatened freshwater fishes of coastal New South Wales and the Murray–

131 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Darling Basin. Report to Fisheries Action Program and World Wide Fund for Nature. NSW Fisheries Final Report Series No. 33. 166 pp. NSW Fisheries.

Mueller, R. P., Brown, R. S., Hop, H. & Moulton, L. (2006). Video and acoustic camera techniques for studying fish under ice: a review and comparison. Reviews in Fish Biology and Fisheries 16, 213–226.

Nakano, S. (1994). Variation in agonistic encounters in a dominance hierarchy of freely interacting red-spotted masu salmon (Oncorhynchus masou ishikawai). Ecology of Freshwater Fish 3, 153–158.

Nakano, S. (1995a). Competitive interactions for foraging microhabitats in a size- structured interspecific dominance hierarchy of two sympatric stream salmonids in a natural habitat. Canadian Journal of Zoology 73, 1845–1854.

Nakano, S. (1995b). Individual differences in resource use, growth and emigration under the influences of a dominance hierarchy in fluvial red-spotted masu salmon in a natural habitat. Journal of Animal Ecology 64, 75–84.

Nakano, S. & Furukawa-Tanaka, T. (1994). Intra- and interspecific dominance hierarchies and variation in foraging tactics of two species of stream-dwelling chars. Ecological Research 9, 9–20.

Naughton, G. P., Caudill, C. C., Keefer, M. L., Bjornn, T. C., Stuehrenberg, L. C. & Peery, C. A. (2005). Late-season mortality during migration of radio-tagged adult Sockeye salmon (Oncorhynchus nerka) in the Columbia River. Canadian Journal of Fisheries and Aquatic Science 62, 30–47.

Nelson, J. S. (1994). Fishes of the world. New York: John Wiley and Sons.

Nelson, T. C., Rosenau, M. L. & Johnston, N. T. (2005). Behavior and survival of wild hatchery-origin winter Steelhead spawners caught and released in a recreational fishery. North American Journal of Fisheries Management 25, 931–943.

Nicol, S. J., Lieschke, J., Lyon, J. & Hughes, V. (2002). Resnagging revolution: river habitat rehabilitation through resnagging. Final report to the Department of Agriculture, Fisheries and Forestry Australia (MD2001 Fish Rehab Program). Victoria, Australia: Department of Natural Resources and Environment.

Noda, M., Gushima, K. & Kakuda, S. (1994). Local prey search based on spatial memory and expectation in the planktivorous reef fish, Chromis chrysurus (Pomacentridae). Animal Behaviour 47, 1413–1422.

Northcote, T. G. (1992). Migration and residency in stream salmonids - some ecological considerations and evolutionary consequences. Nordic Journal of Freshwater Research 67, 5–17.

NSW DNR (2004). New South Wales Department of Natural Resources.

NSW DPI (New South Wales Department of Primary Industries) (2005). Trout Cod (Maccullochella macquariensis) Draft Recovery Plan. Port Stephens: Department of

An ecological approach to re-establishing Australian freshwater cod populations 132 FRDC Report 2003/034

Primary Industries.

NSW Fisheries (2003). Environmental Impact Statement on Freshwater Fish Stocking in NSW. Public Consultation Document. Cronulla: NSW Fisheries.

NSW Fisheries (2004). Eastern (Freshwater) Cod (Maccullochella ikei) Recovery Plan. Cronulla: NSW Fisheries.

O'Connor, J. P., O'Mahony, D. J. & O'Mahony, J. M. (2005). Movements of Macquaria ambigua, in the Murray River, south-eastern Australia. Journal of Fish Biology 66, 392–403.

Odenkirk, J. & Smith, S. (2005). Single- versus multiple-pass boat electrofishing for assessing smallmouth bass populations in Virginia Rivers. North American Journal of Fisheries Management 25, 717–724.

Økland, F., Erkinaro, J., Moen, K., Niemelä, S. L., Fiske, P., McKinley, R. S. & Thorstad, E. B. (2001). Return Migrations of Atlantic salmon in the River Tana: phases of migratory behaviour. Journal of Fish Biology 51, 862–874.

Olla, B. L., Davis, M. W. & Ryder, C. H. (1994). Behavioural deficits in hatchery- reared fish: potential effects on survival following release. Aquaculture and Fisheries Management 25, 19–34.

Olla, B. L., Davis, M. W. & Ryder, C. H. (1998). Understanding how the hatchery environment represses or promotes the development of behavioural survival skills. Bulletin of Marine Science 62, 531–550.

Olsson, M. & Shine, R. (2000). Ownership influences the outcome of male-male contests in the scincid lizard, Niveoscincus microlepidotus. Behavioural Ecology 11, 587–590.

Ovidio, M., Philippart, J. & Baras, E. (2000). Methodological bias in home range and mobility estimates when locating radio-tagged trout, Salmo trutta, at different time intervals. Aquatic living Resources 13, 449–454.

Park, T., Murphy, J. & Ried, D. (2005). The recreational fishery for Murray cod in the Murray Darling Basin Results from the National Recreational and Indigenous Fishing Survey. In Management of Murray Cod in the Murray–Darling Basin: Statement, Recommendations and Supporting Papers. Proceedings of a workshop held in Canberra, ACT 3–4 June 2004 (Lintermans, M. & Phillips, B., eds.), pp. 93–97. Canberra: Murray–Darling Basin Commission.

Patton, T. M., Hubert, W. A., Rahel, F. J. & Gerow, K. G. (2000). Effort needed to estimate species richness in small streams on the Great Plains in Wyoming. North American Journal of Fisheries Management 20, 394–398.

Pennington, M. & Volstad, J. H. (1994). Assessing the effect of intra-haul correlation and variable density on estimates of population characteristics from marine surveys. Biometrics 50, 725–732.

Perrone, M. Jr. (1978). Mate size and breeding success in a monogamous cichlid fish.

133 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Environmental Biology of Fishes 3, 193–201.

Peterson, D. P. & Fausch, K. D. (2003). Testing population-level mechanisms of invasion by a mobile vertebrate: a simple conceptual framework for salmonids in streams. Biological Invasions 5, 239–259.

Phillips, B. (2003). Managing fish translocation and stocking in the Murray–Darling Basin workshop held in Canberra, 25–26 September 2002: Statement, recommendations and supporting papers. Sydney, NSW: WWF Australia.

Polovina, J. J., Howell, E., Kobayashi, D. R. & Seki, M. P. (2001). The transition zone chlorophyll front, a dynamic global feature defining migration and forage habitat for marine resources. Progress in Oceanography 49, 469–483.

Priede, I. G. (1980). An analysis of objectives in telemetry studies of fish in the natural environment. In A Handbook on Biotelemetry and Radio-Tracking (Amlaner, C. J. & Macdonald D. W., ed.). Oxford: Pergamon Press.

Priede, I. G. & Swift, S. M. (1992). Wildlife telemetry: remote monitoring and tracking of animals. New York: Ellis Horwood.

Prosser, I. P., Rutherford, I. D., Olley, J. M., Young, W. J., Wallbrink, P. J. & Moran, C. J. (2001). Large-scale patterns of erosion and sediment transport in river networks, with examples from Australia. Marine and Freshwater Research 52, 81–99.

Pusey, B. J. & Kennard, M. J. (2001). Guyu wujalwujalensis, a new genus and species (Pisces: Percichthyidae) from north-eastern Queensland, Australia. Ichthyological Explorations of Freshwaters 12, 17–28.

Pusey, B. J., Kennard, M. K., Arthur, J. M. & Arthington, A. H. (1998). Quantitative sampling of stream fish assemblages: single- vs multiple-pass electrofishing. Australian Journal of Ecology 23, 365–374.

Railsback, S. F., Lamberson, R. H., Harvery, B. C. & Duffy, W. E. (2001). Reply to "Discussion on 'Movement rules for individual-based models of stream fish' " [S.F. Railsback et al. Ecol. Model. 123 (1999) 73–89]. Ecological Modelling 143, 249–251.

Railsback, S. F., Lamberson, R. H., Harvey, B. C. & Duffy, W. E. (1999). Movement rules for individual-based models of stream fish. Ecological Modelling 123, 73–89.

Reese, E. S. (1989). Orientation behaviour of butterflyfishes (family Chaetontidae) on coral reefs: spatial learning of route specific landmarks and cognitive maps. Environmental Biology of Fishes 25, 79–86.

Reid, D. D., Harris, J. H. & Chapman, D. J. (1997). NSW Inland Commercial Fishery Data Analysis. Fisheries Research and Development Corporation (Project No. 94/027), NSW Fisheries, Cronulla and CRC Freshwater Ecology, Canberra.

Reynolds, L. F. (1983). Migration patterns of five fish species in the Murray–Darling River system. Australian Journal of Marine and Freshwater Research 34, 857–871.

Ricker, W. E. (1975). Computation and interpretation of biological statistics of fish

An ecological approach to re-establishing Australian freshwater cod populations 134 FRDC Report 2003/034 populations. Bulletin of the Fisheries Research Board of Canada 191, 1–382.

Rodríguez, M. A. (2002). Restricted movement in stream fish: the paradigm is incomplete, not lost. Ecology 83, 1–13.

Ross, R. M. (1984). Catheterization: a non-harmful method of sex identification for monomorphic fishes. The Progressive Fish-Culturist 46, 151–152.

Rowland, S. J. (1983a). The hormone induced ovulation and spawning of the Australian freshwater fish golden perch Macquaria ambigua (Richardson) (Percichthyidae). Aquaculture 35, 221–238.

Rowland, S. J. (1983b). Spawning of the Australian freshwater fish Murray cod, Maccullochella peeli (Mitchell) in earthen ponds. Journal of Fish Biology 23, 525– 534.

Rowland, S. J. (1985). Aspects of the biology and artificial breeding of the Murray cod, Maccullochella peelii, and the eastern freshwater cod, M. ikei sp. nov. (Pisces:Percichthyidae) . PhD thesis, Maquarie University.

Rowland, S. J. (1988). Hormone-induced spawning of the Australian freshwater fish Murray cod, Maccullochella peeli (Mitchell) (Percichthyidae). Aquaculture 70, 371– 389.

Rowland, S. J. (1989). Aspects of the history and fishery of the Murray Cod Maccullochella peeli (Mitchell) (Percichthyidae). Proceedings of the Linnaean Society of New South Wales 111, 201–213.

Rowland, S. J. (1993). Maccullochella ikei, an endangered species of freshwater cod (Pisces: Percichthyidae) from the Clarence River system, NSW, and M. peelii mariensis, a new subspecies from the Mary River, Qld. Records of the Australian Museum 45, 121–145.

Rowland, S. J. (1996). Threatened fishes of the world: Maccullochella ikei Rowland, 1985 (Percichthyidae). Environmental Biology of Fishes 46, 350.

Rowland, S. J. (1998a). Age and Growth of the Australian freshwater fish Murray Cod, Maccullochella peelii peelii. Proceedings of the Linnaean Society of New South Wales 120, 163–180.

Rowland, S. J. (1998b). Aspects of the reproductive biology of Murray Cod, Maccullochella peelii peelii. Proceedings of the Linnaean Society of New South Wales 120, 147–162.

Rowland, S. J. (2005). Overview of the history, fishery, biology and aquaculture of Murray cod (Maccullochella peelii peelii). In Management of Murray Cod in the Murray–Darling Basin: Statement, Recommendations and Supporting Papers. Proceedings of a workshop held in Canberra, ACT 3–4 June 2004. (Lintermans, M. & Phillips, B., eds.), pp. 38–61 Canberra: Murray-Darling Basin Commission.

Sale, P. F. (ed.) (1991). The ecology of fishes on coral reefs. San Diego, California: Academic Press.

135 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Salvanes, A. G. V. & Braithwaite, V. A. (2005). Exposure to variable spatial information in the early rearing environment generates asymmetries in social interactions in cod ( morhua). Behavioural Ecology and Sociobiology 59, 250– 257.

Scott, A. (2005). Historical evidence of native fish in the Murray–Darling Basin at the time of European settlement - from the diaries of the first explorers. Canberra: Cooperative Research Centre for Freshwater Ecology.

Shiel, R. J., Walker, K. F. & Williams, W. D. (1982). Plankton of the Lower River Murray, South Australia. Australian Journal of Marine and Freshwater Research 33, 301–327.

Simpson, R. R., Hutchinson, M., Gallagher, T. & Chilcott, K. (2002). Fish stocking in impoundments: a best practice manual for eastern and northern Australia. DPI Agency for food and Fibre Science, Queensland DPI and Fisheries Research and Development Corporation.

Simpson, R. R. & Jackson, P. D. (1996). The Mary River Cod Recovery Plan. Queensland Department of Primary Industries, Fisheries Group.

Simpson, R. R. & Mapleston, A. J. (2002). Movements and habitat use by the endangered Australian freshwater Mary River cod, Maccullochella peelii mariensis. Environmental Biology of Fishes 65, 401–410.

Sinclair, A. R. E. (1983). The function of distance movements in vertebrates. In The Ecology of Animal Movement (Swingland, I. R. & Greenwood, P. J., eds.), pp. 240– 258. Oxford: Oxford University Press.

Skalski, J. R., Lady, J., Townsend, R., Giorgi, A. E., Stevenson, J. R., Peven, C. M. & McDonald, R. D. (2001). Estimating in-river survival of migrating salmonid smolts using radiotelemetry. Canadian Journal of Fisheries and Aquatic Sciences 58, 1987– 1997.

Skalski, J. R., Townsend, R., Lady, J., Giorgi, A. E., Stevenson, J. R. & McDonald, R. D. (2002). Estimating route-specific passage and survival probabilities at a hydroelectric project from smolt radiotelemetry studies. Canadian Journal of Fisheries and Aquatic Sciences 59, 1385–1393.

Smith, K. L. & Jones, M. L. (2005). Watershed-level sampling effort requirements for detecting riverine fish species composition. Canadian Journal of Fisheries and Aquatic Sciences 62, 1580–1588.

Sokal, R. R. & Rohlf, K. J. (1995). Biometry, 3rd edn. New York: W. H. Freeman and Company.

Stuart, I. & Berghuis, A. P. (2001). Upstream passage through a vertical slot fishway in an Australian subtropical river. Fisheries Management and Ecology 9, 111–122.

Stuart, I. & Jones, M. (2002). Ecology and management of in the Barmah–Millewa forest. Final Report to Agriculture, Fisheries and Forestry Australia. Heidelberg, Victoria: Arthur Rylah Institute for Environmental Research.

An ecological approach to re-establishing Australian freshwater cod populations 136 FRDC Report 2003/034

Stuart, I. G. & Mallen-Cooper, M. (1999). An assessment of the effectiveness of a vertical slot fishway for non-salmonid fish at a tidal barrier on a large tropical/subtropical river. Regulated Rivers: Research and Management 15, 575–590.

Sullivan, A. B., Jager, H. I. & Myers, R. (2003). Modelling white sturgeon movement in a reservoir: the effect of water quality and sturgeon density. Ecological Modelling 167, 97–114.

Swales, S. (1994a). Habitat restoration methods - a synthesis. In Rehabilitation of Freshwater Fisheries (Cowx, I. G., ed.), pp. 133–137. Hull International Fisheries Institute, University of Hulll, UK: Fishing News Books.

Swales, S. (1994b). Strategies for the conservation of stream fish populations. In Rehabilitation of Freshwater Fisheries (Cowx, I. G., ed.), pp. 34–47. Hull International Fisheries Institute, University of Hulll, UK: Fishing News Books.

Swingland, I. R. & Greenwood, P. J. (1983). The ecology of animal movement. New York: Oxford University Press.

Tabachnick, B. G. & Fidell, L. S. (1989). Using multivariate statistics. New York: Harper and Row Publishers.

TCRT (Trout Cod Recovery Team) (2004). Recovery Plan for Trout Cod Maccullochella macquariensis (Cuvier, 1829) 2005 – 2009. Heidelberg, Victoria: Department of Sustainability and Environment.

Thévenet, A. & Statzner, B. (1999). Linking fluvial fish community to physical habitat in large woody debris: sampling effort, accuracy and precision. Archiv fur Hydrobiologie 145, 57–77.

Thorfve, S. (2002). Impacts of in-stream acclimatization in post-stocking behaviour of European grayling in a Swedish stream. Fisheries Management and Ecology 9, 253– 260.

Thorpe, J. E. (2004). Life history responses of fishes to culture. Journal of Fish Biology 65, 263–285.

Thorstad, E. B. & Heggberget, T. G. (1998). Migration of adult Atlantic salmon (Salmo salar); the effects of artificial freshets. Hydrobiologia 371/372, 339–346.

Thorstad, E. B., Heggberget, T. G. & Økland, F. (1998). Migratory behaviour of adult wild and escaped farmed Atlantic salmon, Salmo salar L., before, during and after spawning in a Norwegian river. Aquaculture Research 29, 419–428.

Todd, C. R., Nicol, S. J. & Koehn, J. D. (2004). Density-dependence uncertainty in population models for the conservation management of trout cod, Maccullochella macquariensis. Ecological Modelling 171, 359–380.

Trudel, M. & Boisclair, D. (1996). Estimating of fish activity costs using underwater video cameras. Journal of Fish Biology 48, 40–53.

Unmack, P. J. (2001). Biogeography of Australian freshwater fishes. Journal of

137 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Biogeography 28, 1053–1089.

Unmack, P. J., Hammer, M., Adams, M. & Dowling, T. E. (submitted). A phylogenetic analysis of pygmy perches (Teleostei: Percichthyidae) with an assessment of the major historical influences on aquatic biogeography in southern Australia. van Ginneken, V. J. T. & Maes, G. E. (2005). The European eel (Anguilla anguilla, Linnaeus), its lifecycle, evolution and reproduction: a literature review. Reviews in Fish Biology and Fisheries 15, 367–398.

Van Winkle, W., Jager, H. I., Railsback, S. F., Holcomb, B. D., Studley, T. K. & Baldrige, J. E. (1998). Individual-based model of sympatric population of brown and rainbow trout for instream flow assessment: model description and calibration. Ecological Modelling 110, 175–207.

Viddi, F. A. & Lescrauwaet, A-K. (2005). Insights on habitat selection and behavioural patterns of Peale’s Dolphins (Lagenorhynchus australis) in the Strait of Magellan, Southern Chile. Aquatic Mammals 31, 176–183.

Vilhunen, S. (2006). Repeated antipredator conditioning: a pathway to habituation or to better avoidance? Journal of Fish Biology 68, 25–43.

Vilhunen, S. & Hirvonen, H. (2003). Innate antipredator responses of Artic charr (Salvelinus alpinus) depend on predator species and their diet. Behavioural Ecology and Sociobiology 55, 1–10.

Vilhunen, S., Hirvonen, H. & Laakkonen, M. V-M. (2005). Less is more: social learning of predator recognition requires a low demonstrator to observer ratio in Arctic charr (Salvelinus alpinus). Behavioural Ecology and Sociobiology 57, 275– 282.

Wallace, M. P. (2000). Retaining natural behaviour in captivity for reintroduction programmes. In Behavioural Conservation (Gosling, L. M. & Sutherland, W. J., eds.). Cambridge: Cambridge University Press.

Walsh, T. & Nash, W. (2002). Factors influencing the health of the giant freshwater lobster. Unpublished report. Hobart, Tasmania: Inland Fisheries Service.

Walters, C. J. & Holling (1990). Large-scale management experiments and learning by doing. Ecology 71, 2060–2068.

Weisser, W. W. (2001). The effects of predation on dispersal. In Dispersal (Clobert, J., Danchin, E., Dhondt, A. A. & Nichols, J. D., eds.), pp. 180–188. New York: Oxford University Press.

White, G. C. & Garrot, R. A. (1990). Analysis of wildlife radio-tracking data. San Diego: Academic Press, Inc.

Whoriskey, F. G., Prusov, S. & Crabbe, S. (2000). Evaluation of the effects of catch- and-release angling on the Atlantic salmon (Salmo salar) of the Ponoi River, Kola Peninsula, Russian Federation. Ecology of Freshwater Fish 9, 118–125.

An ecological approach to re-establishing Australian freshwater cod populations 138 FRDC Report 2003/034

Winter, J. (1996). Advances in underwater biotelemetry. In Fisheries Techniques (Murphy B. R. & Willis D. W., eds.), pp. 555–590. Maryland: American Fisheries Society.

Woollard, P., Vestjens, W. J. M. & MacLean, L. (1978). The ecology of the eastern water rat Hydromys chrysogaster at Griffith, NSW: food and feeding habits. Australian Wildlife Research 5, 59–73.

Yau, C., Collins, M. A., Bagley, P. M., Everson, I., Nolan, C. P. & Priede, I. G. (2001). Estimating the abundance of Patagonian toothfish Dissostichus eleginoides using baited cameras: a preliminary study. Fisheries Research 51, 403–412.

Ye, Q., Jones, K. & Pierce, B. E. (2000). Murray cod (Maccullochella peelii peelii). Fishery assessment report to PIRSA for the Inland Waters Fishery Management Committee. South Australian Fisheries Assessment Series 2000/17.

Young, W. J., Roberts, J. & Hillman, T. J. (2001). The rivers of the Basin and how they work. In Rivers As Ecological Systems: the Murray–Darling Basin (Young, W. J., ed.), pp. 3–43. Canberra: Murray Darling Basin Commission.

Zalewski, M. & Cowx, I. (1990). Factors affecting the efficiency of electric fishing. In Fishing With Electricity (Cowx, I. G. & Lamarque, P., eds.), pp. 89–111. London: Fishing News Books, Blackwell Scientific Publications.

139 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

APPENDICES

Appendix 1. Intellectual Property Chapters 2, 3, 4, 5, and 6 have been submitted or are in the process of being submitted to peer reviewed journals. No patents were developed during the project.

An ecological approach to re-establishing Australian freshwater cod populations 140 FRDC Report 2003/034

Appendix 2. Staff

Input from Kevin Frawley, Brett Ingram, Mark Kennard, David Shorthouse, Luke Johnston and Katie Ryan improved versions of this document. Mark Dunford provided invaluable GIS support.

The idea for this comparison of wild and hatchery trout cod came from discussion with Ian Wooden.

The following staff provided various support and field assistance during the project:

Parks, Conservation & Lands (Formerly Environment ACT) Mark Jekabsons, Simon Godschalx, Katie Ryan, Ben Broadhurst, Luke Johnston, Kevin Frawley, Mark Dunford, David Shorthouse, Adela Barlin, Claire Wimpenny, Karen Watson, Will Andrew, Murray Evans, Rosie Schmedding, David Miles, Cecelia Bourke

NSW DPI (Formerly NSW Fisheries) Mark Stimson, Les Rava, Vanessa Carracher, Ian Wooden, Lee Baumgartner, Steve Thurstan, Matt McLellan, Peter Maclean, David Patterson

VIC DPI Brett Ingram, Peter Boyd, Peter Ryder, John Douglas

VIC DSE Justin O’Connor, John Koehn, Simon Nicol, John Mahoney, David Crook

State Water Matt Deaton, Lindsay Beck, Mick McNamara and Eddie Taylor mounted the telemetry aerials at Yanco Weir.

Jason Prince assisted with data analysis in Chapter 4.

Gerald Verdouw from SciElex provided technical expertise and advice on the underwater video equipment.

CRCFE Simon Linke, Ann Milligan

MDBC Anthony Chariton, Dean Ansell, Jim Barrett

CSIRO Sue Vink, Ryan McAllister

Brindabella Airlines Jason, Dave Newling, Amy, Geoff

141 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3. Communication outputs

List of written communications 3.1. Newspaper article – “Trout cod research project under way”. Narrandera Argus, 16th October 2003.

3.2. Media release – “Environment ACT leading the way in freshwater cod research and conservation”. Released 15th February 2004, ACT Government.

3.3. Australian Society for Fish Biology Newsletter, 34: 1 (p42) June 2004.

3.4. Media release – “Campaign to recover freshwater cod well advanced”. Released 14th July 2004, Environment ACT.

3.5. Newspaper article – “Groundbreaking cod research project at Narrandera station”. Narrandera Argus, 22nd July 2004.

3.6. Newspaper article – “Cod return soon a reality”. Sunraysia Daily, 23rd July 2004.

3.7. Newspaper article – “Saving the trout cod”. Gundagai Independent 26th July 2004.

3.8. Newspaper article – “Study to revive cod stocks”. Canberra Times, 8th August 2004.

3.9. Australian Society for Fish Biology Newsletter, 34: 2 (p38) December 2004.

3.10. Media release – “Trout Cod movements in the Murrumbidgee catchment”. Released April 2005, Environment ACT.

3.11. “Trout cod in the Murrumbidgee River”. WATERSHED, August 2005.

3.12. Media release – “DPI seminar focuses on saving an Aussie icon – Trout Cod”. Released 21st June 2005, NSW Department of Primary Industries.

3.13. Bulletin of the Ecological Society of Australia, 35: 4 (p 23) December 2005.

3.14. Magazine article – “Fish school of the air”. Fishing World, July 2006.

List of verbal communications 3.15. Abstract for presentation given at the Australian Society for Fish Biology conference, Adelaide, September 2004.

3.16. Abstract for presentation given at the Australian Society for Fish Biology conference, Darwin, July 2005.

3.17. Abstract for presentation given at the Australian Society for Fish Biology conference, Darwin, July 2005.

An ecological approach to re-establishing Australian freshwater cod populations 142 FRDC Report 2003/034

Appendix 3. (cont.)

3.18. Abstract for presentation given at the Australian Society for Fish Biology conference, Hobart, 31st Aug – 1st Sept 2006.

143 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.1. Newspaper article, Narrandera Argus 16th October 2003.

An ecological approach to re-establishing Australian freshwater cod populations 144 FRDC Report 2003/034

Appendix 3.2. Media release, Environment ACT, 15th February 2004.

145 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.3. Australian Society of Fish Biology newsletter, June 2004.

An ecological approach to re-establishing Australian freshwater cod populations 146 FRDC Report 2003/034

Appendix 3.4. Media release, Environment ACT, released 14th July 2004.

147 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.5. Newspaper article, Narrandera Argus, July 22nd 2004.

An ecological approach to re-establishing Australian freshwater cod populations 148 FRDC Report 2003/034

Appendix 3.6. Newspaper article, Sunraysia Daily, 23rd July 2004.

149 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.7. Newspaper article, Gundagai Independent, 26th July 2004.

An ecological approach to re-establishing Australian freshwater cod populations 150 FRDC Report 2003/034

Appendix 3.8. Newspaper article, Canberra Times, 8th August 2004.

151 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.9. Australian Society for Fish Biology newsletter, December 2004.

An ecological approach to re-establishing Australian freshwater cod populations 152 FRDC Report 2003/034

Appendix 3.10. Media release, Environment ACT, released April 2005.

153 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.10. (cont.)

An ecological approach to re-establishing Australian freshwater cod populations 154 FRDC Report 2003/034

Appendix 3.10. (cont.)

155 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.11. Watershed article, August 2005.

An ecological approach to re-establishing Australian freshwater cod populations 156 FRDC Report 2003/034

Appendix 3.12. Media release, NSW Department of Primary Industry, June 2005.

157 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.13. Bulletin of the Ecological Society of Australia, December 2005.

An ecological approach to re-establishing Australian freshwater cod populations 158 FRDC Report 2003/034

Appendix 3.14. Magazine article, Fishing World, July 2006.

159 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.14. (cont.)

An ecological approach to re-establishing Australian freshwater cod populations 160 FRDC Report 2003/034

Appendix 3.14. (cont.)

161 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.15. Abstract of presentation given at the Australian Society for Fish Biology conference, Adelaide, September 2004.

An ecological approach to re-establishing Australian freshwater cod populations 162 FRDC Report 2003/034

Appendix 3.16. Abstract of presentation given at the Australian Society for Fish Biology conference, Darwin, July 2005.

163 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 3.17. Abstract for presentation given at the Australian Society for Fish Biology conference, Darwin, July 2005.

An ecological approach to re-establishing Australian freshwater cod populations 164 FRDC Report 2003/034

Appendix 3.18. Abstract for presentation given at the Australian Society for Fish Biology conference, Hobart, 31st Aug – 1st Sept 2006.

165 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 4. Standard survey effort involving boat electro-fishing by state fisheries agencies in the Murray–Darling Basin. River(s) Length Number of Elapsed Project of site operations operation (km) time Lachlan 1 8–15 2 min Sustainable Rivers Audit Condamine MDBC (2004a). Ovens Murray Macquarie N/A 15 2 min Growns et al. (2003). Namoi Multiple rivers 1 8–15 2 min NSW Rivers Survey Harris & Gehrke (1997) 7 rivers N/A 8–15 2 min Integrated Monitoring of Environmental Flows Chessman (2003) Murrumbidgee ~0.4 10 2 min Standard monitoring: Parks, Conservation & Lands (Lintermans, unpublished data) Murrumbidgee ~0.3–0.4 15 2 min Tharwa groynes: Parks, Conservation & Lands (Lintermans, unpublished data) Murray 1 12 90 s NSW Department of Darling (power on) Primary Industries, Murrumbidgee Integrated Fish Monitoring Project (Gilligan, 2005a, b)

An ecological approach to re-establishing Australian freshwater cod populations 166 FRDC Report 2003/034

Appendix 5. Aerial telemetry methodology – a pilot study over the Cotter catchment

Jason Thiem and Brendan Ebner

INTRODUCTION A pilot study involving aerial radio telemetry was set up to determine its applicability in locating fish attached with radio transmitters, particularly those in inaccessible terrain or outside of a study area. The aim of the study was to determine if the distance at which a fish could be detected using aerial telemetry was influenced by (1) the altitude of the aircraft and, (2) the speed of the aircraft.

METHODS A Cessna 172 was used, fitted with 3-element yagi antennas mounted to each wing via a mounting bracket (produced by Sirtrack) (Figure 1). A total of five frequencies were placed on a scanning receiver (Titley Electronics) on a scan time of ten seconds for each frequency.

Figure 1. A yagi antenna mounted to the wing of a Cessna 172.

A total of six transects were flown, at two altitudes (8000 and 2000 feet ASL) along Cotter Reservoir and Cotter River. Initially two speeds of 80 and 110 knots were to be used, however due to weather conditions and drag induced by the antennas, speed varied constantly between 80 and 100 knots. As such, data at different speeds was pooled. Upon location of a radio signal, a GPS point was taken for each fish. These points were then plotted in ArcView GIS 3.2, with distances measured between the locations of radio signals and each fish’s known location. Distances were measured in the horizontal plane.

A brief trial was also conducted, using a splitter box containing a left-both-right switch that enabled researchers to switch between left and right antennas and gain a more accurate location on a fish once it had been detected. This was conducted at an

167 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 5. (cont.) altitude of 2000 feet ASL and involved ‘zig-zagging’ the river, spacing transects approximately 500 m apart.

RESULTS Of the five fish being scanned for in this study, all five were detected after three passes at an altitude of 8000 feet (Table 1). However only four fish were detected at an altitude of 2000 feet after two passes. Table 1 shows that maximum detection distances were greater at the higher altitudes of 8000 feet, reaching a maximum of 11.86 km for one individual.

Table 1. Maximum and minimum distances (km) that fish fitted with radio transmitters were detected from, in a pilot study over the Cotter Reservoir and Cotter River, using aerial telemetry techniques in a Cessna 172. Distances were calculated in the horizontal plane using ArcView GIS 3.2. Blank spaces denote no record for that transect.

Transect Altitude Fish A Fish B Fish C Fish D Fish E

T1 Max 8000 ft 6.5 km Min 8000 ft 4.9 km n=4

T2 Max 8000 ft 5.3 km 5.1 km 5.15 km 3.89 km Min 8000 ft 4.1 km 3.27 km 3.1 km 3.87 km n=3 n=4 n=3 n=1

T3 Max 8000 ft 3.8 km 3.58 km 3.17 km 4.58 km 3.6 km Min 8000 ft 3.0 km 2.28 km 1.96 km 4.5 km 2.66 km n=2 n=2 n=2 n=1 n=2

T4 Max 8000 ft 4.6 km 11.86 7.2 km 1.94 km 2.49 km km Min 8000 ft 2.97 km 2.43 km 2.17 km 1.90 km 2.44 km n=4 n=6 n=5 n=1 n=1

T5 Max 2000 ft 1.95 km 6.43 km 5.06 km 0.75 km Min 2000 ft 1.53 km 1.3 km 0.71 km 0.71 km n=2 n=4 n=2 n=1

T6 Max 2000 ft 1.74 km 0.36 km 0.76 km Min 2000 ft 1.68 km 0.32 km 0.74 km n=1 n=1 n=1

An ecological approach to re-establishing Australian freshwater cod populations 168 FRDC Report 2003/034

Appendix 5. (cont.)

DISCISSION Previous references to the use of aerial telemetry from light aircraft to detect individuals fitted with radio transmitters recommend obtaining higher altitudes (2000- 2500 m AGL) to increase detection ranges for missing transmitters and lower altitudes (300-1000 m AGL) to increase the accuracy of location estimates (Kenward, 1987; Seddon and Maloney, 2004). Furthermore it is also suggested that higher altitudes may be more appropriate for locating individuals in rugged terrain (Kenward, 1987). Detection of individuals in this study was found to vary greatly between altitudes (Table 1). However replicate transects at both 8000 and 2000 ft indicate that this variability is also present between transects at a single altitude (Table 1). Greater detection distances were recorded for most individuals at a higher altitude and the number of times an individual was detected in a single pass was in general, greater at the higher altitude. However our results indicate that the hypothesis that higher altitudes increase detection distances in aerial telemetry could not be accepted without further work. Personal observations during the course of the work indicate that variations in our ability to detect individuals may have been affected by the transect line or bearing of the plane, the order of a radio frequency on the scanning receiver, the depth of water that individuals were located in and also the interference from the aircraft engine on some frequencies.

Various search techniques have been described when using aerial telemetry as well as the use of different methods to obtain an accurate location once a signal has been detected (see Kenward, 1987; White and Garrott, 1990; Seddon and Maloney, 2004). This study used an adaptation of the grid search method described in Kenward (1987) and applied it to use over a river in what we called a zig-zagging pattern, with passes over the river spaced approximately 500m apart. Observations indicate that our ability to detect the location of an individual in a river could be narrowed down to a 500m section. White and Garrott (1990) suggest that errors from a plane can be in excess of 500m depending on a number of factors. Therefore, follow up by a ground crew may be necessary to obtain more accurate fixes on the location of an individual.

REFERENCES Kenward, R. (1987). Wildlife radio tagging: equipment, field techniques and data analysis. Academic Press, London.

Seddon, P.J. and Maloney, R.F. (2004). Tracking wildlife radio-tag signals by light fixed-wing aircraft. Department of Conservation, Wellington.

White, G.C. and Garrott, R.A. (1990). Analysis of wildlife radio-tracking data. Academic Press, California.

169 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 6. Movements and habitat use of wild M. macquariensis in the Murrumbidgee River (Experiment 1).

100

80

60

40

20 Proportion of movements (%) Proportion of

0 Oct Dec Feb Apr Jun Aug Oct

2003 2004 Figure 1. Proportion of individual movements (%) of wild M. macquariensis in the Murrumbidgee River, with monthly displacement of 0-100 m ( ▲ ), 100 m - 1 km (··∆··), 1 km - 10 km (--▲--) and >10 km (--∆--).

Table 1. Swimming speeds for individual M. macquariensis recorded on the Yanco Weir remote telemetry logger. Spatial locations before logger records for downstream (DS) movements and after logger records for upstream (US) movements are taken from fortnightly or monthly manual radio-tracking. Fish ID Distance between Direction Maximum Minimum locations time taken average speed (km) (h) (km h-1) H20 10.8 DS 333 0.03 10.9 US 4 2.73 W19 30.6 DS 238 0.13 30.6 US 96 0.32 W26 32.6 DS 521 0.06 32.6 US 125 0.26 W28 31.6 DS 422 0.07 31.7 US 224 0.14 W29 26.6 DS 71 0.37 26.6 US 608 0.04

An ecological approach to re-establishing Australian freshwater cod populations 170 FRDC Report 2003/034

Appendix 6. (cont.)

60 (20) (19)

(25) 50 (28) (26) (24) (24)

40

30

20

Proportion homed (%) homed Proportion 10

0 0 102030405060 Time interval (weeks)

Figure 2. The proportion of wild M. macquariensis located within one home-range (upstream or downstream) of their original capture location on a given radio-tracking occasion. Sample sizes of less than 30 are shown in brackets.

171 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 6. (cont.)

Table 2. Individual records of homing ability and displacement distances for M. macquariensis in the Murrumbidgee River near Narrandera

Fish Proximity of Homed? Time until Fixes at Total Proportion ID capture in first home fixes (%) relation to release fix at home site (m) (days) W1 -1270 Y 52 14 17 82 W2 -4015 N 0 W3 -1822 Y 12 16 17 94 W4 -886 Y 12 15 15 100 W5 -4276 Y 40 12 14 86 W6 -10351 N 0 W7 3138 N 0 W8 872 Y 26 16 17 94 W9 -3582 N 0 W10 -8033 N 0 W11 -1190 Y 12 5 16 31 W12 222 Y 7 10 15 67 W13 4658 Y 35 15 17 88 W14 -1190 N 0 W15 -4000 N 0 W16 -4005 N 0 W17 -10332 N 0 W18 4383 N 0 W19 -5136 N 0 W20 963 Y 7 12 12 100 W21 -4013 N 0 W22 -4278 Y 67 6 17 35 W23 -4634 Y 12 17 17 100 W24 -4353 Y 12 17 17 100 W25 381 Y 7 15 15 100 W26 1987 Y 7 17 17 100 W27 3243 N 0 W28 872 Y 7 17 17 100 W29 -4262 Y 21 16 17 94 W30 -4016 N 0 W31 -5349 N 0

An ecological approach to re-establishing Australian freshwater cod populations 172 FRDC Report 2003/034

Appendix 6. (cont.)

Figure 3. a) The capture locations of 31 wild M. macquariensis collected from the Murrumbidgee River at Narrandera by boat electro-fishing. White circles indicate capture locations that individual fish did not return to. Grey circles represent the location that an individual was captured from and returned to following release post- surgery. b) The maximum number of days taken for an individual to return to its capture location once released. c) Each numerical value represents records of an individual at its original capture site expressed as a proportion of all radio-tracks of that individual.

173 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 6. (cont.)

200 2000 W1 W2 W3 W4 0 1000 0 0 -200 0 -1000 -200 -400 -1000 -600 -2000 -400 -2000 -800 -3000 -600 -3000 -1000 -4000 -800 -1200 -4000

-1400 -5000 -5000 -1000 50000 4000 1000 W5 W6 W7 W8 0 40000 800 2000 30000 -1000 600 0 20000 -2000 400 10000 -2000 -3000 200 0 -4000 0 -4000 -10000

-5000 -20000 -6000 -200 2000 3000 W9 W10 W11 W12 2500 0 0 0 2000 -2000 -1000 -1000 1500 -4000 -2000 -2000 1000

Proximity site to release (m) -6000 500 -3000 -8000 -3000 0

-4000 -10000 -4000

5000 1000 W15 W 16 W 13 0 W14 4000 0 0 -2000 3000 -4000 -1000 -1000 -6000 2000 -2000 -2000 -8000 1000 -10000 -3000 -3000 -12000 0 -4000 -4000 -14000 -1000 -16000 -5000 -5000 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 01/09/03 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 01/09/03 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 Date

Figure 4. Movements of wild M. macquariensis (W1–W16) throughout the 13-month study relative to the release site (zero). The first point on each figure represents the capture location of an individual relative the release location and the dashed line joining the first two points represents the displacement distance of an individual prior to release.

An ecological approach to re-establishing Australian freshwater cod populations 174 FRDC Report 2003/034

Appendix 6. (cont.)

4000 5000 10000 1200 W 17 W18 W 19 W 20 2000 1000 4000 0 0 800 3000 -2000 -10000 600 -4000 2000 400 -6000 -20000 1000 -8000 200 0 -30000 -10000 0

-12000 -1000 -40000 8000 W 21 W22 W23 W24 6000 0 0 0 4000 -1000 -1000 -1000 2000 -2000 -2000 -2000 0 -3000 -3000 -3000 -2000

-4000 -4000 -4000 -4000

-6000 -5000 -5000 -5000

500 10000 3500 10000 W25 W26 W28 3000 W27 400 0 0 2500 300 -10000 2000 -10000 200 1500

Proximity to release site (m) site Proximity to release -20000 -20000 1000 100 500 -30000 -30000 0 0 -40000 -500 -40000 10000 14000 20000 W30 W31 W29 12000 01/09/03 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 0 10000 15000 8000 10000 -10000 6000 4000 5000 -20000 2000 0 0 -30000 -2000 -5000 -4000 -40000 -6000 -10000 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 01/09/03 01/09/03 01/11/03 01/01/04 01/03/04 01/05/04 01/07/04 01/09/04 01/11/04 Date

Figure 5. Movements of wild M. macquariensis (W17–W31) throughout the 13-month study relative to the release site (zero). The first point on each figure represents the capture location of an individual relative the release location and the dashed line joining the first two points represents the displacement distance of an individual prior to release.

175 An ecological approach to re-establishing Australian freshwater cod populations FRDC Report 2003/034

Appendix 6. (cont.) Table 3. Summary of the movements of 31 wild M. macquariensis radio-tracked over 13 months. Proximity of capture location to release site is shown as negative for downstream of the release site and positive for upstream of the release site. Total range calculations exclude capture locations. *One record from the downstream logger. ^Area estimate based on less than 10 fixes (omitted from mean). #Radio-tag verified as rejected/mortality through 24 hr radio-tracking. Values in brackets denote number of radio-locations used in estimates relating to each of the multiple home-ranges of that individual. Fish Total Proximity of Home Total Total range Linear home-range (m) Home ID fixes capture to d range (m) post-dispersal -range release site (m) Home Home Home (m2) (m) 1 2 3 W1 17 -1270 Yes 1334 97 97 - - 1377 W2 17 -4015 No 983 934 270 13 (9) - 62.5^ (3) W3 17 -1822 Yes 3997 2305 102 - - 1450 W4 15 -886 Yes 909 29 29 - - 341 W5 14 -4276 Yes 4281 18 18 - - 87 W6 17 -10351 No 40206 37909 159 17 (5) 105 - (2) (5) W7 10 3138 No 4207 14 14 - - - W8 17 872 Yes 903 17 17 - - 55 W9 15 -3582 No 304 25 25 - - 123 W10 19 -8033 No 759 759 150 - - 1478^ W11 16 -1190 Yes 3706 2449 24 - - 53 W12 15 222 Yes 2464 2271 34 (5) 9 (5) - - W13 17 4658 Yes 4742 13 13 - - 62 W14 11 -1190 No 14751 12 12 - - - W15# 17 -4000 No 542 19 - - - - W16 12 -4005 No 605 38 38 - - 132.5^ W17 15 -10332 No 3060 196 196 - - 3192 W18 10 4383 No 345 13 13 - - - W19 18* -5136 No 30815 30736 218 - - 4017 W20 12 963 Yes 957 8 8 - - 22.5^ W21 17 -4013 No 5997 18 18 - - 131 W22 17 -4278 Yes 4659 131 131 - - 1539 W23 17 -4634 Yes 4738 74 74 - - 553 W24 17 -4353 Yes 4422 120 120 - - 1392 W25 15 381 Yes 394 14 14 - - 92 W26 30* 1987 Yes 32891 32891 166 - - 4073 W27 17 3243 No 1548 113 113 - - 891 W28 18* 872 Yes 31687 31687 68 - - 485 W29 18* -4262 Yes 30815 26576 34 - - 418 W30 5 -4016 No 16497 16461 - - - - W31 21 -5349 No 20515 20515 81 - - 1836^ Mean 8840 6660 78 1070 SE 2151 2202 13 302

An ecological approach to re-establishing Australian freshwater cod populations 176 FRDC Report 2003/034

Appendix 6. (cont.)

Figure 6. The daytime home-range occupation by an individual M. macquariensis (W26) radio-tracked over a 13-month period as visualised by Kernel density contours (n= 29 radio-locations). Note: Additional radio-locations were taken for this individual opportunistically throughout the study period and are not temporally spaced at set intervals.

177 An ecological approach to re-establishing Australian freshwater cod populations. FRDC Report 2003/034

Appendix 6. (cont.)

100 100 (a) (b)

80 80

60 60

40 40 Frequency (%) 20 20

0 0 Open Woody Rock Open Undercut Emergent Single woody Single woody Multiple woody Tree roots water debris water banks macrophytes debris branching debris debris unknown not branching Habitat category Woody debris category 35 35 (c) (d) 30 30

25 25

20 20

15 15

Frequency (%) Frequency 10 10

5 5

0 0 >5 >40 0-4.9 5-9.9 0-0.49 10-14.9 15-19.9 20-24.9 25-29.9 30-34.9 35-39.9 0.5-0.99 1.0-1.49 1.5-1.99 2.0-2.49 3.0-3.49 4.0-4.49 2.49-2.99 3.49-3.99 4.49-4.99 Depth category (m) Distance from shore (m) Figure 7. Habitat use by wild M. macquariensis when located within a home-range (dispersal or large-scale movements excluded); use of (a) major habitats, (b) specific structural woody habitat categories, (c) depth of water and, (d) distance to shore based on five metre intervals.

An ecological approach to re-establishing Australian freshwater cod populations 178 FRDC Report 2003/034

Appendix 7. Movements of Individual C27 released in the Cotter River.

Figure 1. The radio-locations of Individual C27 in the Cotter River following release in September 2004. Boxes are used to summarise the spatial position over time with numbers corresponding to the number of days after release. Note: The radio-tag from this individual was retrieved from the shore 208 days after release with evidence of predation or scavenging by a common water rat Hydromys chrysogaster.

179 An ecological approach to re-establishing Australian freshwater cod populations. FRDC Report 2003/034

Appendix 8. A preliminary protocol for using underwater filming as a technique for studying releases of freshwater fish.

Ben Broadhurst and Brendan Ebner

Introduction The integration of underwater filming into freshwater ecological research is in its infancy in Australia. Methods should be transferable to some degree from applications elsewhere including from marine ecology (e.g. Mills et al., 2005). However, a number of characteristics of Australian freshwater ecosystems require consideration if the technique is to be applied successfully. In particular, periodic or permanent turbidity levels generally found in inland rivers are likely to affect visual detection of fish (Lagler, 1968). In these and other types of environments, it is possible to use alternatives including a dual-frequency identification sonar (DIDSONTM) acoustic camera. These units typically have a range of about 16 m ( Mueller et al., 2006) and can be used for individual species detection, although fishes under 75 mm in length can not be reliably detected or identified (Baumgartner et al., 2006). In Australia, water clarity is likely to be best in slightly saline areas, in relation to aquifers or in upland streams with well-vegetated catchments. However, within these habitats underwater video is still reliant on local factors that effect light availability and water clarity. These factors include shading from the riparian zone, sediment addition and sediment suspension. This appendix serves to a) outline factors that effect the application of underwater video monitoring in fresh waters and b) makes some preliminary recommendations taking into account these factors prior to applying underwater video, in relation to monitoring releases of freshwater fish in Australia.

Important factors Visibility Visibility in fresh waters can be affected by many factors including turbidity, turbulence, light levels and physical obstructions (Dolloff et al., 1996; Lucas & Baras, 2000; Frezza et al., 2003). These factors can remain constant in a waterway or vary in response to other factors including discharge and rainfall. Estimates of how these factors effect visibility is often site specific and variable over time and space. Therefore it is useful to measure visibility at a site prior to filming. Visibility can be reliably measured using a periscope to record visibility with a horizontal visual disc, a variation to the more common technique of viewing a Secchi disc in the vertical plane (Steel & Neuhausser, 2002). The minimum threshold for viewing will depend on the objectives of the study. Should these minimum requirements be met, there may also be a need to calibrate for variable visibility during the study. For instance, increased visibility enables a greater volume of water to be monitored, and this may lead to biased data among filming events or among study sites or cameras.

An ecological approach to re-establishing Australian freshwater cod populations 180 FRDC Report 2003/034

Appendix 8. (cont.)

Figure 1. Decision framework for video monitoring the release of fish into fresh waters.

Spatial resolution Figure 1 is used to illustrate different scales of viewing to provide a basis for considering some of the pragmatic issues relating to using underwater filming. Typical levels of visibility at a site govern to a large degree the maximum resolution of a camera. In high visibility waterways (visibility >3 m) it is possible to monitor multiple micro-habitats per camera. Micro-habitats are defined as areas within meso- habitats of homogenous substrate, depth and flow features (Frissel et al., 1986). In waterways with medium visibility (1–3 m), it may only be viable to monitor one micro-habitat per camera. In waters with low visibility (<1 m), only part of a micro- habitat may be monitored. We acknowledge that subjective thresholds have been used in this discussion, and that microhabitats can vary according to the ecosystem and to researcher definitions.

It is likely that the scale at which fish behaviour operates is dependent to a large degree on body size. For example ecological interactions involving fingerlings (e.g. competitive exclusion, predation) are more likely to operate at scales of tens of centimetres to a metre, comparable with their home-range sizes (Ebner, Clear, Beitzel and Godschalx, unpublished data). In contrast, interactions involving larger on-grown individuals are more likely to occur over metres to tens of metres (e.g. Chapter 4). For instance, a camera with a field of view spanning a few metres may only partly capture aggressive interactions involving two conspecifics with territories encompassing tens of metres. Banks of several cameras forming a continuous view of a larger area (~ 10 m) are probably apt for viewing the behaviour of larger on-grown fish.

Taxonomic resolution The identification of individuals to species level is a function of the quality of data, the skills of the observer, the speed of an individual’s travel and the number of morphometrically similar species present. This will be related to camera quality, visibility, lighting, and the size of individuals and the species that are present. Where species richness is high and/or congeneric species are present it may not be possible to distinguish all taxa to species level. Grouping similar species may be appropriate at these sites (for instance as was the case with the two species that were observed but not collected in the electro-fishing survey in chapter 2).

181 An ecological approach to re-establishing Australian freshwater cod populations. FRDC Report 2003/034

Appendix 8. (cont.)

Preliminary protocol for underwater filming in Australian freshwaters 1. Determine your study objectives and determine if these can be met with a less costly technique (note there is considerable time and expertise in processing and analysing video data). 2. Define the variables of interest. 3. Conduct a pilot study that includes filming appropriate fish (e.g. size, behaviour) and processing of film and data analysis not just the field component of the technique. 4. Develop an understanding of visibility and factors that effect visibility at the study sites. (measure visibility by Secchi depth prior to filming, obtain predictive data that indicate visibility e.g. discharge or rainfall).

Recommendations • A trial of underwater filming is used to monitor the short-term fate of fingerling releases in Australia.

• That expertise be developed and shared for using underwater filming to reveal aspects of the behaviour of Australian freshwater fishes. In particular this technique is likely to be of high value where behavioural interactions are the focus (e.g. competitive interactions with invasive species, rapid mortality following release of threatened species).

References Baumgartner, L. J., Reynoldson, N., Cameron, L. & Stanger, J. (2006). Assessment of a Dual-Frequency Identification Sonar (DIDSON) for application in fish migration studies. (NSW Department of Primary Industries, Cronulla).

Dolloff, A., Kershner, J. & Thurow, R. (1996). Underwater observation. In, Fisheries techniques: second edition (Murphy, B. R. & Willis, D. W., eds), pp. 533–554. (American Fisheries Society, Maryland).

Frezza, T. L., Carl, L. M. & Reid, S. M. (2003). Evaluation of a portable underwater video camera to study fish communities in two Lake Ontario tributaries. Journal of Freshwater Ecology 18(2), 269–276.

Frissel, C. A., Liss, W. L., Warren, C. E. & Hurley, M. D. (1986). A hierarchical framework for stream habitat classification: viewing streams in a watershed context. Environmental Management 10(2), 199–124.

Lagler, K. F. (1968). Capture, sampling and examination of fishes. In, Methods for assessment of fish production in fresh waters. IBP Hanbook No. 3. (Ricker, W. E., ed), pp. 7–45. (Blackwell Scientific Publications, Oxford).

Mills, D. J., Verdouw, G. & Frusher, S. D. (2005). Remote multi-camera system for in situ observations of behaviour and predator/prey interaction of marine benthic macrofauna. New Zealand Journal of Marine and Freshwater Research 39, 347–352.

An ecological approach to re-establishing Australian freshwater cod populations 182 FRDC Report 2003/034

Appendix 8. (cont.)

Mueller, R. P., Brown, R. S., Hop, H. & Moulton, L. (2006). Video and acoustic camera technique for studying fish under ice: a review and comparison. Reviews in Fish Biology and Fisheries 16, 213–226.

Murphy, B. R. & Willis, D. W. (1996). Fisheries techniques: second edition. American Fisheries Society, Maryland.

Ricker, W. E. (1968). Methods for assessment of fish production in fresh waters. IBP Hanbook No. 3. Blackwell Scientific Publications, Oxford.

Steel, E. A. & Neuhausser, S. (2002). Comparison of methods for measuring visual water clarity. Journal of the North American Benthological Society 21(2), 326–335.

183 An ecological approach to re-establishing Australian freshwater cod populations.