Wat er Assessment Aquatic Ecology Report Series
Response of the Lake Crescent and Lake
Sorell ecosystems to water level variability during 2009: a period of drought and post-
drought recovery
October 2010 ISSN: 1835-9523 Report No. WA 10/01
Water Assessment Branch W at er and Marine Resources Division D e p a r t m ent of Primary Industries, Parks, Water and Environment
Copyright Notice: Material contained in the report provided is subject to Australian copyright law. Other than in accordance with the Copyright Act 1968 of the Commonwealth Parliament, no part of this report may, in any form or by any means, be reproduced, transmitted or used. This report cannot be redistributed for any commercial purpose whatsoever, or distributed to a third party for such purpose, without prior written permission being sought from the Department of Primary Industries, Parks, Water and Environment, on behalf of the Crown in Right of the State of Tasmania.
Disclaimer: Whilst DPIPWE has made every attempt to ensure the accuracy and reliability of the information and data provided, it is the responsibility of the data user to make their own decisions about the accuracy, currency, reliability and correctness of information provided. The Department of Primary Industries, Parks, Water and Environment, its employees and agents, and the Crown in the Right of the State of Tasmania do not accept any liability for any damage caused by, or economic loss arising from, reliance on this information.
Preferred Citation: DPIPWE (2010). Response of the Lake Crescent and Lake Sorell ecosystems to water level variability during 2009: a period of drought and post-drought recovery. Water Assessment Aquatic Ecology Report Series, Report No. WA 10/01. Water and Marine Resources Division. Department of Primary Industries, Parks, Water and Environment, Hobart, Tasmania.
Contact Details: Department of Primary Industries, Parks, Water and Environment Water Assessment 13 St Johns Avenue, New Town Phone: 03 6233 6833 Web: www.dpiw.tas.gov.au Email: [email protected]
Cover Page Images: Top left: Lake Crescent from Tea Tree Point boat ramp, April 2009; top right: Boathouse Shore, Lake Crescent, May 2009; bottom left: Clyde Marsh, November 2009; bottom right: Sexually mature female golden galaxias (Galaxias auratus) from Lake Sorell, July 2009.
The Department of Primary Industries, Parks, Water and Environment The Department of Primary Industries, Parks, Water and Environment provides leadership in the sustainable management and development of Tasmania’s resources. The Mission of the Department is to advance Tasmania’s prosperity through the sustainable development of our natural resources and the conservation of our natural and cultural heritage for the future. The Water and Marine Resources Division provides a focus for water management and water development in Tasmania through a diverse range of functions including the design of policy and regulatory frameworks to ensure sustainable use of the surface water and groundwater resources; monitoring, assessment and reporting on the condition of the State’s freshwater resources; facilitation of infrastructure development projects to ensure the efficient and sustainable supply of water; and implementation of the Water Management Act 1999, related legislation and the State Water Development Plan.
Response of Lake Crescent and Lake Sorell to water level variability during 2009
Summary
The interconnected Lake Crescent and Lake Sorell are impounded natural lakes which are located 1 km apart in the upper reaches of the Clyde River catchment (~800 m a.s.l.) in the south-east of the Tasmanian Central Plateau. This lake system contains several distinctive and significant environmental values, including extensive littoral wetlands (one of which is Ramsar-listed), and the freshwater fish golden galaxias (Galaxias auratus Johnston) which is endemic to lakes Crescent and Sorell and listed under Tasmanian and Commonwealth threatened species legislation. Because of these and other environmental values, and their shallow and highly productive states, lakes Crescent and Sorell are considered to be a regionally unique lake system.
Between 2000 and 2008, a substantial volume of scientific work was undertaken to examine the influences of low water levels and patterns of water level fluctuations on the Crescent-Sorell ecosystem. This work showed that these components of the hydrological regime of this system directly and indirectly affect three broad aspects of the Crescent-Sorell ecosystem: (1) important aquatic habitats (i.e. wetlands and rocky shores), (2) water quality, and (3) G. auratus populations. The findings of studies during 2000-2003 were used to guide the formulation of water level operating rules in the water management plan (WMP) for lakes Crescent and Sorell, which took effect in November 2005.
Since the late 1990s, the Crescent-Sorell ecosystem has been severely stressed by prolonged periods of low water levels, which have been primarily caused by below average and unseasonal rainfall (including significant El Niño-induced drought events). While there was a slight increase in water levels in both lakes during 2003-2005, between 2007 – early 2009, water levels fell substantially in both lakes. These low water levels were thought to have the potential to further degrade water quality and habitats in the lakes, and cause additional declines in G. auratus populations.
This study took advantage of a unique opportunity to examine several components of the Crescent- Sorell ecosystem during a period (i.e. 2009 – early 2010) of extreme water level variation, which included very low water levels in summer-autumn 2009, followed by rapid and substantial seasonal rises in levels during winter-spring 2009. These conditions also provided quite different circumstances to undertake the second environmental water level manipulation in this lake system; with the first being undertaken in 2007 when water level minima were higher and only moderate natural seasonal rises in levels occurred in Lake Crescent and Lake Sorell. Similar to during 2007, during 2009 the response of G. auratus populations (at several life stages) in Lake Crescent and Lake Sorell to the water level manipulation was monitored. Using long-term data sets, which included data collected during the present study, several previously established correlations and relationships between water levels and parameters of the Crescent-Sorell ecosystem were re-
i Response of Lake Crescent and Lake Sorell to water level variability during 2009 examined. Furthermore, the resistance/resilience of some environmental values in lakes Crescent and Sorell in relation to impacts of low water levels and recovery during a period of substantial rises in levels was investigated.
This study confirmed many findings from previous studies in lakes Crescent and Sorell. It also provided new information in some areas which is highly relevant to the management of water levels in these lakes, and hence water resources in the Clyde River catchment. Principally, this study furthered understanding of: (1) water level patterns in Lake Crescent and Lake Sorell, (2) the effects of water levels on water quality and the ecosystems of these lakes, (3) relationships between water levels and the inundation and condition of littoral habitats in these lakes, (4) the spawning strategy of G. auratus, (5) relationships between water levels and the abundance of juvenile G. auratus, (6) the status of G. auratus populations, and (7) the environmental benefits of water level manipulations in these lakes.
Based on the findings of the present study and previous studies, components of the water level regimes of Lake Crescent and Lake Sorell which are important to the management of their ecosystems were synthesized. In this synthesis, water level thresholds which related to: (1) the inundation of important littoral habitats (rocky shores and wetlands) in Lake Crescent and Lake Sorell, (2) maintaining water quality (particularly low turbidity levels) in Lake Crescent and Lake Sorell, and (3) allowing mass spawning of the G. auratus population in Lake Crescent, were identified. Furthermore, the importance of maintaining water level fluctuations in lakes Crescent and Sorell that follow historical seasonal patterns (which relate to climatic conditions in the area) is also specified. Based on this synthesis, to sustainably manage the Crescent-Sorell ecosystem it is recommended that the:
Critical minimal operating water level in Lake Crescent is 802.200 m AHD Critical minimal operating water level in Lake Sorell is 803.200 m AHD Preferred minimum operating water level in Lake Crescent is 802.700 m AHD Preferred minimum operating water level in Lake Sorell is 803.500 m AHD Preferred operating range for water levels in Lake Crescent is 802.200-803.800 m AHD Preferred operating range for water levels in Lake Sorell is 803.500-804.360 m AHD.
The findings of the present study, and those of DPIW (2008), have demonstrated that appropriately timed releases of water from Lake Sorell to Lake Crescent can be used to manage the Crescent- Sorell ecosystem. This innovative approach has been trialled twice, under quite different water level conditions in these lakes, and positive responses in environmental parameters in the system on both occasions show that it provides a powerful tool for managing these lakes. Therefore, additional to the environmentally important components of the water level regimes of lakes Crescent and
ii Response of Lake Crescent and Lake Sorell to water level variability during 2009
Sorell, where appropriate circumstances occur, it is also recommended that water level manipulations be routinely employed to assist management of the ecosystems of these lakes. The potential benefits of water level manipulations and key factors which should be considered when making decisions regarding their use are detailed.
Since 2000, studies in Lake Crescent and Lake Sorell have provided baseline data for several parameters of the ecosystems of these lakes, and in some instances long-term data sets have been compiled. Recently, these data sets have been invaluable to decision making associated with routine management of water levels in lakes Crescent and Sorell, particularly during periods of low water levels. Furthermore, knowledge of the status of water levels, water quality, littoral wetlands and G. auratus populations was crucial to formulating water level manipulation strategies in 2007 and 2009, and this information will also be important for future decisions regarding manipulations. Therefore, to ensure management of water resources in lakes Crescent and Sorell (and the Clyde River catchment) is environmentally sustainable, it is recommended that water levels, water quality, littoral wetlands and G. auratus populations continue to be monitored. The details of monitoring strategies are outlined and further applied research topics are proposed to address knowledge gaps that relate to the management of water levels in lakes Crescent and Sorell.
iii Response of Lake Crescent and Lake Sorell to water level variability during 2009
Table of Contents
Summary ...... i Acknowledgments ...... iv Glossary and abbreviations ...... iv
1. Introduction ...... 1 1.1 Background information ...... 1 1.2 Recent environmental conditions in Crescent-Sorell system ...... 7 1.3 Study approach and aims ...... 9
2. Methodology ...... 12 2.1 Study site ...... 12 2.2 Hydrological manipulation ...... 15 2.3 Physico-chemical parameters ...... 16 2.4 Littoral habitat surveys ...... 17 2.5 Littoral golden galaxias egg sampling ...... 18 2.6 Pelagic juvenile golden galaxias sampling ...... 19 2.7 Littoral adult golden galaxias sampling ...... 20 2.8 Data analysis ...... 21
3. Results ...... 25 3.1 Water level regime ...... 25 3.2 Physico-chemical parameters ...... 35 3.3 Habitat conditions and water level – induced alterations ...... 50 3.4 Water levels and area of inundated littoral rocky substrate in Lake Crescent ...... 65 3.5 Water levels, and physical structure and processes of littoral areas of rocky substrate ...... 74 3.6 Spawning activity of golden galaxias ...... 78 3.7 Golden galaxias larval emergence and abundance ...... 83 3.8 Golden galaxias larval abundance and water level fluctuations ...... 85 3.9 Abundance of golden galaxias young-of-the-year and water level fluctuations ...... 89 3.10 Relative abundance of golden galaxias populations ...... 91 3.11 Size and age structure of golden galaxias populations ...... 92
4. Discussion ...... 97 4.1 Water level patterns in Lake Crescent and Lake Sorell ...... 97 4.2 Effects of water levels on physico-chemical parameters ...... 98 4.3 Effects of poor water quality on the Crescent-Sorell ecosystem ...... 101 4.4 Water levels, and the inundation and condition of littoral habitats ...... 103 4.5 Spawning activity of golden galaxias: timing, habitat availability and water level effects ...... 106
ii Response of Lake Crescent and Lake Sorell to water level variability during 2009
4.6 Abundance of young-of-the-year golden galaxias and water level fluctuations ...... 108 4.7 Status of golden galaxias populations ...... 109 4.8 Environmental benefits of water level manipulation during 2009 ...... 111
5. Conclusion and Recommendations ...... 115 5.1. Key findings ...... 115 5.2 Environmental management of water levels in lakes Crescent and Sorell ...... 119
6. References ...... 129
7. Appendices ...... 135
iii Response of Lake Crescent and Lake Sorell to water level variability during 2009
Acknowledgments
Scott Hardie designed and managed this study, and prepared this report. Adam Uytendaal assisted with the design of the water quality sampling regime and interpretation of resulting data. Aerial photography was undertaken by the Information and Land Services Division (ILS) of the Department of Primary Industries, Parks, Water and Environment (DPIPWE), and Malcolm Crawford and Andrew Menzies co-ordinated this work and processed the images. Spatial data analyses which were used to derive relationships between water levels and the inundation of the rocky shores in Lake Crescent were undertaken by Mark Chilcott (ILS, DPIPWE). Mark and Stuart Fletcher (ILS, DPIPWE) guided the direction of these analyses. Kate Hoyle (DPIPWE) undertook spatial data analyses which were used to estimate the surface areas, and extents of dewatering and inundation of the entire lakes and selected littoral habitats. Kate also prepared maps of turbidity readings in littoral wetlands of the lakes. Lois Koehnken (Technical Advice on Water) helped synthesise the conceptual understanding of the physical structure and processes on rocky shores in the lakes. Brett Littleton (ILS, DPIPWE) prepared the conceptual diagram of the rocky shores. Field support was provided by Chris Wisniewski (coordination), Terry Byard, Kevin MacFarlane, Michael Mayrhofer, Daniel Elson, and Andrew Taylor (Inland Fisheries Service; IFS), and Melanie Evans (DPIPWE). David Horner (DPIPWE) and Melanie processed egg and larval fish samples. The Water Assessment Branch would like to thank Gunns Ltd for access to the south-eastern shore of Lake Crescent during autumn 2009 (when boating on the lake was not possible). This study was conducted with financial support from the Tasmanian State Government via DPIPWE and IFS.
Glossary and abbreviations
Aeolian: Pertaining to the geomorphic effects of moving air, especially wind currents.
Benthic habitat: Area on the bottom of a lake or river.
Cohort: A group of fish that were born during the same breeding; thus, are of approximately the same body length (and age).
Complex habitat: Environments in freshwater ecosystems which contain many interstitial spaces. These are typically formed by clusters of rocks, wood debris or aquatic vegetation.
DPIPWE: Department of Primary Industries, Parks, Water and Environment.
Epibenthic habitat: Area directly over or above the bottom of a lake or river.
Euphotic depth: The depth at which light intensity falls to 1% of the value at the surface, thus, plants have no net energy gain from photosynthesis, and therefore cannot grow.
IFS: Tasmanian Inland Fisheries Service.
iv Response of Lake Crescent and Lake Sorell to water level variability during 2009
Galaxiid: Freshwater fish belonging to the Family Galaxiidae. These fishes are scaleless and generally small-sized (<250 mm in length) and are indigenous to several land masses in the Southern Hemisphere including Australia, New Zealand, South America, Africa and New Caledonia.
GIS: Geographical Information System; a set of tools that captures, stores, analyzes, manages and presents data that are linked to locations (i.e. spatial data).
Lentic: Pertaining to or living in standing water (i.e. habitats such as lakes and lagoons).
Littoral habitat: Area around the shore of a lake or river.
Lotic: Pertaining to or living in flowing water (i.e. habitats such as creeks and rivers).
Macroinvertebrate: Invertebrate (without a backbone) animals which can be seen with the naked eye. In lakes and rivers, common macroinvertebrates are insects, crustaceans, worms and snails.
Macrophyte: Vascular plant that is visible with the naked eye. In this report, this terms refers to aquatic plants that grow in freshwater environments.
Physico-chemical variables: Parameters which relate to the physical and chemical properties of water (e.g. water temperature and turbidity).
Recruitment: The influx of juvenile fish into the adult proportion of a population.
Substrate: The surface on the bottom of a lake or river.
Threatened species Recovery Plan: A strategy under which endangered flora and fauna (individual species or groups of species) are managed in Tasmania.
WMP: Water Management Plan; a strategy under which water resources associated with freshwater ecosystems in Tasmania are managed by DPIPWE.
YOY: Young-of-the-year; juvenile fish in the first year of their life (i.e. <1 year old).
v Response of Lake Crescent and Lake Sorell to water level variability during 2009
1. Introduction
1.1 Background information
1.1.1 Lakes Crescent and Sorell, and management of water resources in the Clyde River catchment
The interconnected Lake Crescent (42°10’S, 147°10’E) and Lake Sorell (42°6’S, 147°10’E) are impounded natural lakes which are located 1 km apart in the upper reaches of the Clyde River catchment (~800 m a.s.l.) in the south-east of the Tasmanian Central Plateau (Figure 1.1). The Clyde catchment covers an area of 1130 km2, while the combined catchment of the lakes is 207 km2 (inclusive of the lakes) (CFEV, 2005). Lakes Crescent and Sorell are quite large in comparison to other lakes in Tasmania (surface areas of approximately 23 and 56 km2, respectively; CFEV, 2005), relatively shallow (mean depths <3.5 m at full supply levels), and turbid (mean total turbidities between November 1991 and January 2010: Lake Crescent = 131 NTU; Lake Sorell = 89 NTU). The Conservation of Freshwater Ecosystem Values (CFEV) database (CFEV, 2005) considers lakes Crescent and Sorell to be mesotrophic upland lakes, which are shallow and productive (in terms of the generation of biomass by primary and secondary consumers in the aquatic ecosystem). Furthermore, according to this database, the waterbodies of both lakes have very high conservation value.
Extensive littoral wetlands are a unique feature of the Crescent-Sorell system, especially in comparison to other lakes on the Tasmanian Central Plateau. Three areas of littoral wetland (Kemps/Kermodes Marsh, Robertson Marsh and Silver Plains Marsh) occur on the western shores of Lake Sorell, and collectively these areas account for 10% of the surface area of the lake at the full supply level (CFEV, 2005; Figure 1.1; Appendix 1). Two areas of littoral wetland (Clyde Marsh and Interlaken Lakeside Reserve) occur of the western shores of Lake Crescent, and collectively these areas account for 18% of the surface area of the lake at the full supply level (CFEV, 2005; Appendix 1). The wetlands of lakes Crescent and Sorell are shallow, temporary, freshwater systems, and are some of the largest areas of this type of wetland in Tasmania. The wetlands hold significant conservation value, with one being listed under the Ramsar convention (Ramsar, 2002), two being listed under the Directory of Important Wetlands in Australia and two being considered of State Significance (Heffer, 2003c; Heffer, 2003b). According to the CFEV database (CFEV, 2005), the littoral wetlands of the Crescent-Sorell system have high to very high conservation value.
1 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 1.1 Map of the Clyde River catchment, Tasmania showing Lake Crescent and Lake Sorell, and their tributaries and littoral wetlands (darkest shading). Locations of the townships of Bothwell and Hamilton in the catchment are also shown.
2 Response of Lake Crescent and Lake Sorell to water level variability during 2009
The water resources of the Crescent-Sorell system has been used to service the water supply needs of two downstream townships (Bothwell and Hamilton) and agricultural water users in the Clyde Valley since the early 1800s (Mason-Cox, 1994). To provide greater reliability of water to users in the Clyde Valley, both lakes were impounded in the 1830s (Mason-Cox, 1994), and since then their dams and associated infrastructure have undergone various modifications (e.g. Wisniewski, 2003). Since November 2005, the water resources in lakes Crescent and Sorell and the Clyde River have been managed by the Department of Primary Industries, Parks, Water and Environment (DPIPWE) largely in accordance with Water Management Plans (WMPs) (DPIWE, 2005b; DPIWE, 2005a) which were developed by DPIPWE through a comprehensive consultative process involving key stakeholders and the community in the catchment.
The environmental provisions within the current water management plans for lakes Crescent and Sorell (DPIWE, 2005a) and the Clyde River (DPIWE, 2005b) are supported by the findings of a considerable volume of technical work and applied research which was/has been undertaken prior to, and after the adoption of the WMPs (Davies and Pinto, 2000; DPIW, 2008; Hardie, 2003a; Hardie, 2003b; Hardie, 2007; Hardie et al., 2004; Hardie et al., 2005; Hardie et al., 2006a; Hardie et al., 2007a; Hardie et al., 2007b; Smith and Mendel, 2009; Uytendaal, 2003; Uytendaal, 2006; Uytendaal et al., 2003; Heffer, 2003c; Heffer, 2003b). Collectively, this work has been conducted by DPIPWE, the Tasmanian Inland Fisheries Service (IFS), private consultants, and researchers at the University of Tasmania. The environmental objectives of the water resource management strategies in the Clyde catchment (DPIWE, 2005b; DPIWE, 2005a) largely focus on maintaining three broad aspects of the Crescent-Sorell ecosystem: (1) important aquatic habitats (i.e. wetlands and rocky shores), (2) water quality, and (3) populations of the threatened freshwater fish golden galaxias (Galaxias auratus Johnston). The status of these three components of the Crescent-Sorell ecosystem and their responses to water level variability during 2009 are the focus of this study.
1.1.2 Golden galaxias (Galaxias auratus)
Galaxias auratus (Figure 1.2) is a fundamental environmental value and functional component of the unique Crescent-Sorell ecosystem (Hardie, 2003a; Hardie, 2007; Uytendaal, 2006), and the need to provide appropriate environmental conditions for this species is a significant component of water management strategies in the Clyde catchment (DPIWE, 2005a; DPIWE, 2005b). This species is endemic to the interconnected lakes Crescent and Sorell and their associated wetlands, and, to a lesser extent, the lower reaches of their tributaries and upper reaches of the Clyde River (i.e. the outflow of Lake Crescent; Figure 1.1) (Hardie, 2003a; Hardie et al., 2004). Currently, only three populations of G. auratus exist: two natural populations in lakes Crescent and Sorell, and one translocated population in a small off-stream agricultural water storage (i.e. farm dam) in the Clyde River catchment (Hardie, 2003a; IFS, unpubl. data); one other translocated population in a farm dam in the Clyde catchment is thought to have recently died out due to low water levels in the dam,
3 Response of Lake Crescent and Lake Sorell to water level variability during 2009 although this is yet to be confirmed (IFS, unpubl. data). Given the relatively small size of the viable translocated population (i.e. waterbody has a surface area of <1 ha), the wild populations in lakes Crescent and Sorell provide the strong-hold for this species. Galaxias auratus is listed as ‘rare’ under the Tasmanian Threatened Species Protection Act 1995, ‘endangered’ under the Commonwealth Environmental Protection and Biodiversity Conservation Act 1999, and is managed in accordance with a Tasmanian Threatened Species Recovery Plan (Threatened Species Section, 2006). The primary threat to this species is low water levels and alterations to seasonal water level fluctuation patterns. Predation and/or competition with the introduced fish species, brown trout (Salmo trutta L.) and European carp (Cyprinus carpio L.), and degradation of shoreline habitats (including wetlands) also threaten G. auratus.
Figure 1.2 Sexually mature female golden galaxias (Galaxias auratus) from Lake Sorell, July 2009.
The life history of G. auratus has recently been intensively studied (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2004; Hardie et al., 2005; Hardie et al., 2006a; Hardie et al., 2007a; Hardie et al., 2007b); this work has resulted in G. auratus having the most thoroughly documented life cycle of any galaxiid species in Tasmania. Galaxias auratus has a non-diadromous (i.e. non- migratory) life history; thus, it completes its life cycle in lentic waters using littoral-benthic, epibenthic and pelagic habitats at different life stages. Spawning takes place in late autumn – early spring (peak activity in winter – early spring) on rocky shores and possibly in wetland habitat when available. Spawning mostly occurs at water temperatures c. <7 C and is triggered by rising water levels. Fertilised eggs, which are ~1.5 mm in diameter, transparent and adhesive, are scattered over littoral cobble substrata or aquatic vegetation at depths of c. 0.2-0.6 m. Fertilised eggs are thought to incubate for approximately 30-45 days in the wild. Larval hatching peaks during late winter – spring, and newly hatched larvae are 6-7 mm in length and free swimming. Juveniles occupy pelagic habitats until mid-summer to autumn (when they are ~40 mm in length) after which
4 Response of Lake Crescent and Lake Sorell to water level variability during 2009 they move into adult littoral habitats. Whilst G. auratus may attain up to 10 years of age, most live for <3 years and many males live for only 1-2 years.
1.1.3 Effects of water level regimes on the Crescent-Sorell ecosystem
Similar to flow regimes in lotic (e.g. riverine) ecosystems (Arthington et al., 2010; Arthington and Pusey, 2003; Poff et al., 1997), water level regimes in lentic (e.g. lake) ecosystems (Wantzen et al., 2008) can have a profound influence on their character and condition. This is particularly the case in shallow lakes, where substantial areas of substrate can be exposed to wind-induced wave action or dewatered by reductions in water level (Scheffer, 2004). Water level fluctuations can affect habitat availability (to aquatic fauna) (Winfield, 2004), water quality (Scheffer, 2004; Scheffer et al., 1993; Scheffer and van Nes, 2007), invertebrate communities (Thompson and Ryder, 2008; Valdovinos et al., 2007; Scheifhacken et al., 2007) and fish populations (Gafny et al., 1992; Havens et al., 2005; Probst et al., 2009; Sammons et al., 1999; Winfield et al., 2004).
Between 2000 and 2008, a substantial volume of scientific work was undertaken to examine the influences of low water levels and patterns of water level fluctuations on the Crescent-Sorell ecosystem. These aspects of the hydrological regime of this system directly and indirectly affect important habitats and aquatic vegetation communities (i.e. rocky shores and macrophytes in wetlands) (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b; Heffer, 2003c; Smith and Mendel, 2009), water quality (Uytendaal, 2003; Uytendaal, 2006), and G. auratus populations (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b).
While the extensive littoral wetlands of the Crescent-Sorell system are important ecosystems in their own right, structural complex habitats (such as macrophytes and rocky substrates) are also important to the entire ecosystem of the interconnected lakes because they contain diverse and abundant macroinvertebrates communities (Hardie, 2003b) and are critical to G. auratus populations. Galaxias auratus prefers these habitats for feeding and refuge (Hardie, 2003a; Hardie et al., 2006a; Stuart-Smith et al., 2006; Stuart-Smith et al., 2007b; Stuart-Smith et al., 2007a; Stuart- Smith et al., 2008), and most importantly, spawns in these areas (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b).
The condition and extent of inundation, hence availability to aquatic fauna, of complex habitats in lakes Crescent and Sorell is strictly controlled by water levels. Firstly, littoral wetlands in both lakes become largely disconnected from the main bodies the lakes at defined water level thresholds. In Lake Sorell, at levels <803.60 m AHD, the three areas of littoral wetlands (which account for 10% of the surface area of the lake at the full supply level; CFEV, 2005; Appendix 1) begin to be dewatered, whereas the two littoral wetlands (which account for 18% of the surface area of the lake at the full supply level; CFEV, 2005; Appendix 1) in Lake Crescent begin to be dewatered at levels <802.70 m AHD (Appendix 2; Heffer, 2003). Secondly, in Lake Crescent, at levels <802.20 m AHD,
5 Response of Lake Crescent and Lake Sorell to water level variability during 2009 areas of rocky shore habitat (which account for <2% of the surface area of the lake at the full supply level; Appendix 3) become dewatered (Appendix 2; Hardie et al., 2007b). Conversely, in many areas in Lake Sorell, rocky substrates extend across the lake’s basin; therefore, access to this important habitat is not as dependent on water levels in this lake.
Water quality and phytoplankton communities in lakes Crescent and Sorell are strongly influenced by water levels (Uytendaal, 2003; Uytendaal, 2006; Uytendaal et al., 2003). Low water levels cause sediment resuspension to become a driving factor which governs levels of suspended sediment, turbidity, light penetration, nutrient concentrations in the water column and phytoplankton biomass (Uytendaal, 2003). Water quality conditions in the lakes since 1998 provide a clear example of the profound influence of low water levels in the lakes. During this period, high levels of physical disturbance of sediments and the resulting reduction in euphotic depth has led to a complete loss of in-lake aquatic macrophytes in Lake Sorell, which has removed a valuable resource for aquatic fauna, and eliminated an important stabilising habitat for sediments and water clarity (Uytendaal, 2003; Uytendaal, 2006). Based on relationships between water levels and turbidity levels, and shear stress modelling in lakes Crescent and Sorell, three water level thresholds have been identified as being critical to water quality in the lakes (Uytendaal, 2003). Water levels <802.7 m AHD and <803.5 m AHD cause total turbidity levels to increase in Lake Crescent and Lake Sorell, respectively; and, in Lake Sorell water levels <803.2 m AHD cause colloidal turbidity to increase (Appendix 4).
Hydrological variables have a significant influence on the spawning and subsequent recruitment of G. auratus populations (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b). Three key hydrological factors affect recruitment of this species: (1) water level-dependent availability of spawning habitats, (2) rising levels during spawning periods, and (3) the magnitude of rises during spawning and the incubation of eggs. Hardie (2007) presents detailed explanations of these processes; however, brief descriptions are included here.
Firstly, as mentioned previously, water levels control the availability (to G. auratus) of important littoral habitats in the Crescent-Sorell system. For G. auratus to be able spawn, it is essential that suitable habitats (i.e. rocky shores or macrophytes) are present in the inundated region of the lakes during the spawning season (winter – early spring). In Lake Crescent, during periods of relatively low water levels (i.e. when adjacent wetlands are not inundated), the four defined areas of rocky shore habitat on the western and eastern margins of the lake (Appendix 3) are the only habitats that are available to G. auratus for spawning. When available, G. auratus is likely to also use aquatic vegetation in the wetlands and the main bodies of both lakes for spawning, as this species has been found to spawn on macrophytes in a farm dam in the Clyde River catchment (Hardie et al., 2007b).
6 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Secondly, not only do suitable habitats need to be present, but appropriate hydrological conditions need to occur for spawning to take place. Analysis of a 3-year monthly biological record of populations of G. auratus in lakes Crescent and Sorell (during 2000-2002) along with environmental data, such as water levels and water temperatures, showed that this species spawns during periods of rising water levels (Hardie et al., 2007b). These conditions are likely to provide important stimuli to induce G. auratus to spawn. Subsequent work in 2007 confirmed the importance of these conditions for spawning (DPIW, 2008).
Thirdly, in Lake Crescent, monitoring of the abundance of juveniles in breeding seasons over six breeding seasons has shown that early recruitment strength (i.e. the number of juveniles which emerge and survive) is strongly and positively related to the magnitude of the rise in water level during the spawning period of G. auratus (May-September) (Hardie, 2007; DPIW, 2008). Therefore, for G. auratus in Lake Crescent to spawn during periods of low water levels, it is not a simple matter of the water level being above the level which inundates rocky shores at spawning time (802.20 m AHD), but the amount the level rises above this threshold during this period is also very important and controls the abundance of subsequent cohorts (DPIW, 2008; Hardie, 2007).
Based on these relationships between water levels and G. auratus populations, during 2007, water levels in the Crescent-Sorell system were manipulated to assist spawning of G. auratus population in Lake Crescent (DPIW, 2008). DPIW (2008) provides a detailed account of that work, and the present study builds on the findings of several aspects of that work.
1.2 Recent environmental conditions in Crescent-Sorell system
1.2.1 Hydrology and habitats
Between the late 1990s and December 2008, the Crescent-Sorell ecosystem had been severely stressed by prolonged periods of low water levels (Appendix 2) which had been associated with below average and unseasonal rainfall (including significant El Niño-induced drought events; http://www.bom.gov.au/climate/enso/enlist/index.shtml, accessed 5 October 2010), water use by downstream users (DPIPWE, unpubl. data), and water level alterations in the lakes to assist management of a pest fish species, European carp (Cyprinus carpio L.) (IFS, unpubl. data). Water levels in both lakes during this period were mostly below, or only marginally above, the preferred operating ranges for the lakes according to their WMP (DPIWE, 2005a; Appendix 5). These reductions in water levels have caused major alterations to habitats in the lakes including: long-term drying of extensive littoral wetlands (Heffer, 2003c; Smith and Mendel, 2009; Appendices 1 and 2); frequent dewatering of rocky shores (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Appendices 2 and 3), disappearance of in-lake macrophytes (Uytendaal, 2003; Uytendaal, 2006), and poor water quality (i.e. elevated turbidity levels) (DPIW, 2008; Uytendaal, 2003; Uytendaal, 2006; Appendix 4).
7 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Following record (post-1970) water level minima in 2000 in Lake Sorell and 2001 in Lake Crescent, there was a slight increase in levels in both lakes during 2003-2005, but levels fell substantially in both lakes during the 2006/07 summer-autumn period (Appendix 2). These poor hydrological conditions in early 2007 led to the water level manipulation study that was undertaken during winter – early summer 2007 (DPIW, 2008). That study successfully manipulated the water levels in the lakes to assist spawning of the galaxiid population in Lake Crescent, and documented the condition of the lakes and the status of the G. auratus populations at that time. However, following that work, water levels continued to decrease in both lakes during late 2007 – 2008, with levels falling substantially below the critical minimum level of 802.20 m AHD in Lake Crescent (the level at which the last available spawning habitat, rocky shores, is dewatered) and remained below that level during the 2008 breeding season (Appendix 2). Similar low water levels looked likely to occur again in the 2009 breeding season, with predictions of very low levels in both lakes during summer- autumn 2009 (DPIPWE, unpubl. data). As a result of the reduction in water levels in both lakes, turbidity levels also began to increase markedly (Appendix 4). Additionally, these low water levels also extended the period during which the wetlands in the system had not been properly inundated from high water levels in the lakes (Appendices 1 and 2).
1.2.2 Golden galaxias populations
Additional to low water levels and subsequent poor habitat conditions, monitoring of G. auratus populations in lakes Crescent and Sorell between 2001 and 2008 (at mostly annual intervals) shows that, during this period, relative abundance of both populations has fluctuated (Appendix 6). It is unclear what causes inter-annual variability in the catches of juvenile and adult fish in the Lake Sorell population. However, in the Lake Crescent population, the variability is closely associated with water levels fluctuations, as there is a strong positive relationship between the magnitude of the seasonal water level rise and recruitment strength of larval – early stage juveniles (DPIW, 2008; Hardie, 2007).
This is supported by length frequency data of the fish collected during population monitoring (Appendix 7). During late summer – autumn (when population monitoring was undertaken), young- of-the year (YOY) fish (<70 mm in length) from the previous breeding season move into adult littoral habitats, thus, are able to be captured by netting in these areas. Appendix 7 shows that in Lake Sorell during 2000-2007, recruitment (as indicated by the YOY cohort) has been relatively strong in all years that were sampled; however, in Lake Crescent, recruitment has been more variable (e.g. poor in 2000/01 and strong in 2003/04). Most critically, there was obviously poor recruitment in 2006/07 (Appendix 7) and this was also likely to be the case in 2008/09 when water levels in Lake Crescent were well below the critical minimum level (802.20 m AHD). So, even if water levels were to increase during 2009, with two of the past three breeding seasons having resulted in poor recruitment in Lake Crescent, it was anticipated that there would be further reductions in the
8 Response of Lake Crescent and Lake Sorell to water level variability during 2009 abundance of the adult proportion of the G. auratus population in Lake Crescent over the next few years as natural mortality of the older cohorts occurs.
1.3 Study approach and aims
In response to increased pressure on water resources in Australia (and many other regions), recently, scientists have begun to put greater emphasis on the need to use adaptive management approaches (Bunn et al., 2010; King et al., 2010; Souchon et al., 2008) to achieve positive environmental outcomes from water management strategies. This type of approach has also been advocated by Australian Government agencies that focus on water resource management, such as the National Water Commission (Parsons et al., 2009). However, to use adaptive management approaches, knowledge in three key areas is required: (1) hydrological-ecological relationships (Anderson et al., 2006; King et al., 2009), (2) critical hydrological thresholds for environmental values and processes (Ficetola and Denoël, 2009), and (3) the resistance and resilience of freshwater dependent ecosystems and their components (Bêche et al., 2009; Crook et al., 2010).
To date, the substantial volume of scientific and managerial/operational work that has been undertaken in the Crescent-Sorell system has been consistent with the first two research areas outlined above (e.g. water level and phyiscal habitat requirements of galaxiid populations are understood and this knowledge is used to manage water levels and the populations; DPIW, 2008; Hardie, 2007; Hardie et al., 2007b). However, less has been done to examine the resilience/resistance of components of this freshwater ecosystem. Greater understanding of the resilience/resistance of the system will provide further insight into the environmental risks posed by various water level scenarios within the lakes.
The low water levels in Lake Crescent and Lake Sorell during late 2008 – early 2009 (and the relatively low water levels in preceding years) were thought to have the potential to further degrade the water quality and habitat conditions of the lakes, and cause additional declines in the G. auratus populations in these lakes. For these reasons, this study conducted monitoring to examine the impacts of the extreme environmental conditions which occurred in summer-autumn 2009. It also took a multi-purpose adaptive management approach to these low water level conditions and the subsequent and unprecedented (since 1970) wet winter-spring (Appendix 8), which resulted in large seasonal rises in water levels in both lakes. This extreme variability in water levels during 2009 provided a unique opportunity to examine the condition of the Crescent-Sorell system during drought-induced very low water levels and its response (i.e. post-drought recovery) when water levels rose markedly in winter-spring. Studying the ecosystem as its state shifted between the two hydrological extremes provided an assessment of its resistance/resilience to such conditions.
9 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Therefore, the objectives of this study were to monitor several parameters in the Crescent-Sorell system during 2009 in: (1) summer-autumn as water levels fell to seasonal minima in both lakes, and (2) winter-spring while significant natural seasonal rises were occurring. Additionally, to further assist recovery in the Crescent-Sorell system, water was released from Lake Sorell to Lake Crescent during winter-spring and the effects of this water level manipulation were also monitored. The water level manipulation strategy during 2009 had similar objectives to those of the initial trial of this management approach in 2007 regarding G. auratus populations (provide suitable habitats and hydrological conditions for spawning; DPIW, 2008); however, as water levels in both lakes were considerably higher during the period of the release (cf. those during 2007), this strategy intended to provide suitable conditions for G. auratus populations in both lakes to undertake large-scale spawning events. Furthermore, the water release also aimed to increase the extent of inundation of the littoral wetlands in Lake Crescent (see DPIPWE, 2010a)1. By monitoring the influence of the water level manipulation during 2009, under higher water level conditions than those in 2007, it was also envisaged that further knowledge would be gained about how best to use this strategy in the future.
Therefore, based on the approach outlined above, the specific aims of this study were to:
1. Document the hydrological influence of the water level manipulation on Lake Crescent and Lake Sorell during winter-spring 2009.
2. Examine hydrological variability in both lakes during 2009 and compare this period to historical patterns in water levels.
3. Monitor physico-chemical variables in both lakes during 2009 and examine the influence of water levels on these variables. Compare physico-chemical data collected in 2009 with historical data.
4. Examine the effects of the low and high water levels in 2009 on selected habitats in both lakes.
5. Re-survey areas of rocky shore habitat in Lake Crescent and derive relationships between water levels and the inundation of this habitat.
6. Synthesise conceptual understanding of physical structure and processes on rocky shores in both lakes.
1The response of aquatic vegetation communities in Interlaken Lakeside Reserve, the Ramsar-listed wetland in Lake Crescent, was monitored by DPIPWE during February 2010, and the findings of that work are documented in a separate report.
10 Response of Lake Crescent and Lake Sorell to water level variability during 2009
7. Monitor spawning activity in G. auratus populations in both lakes during 2009 by sampling eggs, larvae and adults during the breeding season.
8. Monitor the occurrence and abundance of pelagic young-of-the-year G. auratus in both lakes during the 2009 breeding season, and estimate early year class strength (at the larval and juvenile stages) and its relationship with water level fluctuations using historical data.
9. Examine the overall abundance and size structure of G. auratus populations in both lakes in autumn 2009 and 2010, and compare these data with historical data.
10. Based on the findings of this study (and other recent studies in the Crescent-Sorell system), review environmental water level management strategies for Lake Crescent and Lake Sorell.
11 Response of Lake Crescent and Lake Sorell to water level variability during 2009
2. Methodology
2.1 Study site
Unlike many lakes on the Tasmanian Central Plateau which are of glacial origin, Lake Crescent and Lake Sorell (Figure 2.1) are thought to be of non-glacial origin, occupying deflation hollows formed by wind erosion beyond ice limits during glacial phases (Kiernan, 1990; Sharples, 1997). It is likely that the basins of the lakes have progressively formed during arid times between the Late Pleistocene (Last Glacial Stage) and the Middle to Late Holocene (Sharples, 1997). Lakes Crescent and Sorell are located in an area of mostly uniform geology, soils, and vegetation which is predominantly dry sclerophyll forest (Pemberton, 1986). The catchment of the lakes is dominated by Jurassic dolerites. Quaternary marsh, lake and lagoonal sediments predominate on the western shores of both lakes (Nankivell and Hollick, 1986). Quaternary aeolian sand dune (lunettes) and sheet deposits are present on the southern and eastern shores of Lake Crescent and on the eastern shore of Lake Sorell (Pemberton, 1986; Sharples, 1997). The lunettes of the Crescent- Sorell system are a unique feature of these interconnected lakes and are some of the most extensive on the Tasmanian Central Plateau (Hydro Tasmania, 2003; Sharples, 1997).
The Crescent-Sorell system has six main ephemeral tributaries (Figure 1.1). The largest tributary, Mountain Creek (annual discharge = 11 700 ML) which flows in at the northern end of Lake Sorell, has an annual discharge of >3 times that of any other tributary in the system, and ~8 times that of the main tributary (besides the Interlaken Canal) of Lake Crescent, Agnews Creek (Uytendaal, 2003). The single outflow of the system, the Clyde River, flows out of Lake Crescent to the south- west. The hydrology of this system is primarily controlled by climatic variables, including relatively low rainfall (annual mean for 1984 to 2009 = 660.2 mm; Appendix 9) and relatively high evaporation rates during summer (summer mean for 1970 to 2010 = 4.3 mm day-1; http://www.longpaddock.qld.gov.au/silo/, accessed 5 October 2010). Water releases for downstream users (annual maximum = 10 000 ML) also alter water levels (DPIWE, 2005a). Historically, water level fluctuations in the lakes have occurred seasonally (minima occur March-April and maxima in October-November) and typically range from c. 0.2 to 0.9 m in both lakes (Appendix 2). The greater catchment area of Lake Sorell (c. 98 km2; Uytendaal, 2003a) compared to Lake Crescent (c. 33 km2; Uytendaal, 2003a) and greater inflows in its tributaries, cause seasonal water level fluctuations in Sorell to often be of longer duration and greater magnitude than those in Crescent. In recent times, the operation of sluice gates in the outflow of Lake Sorell (which have typically been closed between autumn and spring) has also contributed to this difference in the water level regimes of the lakes.
12 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 2.1 Lakes Crescent and Sorell, Tasmania, their associated littoral wetlands (depicted by dark shading) and major tributaries, and the location of littoral (●) and pelagic (▲) sampling sites for Galaxias auratus eggs and larvae, respectively. Lake perimeters are at full supply levels.
13 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Physico-chemical properties of the water columns of both of these turbid lakes are typically well homogenised due to their shallowness, and prevailing wind action which is a significant feature of lakes on the Tasmanian Central Plateau. Lakes Crescent and Sorell are very similar physically and chemically (Cheng and Tyler, 1973). However, the trophic status of these lakes differs, with Lake Sorell being mesotrophic and Lake Crescent moderately eutrophic (Cheng and Tyler, 1976a; Uytendaal, 2003; Uytendaal, 2006), and Lake Crescent typically having a phytoplankton standing crop biomass 10 times that of Lake Sorell (Cheng and Tyler, 1976b; Uytendaal, 2003; Uytendaal, 2006).
The fish assemblages of both lakes include two native species, the endemic G. auratus and indigenous short-finned eel (Anguilla australis Richardson), and three alien species, S. trutta, rainbow trout (Oncorhynchus mykiss Walbaum) and C. carpio. Galaxias auratus populations are self-sustaining, while A. australis populations are maintained by periodic stocking of elvers (juvenile eels) because natural migrations from the Derwent River and Clyde River are impeded by substantial barriers (i.e. Meadowbank Dam, and sluice gates and screens at the outflow of Lake Crescent). Since 1992, elvers have been stocked periodically into both lakes at a mean rate of 75 kg lake-1 (IFS, unpubl. data), and by-catch from sampling regimes targeting other fish species (C. carpio and G. auratus) in the lakes indicates that A. australis populations are reasonably abundant in both lakes (DPIPWE and IFS, unpubl. data).
Salmonid populations in Lake Sorell and Lake Crescent (which are dominated by S. trutta) pose a threat to G. auratus populations via competition and predation (Stuart-Smith, 2001; Stuart-Smith et al., 2004; Stuart-Smith et al., 2007b), and are largely maintained by natural recruitment and stocking, respectively. However, since 1996, annual recruitment in Lake Sorell has generally been poor due to low and unseasonal inflows in its tributaries (IFS, unpubl. data). Furthermore, between 2000 and 2005, stocking rates were reduced in Lake Crescent and no stocking in this lake has occurred since 2005 (Appendix 10). Because of this, and considerable by-catches of salmonids during carp eradiation activities in both lakes (IFS, 2004; IFS, 2010), salmonid populations in lakes Crescent and Sorell have been in relatively low abundance over the past decade. Thus, their impact on the G. auratus populations in recent years is likely to have been significantly reduced.
Cyprinus carpio were discovered in both Lake Crescent and Lake Sorell in 1995, and since then the populations in both lakes have been subject to an intensive eradication program (Anon, 1995; Diggle et al., 2004; IFS, 2004; IFS, 2010). Because of this program, C. carpio populations in lakes Crescent and Sorell have been in relatively low abundance since 2001 (IFS, 2010). Currently, the population in Lake Crescent is thought to have been completely eradicated or in very low abundance, while the population in Lake Sorell is in relatively low abundance, despite significant spawning events during spring – early summer 2009 (IFS, 2010).
14 Response of Lake Crescent and Lake Sorell to water level variability during 2009
2.2 Hydrological manipulation
Following the successful use of a water level manipulation in the Crescent-Sorell system to assist spawning of the G. auratus population in Lake Crescent during 2007 (DPIW, 2008), a second manipulation was employed in 2009. Similar to 2007, this involved releasing water from Lake Sorell into Lake Crescent during the breeding season of G. auratus. However, due to differences in the hydrological state of the lakes between the two years (cf. 2007, during 2009, in both lakes there were very low water levels in summer-autumn and greater natural inflows during winter-spring), the 2009 manipulation had slightly different objectives. This manipulation aimed to: (1) assist spawning of the G. auratus population in Lake Crescent, (2) increase the likelihood of inundation of the littoral wetlands in Lake Crescent, and increase the extent, magnitude and duration of an inundation, (3) help manage water quality in both lakes, particularly by improving water quality in Lake Crescent, (4) aid management of pest fish (C. carpio) populations in both lakes, and (5) aid management of water resources in the lakes in regard to downstream releases during the 2009/10 irrigation season. The first three objectives (i.e. those relating to the environment) will be focused upon in this report, but it is important to recognise that such a strategy can also have operational benefits for management of pest fish populations in the lakes and water resources in the Clyde River catchment.
Prior to the water level manipulation in the Crescent-Sorell system during 2009, the water levels in both Lake Crescent and Lake Sorell had risen substantially (Lake Crescent = 0.535 m; Lake Sorell = 0.620 m) from their minima in April due to natural inflows. By early July 2009, the catchment of the lakes was quite saturated and strong inflows looked likely to continue from the tributaries of both lakes. These circumstances were thought to be alleviating the poor habitat conditions which were present in both lakes in summer-autumn 2009 and looked likely to provide suitable conditions for G. auratus to spawn in both lakes during winter-spring. To help determine if a water level manipulation was warranted in the system, a survey of the G. auratus populations was undertaken in July 2009 to assess the level of spawning that had occurred in both lakes (section 2.7). Briefly, this survey involved subjectively assessing the reproductive state of littoral catches of G. auratus in Lake Crescent and Lake Sorell. This assessment determined that approximately 33% of the population in Lake Sorell appeared to have already spawned (i.e. were classed as ‘spent’), whereas only 5% of the Crescent population appeared to have spawned. These estimates were supported by sampling of eggs at known spawning sites in both lakes, with many eggs being collected in Lake Sorell and no eggs being collected in Lake Crescent.
Given the findings of the field assessment, which indicated that a reasonable amount of spawning had occurred in Lake Sorell, but only a minimal amount had occurred in Lake Crescent, it was decided that water should be released from Lake Sorell to Lake Crescent to assist spawning in Lake
15 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Crescent (and also achieve the other four aims which have been previously been mentioned). Given the dynamic hydrological conditions in the catchment of the lakes during winter-spring 2009 (due to significant rainfall and stream inflows), volumes of water which would allow important water level thresholds to be reached were calculated (Lakes Sorell and Crescent Allocation Model; DPIPWE, unpubl. data), but not strictly used to guide management of the release. Instead, the water level manipulation was managed in line with the objectives of the strategy in real-time based on the findings of environmental monitoring during the release and natural changes in the hydrological conditions in the lakes.
In addition to the 2007 water release, the 2009 water level manipulation provided an opportunity to further examine the ability to release significant volumes of water from Lake Sorell into Lake Crescent through the sluice gates of Lake Sorell and the Interlaken Canal, and logistical issues associated with this exercise. It was envisaged that the contrast in the water levels of the lakes between winter-spring 2007 and winter-spring 2009 would allow the influence of water levels in the lakes on the delivery of water from Lake Sorell to Lake Crescent to be assessed.
To monitor the volumes of water that were released during the manipulation, discharge in the Interlaken Canal immediately below the outflow of Lake Sorell was gauged periodically. The effect of the release on the seasonal water level fluctuations in both lakes was examined using the Water Availability Model for lakes Crescent and Sorell (DPIPWE, unpubl. data) and relationships between water levels and volumes in the lakes (Appendices 11 and 12). Water level data for Lake Crescent and Lake Sorell prior to, and during this study were provided by IFS (IFS, unpubl. data). These were collected by periodic readings of surveyed gauge boards in each lake and recorded as metres of elevation according to the Australian Height Datum (m AHD).
2.3 Physico-chemical parameters
Physico-chemical water variables were measured using several methods during this study. On each fish sampling occasion at pelagic and littoral fish sampling sites (Figure 1), the following variables were measured in situ: water temperature and conductivity (WTW Conductivity Meter (LF 330)), dissolved oxygen (YSI Oximeter (YSI 550DO)), pH (WTW pH Meter (pH 320)), and total turbidity (HACH Turbidimeter (2100P)). While surveying the extent of inundation of the wetlands in both lakes during spring 2009 (section 2.4), total and colloidal turbidity were measured at sites distributed across the wetlands. Additional long-term total and colloidal turbidity measurements were taken periodically (mostly monthly) at pelagic sites in each lake between January 2008 and June 2010 (IFS, unpubl. data), where colloidal turbidity was determined on filtrate passed through a Whatman GD filter paper (1 µm nominal pore size). Turbidity data prior to 2008, which were used to examine long-term trends, were from Uytendaal (2003; 2006), and DPIPWE (unpubl. data) and IFS (unpubl. data).
16 Response of Lake Crescent and Lake Sorell to water level variability during 2009
In both lakes, short-term physico-chemical variability, which was likely to occur while water levels were very low during summer-autumn 2009, was also measured using YSI Sonde loggers (YSI 6920 multi-probe data logger). Continuous loggers were deployed in Lake Crescent between 11 February and 1 April 2009, and in Lake Sorell between 11 February and 1 March 2009. These loggers recorded water temperature, conductivity, dissolved oxygen, and total turbidity at 30-min intervals. The loggers were fixed to a solid base (concrete block with up-right mounting for the logger) which ensured the logger probes were 290 mm above the substrate. In Lake Crescent, the logger was deployed at an open-water site approximately 400 m south-east of Tea Tree Point, where the water depth ranged from 590-640 mm during the deployment. In Lake Sorell, the logger was deployed at an open-water site approximately 500 m north-east of Point of Chillon, where the water depth was approximately 1.5-2 m during the deployment. The depths at these sites approximated the average depths in the main basins of the lakes during summer-autumn 2009.
2.4 Littoral habitat surveys
Extreme variability in the water levels in lakes Crescent and Sorell during 2009 provided unique opportunities to determine the extent of important littoral habitats in the lakes and document the seasonal inundation of these habitats by substantial rises in water levels. Three types of habitat surveys were undertaken: (1) aerial photography of the entire Crescent-Sorell system and rocky shores in Lake Crescent, (2) on-ground surveys of rocky shores in Lake Crescent, and (3) on- ground surveys of the Interlaken Lakeside Reserve wetland in Lake Crescent and Silver Plains Marsh in Lake Sorell.
Aerial photographs of the Crescent-Sorell system were taken to document the surface area of the lakes and some of their important habitat features (i.e. wetlands, rocky shores, sand shores, etc.) under low and high water levels. Aerial photographs on the entire system were taken by the Information and Land Services Division of DPIPWE on 20-21 March 2009 and 7 November 2009. The ground sample distance (pixel size) of the photographs was 0.4 m. They were geo-referenced and ortho-rectified using 1:25 000 topographical map data; therefore, their accuracy was similar to these data (mostly 6-12 m, but with a maximum error of 18 m in some places). Digital copies of the images were created as 24-bit colour TIFF files (resolution of images taken in March and November was 2309 dpi (11 µm) and 2302 (12.5 µm), respectively). Image data coordinates were MGA Zone 55 and the projection datum was GDA94. Additional photographs of the rocky shores in Lake Crescent were also taken on 20 March 2009. These photographs had a ground sample distance of 0.125 m and were used to assist mapping of the extent of this habitat in this lake.
On-ground surveys of the four defined rocky shores in Lake Crescent (Tea Tree Point shore, Lakeside Island shore, Triffitt Point shore and Boathouse Shore) were undertaken on 18-19 May and 8-9 October 2009 to map the extent, and characteristics and condition (i.e. silt cover and
17 Response of Lake Crescent and Lake Sorell to water level variability during 2009 density of rocky substrata) of these areas. Both surveys were conducted under favourable weather conditions (mostly calm and fine, hence minimal wave action). At the time of the survey in May, the water level in Lake Crescent was 801.575 m AHD, while in October it was 803.150 m AHD; thus, the autumn and spring surveys were undertaken while water levels were relatively low and high, respectively. During the autumn survey, rocky substrate in many areas of the shores was completely dewatered, while in some areas this habitat was inundated. To determine the extent of the surface area of the rocky substrate along each shore, during the survey in autumn, the dewatered and inundated edges of the rocky substrate were located along transects that were run at 90° to the high water mark of the shore at intervals of 10-50 m. A staff was used to help locate the inundated edge and at these locations the water depth (± 10 mm) was measured. The location of both the dewatered and inundated edges was recorded with a global positioning system (GPS), and the water level on each day of the surveying was used as the reference level for inundated locations. During the survey in spring, similar survey methods were used, but only the inundated edge was surveyed on all four shores.
Surveys of the Interlaken Lakeside Reserve wetland and Silver Plains Marsh were undertaken to: (1) assess the extent of inundation of these wetlands while water levels in the lakes were relatively high, (2) examine the accuracy of the previously surveyed in-lake edges of the wetlands (which were undertaken by Heffer (2003c) when the wetlands were dewatered), and (3) examine turbidity levels in the wetlands. Interlaken Lakeside Reserve was surveyed on 9 October 2009 when the water level in Lake Crescent was 803.160 m AHD, while Silver Plains Marsh was surveyed on 5 November 2009 when the water level in Lake Sorell was 804.100 m AHD. Both surveys were conducted under favourable weather conditions (i.e. fine and light winds, thus limited wave action). Eight transects were surveyed in the Interlaken Lakeside Reserve and seven were surveyed in Silver Plains Marsh. Along each transect, the high water mark and the in-lake edge of the wetlands were located, and water depth (± 10 mm) was measured at intervals of approximately 20-75 m. Water samples were collected at various locations along the transects and used to measure total and colloidal turbidity (see section 2.3 for methods).
2.5 Littoral golden galaxias egg sampling
During the 2009 breeding season, spawning activity was monitored at the egg stage following similar methods to those employed during the 2007 water level manipulation study (DPIW, 2008). Spawning activity was monitored in littoral areas where G. auratus had previously been found to spawn in Lake Crescent (Boat House Shore) and Lake Sorell (Dogs Head Shore) (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b) (Figure 2.1). On both shores, 200 m of shoreline which contained (or had the potential to contain, given inundation by rising water levels) areas of suitable spawning substrata (cobble of c. 20-250 mm diameter; Hardie et al., 2007b) was
18 Response of Lake Crescent and Lake Sorell to water level variability during 2009 designated as a spawning monitoring site for each lake. The site in Lake Sorell was monitored in July and September 2009, whereas the site in Lake Crescent was monitored in July and then at monthly intervals between September and November 2009. The period over which the monitoring took place was shorter in Lake Sorell than in Lake Crescent because the spawning season occurred earlier in Sorell than in Crescent.
On each sampling occasion at each site, 10 kick-net sweeps of 60 s duration were performed over c. 1 m2 areas of suitable spawning substrata (at depths <1.0 m; Hardie et al., 2007b) using a hand net (200 × 215 mm, 350 mm tail, 500 m mesh). Prior to sampling, the water depth (± 10 mm) at each netting site (1 m2 area) was recorded. To perform kick-net sweeps, the collector disturbed substrata by foot and continuously moved the fully inundated opening of the net within the designated sampling area. The material collected in kick-net samples was preserved in 70% ethanol and the total number of G. auratus eggs in each sample (eggs sweep-1) was determined in the laboratory. These results estimated the density of eggs incubating at the site, and hence indicated recent spawning activity and hatching of larvae. All egg samples have been lodged at the Tasmanian Museum and Art Gallery, Hobart, Tasmania.
2.6 Pelagic juvenile golden galaxias sampling
During 2009 – early 2010, young-of-the-year (YOY) G. auratus were monitored following similar methods to those employed during the 2007 water level manipulation study (DPIW, 2008) and previous research on the early-stage recruitment of this species (Hardie, 2003a; Hardie, 2007). Pelagic YOY G. auratus in Lake Crescent and Lake Sorell were sampled in July 2009 and then monthly between September 2009 and January 2010 to examine their occurrence and abundance. Fish were collected at four sites in each lake (Figure 2.1) by towing a conical ichthyoplankton net (400 mm diameter, 1.25 m tail, 500 μm mesh) for 10 min approximately 15 m behind a boat at a speed (~2 m s-1) that ensured the net was sampling the top 1 m of the water column. For the purposes of this report, fish captured in pelagic habitats between July 2009 and January 2010 (i.e. ≤40 mm fork length, FL) are referred to as ‘larvae’ and all other YOY (typically 41-70 mm FL) as ‘juveniles’ (see Hardie, 2007). All material collected in ichthyoplankton samples was euthanased in an anesthetic solution (Aqui-s® (Isoeugenol)) and preserved in 70% ethanol. Galaxias auratus larvae were identified following an unpublished taxonomic key (Frijlink, 1999), the number of larvae and juveniles in each sample were counted (fish tow-1), and the first 100 larvae and all juveniles in each sample were measured (mm FL). All pelagic fish samples have been lodged at the Tasmanian Museum and Art Gallery, Hobart, Tasmania.
19 Response of Lake Crescent and Lake Sorell to water level variability during 2009
2.7 Littoral adult golden galaxias sampling
During 2009 and 2010, the relative abundance and length structure of G. auratus populations in lakes Crescent and Sorell were examined following similar methods to those employed previously to annually monitor these populations (DPIW, 2008; Hardie, 2003a). Similar to previous years, annual monitoring of the G. auratus populations was conducted in late summer – autumn because at this time of year: (1) YOY G. auratus from the previous breeding season have moved into littoral habitats and are able to be captured along with adults (DPIW, 2008; Hardie, 2003a; Hardie, 2007) using fyke netting (the preferred method to capture this species; Hardie et al., 2006), (2) the catchability of adult G. auratus in littoral habitats is relatively stable and not influenced by reproductive behaviour, which can cause catchability to be highly variable with significant increases during spawning (winter – early spring; Hardie et al., 2005), and (3) the weather on the Tasmanian Central Plateau is generally more stable, hence, sampling conditions are logistically favourable and catch rates are more consistent (S. A. Hardie, unpubl. obs.).
Juvenile and adult G. auratus were sampled twice during 2009 and once during 2010 from littoral habitats (depths ≤1.2 m) in Lake Crescent and Lake Sorell. The first sampling occasion in April 2009 and the third in March 2010 were undertaken to assess the overall status of the population; in line with previous annual monitoring between 2001 and 2008 (see Appendices 6 and 7 for data resulting from previous surveys). The second sampling occasion in July 2009 was undertaken to examine the reproductive status of the populations to assist decision making regarding the water level manipulation strategy (section 2.2).
In Lake Crescent, fish were collected at four, three and six sites on the first, second and third sampling occasions, respectively. In Lake Sorell, fish were collected at six, three and six sites on the first, second and third sampling occasions, respectively. Only four sites were sampled in Lake Crescent during the first sampling occasion due to logistical difficulties associated with accessing sites when the water levels in the lake were very low (i.e. boating on the lake was not possible at the time). The sites used in each lake typically represented the dominant littoral habitats (fine sediment, sand and rock) in both lakes at the time of each sampling occasion.
On each sampling occasion, fish were sampled overnight (mean soak time of 21 h) using four fyke nets (3 m × 0.6 m wing, 570 mm D-shaped entrance, 5 mm stretched mesh) at each site. All nets had 80 mm stretched mesh screens on the entrance to avoid capture of platypus (Ornithorhyncus anatinus) (Grant et al., 2004), waterbirds (Beumer et al., 1981) and larger fish species (e.g. salmonids). Catches from littoral habitats were anaesthetised in an anesthetic solution (Aqui-s® (Isoeugenol)), counted and a random sub-sample of 50 fish from each site were measured (mm FL). Fish were then revived in fresh water and released at the site of capture. Additionally, during sampling in July 2009, sub-samples of fish that were measured at each site were also examined
20 Response of Lake Crescent and Lake Sorell to water level variability during 2009 externally to determine their reproductive state. This was classified using three broad physiological categories (indeterminate, ripe or spent; the latter two categories were thought to correspond to developmental stages 4 and 5 of Hardie et al., 2007b).
2.8 Data analysis
Generally, where possible and appropriate, comparisons between data collected during this study were made to historical data (i.e. physico-chemical conditions, abundance of G. auratus larvae and juveniles/adults, and relationships between G. auratus populations and water levels, etc.). This enabled the findings of the present study to be viewed in the context of historical conditions in the lakes. Importantly, this also allowed the results of the work undertaken in 2009 – early 2010 to be used to re-examine previously established trends and relationships.
2.8.1 Habitat mapping and spatial modelling
Geographical Information System (GIS) modelling is well suited to determining the extent of suitable habitats for freshwater fish (e.g. Douglas et al., 2009; Rowe et al., 2002). Therefore, littoral areas of rocky substrate in Lake Crescent were mapped, and a GIS model was constructed using ArcInfo 9.3.1 (ESRI, Redlands, California, USA) to determine relationships between water levels and the extent of inundation of this habitat (hence the availability of this habitat to G. auratus). The model was used to calculate the total area of rocky shore inundated and ‘optimal’ spawning habitat for G. auratus (i.e. rocky shore at depths of 0.2-0.6 m; Hardie et al., 2007b) at 14 water levels between 803.800 m AHD (full supply level; FSL) and 801.200 m AHD (which is a very low level, well below the operation sill level of 801.700 m AHD) at intervals of 0.2 m. The relationships derived from these analyses were used to investigate how a decline in water level affects the amount of inundated rocky substrate habitat, and to also determine if there are important water level thresholds where marked changes in habitat availability occur. Furthermore, this exercise enabled the influence of changes in water level on the area of available spawning habitat to G. auratus to be examined quantitatively.
Specifically, the aims of the rocky shore habitat mapping and GIS modelling were to:
(a) Determine total area of littoral rocky substrate in Lake Crescent and the proportion of the lake’s substrate that is comprised of this habitat.
(b) Accurately map areas of littoral rocky substrate in Lake Crescent.
(c) Construct a 3-demensional (3D) GIS terrain model and use it to determine the relationship between water levels and the proportion of littoral rocky substrate inundated.
(d) Use the 3D model to determine the relationship between water levels and the amount of optimal spawning habitat for G. auratus (i.e. rocky shore at depths of 0.2-0.6 m).
21 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Additional to aerial photographs of lakes Crescent and Sorell, and spatial data from on-ground surveys of the rocky shores in Lake Crescent (section 2.4), mapping of the rocky shores in Lake Crescent and spatial modelling of the inundation of these areas also utilised other data sources. These included the CFEV database drainage layer for Tasmania (CFEV, 2005), bathymetric data for Lake Crescent (based on a survey on the lake in April 1995; Hydro Tasmania, unpubl. data), various DPIPWE spatial data sets (1:25,000 scanned TASMAP maps, hydrography, roads, survey data, aerial photogrammetry and contours), miscellaneous surveyors notes (Hydro Tasmania, unpubl. data), and Surcom survey control database. Full details of the methods used to map rocky shores in Lake Crescent and model inundation of these areas are provided in DPIPWE (2010b); however, a brief description is provided below.
Mapping littoral areas of rocky substrate
A transverse Mercator projection MGA94 Zone 55 is used for all mapping purposes in Tasmania, so all data were placed on this datum and projection. The boundary (i.e. high water mark) of Lake Crescent, including its littoral wetlands, was derived from the CFEV drainage layer and compared to bathymetric data, aerial photographs and on-ground survey points. The definition of the boundary of the lake was found to be in close agreement in all data sets, so the CFEV boundary was accepted to be the boundary of the lake, which is the 803.800 m AHD contour.
There are four defined shores containing rocky substrate in Lake Crescent: Boathouse Shore, Lakeside Island shore, Tea-tree Point shore and Triffitt Point shore (Hardie, 2003a; Hardie et al., 2007b). A number of GPS points were captured during the on-ground survey in May 2009, identifying the dewatered edge and the edge of the rocky shore (which was inundated in some areas at the time of the survey). These points were loaded into a geodatabase, and polygons were created from the points to define the extent of the rocky substrate on all four shores. The extent of the rocky shores according to aerial photographs that taken while water levels were low (in March 2009) compared favourably with the on-ground survey data, thus, confirming the extent of rocky substrate in these areas. The areas of these defined regions of rocky substrate in Lake Crescent were calculated and mapped (as 2D surfaces).
Relationships between water levels and inundation of littoral rocky substrate
A terrain model is a multi-resolution, triangulated irregular network (TIN)-based surface. The following data were used to construct a terrain model of Lake Crescent: the FSL boundary of the lake (SFType hard clip), on-ground survey points along the dewatered edge of the rocky shores which were captured at a water level of 801.575 m AHD (SFType mass points), bathymetric contours (SFType mass points) and bathymetric soundings (SFType mass points).
While building the terrain, some bathymetric contours in the rocky shore areas were found to have values that did not correspond to on-ground survey data. These contours had been constructed
22 Response of Lake Crescent and Lake Sorell to water level variability during 2009 from few data points in the bathymetric survey (which in some cases appeared to be measurements that were taken above the water level in the lake at the time of the bathymetric survey), and values that did not agree with data from the on-ground surveys. These contours were removed and, in their place, on-ground survey data and sounding points were used to build the terrain model. This resulted in a terrain with smooth contours that closely represented the physical surface of the lake shore, especially in the regions of interest, along the rocky shores.
The terrain model was used to generate a raster (grid) of the lake surface using 1 × 1 m pixels in the areas of rocky shore habitat. The rocky shore polygons were used as ‘masks’ to create floating point elevation grids for just the areas of rocky shore habitat. The rasters were reclassed into 0.2 m bands (or contours) between 803.800 m AHD and 801.200 m AHD. The reclassed raster was converted to a polygon feature class. This was done in each of the four areas, and the resulting polygons merged into a single feature class. Extensive use of ArcGIS Model builder was made to combine geoprocessing tools together in each step. This procedure resulted in a set of polygons representing a drop in water level of 0.2 m between 803.800 m AHD and 801.200 m AHD on the four rocky shores in Lake Crescent.
Using the single polygon feature class for all of the rocky shores, a model was built to select subsets of the polygons to determine the total area of littoral rocky substrate inundated at 14 water levels between 803.800 m AHD and 801.200 m AHD on each rocky shore in Lake Crescent. Similarly, a model was also built to select subsets of the polygons to determine the area of littoral rocky substrate at depths of 0.2-0.6 m (i.e. optimal spawning habitat for G. auratus) that was inundated in the lake at the same set of water levels.
Additional habitat mapping and inundation estimates
Additional to mapping of rocky shores in Lake Crescent and spatial modelling of the inundation of these areas of habitat, estimates of the surface area of the entire wetted proportion of the lakes, and selected littoral habitats were made using combinations of aerial photographs, GPS survey data, and CFEV spatial data for the waterbodies and wetlands. This was done for littoral wetlands and sandy shores in both lakes, and rocky shores in Lake Crescent, during autumn and spring 2009, when water levels in the lakes were relatively low and high, respectively.
2.8.2 Golden galaxias larval abundance and population status
Larval G. auratus catch data from pelagic samples (fish tow-1) were combined with similar data that were collected in both lakes using the same sampling methods during the 2000-2004 and 2007 breeding seasons (DPIW, 2008; Hardie, 2007); in these years, water level conditions varied, but levels were generally relatively low (i.e. littoral wetlands in both lakes were dry and rocky shores in Lake Crescent were periodically dewatered). In the combined data set, larval G. auratus from pelagic habitats were assigned to cohorts based on the calendar year of their birth, which was
23 Response of Lake Crescent and Lake Sorell to water level variability during 2009
clearly defined by length classes in temporal catches. The combined data were log10(x + 1) transformed to homogenise variances and normalise residuals.
For the 2000-2004 larval catches, sites sampled between late August and late December of each year in each lake were used to calculate mean larval abundances during the breeding seasons, whereas, for the 2009 breeding season, this was done using sites sampled between early September 2009 and early January 2010 (during the 2009 breeding season, logistical issues prevented sampling being undertaken at the same times as in previous years). Mean larval abundances during the breeding seasons were used to test for differences between years in each lake (mean larval abundances during this period have been found to be a robust measure of early year class strength in these G. auratus populations; Hardie, 2007). To do this, repeated measures analyses of variance (ANOVAs) were performed with lakes as fixed factors, years as within-subject factors, and sample sites were the subjects. Inter-annual variation in seasonal larval abundances in both lakes was assessed by comparing mean catches during 2001 to all other years (i.e. using 2001 as the reference season).
Juvenile/adult catch data from littoral samples collected during 2009 and 2010 were combined with similar data that were collected in both lakes using the same sampling methods during late summer – autumn 2001, 2002, 2004, 2005, 2007 and 2008 (Hardie, 2003a, 2007; DPIW, 2008; IFS, unpubl. data). In the combined data set, YOY G. auratus were assigned to cohorts based on the calendar year of their birth, which was clearly defined by length classes in temporal catches. The combined data were log10(x + 1) transformed to homogenise variances and normalise residuals.
Total catch data (juveniles and adults) for sites sampled in each year in each lake were used to test for differences in mean abundances between years in each lake. To do this, repeated measures ANOVAs were performed with lakes as fixed factors, years as within-subject factors, and sample sites were the subjects. Inter-annual variation in total catch abundances in both lakes was assessed by comparing mean catches during 2002 to all other years (i.e. using 2002 as the reference season).
Additionally, in both lakes across years, regression analysis was used to assess relations between mean abundances of larvae during the breeding season, the proportion of YOY in littoral catches in the following summer-autumn, and lake hydrology (i.e. magnitude of rise between May and September). All data analyses (other than spatial analyses) were carried out in R version 2.9.2 (R Development Core Team, 2009). The significance level for hypothesis tests was P = 0.05.
24 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3. Results
3.1 Water level regime
3.1.1 Environmental water release
Similar to 2007, water was released from Lake Sorell into Lake Crescent during 2009 to assist spawning of G. auratus and inundation of littoral wetlands in Crescent, as well as helping to manage pest fish, water quality and water resources in both lakes. During 2009, a total of 10.03 × 106 m3 (10025 ML) was estimated to have been released from Lake Sorell between July 23 and October 30 (Figure 3.1). This water release was >3 times greater than the release used in the 2007 water level manipulation (3090 ML was released in 2007; DPIW, 2008) (Appendix 13).
Discharge in the Interlaken Canal (Figure 3.1) was gauged on 12 occasions during the 99-day period of the release, and gradually increased over most of this time from a minimum of 47 ML d-1 to a maximum of 181 ML d-1 (mean discharge = 93 ML d-1). Conversely, during the release, the difference in the water levels between the lakes gradually decreased from 1.36 m to 0.89 m (mean difference = 1.18 m); however, with the substantial seasonal water level rise in Lake Sorell, the potential level above the sill of the sluice gates of that lake gradually increased from 0.40 m to 1.15 m (the actual levels above the sill were influenced by the addition of planks to control the discharge of the release). Strong natural rises in water levels occurred in both lakes during the water level manipulation due to relatively heavy rainfall in the catchment (Appendix 9) and solid baseflows in tributaries (e.g. Mountain Creek; Figure 3.2).
When the water release began on 23 July 2009, Lake Crescent had already risen 0.535 m from its seasonal minimum of 801.510 m AHD which was recorded on April 3; however, the water level at this time (802.045 m AHD) was 0.155 m below the edge of the rocky shores and 0.655 m below the edge of the wetlands (Figure 3.3). In Lake Sorell, at the beginning of the water release, the water level had already risen 0.630 m from its seasonal minimum of 802.780 m AHD which was recorded on April 20; however, the water level at this time (803.410 m AHD) was 0.190 m below the edge of the wetlands (Figure 3.3).
25 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.1 Sluice gates in the Interlaken Canal at the outflow of Lake Sorell, 9 October 2009. The gauged discharge at this time was ~170 ML d-1. Discharge gaugings were undertaken approximately 100 m downstream of the sluice gates.
Figure 3.2 Mountain Creek immediately upstream of Lake Sorell, 4 September 2009. According to a discharge rating curve for the creek, the discharge at this time was ~50 ML d-1.
26 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
Observed Modelled 804.5 Full supply
804.0
Wetlands
803.5 Water levelAHD)Water (m
Sill 803.0
802.5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Month
Lake Crescent
Full supply
Observed Modelled 803.5
803.0
Wetlands
802.5
Rocks Water levelAHD)Water (m
802.0
Sill
801.5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Month
Figure 3.3 Actual (observed) daily water levels in lakes Crescent and Sorell between January and October 2009, and levels that were estimated (modelled) to have occurred in both lakes without the release during the manipulation period. Levels at which littoral wetlands connect to the lakes and rocky shores begin to be inundated in Lake Crescent are indicated. Maximum (full supply) and minimum (sill) operating levels of sluice gates in both lakes are also shown.
27 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Hydrologically, the water level manipulation had minimal effect on Lake Sorell, but a significant effect on Lake Crescent (Figure 3.3). In Lake Sorell, in comparison to modelled ‘without release’ levels, there was only a minor decrease in the magnitude of the seasonal water level rise up to 30 October (observed = 1.380 m v. modelled = 1.530 m), and at the end of the manipulation, the water level was estimated to have been 190 mm lower than what would have occurred at that time without the release of water from that lake. The water level reached the surveyed edge of the wetlands in Lake Sorell on 14 August and the release had little effect on the inundation of the wetlands during the release period, with the wetlands being inundated by depths >300 mm for 52 days prior to the end of the release (the release reduced the period of inundation to these depths by 4 days).
In Lake Crescent, without the release of water from Lake Sorell, the magnitude of the seasonal rise in water level would have been much less (observed = 1.780 m v. modelled = 1.240 m) (Figure 3.3). At the end of the manipulation, the water level in Lake Crescent was estimated to have been 0.540 m higher than what would have occurred at that time without the release of water from Lake Sorell. Most importantly, the water level manipulation increased the magnitude of the rise above the critical minimum (i.e. edge of the rocky shores) by 0.540 m, and provided a continual, steady rise in water level throughout the manipulation; this would not have happened under the climatic conditions that occurred during October. The water level reached the edge of the rocky shores in Lake Crescent on 11 August and the release lengthened the time during which the rocky shores were inundated by depths of >0.2 m by approximately 1 week (observed = 72 days v. modelled = 64 days. (The timing of the additional period of inundation to these depths was also important for the G. auratus population as this is the time when most of the spawning occurred (section 3.6)). In Lake Crescent, the water level reached the surveyed edge of the wetlands on 9 September and the release substantially increased the depth of inundation of the wetlands during the release period: the wetlands were inundated by depths >300 mm for 44 days prior to the end of the release, whereas, without the release, they would not have been inundated to this depth at all during this period.
Additional to the benefits to the G. auratus population and littoral wetlands in Lake Crescent, the water release also helped avoid near-spill conditions occurring in Lake Sorell. An uncontrolled spill of water from Lake Sorell could allow the pest fish C. carpio to move between the lakes which have been managed as closed populations (by containment screens on sluice gates; Wisniewski, 2003) under the control and eradication program for this species for several years. Partly due to the potential for Lake Sorell to spill, the discharge from Lake Sorell was increased during the later stages of the water release (Figure 3.3). This action provided greater insurance that this situation was avoided, as without the release, the water levels in Lake Sorell would have peaked at 50 mm below the full supply level, but instead the release increased this margin to 200 mm.
28 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.1.2 Historical water level patterns and levels during 2009
To examine the water level regimes that occurred in lakes Crescent and Sorell during 2009, it is important to consider the recent hydrological conditions in the context of long-term water level patterns in the lakes. Additionally, an understanding of the ecologically-influential aspects of the historical water level regimes of the lakes also provides important contextual information to assess the water levels and status of the Crescent-Sorell ecosystem during the period of the current Water Management Plan (WMP) (i.e. 2005 to present; DPIWE, 2005a).
Since 1970, water levels have had ranges of approximately 2.635 and 2.160 m in Lake Crescent and Lake Sorell, respectively (Figure 3.4). Minima and maxima during this period were 801.510 m AHD (April 2009) and 804.145 m AHD (August 1996) in Lake Crescent, respectively, and 802.620 m AHD (April 2000) and 804.780 m AHD (December 1975) in Lake Sorell, respectively. Since 1970, water level fluctuations in Lake Crescent and Lake Sorell have followed similar patterns at annual and inter-annual scales (Figures 3.4 and 3.5). Overall, since 1970, there have been substantial decreases in long-term trends in water levels in both Lake Crescent and Lake Sorell (Figure 3.5), and since 1996, inter-annual variability in levels has increased. Both lakes maintained relatively high levels between 1970 and the early 1980s when a significant El Niño- induced drought in 1982-1984 resulted in quite low levels in both lakes. After this drought event, both lakes again had relatively high levels between the mid-1980s and 1996. However, since 1996, levels have generally been relatively low in both lakes largely due to prolonged drought conditions in the catchment which began in the late 1990s. During the past decade, levels in both systems were particularly low in 2000-2001, and 2007 – early 2009, with a marginal increase in the levels in both lakes during the mid-2000s.
Because water level fluctuations in lakes Crescent and Sorell can dewater important littoral habitats (wetlands in both lakes and rocky shores in Crescent; section 1.1.3), it is important to understand long-term patterns of water level fluctuations in these lakes in relation to the surveyed edges of these habitats (which also approximate water level thresholds that are used guide management of the water resources associated with these lakes under the current WMP; DPIWE, 2005a). Since 1996, there has been an obvious shift in the water level regimes of the lakes from one which was mostly comprised of relatively high levels which frequently inundated littoral wetlands and rocky shores, to one of much lower levels which has frequently dewatered these littoral habitats (Figure 3.4).
29 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
Full supply
804 Wetlands
Sill 803
802
Lake Crescent Water level (m AHD) level (m Water 804 Full supply
803 Wetlands
Rocks 802 Sill
Jan-70Jan-71Jan-72Jan-73Jan-74Jan-75Jan-76Jan-77Jan-78Jan-79Jan-80Jan-81Jan-82Jan-83Jan-84Jan-85Jan-86Jan-87Jan-88Jan-89Jan-90Jan-91Jan-92Jan-93Jan-94Jan-95Jan-96Jan-97Jan-98Jan-99Jan-00Jan-01Jan-02Jan-03Jan-04Jan-05Jan-06Jan-07Jan-08Jan-09Jan-10
Figure 3.4 Water levels in lakes Crescent and Sorell between January 1970 and June 2010, and habitat thresholds. Levels at which littoral wetlands connect to the lakes, rocky shores begin to be inundated in Lake Crescent are indicated. Maximum (full supply) and minimum (sill) operating levels of sluice gates in both lakes are also shown.
30 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
804.5
804.0
803.5
803.0
802.5
802.0 Lake Crescent
804.5
Water level (m AHD) level (m Water 804.0
803.5
803.0
802.5
802.0
1970 1980 1990 2000 2010 Year
Figure 3.5 Mean annual water levels in lakes Crescent and Sorell between 1970 and 2009 plotted with an aspect ratio following the ‘banking rule’ of Cleveland et al., (1988) to visualise long-term trends.
31 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Based on mean monthly water levels, the levels in Lake Crescent have been at or below the sill of the current sluice gates, the edge of the rocky shores, the edge of the wetlands and the current full supply level <1%, 9%, ~29% and ~84% of the time, respectively (Figure 3.6). Similarly, the levels in Lake Sorell have been at or below the sill of the current sluice gates, the edge of the wetlands and the current full supply level ~6%, ~30% and 78% of the time, respectively (Figure 3.6). Therefore, since 1970, the downward trend in the water levels in both Lake Crescent and Lake Sorell – especially the relatively low levels in both lakes since 1996 – have caused levels in both lakes to dewater important littoral habitats in these lakes for significant periods.
Besides the influence of the water level manipulation on both Lake Crescent and Lake Sorell during 2009 (section 3.1.1), the water level conditions that occurred during 2009 in both lakes were quite extraordinary in comparison to annual water level fluctuations since 1970 (Figure 3.4). During 2009, substantial seasonal rises occurred in both lakes after c. 3 years of relatively low levels. In Lake Crescent, water levels rose approximately 1.78 m, with a minimum of 801.510 m AHD (April 3; which was the lowest recorded level since 1970) and maximum of 803.290 m AHD (October 28). In Lake Sorell, water levels rose approximately 1.38 m, with a minimum of 802.780 m AHD (April 20) and maximum of 804.160 m AHD (October 2).
Comparisons between the long-term mean monthly water levels and those that occurred in 2009 in both lakes clearly illustrate the large magnitudes of the seasonal rises during 2009 (Figure 3.7). Mean monthly levels in Lake Crescent were well below the average long-term levels between January and September, but rose above the averages for October, November and December. In this lake, based on mean monthly levels, the magnitude of the rise during 2009 was approximately 2.9 times greater than the long-term mean pattern. In Lake Sorell, the mean monthly water levels during 2009 were below the average long-term levels between January and August, but rose above the averages for all months between September and December. In this lake, based on mean monthly levels, the magnitude of the rise during 2009 was approximately 2.2 times greater than the long-term mean pattern.
These dramatic and relatively rapid changes in water levels in lakes Crescent and Sorell during 2009 provided a unique opportunity to examine the response of the Crescent-Sorell ecosystem (especially the G. auratus populations) to this transition phase from severe drought conditions and low water levels, to a wetter climatic period (albeit <1 year in length) and a recovery in water levels.
32 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell 5
4
3
2
1
0 Lake Crescent
5 Frequency (%) Frequency 4
3
2
1
0 801 802 803 804 805 Water level (m AHD)
Figure 3.6 Histograms of mean monthly water levels in lakes Crescent and Sorell between 1970 and 2009. Levels at which littoral wetlands connect to the lakes (green lines), rocky shores begin to be inundated in Lake Crescent (red line) are indicated. Maximum (full supply) and minimum (sill) operating levels of sluice gates (grey lines) are also shown.
33 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Crescent Lake Sorell
1970-2009 2009 804.0
803.5
803.0
802.5 Water level (m AHD) level (m Water
802.0
801.5
J F M A M J J A S O N D J F M A M J J A S O N D Month
Figure 3.7 Cumulative mean monthly water levels in lakes Crescent and Sorell (January-December) between 1970 and 2009, and during 2009.
34 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.2 Physico-chemical parameters
3.2.1 Conditions during the 2007 and 2009 Galaxias auratus breeding seasons
In situ (spot sample) measurements of physico-chemical variables during the 2007 (July-December) and 2009 (July – early January) G. auratus breeding seasons provide an indication of the water quality in lakes at these times (Table 3.1). The data from these measurements are relatively coarse in terms of the temporal and spatial frequencies at which they were recorded; however, importantly, as they were taken during egg and/or larval fish sampling, they document the broad-scale conditions experienced by these early life stages in both populations.
Water levels in lakes Crescent and Sorell varied markedly between the 2007 and 2009 breeding seasons (Figure 3.4), with low levels and only moderate seasonal rises in both lakes during 2007, whereas, as previously documented, levels in both lakes were rising from very low levels at the start of the 2009 season and rose substantially due to natural inflows to reach relatively high levels late in the season. These differences in water level conditions in the lakes caused some physico-chemical variables to vary between 2007 and 2009. It should be noted that during breeding seasons, many of the physico-chemical variables measured show distinct seasonal trends; therefore, comparisons of mean values between the seasons should be treated with caution (in some instances the range of the measurements is also important information).
With these issues in mind, the following are some observations on physico-chemical conditions in lakes Crescent and Sorell during the 2007 and 2009 G. auratus breeding seasons. Water temperature (total range = 3.8-18.7°C) and pH (means of c. 7.0) were similar in both lakes between and within the breeding seasons (Table 3.1). Dissolved oxygen concentrations were similar in the lakes within the breeding seasons, but differed between the seasons with slightly higher mean concentrations during 2009 (means of <10 mg L-1 in 2007 in both lakes v. means = 11.1 mg L-1 in both lakes in 2009). The two variables which varied most in the in situ measurements between the breeding seasons and the lakes were electrical conductivity and total turbidity.
In both seasons, Lake Crescent had greater mean conductivity values than Lake Sorell, and both lakes also had higher mean conductivities in 2007 in comparison to 2009 (Table 3.1). These differences between the seasons in each lake are likely to be due to substantial inflows from the catchment in 2009 diluting concentrations of dissolved salts in both lakes. This dilution effect is indicated by differences in the ranges of the conductivity readings between the two seasons in Lake Crescent (conductivity range: 2007 = 33.7 NTU; 2009 = 96.2 NTU) and to a lesser extent in Lake Sorell (conductivity range: 2007 = 15.2 NTU; 2009 = 24.9 NTU), with high readings occurring early in the breeding season and lower readings occurring later in the season after water levels had risen.
35 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Table 3.1 Mean ± SE (minimum-maximum) values for physico-chemical variables measured in situ at pelagic sites in lakes Crescent and Sorell during the 2007 and 2009 Galaxias auratus breeding seasons (July-January).
Physico-chemical variable 2007 2009 Lake Crescent
Water temperature (°C) 11.5 ± 1.1 (3.9-18.1) 10.0 ± 1.0 (4.4-16.4) Dissolved oxygen (mg L-1) 8.9 ± 0.4 (4.5-11.3) 11.1 ± 0.3 (9.5-13.1) Dissolved oxygen (% saturation) 81.5 ± 2.9 (46.5-91.7) 96.6 ± 0.5 (91.0-101.0) pH 7.2 ± 0.1 (6.8-7.5) 6.9 ± 0.1 (6.5-7.4) Electrical conductivity (μS cm-1) 166.4 ± 2.4 (154.1-187.8) 153.6 ± 6.9 (127.8-224.0) Total turbidity (NTU) 234 ± 7 (191-339) 186 ± 19 (95-346) Lake Sorell
Water temperature (°C) 11.1 ± 1.0 (3.8-18.7) 10.1 ± 1.0 (4.4-16.3) Dissolved oxygen (mg L-1) 9.7 ± 0.3 (6.6-11.8) 11.1 ± 0.3 (9.4-13.0) Dissolved oxygen (% saturation) 89.0 ± 1.1 (69.4-98.0) 96.8 ± 0.5 (94.0-102.0) pH 7.1 ± 0.1 (6.6-7.5) 7.0 ± 0.1 (6.6-7.5) Electrical conductivity (μS cm-1) 103.6 ± 0.9 (98.6-113.8) 92.2 ± 1.5 (83.2-108.1) Total turbidity (NTU) 118 ± 3 (84-139) 138 ± 12 (80-251)
Similar to electrical conductivity, total turbidity also varied between the lakes and the breeding seasons. In both seasons, Lake Crescent was much more turbid than Lake Sorell, with mean turbidities being 116 and 48 NTU greater in Lake Crescent (v. Lake Sorell) in 2007 and 2009, respectively. Lake Crescent was on average more turbid in 2007 than in 2009, whereas the opposite occurred in Lake Sorell. This variability between the seasons in both lakes was caused by differences in water levels in the lakes. Temporal trends in turbidity levels in the lakes and their relationships to water levels are examined in section 3.2.3.
3.2.2 Extreme conditions during low water levels in summer-autumn 2009
Continuously recording physico-chemical loggers were deployed at mid-basin sites in Lake Crescent (between 11 February and 1 April 2009) and Lake Sorell (between 11 February and 1 March 2009). During the periods when the continuous loggers were deployed, water levels in Lake Crescent (range = 801.510-801.700 m AHD) and Lake Sorell (range = 802.850- 802.900 m AHD) were very low in comparison to levels in both lakes since 1970 (Figure 3.4); therefore, the continuous measurements provide an indication of the water quality in the lakes during low water level conditions. Moreover, these data document fine-scale temporal variability in the physico-chemical conditions in the lakes, particularly variability associated with changes in ambient meteorological conditions, such as strong wind events and fluctuations in air temperature.
All physico-chemical parameters differed considerable between the lakes, either in their magnitude and/or degree of temporal variation (Table 3.2; Figures 3.8-3.11), and water quality in Lake Crescent was much more dynamic than in Lake Sorell. While water temperatures in both lakes
36 Response of Lake Crescent and Lake Sorell to water level variability during 2009 during February had a similar overall pattern, diurnally, temperature fluctuated much more in Lake Crescent than in Lake Sorell (coefficients of variation (CV) for entire deployments periods: Lake Crescent = 0.222; Lake Sorell = 0.093) (Table 3.2; Figure 3.8). For example, on February 13, water temperature varied by 10.8°C in Lake Crescent, while on the same day in Lake Sorell the range was 3.1°C (Figure 3.8). Furthermore, in Lake Crescent, during the entire deployment period, water temperature ranged from 4.9 to 22.8°C (Figure 3.8).
Table 3.2 Summary of physico-chemical variables measured by continuous loggers at mid-basin sites in Lake Crescent (between 11 February and 1 April 2009) and Lake Sorell (between 11 February and 1 March 2009). Note suspended sediment concentration, euphotic depth and Secchi depth were derived from previously established relationships between these variables and total turbidity in these lakes (Uytendaal, 2003; Uytendaal, 2006). Standard deviation = SD and coefficient of variation = CV.
Physico-chemical variable Mean min max SD CV Lake Crescent
Water temperature (°C) 14.1 4.9 22.8 3.2 0.222 Dissolved oxygen (mg L-1) 9.5 7.8 11.8 0.8 0.082 Dissolved oxygen (% saturation) 92.9 81.3 116.2 7.5 0.082 Electrical conductivity (μS cm-1) 328.4 295.1 348.8 7.3 0.022 Total turbidity (NTU) 433 165 1553 190 0.438 Suspended sediment (mg L-1) 1026 540 3052 343 0.335 Secchi depth (cm) 5.8 2.1 10.6 1.6 0.269 Euphotic depth (cm) 7.5 1.7 17.3 2.9 0.383 Lake Sorell
Water temperature (°C) 15.9 13.1 18.8 1.5 0.093 Dissolved oxygen (mg L-1) 8.9 7.9 9.7 0.4 0.045 Dissolved oxygen (% saturation) 90.2 82.9 97.3 2.5 0.028 Electrical conductivity (μS cm-1) 127.2 123.7 130.4 1.5 0.012 Total turbidity (NTU) 259 206 404 38 0.148 Suspended sediment (mg L-1) 600 530 793 51 0.084 Secchi depth (cm) 8.1 5.7 9.4 0.8 0.099 Euphotic depth (cm) 14.8 10.2 17.5 1.6 0.108
Similar to water temperature, dissolved oxygen concentrations had similar overall patterns in the lakes, but levels varied more in Lake Crescent than in Lake Sorell (CV for entire deployments periods: Lake Crescent = 0.082; Lake Sorell = 0.045) (Table 3.2; Figure 3.9). However, overall, oxygen concentrations remained at relatively high levels in both lakes during the deployment periods (i.e. >7.8 mg L-1; >81% saturation). Electrical conductivity remained quite constant in both lakes during February-April 2009, but levels were c. 2.5 times greater in Lake Crescent than in Lake Sorell (Table 3.2; Figure 3.10). As has been found previously in the Crescent-Sorell system during periods of low water levels (DPIW, 2008; Uytendaal, 2003; Uytendaal, 2006), Lake Crescent was
37 Response of Lake Crescent and Lake Sorell to water level variability during 2009 markedly more turbid than Lake Sorell (mean total turbidity: Crescent = 433 NTU; Sorell = 259 NTU) (Table 3.2; Figure 3.11). Additionally, while turbidity varied significantly in Lake Sorell (total turbidity range = 206-404 NTU; CV = 0.148), the variation in Lake Crescent was extremely large (total turbidity range = 165-1553 NTU; CV = 0.438) (Table 3.2; Figure 3.11) and far exceeded the variability recorded in this lake during 2001 when water levels were low (Uytendaal, 2003; Uytendaal, 2006).
Suspended sediment concentrations, and euphotic and Secchi depths (which indicate the degree of light attenuation in the water column) during February-April 2009 were derived from previously established relationships between these variables and total turbidity in these lakes (Uytendaal, 2003; 2006; Table 3.2). Summary statistics of these values reflect the higher and more variable turbidity levels in Lake Crescent (v. Lake Sorell), with suspended sediment concentrations being much higher in Lake Crescent, and euphotic and Secchi depths being lower in this lake. Mean euphotic depths in both lakes were <15 cm, which indicates that light penetration into the water column of the lakes was very limited in February-April 2009. Based on the suspended solid concentrations and the mean volumes of the lakes during the deployment of the loggers, during February-April 2009, there were on average 12 308 t (maximum = 36 625 t) and 49 526 t (maximum = 65 393 t) of sediment suspended in the water columns of Lake Crescent and Lake Sorell, respectively.
The differences between the physico-chemical conditions in the lakes during February-April 2009 are thought to be primarily related to differences in the depth of the water columns of the lakes at this time. Water depth at the mid-basin site in Lake Crescent where the continuous logger was deployed ranged from 590-640 mm, whereas the site in Lake Sorell was about 1.5-2 m deep. As has previously been found in these lakes, wind-induced resuspension of sediments governed the turbidity levels in the lakes during 2009 when water levels were low. Under such conditions, large areas of substrate in the lake basins would have been exposed to significant shear stress caused by wave action, thus, enabling sediment on the bed of the lakes to be disturbed and suspended in the water columns of the lakes. This is illustrated by a comparison of average wind speed data from Liawenee on the Tasmanian Central Plateau during the logger deployment period and total turbidity measurements in Lake Crescent (Figure 3.12). Figure 3.12 shows that during all of the periods where total turbidity levels in Lake Crescent were >600 NTU, the mean wind speed at Liawenee was typically >25-30 km h-1; thus, spikes in turbidity were caused by strong wind events on the Tasmanian Central Plateau. Similarly, wind action also appears to strongly influence dissolved oxygen concentrations in Lake Crescent, as, generally, oxygen concentrations in the lake increased during, and slightly in lag of, strong wind events (Figure 3.13); this relationship would be effected by water temperature which is inversely related to dissolved oxygen concentrations (Figures 3.8 and 3.9).
38 Response of Lake Crescent and Lake Sorell to water level variability during 2009
25
Crescent Sorell
20
15
Water temperatureWater (°C) 10
5
11-Feb 17-Feb 23-Feb 01-Mar 07-Mar 13-Mar 19-Mar 25-Mar 31-Mar Date
Figure 3.8 Water temperature measured by continuous loggers at mid-basin sites in lakes Crescent and Sorell between 11 February and 1 April 2009.
12
Crescent Sorell
11
)
1
l
10
9 Dissolvedoxygen (mg
8
7 11-Feb 17-Feb 23-Feb 01-Mar 07-Mar 13-Mar 19-Mar 25-Mar 31-Mar Date
Figure 3.9 Dissolved oxygen concentration measured by continuous loggers at mid-basin sites in lakes Crescent and Sorell between 11 February and 1 April 2009.
39 Response of Lake Crescent and Lake Sorell to water level variability during 2009
350
300
) Crescent 1
Sorell
m c
S S 250
200 Conductivity (
150
100 11-Feb 17-Feb 23-Feb 01-Mar 07-Mar 13-Mar 19-Mar 25-Mar 31-Mar Date
Figure 3.10 Electrical conductivity measured by continuous loggers at mid-basin sites in lakes Crescent and Sorell between 11 February and 1 April 2009.
1600 Crescent Sorell 1400
1200
1000
800
Turbidity(NTU) 600
400
200
0 11-Feb 17-Feb 23-Feb 01-Mar 07-Mar 13-Mar 19-Mar 25-Mar 31-Mar Date
Figure 3.11 Total turbidity measured by continuous loggers at mid-basin sites in lakes Crescent and Sorell between 11 February and 1 April 2009.
40 Response of Lake Crescent and Lake Sorell to water level variability during 2009
1600 70 Turbidity Wind speed
1400 60
1200
50
)
1
h 1000 40
800
30
600
Total turbidity (NTU) turbidity Total Mean wind speed (km wind(km Mean speed 20 400
10 200
0 0
11-Feb 17-Feb 23-Feb 01-Mar 07-Mar 13-Mar 19-Mar 25-Mar 31-Mar
Date
Figure 3.12 Total turbidity measured by a continuous logger at a mid-basin site in Lake Crescent and mean wind speed at Liawenee on the Tasmanian Central Plateau between 11 February and 1 April 2009.
70 Dissolved oxygen Wind speed 12 60
11 50
)
)
1
1
l h
40 10
30
9
Dissolved oxygen (mg Dissolved (mg oxygen Mean wind speed (km wind(km Mean speed 20
8 10
7 0
11-Feb 17-Feb 23-Feb 01-Mar 07-Mar 13-Mar 19-Mar 25-Mar 31-Mar
Date
Figure 3.13 Dissolved oxygen concentration measured by a continuous logger at a mid-basin site in Lake Crescent and mean wind speed at Liawenee on the Tasmanian Central Plateau between 11 February and 1 April 2009.
41 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.2.3 Long-term turbidity trends and the influence of water levels
Examining in situ measurements of total and colloidal turbidity taken in lakes Crescent and Sorell during this study (2009-2010) with historical data allows long-term trends in the water clarity in the lakes to be investigated. It should be noted that these long-term turbidity data have mostly been collected at intervals of 1 month or greater which typically have not captured the higher turbidity levels that occur during strong wind events (section 3.2.2); some of these events caused adverse conditions which make boating on the lakes unsafe, thus collecting water samples is not possible at these times. Because of this, long-term data provide a conservative measure of turbidity in the lakes, which primary depict trends in background turbidity levels.
Since autumn 1998, total and colloidal turbidities have increased from levels that are thought to reflect historical conditions in Lake Crescent (total turbidity c. 7-60 NTU; colloidal turbidity c. 5- 14 NTU) and Lake Sorell (total turbidity c. 5-40 NTU; colloidal turbidity c. 8-11 NTU) to much higher background levels (means for January to June 2010: Lake Crescent total turbidity = 122 NTU and colloidal turbidity = 50 NTU; Lake Sorell total turbidity = 83 NTU and colloidal turbidity = 64 NTU) (Figures 3.14 and 3.15). The initial increase in turbidities during autumn 1998 coincided with water levels falling below the preferred minimum levels in both lakes (DPIWE, 2005a; Figures 3.14 and 3.15). Since this time, water levels have remained below these thresholds for substantial periods. Furthermore, two episodes of very low water levels occurred in both lakes in 2000-2001 and 2008- 2009, during which water levels fell well below critical minimum thresholds (DPIWE, 2005a; Figures 3.14 and 3.15). During these episodes, total and colloidal turbidity levels in both lakes increased markedly (i.e. mean monthly total turbidities were >240 NTU and colloidal turbidities were >60 NTU), especially in Lake Crescent during 2008-2009 when mean monthly total turbidity peaked at 872 NTU and colloidal turbidity peaked at 120 NTU (Figures 3.14 and 3.15).
Associations between patterns in water levels and total and colloidal turbidity levels in lakes Crescent and Sorell are obvious (Figures 3.14 and 3.15). Furthermore, there are strong, negative log-linear relationships between mean monthly total turbidities and water levels in both lakes (Figure 3.16). However, relationships between colloidal turbidity and water levels in both lakes are not as strong due to significant hysteresis in the relationships, especially in Lake Sorell (Figure 3.17). This is because unlike larger-size particles, colloids do not settle out of the water columns of the lakes during periods of high water levels, hence times when there is less erosion of the lake beds. Therefore, since 1996, the sequential episodes of low water levels have caused step- like increases in colloidal turbidity levels in the lakes (Figures 3.15 and 3.17). Two main processes decrease colloidal turbidity levels in the lakes, dilution (via inflows) and flushing (via releases of water into the Clyde River). Reductions in colloidal turbidity levels in both lakes in late 2009 provide clear examples of decreases associated with dilution due to the relatively large, positive seasonal yields in both lakes (Figure 3.15).
42 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
Water level 350 Total turbidity
804.5 300
250 804.0
200
Pref. min
803.5 150
Total turbidity (NTU) turbidity Total Water level (m AHD) level (m Water Crit. min 100 803.0
50
802.5 0
1995 2000 2005 2010
Year
Lake Crescent
Water level Total turbidity 900
804.0 800
700 803.5
600
803.0 500 Pref. min
802.5 400
Total turbidity (NTU) turbidity Total Water level (m AHD) level (m Water Crit. min 300 802.0 200
801.5 100
0
1995 2000 2005 2010
Year
Figure 3.14 Mean monthly total turbidity and water levels in lakes Crescent and Sorell, November1991 – June 2010. The preferred and critical minimum operating levels under the Water Management Plan for the lakes (DPIWE, 2005a) are shown. Note different y-axis scales.
43 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
Water level 140 Colloidal turbidity
804.5 120
100 804.0
80
Pref. min
803.5 60 Water level (m AHD) level (m Water
Crit. min Colloidal(NTU) turbidity 40 803.0
20
802.5 0
1995 2000 2005 2010
Year
Lake Crescent
Water level 140 Colloidal turbidity
804.0 120
803.5 100
803.0 80
Pref. min 60
802.5 Water level (m AHD) level (m Water
Crit. min Colloidal(NTU) turbidity 40 802.0
20 801.5
0
1995 2000 2005 2010
Year
Figure 3.15 Mean monthly colloidal turbidity and water levels in lakes Crescent and Sorell, November1991 – June 2010. The preferred and critical minimum operating levels under the Water Management Plan for the lakes (DPIWE, 2005a) are shown. No colloidal turbidity data are available prior to October 1996.
44 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Crescent Sorell
3.0
2.5
2.0
(Total turbidity (NTU)) turbidity (Total 1.5
10
g
o L
1.0
2.9040 2.9045 2.9050 2.9055
Log10(Water level (m AHD))
Figure 3.16 Relationships between mean monthly total turbidity (TT) and water levels (WL) in lakes Crescent and Sorell, November1991 – June 2010. Linear models of the log-log relationships are indicated by dashed lines (Lake Crescent: log10 TT = 3936 - log10(1354 WL), R2 = 0.73; Lake Sorell: log10 TT = 4234 - log10(1457 WL), R2 = 0.66).
45 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
120 2009
100
80 2010
60 2000 40
20 1996
0 Lake Crescent 2009 120
100 Colloidal(NTU) turbidity 80
60 2001
40 2010
20 1996
0 801.5 802.0 802.5 803.0 803.5 804.0 804.5 Water level (m AHD)
Figure 3.17 Relationships between mean monthly colloidal turbidity and water levels in lakes Crescent and Sorell, October 1996 – June 2010. Lines between sequential months are shown. The temporal sequence of data from both lakes is indicated by labels for years on selected data points.
46 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.2.4 Turbidity in littoral wetlands during 2009
Prior to this study, no water quality measurements had been formerly undertaken in the littoral wetlands of lakes Crescent and Sorell when they have been inundated by water from the main basins of the lakes. The relatively high water levels in the lakes during winter-spring 2009, which resulted in extensive inundation of the wetlands, provided an opportunity to examine turbidity levels in the littoral wetlands of the lakes.
Eight water samples were collected in the Interlaken Lakeside Reserve wetland on 9 October 2009 and 10 samples were collected in Silver Plains Marsh on 5 November 2009. The water levels in Lake Crescent and Lake Sorell at the time of the samplings were 803.160 m AHD and 804.100 m AHD, respectively. The sites where the samples were collected were distributed across the wetlands and the mean depth of the sites in the Interlaken Lakeside Reserve was 281 mm (range = 105-370 mm), whereas in Silver Plains Marsh the mean was 252 mm (range = 190- 350 mm).
Turbidity (both total and colloidal) levels differed markedly between the main basins of the lakes and the wetlands. Mean (± S.E.) total and colloidal turbidity in Lake Crescent were 157.5 ± 6.5 NTU and 50.4 ± 0.1 NTU, respectively, whereas in the Interlaken Lakeside Reserve wetland these values were 36.4 ± 9.3 NTU and 21.9 ± 4.1 NTU, respectively. Mean (± S.E.) total and colloidal turbidity in Lake Sorell were 114.8 ± 10.0 NTU and 64.6 ± 0.9 NTU, respectively, whereas in the Silver Plains Marsh these values were 44.2 ± 6.9 NTU and 36.6 ± 5.7 NTU, respectively. The percent of colloidal material contributing to total turbidity levels in the wetlands was quite high in both wetlands (based on mean total and colloidal turbidities: Interlaken Lakeside Reserve = 60%; Silver Plains Marsh = 83%), especially in comparison to levels in the lakes (based on mean total and colloidal turbidities: Lake Crescent = 32%; Lake Sorell = 56%).
Spatially, turbidity levels varied greatly in both wetlands; especially in comparison to their respective lakes where turbidity levels are generally spatially uniform (Figures 3.18 and 3.19). The ranges of total and colloidal turbidities in Interlaken Lakeside Reserve were 10.6-86.2 NTU and 7.6-37.0 NTU, respectively, and in Silver Plains Marsh the ranges of these parameters were 12.1-71.3 NTU and 10.5-57.6 NTU, respectively. Spatial variability in turbidity levels in both wetlands shows that the influence of lake water on water quality in these littoral systems is not uniform. In the Interlaken Lakeside Reserve wetland, total and colloidal turbidity levels were quite high in the south-western area of the marsh and along its in-lake edge, but levels of these parameters were lower in the outer and central northern areas (Figure 3.18). Similar spatial patterns were evident in Silver Plains Marsh, with higher total and colloidal turbidity levels occurring in the northern area of the marsh and along its in-lake edge (Figure 3.19).
47 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.18 Total and colloidal turbidity readings in Interlaken Lakeside Reserve, Lake Crescent, 9 October 2009. Lake perimeters are at full supply levels. Colloidal turbidities are shown in parentheses. Variation in the diameter of the sampling site symbols reflects the magnitude of total turbidity readings.
48 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.19 Total and colloidal turbidity readings in Silver Plains Marsh, Lake Sorell, 5 November 2009. Lake perimeters are at full supply levels. Colloidal turbidities are shown in parentheses. Variation in the diameter of the sampling site symbols reflects the magnitude of total turbidity readings.
49 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.3 Habitat conditions and water level – induced alterations
As previously mentioned (section 3.1.2), the water level conditions that occurred during 2009 in Lake Crescent and Lake Sorell were quite extraordinary in comparison to historical water levels and annual water level fluctuations since 1970 (Figure 3.4). After three previous years of relatively low levels, water levels were very low in both lakes during summer-autumn 2009, especially in Lake Crescent where they reached the lowest recorded level since 1970 (801.510 m AHD). Following these drought-induced low levels in early 2009, substantial seasonal rises occurred in both lakes in winter-spring due to heavy rainfall in the catchment. The large seasonal water level fluctuations during 2009 not only significantly influenced physico-chemical variables (water quality) in the lakes (section 3.2), but also caused large areas of different types of littoral habitats to be dewatered and then subsequently inundated. The following sections of this report (sections 3.3.1 and 3.3.2) document changes in the habitat conditions in the Crescent-Sorell system caused by water level fluctuations during 2009.
3.3.1 Effects of low water levels during summer-autumn 2009
Low water levels in summer-autumn 2009 substantially reduced the surface area of Lake Crescent (14.96 km2; 65% of total area at full supply level) and Lake Sorell (45.71 km2; 82% of total area at full supply level) (Figure 3.20), and decreased the diversity of wetted habitats in the lakes. These conditions dewatered: (1) extensive littoral wetlands in both lakes (Figures 3.20 and 3.21), (2) basically all sandy shore habitat in both lakes (Figures 3.20, 3.22 and 3.23), and (3) the majority of the rocky shore habitat in Lake Crescent (Figures 3.20 and 3.24-3.28) and a substantial proportion of this habitat in Lake Sorell (Figure 3.20). The dewatering of these littoral habitats meant that the substrate in the inundated areas of Lake Crescent was dominated by consolidated sediment, whereas in Lake Sorell consolidated sediment was prominent, but areas of rocky substrata were also present. The water levels also fell below sill levels in both lakes, which prevented water from being released from Lake Sorell to Lake Crescent, and out of Lake Crescent into the Clyde River (see Figure 3.29).
The littoral wetlands of Lake Crescent (total area = 4.18 km2) and Lake Sorell (total area = 5.77 km2) had already been dewatered for some time prior to early 2009: in both lakes extensive inundation last occurred in 1996 and minor inundation occurred periodically between 2003 and 2006 (Figure 3.4). Therefore, the low water levels in summer-autumn 2009 extended the period during which these areas had been dewatered. The very low levels in Lake Crescent appeared to also substantially reduce the moisture content of soil in the wetlands (S. A. Hardie, pers. obs.) by lowering groundwater levels in these areas.
50 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Because the sandy shores in both lakes were largely dewatered (dewatered areas: Lake Crescent = 2.00 km2; Lake Sorell = 0.64 km2), aeolian (wind-related) processes were able to re- work these areas (Figure 3.22) and the lunette systems (Figure 3.23) that are located along these shores at the perimeters of the lakes (at full supply levels). This marked increase in aeolian activity appeared to be a relatively large-scale ‘event’ that had not occurred for some time (due to higher water levels in the lakes), with live terrestrial vegetation being covered by sand in many areas of the lunettes (Figure 3.23).
Based on water level records since 1970, the extent of dewatering of the rocky shores in Lake Crescent (total area = 0.30 km2) during summer-autumn 2009 was unprecedented. These habitats were almost completely dewatered at this time (83% dewatered), which allowed their extent to be re-surveyed (see section 3.4), This detailed work on the rocky shores also provided insights into the physical structure of these habitats and processes that influence the condition of these areas (section 3.5).
The structure of rocky shores varies across and within the four defined areas of this habitat in Lake Crescent (section 3.4). Overall, the topography of rocky shores is quite uniform with a gentle slope from the perimeter of the lake to the in-lake edge of these habitats (which is typically well defined; Figures 3.24, 3.27 and 3.28), but in some areas there are distinct ‘steps’ and irregularities in the gradient (Figures 3.24 and 3.25). Substrate on the rocky shores is generally dominated by cobbles and boulders with some extrusions of bedrock; however, particle size diversity, and hence habitat structural complexity, gradually decreases from the outer perimeter to the in-lake edge of the shores (Figures 3.24 and 3.25). Furthermore, in many places, especially near the in-lake edge of rocky shores, rocky substrata are embedded in sand- or sediment-based substrate (Figures 3.26-3.28). This variation in the structure of the rocky shores has ramifications for aquatic fauna which use these habitats, including G. auratus which prefers cobble-dominated substrata with a high degree of interstitial space for spawning.
51 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.20 Mosaic image of aerial photographs of lakes Crescent and Sorell that were taken on 20 and 21 March 2009. The water levels in Lake Crescent and Lake Sorell at this time were 801.570 m AHD and 802.840 m AHD, respectively.
52 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.21 Interlaken Lakeside Reserve wetland, Lake Crescent, 18 May 2009. The wetland was completely dewatered at this time.
Figure 3.22 Table Mountain Shore, Lake Crescent, 3 March 2009. Note dark-coloured organic sediment covering the sandy substrata of the shore, which is likely to have originated from the main basin of the lake.
53 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.23 Lunette and terrestrial vegetation, southern end of Sandbank Shore, Lake Crescent, 19 May 2009. Note recent aeolian activity (i.e. movement of sand particles by wind) has partly buried some of the vegetation.
Figure 3.24 Boathouse Shore, Lake Crescent, 19 May 2009. An extensive amount of rocky substrate is dewatered.
54 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.25 Boathouse Shore, Lake Crescent, 19 May 2009. Two distinct layers in the bathymetry of the rocky shore are shown, with the in-lake edge of the upper layer being the interface between the zone of boulders and the relatively uniform area of cobble which extends to the wetted edge of the lake.
Figure 3.26 Boathouse Shore, Lake Crescent, 19 May 2009. An area of rocky shore that contains sparse rock embedded in sand-based substrate is shown in the foreground.
55 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.27 Boathouse Shore, Lake Crescent, 19 May 2009. The in-lake edge of the rocky shore is quite distinct. Beyond the in-lake edge of the rocky shore, only sparse rock is present and it is mostly embedded in sediment-based substrate.
Figure 3.28 Tea Tree Point Shore, Lake Crescent, 18 May 2009. This region of rocky shore has a defined in-lake edge, and contains areas of bedrock and rocks embedded in sediment-based substrate.
56 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.29 Lake Crescent, the in-lake edge of Clyde Marsh and the entrance to the canal which forms the outflow of the lake into the Clyde River. View from the boat ramp at Tea Tree Point, 1 April 2009. Note platypus (Ornithorhynchus anatinus) tracks in soft sediment in the foreground of the photograph. The water level in the lake at this time was ~801.510 m AHD, which was the lowest recorded level since January 1970.
3.3.2 Effects of seasonal water level rise and high levels during winter-spring 2009
The large seasonal rises in water levels in lakes Crescent and Sorell during winter-spring 2009 (Figure 3.4) increased the surface area of Lake Crescent (22.62 km2; 98% of total area at full supply level) and Lake Sorell (53.13 km2; 96% of total area at full supply level) (Figure 3.30), and increased the diversity of wetted habitats in the lakes. These conditions inundated relatively large proportions of: (1) the littoral wetlands in Lake Crescent and Lake Sorell (Figures 3.30-3.34), (2) sandy shore habitat in Lake Crescent and Lake Sorell (Figures 3.30, 3.35 and 3.36), and (3) rocky shore habitat, especially in Lake Crescent (Figures 3.30, 3.37 and 3.38). Inundation of these littoral habitats meant that substrate in the inundated areas of both lakes included aquatic vegetation (in the wetlands), sand, rocks and consolidated sediment. The water levels also rose above sill levels in both lakes, which enabled water to be released from Lake Sorell to Lake Crescent (Figure 3.1), and out of Lake Crescent into the Clyde River (see Figure 3.39).
During late 2009 – early 2010, the littoral wetlands in lakes Crescent and Sorell were inundated to the greatest extent since 1996 (proportion inundated: Lake Crescent = 89%; Lake Sorell = 73%; Figures 3.4 and 3.30-3.34). This was somewhat remarkable, given the very low water levels that occurred in both lakes in early 2009, and illustrated how quickly the conditions in the Crescent- Sorell system can change due to local climatic conditions. In Lake Sorell, the water level reached
57 Response of Lake Crescent and Lake Sorell to water level variability during 2009 the surveyed edge of the wetlands (803.600 m AHD) on 16 August 2009 and, up to 20 September 2010, it had remained above edge of the wetlands (Figure 3.4; period of 400 days which is continuing). According to a bathymetric survey of the wetlands in the Crescent-Sorell system (Heffer, 2003c), a water level of 803.900 m AHD is likely to inundate 90% of the surface area of three of the four wetlands in Lake Sorell to a depth of 10 mm. During late 2009 – early 2010, water levels in Lake Sorell remained above 803.900 m AHD for 119 days. This proposed depth of inundation is supported by the results of a depth survey of Silver Plains Marsh on 4 November 2009 when the water level in Lake Sorell was 804.100 m AHD (Figure 3.34). During this survey, depth measurements at 109 locations showed the mean (± SE) depth in the wetland was 238 ± 7 mm (range = 0-450 mm).
In Lake Crescent, the water level reached the surveyed edge of the wetlands (802.700 m AHD) on 10 September 2009 and it remained above the edge of the wetlands until 1 April 2010 (Figure 3.4; period of 203 days). According to a bathymetric survey of the wetlands in the Crescent-Sorell system (Heffer, 2003c), a water level of 803.000 m AHD is likely to inundate 90% of the surface area of both wetlands in Lake Crescent to a depth of 10 mm. During late 2009 – early 2010, water levels in Lake Crescent remained above 803.000 m AHD for approximately 109 days. This proposed depth of inundation is supported by the results of a depth survey of the Interlaken Lakeside Reserve wetland on 9 October 2009 when the water level in Lake Crescent was 803.160 m AHD (Figure 3.33). During this survey, depth measurements at 66 locations showed the mean (± SE) depth in the wetland was 289 ± 11 mm (range = 0-520 mm). The extensive inundation of the littoral wetlands in the Crescent-Sorell system provided suitable conditions for aquatic vegetation in these areas to regenerate (Figure 3.32; DPIPWE, 2010) and allowed aquatic (e.g. macroinvertebrates, fish, waterbirds, platypus, etc.; Figure 3.32) and terrestrial (e.g. grazing marsupials) fauna to use these productive, wetted habitats.
Inundation of the majority of the sandy shores in both lakes (proportion inundated: Lake Crescent = 90%; Lake Sorell = 93%) reduced the influence of aeolian processes (wind) on these areas and their associated lunette systems. Instead wave action and patterns of surface water flow on the sandy shores and lunettes re-worked sediments in the areas. These conditions recharged water levels in lagoons in the lunettes (Figure 3.36) and displaced large qualities of sand in the channels of creeks in these areas (Figure 3.35).
Based on water level records since 1970, the extent of the seasonal inundation of rocky shores in Lake Crescent and Lake Sorell during winter-spring 2009 was unprecedented. In Lake Crescent, these habitats were almost completely dewatered during autumn 2009 (Figure 3.24), but the large magnitude of the water level rise in the lake during winter-spring (1.780 m) almost completely inundated these areas (97% inundated; Figure 3.37). In both lakes, the high water levels inundated terrestrial vegetation that had encroached onto the outer margins of rocky shores in some areas
58 Response of Lake Crescent and Lake Sorell to water level variability during 2009
(Figure 3.38). Similar to surveys of the rocky shores in Lake Crescent in May 2009, the extent of the inundation in these areas was surveyed in October 2009, and this work provided further insights into the physical processes that influence the condition of these habitats (section 3.5).
59 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.30 Mosaic image of aerial photographs of lakes Crescent and Sorell that were taken on 7 November 2009. The water levels in Lake Crescent and Lake Sorell at this time were ~803.260 m AHD and 804.090 m AHD, respectively.
60 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.31 Clyde Marsh, Lake Crescent, 5 November 2009. View from the sluice gates at the outflow of the lake into the Clyde River.
Figure 3.32 Black swans (Cygnus atratus) in Clyde Marsh, Lake Crescent, 5 November 2009. Note the dense cover of aquatic macrophytes in the wetland. View from the sluice gates at the outflow of the lake into the Clyde River.
61 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.33 Interlaken Lakeside Reserve wetland, Lake Crescent, 9 October 2009. Note this photo was taken during the survey that assessed the extent to which this wetland was inundated in spring 2009.
Figure 3.34 Silver Plains Marsh, Lake Sorell, 5 November 2009. Note this photo was taken during the survey that assessed the extent to which this wetland was inundated in spring 2009.
62 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.35 Mouth of Agnews Creek into Lake Crescent on Sandbank Shore, 21 July 2009. Note the substantial volumes of sand that appear to have been moved by flows in the creek.
Figure 3.36 An unnamed lagoon in the lunette system on Sandbank Shore, Lake Crescent, 9 October 2009. Note the relatively large sand dune which appears to be encroaching into basin of the lagoon.
63 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.37 Boathouse Shore, Lake Crescent, 8 October 2009. Note most of the rocky substrate on this shore was inundated.
Figure 3.38 Hatchery Shore, Lake Sorell, 4 September 2009. View from the mouth of Mountain Creek into the lake. Note the rising water level in the lake is inundating terrestrial vegetation which had encroached onto the rocky shore during the extended period of low water levels.
64 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.39 Lake Crescent, the in-lake edge of Clyde Marsh and the entrance to the canal which forms the outflow of the lake into the Clyde River. View from the boat ramp at Tea Tree Point, 4 November 2009. The water level in the lake at this time was ~803.290 m AHD, which was the highest level during 2009 (and since January 1997).
3.4 Water levels and area of inundated littoral rocky substrate in Lake Crescent
As identified in a previous survey in March 2002 (Hardie, 2003a; Hardie et al., 2007b), four defined shores contain rocky substrate in Lake Crescent: Boathouse Shore, Lakeside Island shore, Tea-tree Point shore and Triffitt Point shore (Table 3.3; Figure 3.40). Aerial photographs of these areas in March 2009, and on-ground surveys of rocky substrate on these shores in May 2009, provided accurate estimates of the extent of littoral areas of this substrate in Lake Crescent. These estimates are thought to be more precise than those derived in the previous survey in 2002. This is largely because lower water levels in 2009 (levels of ~801.570 m AHD in 2009 cf. ~802.440 m AHD in 2002) provided better conditions for determining the extent of the rocky substrate (i.e. greater areas of rocky substrate were dewatered and areas that were inundated were quite shallow, thus, could be accurately surveyed).
The surface area of rocky substrate on the four shores in Lake Crescent varies, with Triffitt Point shore (25 090 m2) and Boathouse Shore (132 970 m2) having the least and greatest areas, respectively (Table 3.3; Figure 3.40). The total surface area of littoral rocky substrate in Lake Crescent was found to be 296 610 m2 (c. 0.30 km2), which accounts for 1.3% of the surface area of Lake Crescent (including its littoral wetlands).
65 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Table 3.3 Surface area of littoral regions of rocky substrate in Lake Crescent.
Shore Surface area (m2) Triffitt Point 25 090 Tea-tree Point 45 500 Lakeside Island 93 050 Boathouse Shore 132 970 Total 296 610
A 3D terrain model of Lake Crescent clearly illustrates the bathymetry of the lake (Figure 3.41). The main basin of Lake Crescent is relatively uniform and flat (level of c. 800.400 m AHD) and its littoral zone (range of levels = 800.400-803.800 m AHD) varies in width, with sandy shores (see Figure 3.20) and wetlands (see Figures 2.1 and 3.30) having expansive areas of relatively flat and shallow substrate, while rocky shores have substrate of steeper gradient which causes the littoral zone to constrict on these shores (Figure 3.41).
The 3D terrain model of Lake Crescent was used to estimate the total area of rocky shore inundated and ‘optimal’ G. auratus spawning habitat (i.e. rocky shore at depths of 0.2-0.6 m; notwithstanding the effects of sedimentation on this habitat) at 14 water levels between 803.800 m AHD and 801.200 m AHD at intervals of 0.2 m (Table 3.4). An illustration of this exercise for Tea-Tree Point shore at a water level of 802.600 m AHD is shown in Figure 3.42. In Figure 3.42, the extent of the rocky substrate on this shore is the region inside the thick black line, the total area of rocky substrate inundated at this water level is shaded purple, and optimal spawning habitat for G. auratus at this level is shaded orange. Due to the relatively steep gradient of the rocky substrate on Tea-Tree Point shore, at a water level of 802.600 m AHD, the area of spawning habitat for G. auratus is much less than the total area of this substrate that is inundated (40 461 m2 v. 152 780 m2, respectively) (Table 3.4; Figure 3.42). Similar patterns were also evident on the other three rocky shores in Lake Crescent at a range of water levels; thus, the depth range at which G. auratus deposits its eggs restricts the area of potential spawning habitat for this species at any given water level.
66 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.40 Littoral areas of rocky substrate in Lake Crescent. Lake perimeter is at full supply level.
67 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.41 Three-dimensional terrain model of Lake Crescent basin with 0.2 m depth contours. The outlines of the four littoral areas of rocky substrate are shown (brown line). Lake perimeter is at full supply level.
68 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Table 3.4 Surface area of inundated littoral rocky substrate and spawning habitat for Galaxias auratus at water levels between 801.200 m AHD and 803.800 m AHD in Lake Crescent.
Water level (m AHD) Total area of inundated rocky Area of inundated spawning substrate (m2) habitat (m2)
803.8 296 610 47 423 803.6 265 613 44 982 803.4 240 861 44 140 803.2 218 189 43 097 803.0 195 878 42 175 802.8 174 048 41 473 802.6 152 780 40 461 802.4 131 872 39 909 802.2 111 307 42 892 802.0 91 411 48 875 801.8 71 397 35 344 801.6 48 519 18 717 801.4 22 522 9 370 801.2 13 174 3 804
69 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.42 Total area of rocky substrate and the area of optimal spawning habitat for Galaxias auratus at Tea-Tree Point, Lake Crescent at a water level of 802.600 m AHD. Area estimates are based on a 3- demensional terrain model of the lake basin. Lake perimeter is at full supply level.
70 Response of Lake Crescent and Lake Sorell to water level variability during 2009
There is a positive linear relationship between water levels and the proportion of inundated littoral rocky substrate in Lake Crescent (Table 3.4; Figure 3.43). Approximately 38% and 55% of the total area of littoral rocky substrate is inundated at the critical minimum (802.200 m AHD) and preferred minimum (802.700 m AHD) operating levels for this lake under the current Water Management Plan for the Crescent-Sorell system (DPIWE, 2005a) (Figure 3.43). Interestingly, the relationship between water levels and the area of spawning habitat for G. auratus is non-linear, with water levels of 802.000-803.800 m AHD inundating approximately 40 000-50 000 m2 of spawning habitat and water levels <802.000 m AHD significantly reducing the area of inundated habitat (Figure 3.44). A decrease in water level of 0.8 m between 802.000 m AHD to 801.200 m AHD reduces the amount of spawning habitat by approximately 92%.
Based solely on the area of inundated spawning habitat, a level of 802.000 m AHD is a significant threshold, below which reductions in water level cause a significant decrease in available spawning habitat for the G. auratus population in Lake Crescent. However, inundation (i.e. availability) is not the only issue governing the use of rocky substrata by G. auratus for spawning or the subsequent success of spawning events: the ‘quality’ (degree of sedimentation and habitat structural complexity) of the substrate is also thought to be important. Sedimentation on littoral areas of rocky substrate is thought to have a significant influence on the use of rocky substrata by G. auratus for spawning and the successful incubation of eggs (DPIW, 2008; Hardie, 2003a; Hardie et al., 2007b), and sedimentation rates are greater at lower water levels (sections 3.2 and 3.5). Additionally, G. auratus prefers rocky substrate with a high degree of structural complexity (interstitial space) for spawning (DPIW, 2008; Hardie, 2003a; Hardie et al., 2007b), and the complexity of rocky substrate on the rocky shores in Lake Crescent is less in deeper, more in-lake areas of the shores (section 3.5). Therefore, interpretation of the relationship between water levels and area of spawning habitat for G. auratus should be viewed in this context (see Figure 3.44).
Thus, while an area of 40 000-50 000 m2 is thought to provide sufficient spawning habitat for the G. auratus population in Lake Crescent, and water levels of 802.000-803.800 m AHD inundate this amount of habitat (Figure 3.44), there is a progressive increase in the quality of habitat with an increase in water level. Figure 3.44 illustrates this, where water levels less than the critical minimum (i.e. <802.200 m AHD; region a) provide poor quality habitat, levels between the critical minimum and the preferred minimum (i.e. 802.200-802.700 m AHD; region b) provide moderate quality habitat, and levels greater than the preferred minimum (i.e. >802.700 m AHD; region c) provide good quality habitat. Therefore, the critical minimum water level of 802.200 m AHD for Lake Crescent under the current WMP (DPIWE, 2005a) for the Crescent-Sorell system, is an important threshold for the management of this lake, as below this level, decreases in water levels reduce the area of available rocky shore spawning habitat for G. auratus and the quality of the inundated habitat is poor.
71 Response of Lake Crescent and Lake Sorell to water level variability during 2009
100
80
60 55% inundated
40 38% inundated
Proportion of rocky shore inundated(%) shore rocky of Proportion 20
Crit. min. Pref. min.
0
801.2 801.6 802.0 802.4 802.8 803.2 803.6 Water level (m AHD)
Figure 3.43 Relationship between water levels and the proportion of littoral rocky substrate inundated in Lake Crescent. The preferred and critical minimum operating levels under the Water Management Plan for the lakes (DPIWE, 2005a) are shown.
72 Response of Lake Crescent and Lake Sorell to water level variability during 2009
50000
(a) (b) (c)
) 2
40000
30000
20000
Area of spawning habitat inundated (m spawninginundated (m of habitat Area 10000 Crit. min. Pref. min.
0 801.2 801.6 802.0 802.4 802.8 803.2 803.6 Water level (m AHD)
Figure 3.44 Relationship between water levels and the surface area of spawning habitat for Galaxias auratus in Lake Crescent. The preferred and critical minimum operating levels under the Water Management Plan for the lakes (DPIWE, 2005a) are shown. A surface area of approximately 40,000-50,000 m2 (grey shaded region) is thought to provide sufficient spawning habitat for the population. A progressive increase in the quality (i.e. degree of sedimentation and habitat structural complexity) of the habitat is indicated, where water levels less than the critical minimum (i.e. <802.20 m AHD; region a) provide poor quality habitat, levels between the critical minimum and the preferred minimum (i.e. 802.20-802.70 m AHD; region b) provide moderate quality habitat, and levels greater than the preferred minimum (i.e. >802.70 m AHD; region c) provide good quality habitat.
73 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.5 Water levels, and physical structure and processes of littoral areas of rocky substrate
Between 2000 and 2009, a substantial volume of work has been undertaken in rocky shore habitats in lakes Crescent and Sorell. This has included periodically sampling G. auratus eggs and juveniles/adults (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b; section 3.6), on- ground surveys of the extent and condition of rocky substrate in 2002 (Hardie, 2003a) and 2009 (section 3.4), measuring volumes of sediment in flux in these areas in 2007 (DPIW, 2008), and taking aerial photographs of these habitats in 2009 (section 3.4). This work, which has been conducted under varying water level conditions, has provided detailed knowledge of the physical structure of these habitats in lakes Crescent and Sorell, and the processes which influence them; especially in Lake Crescent where the limited extent of rocky shores is a significant management issue. This conceptual understanding for Lake Crescent is synthesised in Table 3.5 and Figure 3.45, where the influence of two water level regimes (water levels fluctuations above the preferred minimum (802.7 m AHD) and fluctuations below this level) on five lateral substrate zones is documented. This conceptual understanding focuses on the structure of the rocky shores in their current state and the processes acting on them. It does not consider long-term changes that would have occurred during the evolution of the lake basin and the progressive raising of the water level in the lake since the 1830s.
The substrate in margins of littoral areas of rocky substrate in Lake Crescent nearest the perimeter of the lake at full supply level (Figure 3.45; zone 1) has a relatively steep gradient and mostly contains large-sized dolerite rocks (boulders) and bedrock (Table 3.5; Figure 3.45). Immediately in- lake of this area where the gradient of the substrate becomes more gentle (Figure 3.45; zone 2), there is a diverse range of particle sizes (i.e. pebble, cobble and boulder), and hence a high degree of habitat structural complexity. From this point to beyond the edge of the rocky shores (Figure 3.45; zone 5), particle size diversity, and hence habitat structural complexity, gradually decreases until the substrate is comprised on barren sediment (Table 3.5; Figure 3.45). The rock-sediment interface of the shores (Figure 3.45) is well defined in many areas, but less in others. Rocky substrata occur below the interface in some areas, but the cover of rock is typically very sparse and most rocks are quite embedded in the sediment-based substrate. Beyond the edge of the rocky shores (Figure 3.45; zone 5), the substrate is mostly consolidated sediment, which is covered by a layer of unconsolidated sediment that becomes suspended in the water column under turbulent conditions. This layer of mobile sediment is an important feature of the rocky shores in Lake Crescent, as it influences the condition (i.e. degree of sedimentation) on areas of rocky substrata near the rock- sediment interface.
74 Response of Lake Crescent and Lake Sorell to water level variability during 2009
The physical structure of the substrate on rocky shores in Lake Crescent is largely controlled by six key processes: (1) wetting and drying, (2) wind, (3) frost, (4) wave action, (5) sediment erosion, (6) sediment accumulation (Table 3.5). Water level fluctuations directly and indirectly influence all of these processes and are important for maintaining the structure of rocky shores, as they enable natural levels of shoreline sediment erosion and accumulation. However, under relatively low and high water level regimes (i.e. fluctuation ranges below and above the preferred minimum level, respectively; regimes A and B in Table 3.5 and Figure 3.45), these processes influence different lateral regions of the shores. The effects of low and high water levels on rocky shores are explained in detail in Table 3.5. Broadly, low water levels reduce the surface area of the rocky shores which is inundated and expose substrata in the in-lake margins of the rocky shores (Figure 3.45; zones 3 and 4), and also the main basin of the lake, to wave action. This increases sediment erosion in these areas, frequently dewaters the rock-sediment interface, and increases the influence of the mobile sediment layer on the remaining inundated areas of rocky substrate. Conversely, high water levels increase the surface area of the rocky shores which is inundated, reduce exposure of the in- lake regions of the rocky shores (Figure 3.45; 3 and 4) to wave action, and reduce the influence of the mobile sediment layer on the inundated areas of rocky shores.
75 Response of Lake Crescent and Lake Sorell to water level variability during 2009 Table 3.5 Physical structure and processes acting on littoral areas of rocky substrate in Lake Crescent under two water level regimes: (A) preferred operating range under the current Water Management Plan for lakes Crescent and Sorell, and (B) low levels as have occurred during drought conditions in recent years. Alterations to physical structure and processes associated with the shift from regime A (high water levels) to low water levels are listed under regime B. This conceptual information is based on processes that have previously been identified in such habitats by other researchers (Furey et al., 2004; Hakanson, 1977; Imboden, 2004) and observations in these habitats in Lake Crescent between 2000 and 2009.
Water level regime A: preferred water level operating range Water level regime B: low water levels Zone 1 Zone 1 Steep gradient edge of lake. Permanently dewatered. Substrate dominated by boulders and bedrock, with some cobble. High interstitial Terrestrial vegetation in riparian zone encroaches onto substrate. space between rocks due to erosion. Accumulation of organic material from riparian zone. Frequently dewatered by low water levels. Zone exposed to wave action. An area of sediment erosion. Substrate frequently exposed to wetting and drying, wind and frost. Zone 2 Zone 2 Generally low gradient substrate, but with variable topography. Permanently dewatered. Diverse rocky substrata including pebble, cobble, boulder with some bedrock. High Terrestrial vegetation in riparian zone encroaches onto substrate. interstitial space between rocks. Substrate frequently exposed to wetting and drying, wind and frost. Often dewatered by low water levels. Zone exposed to wave action. An area of sediment erosion. Substrate only exposed to wetting and drying, wind and frost at extreme low water levels (infrequent under regime A). Zone 3 Zone 3 Low gradient substrate. Increasing interstitial space with depth. Rocky substrata comprised of cobble and boulder. Low interstitial space due to Frequently dewatered by low water levels. sedimentation. Zone more exposed to wave action. Sediment erosion rates increase. Permanently inundated. Rock-sediment interface moves further into the lake basin, exposing some rocky Zone marginally exposed to wave action. An area of sediment erosion and substrata that were previously covered by sediment. accumulation. Substrate frequently exposed to wetting and drying, wind and frost. Surface area of rocky substrate controlled by sediment deposition patterns which in Smothering of substrate occurs due to remobilization of fine sediments from deeper turn are controlled by water levels and wind-induced wave action. areas of the lake that are not usually affected by wind-induced wave action. Zone 4 Zone 4 Low gradient substrate. Increase in interstitial space as wave action increases. Sparse rocky substrata, mostly boulders. Very low interstitial space. Infrequently dewatered by low water levels. Permanently inundated, but affected by wind-induced wave action. Zone more frequently exposed to wave action. Sediment erosion rates increase. An area of sediment accumulation and reworking. Rock-sediment interface dewatered in most areas. Surface area of rocky substrata heavily influenced by distribution of sediments due to Substrate infrequently exposed to wetting and drying, wind and frost. reworking. Area affected by reworking of fine sediment derived from deeper areas of lake which smothers substrate. Zone 5 Zone 5 Low gradient substrate. Zone becomes exposed to wave action. Sediment erosion rates increase. No rocky substrata only mobile sediment and silt, clay and consolidated sediments. Area affected by reworking of fine sediment derived from deeper areas of lake which Permanently inundated. smothers substrate. An area of sediment accumulation and reworking.
76 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.45 Conceptual diagram of a cross-section of a littoral area of rocky substrate in Lake Crescent. Physical structure of substrate and physical processes acting on five littoral zones are shown for two water level regimes: (A) preferred operating range under the current Water Management Plan for lakes Crescent and Sorell, and (B) low levels as have occurred during drought conditions in recent years. This diagram is based on processes that have previously been identified in such habitats by other researchers (Furey et al., 2004; Hakanson, 1977; Imboden, 2004) and observations in these habitats in Lake Crescent between 2000 and 2009.
77 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.6 Spawning activity of golden galaxias
3.6.1 Spawning habitat and habitat conditions
The macro-habitats (rocky shores) and micro-habitats (cobble substrata of c. 20-250 mm diameter; <0.9 m deep) where G. auratus eggs were collected in Lake Crescent and Lake Sorell during 2009 were similar to those observed in the 2000, 2001, 2002 (Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b), 2005 (S. A. Hardie, unpubl. obs.), and 2007 (DPIW, 2008) breeding seasons in these lakes. Spatial patterns of egg deposition on rocky substrata were also consistent with the findings of previous work, with eggs scattered on the top and sides of rocks, mostly individually or in small clusters of 2-5 eggs.
At the spawning monitoring sites in both lakes (Figures 3.46-3.49), the zone where spawning was taking place, hence eggs were being deposited, moved a significant distance laterally (i.e. in-shore) during the breeding season (see Figure 3.50), especially in Lake Crescent where eggs were deposited over a lateral distance of approximately 15-40 m. This distribution pattern was associated with the relatively large magnitudes and rates of the rises in water levels in both lakes during winter- spring 2009, with eggs being continually deposited at a reasonably constant depth range during the spawning periods in both lakes (see Figure 3.50).
Sweep net G. auratus egg samples (all samples including those not containing eggs) were collected at depth ranges of 170-650 mm in Lake Crescent and 250-900 mm in Lake Sorell. In Lake Crescent, the mean (± SE) depth at which eggs were found was 370 ± 26 mm (range = 170- 650 mm), whereas in Lake Sorell, it was 494 ± 50 mm (range = 250-780 mm). The difference in depths at which eggs were found in each lake was likely to have been associated with the difference in the bathymetry of the two sites: the site in Lake Sorell was deeper and had a steeper gradient than the site in Lake Crescent.
The depth range at which eggs were collected during this study (and also the 2007 study) encompasses the depths at which newly deposited eggs (i.e. those in the early stages of development) occur and those incubating following earlier spawning events (i.e. those in later stages of development) (see Figure 3.50). Collectively, during the 2007 (DPIW, 2008) and 2009 (this study) breeding seasons, G. auratus eggs were collected at a depth range of 150-840 mm. This range is a slightly greater than that observed during the 2000-2002 breeding seasons (200- 600 mm; Hardie et al., 2007b). However, this difference in depth ranges is likely to reflect the larger magnitudes and rates of the seasonal rises in water levels which occurred during 2009, and therefore provides a more accurate estimate of the depth range at which eggs of this species are deposited and incubate.
78 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.46 Galaxias auratus spawning monitoring site on Boathouse Shore, Lake Crescent, 21 July 2009. Note the large area of rocky substrate which is dewatered.
Figure 3.47 Galaxias auratus spawning monitoring site on Boathouse Shore, Lake Crescent, 8 October 2009. Note most of the rocky substrate on this shore is inundated.
79 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.48 Galaxias auratus spawning monitoring site on Dogs Head Shore, Lake Sorell, 22 July 2009.
Figure 3.49 Galaxias auratus spawning monitoring site on Dogs Head Shore, Lake Sorell, 8 October 2009. Note the inundated terrestrial vegetation at this site.
80 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Early winter
New eggs
New eggs
Early spring
New eggs
Hatching eggs
Figure 3.50 Conceptual diagram of a cross-section of a littoral area of rocky substrate in lakes Crescent and Sorell, and the progressive movement of Galaxias auratus egg deposition sites with rising water levels during the spawning period. ‘New’ eggs are those which have recently been deposited during spawning and ‘hatching’ eggs are those which are in the final stages of incubation. This diagram is based on field observations during egg sampling in these habitats between 2000 and 2009.
81 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Whilst the depth range where eggs have been collected is quite large (690 mm), field observations suggest that most eggs are actually spawned (i.e. deposited) in a narrower depth range (see Figure 3.50). Collectively, the range of the mean depths at which eggs were collected in 2007 and 2009 is 273-494 mm. This range is thought to provide a better estimate of the depths at which G. auratus spawns.
Similar to the findings of the 2007 study, sweep net eggs samples collected in Lake Crescent generally contained greater amounts of sediment and detritus than those from Lake Sorell. During the spawning period, the substantial rises in water level in Lake Crescent inundated large areas of rock at the spawning monitoring site in Lake Crescent. These conditions provided cleaner rocky substrata in the shallows as water levels rose, but there was still considerably more mobile sediment in these areas in comparison to Lake Sorell.
In many areas of lakes Crescent and Sorell, terrestrial vegetation had encroached onto the margins of the rocky shores during the prolonged period of low water levels prior to the 2009 G. auratus breeding season (Figures 3.38 and 3.49). This vegetation was present in the spawning monitoring sites in both lakes and occurred in the zone where spawning was taking place; however, no eggs were found on this substrate.
3.6.2 Spawning timing and egg abundance
In the present study, field sampling to monitor the spawning season of the G. auratus populations in lakes Crescent and Sorell did not commence until July 2009, which could have been after the onset of spawning in these populations based on the findings of previous studies (DPIW, 2008; Hardie, 2003a; Hardie et al., 2007b). Despite this, examination of the littoral catches of G. auratus in Lake Crescent and Lake Sorell in July provided a reasonable assessment of the reproductive status of both populations at this time. This survey determined that approximately 33% of the population in Lake Sorell appeared to have already spawned (i.e. were classed as ‘spent’), whereas only 5% of the Lake Crescent population appeared to have spawned (see Figure 3.51).
These estimates of the proportions of the populations which had spawned by late July 2009 were supported by patterns in the occurrence and abundance of eggs sampled at known spawning sites in both lakes. According to egg sampling, G. auratus spawned earlier in Lake Sorell than in Lake Crescent (Figure 3.52), with eggs occurring in 100% of the samples collected in Sorell in July and 40% of samples in September. Eggs were also relatively abundant in Lake Sorell during these months (mean catch (± SE) in July = 8 ± 3 and September = 1 ± 1 eggs sweep-1) (Figure 3.52). Conversely, in Lake Crescent, eggs where not found until September, and were common at the monitoring site between September and November when they occurred in ≥70% of the samples. Eggs were most abundant in Lake Crescent during September (mean catch (± SE) = 26 ± 12 eggs sweep-1) and October (77 ± 25 eggs sweep-1) (Figure 3.52).
82 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Figure 3.51 ‘Spent’ (i.e. recently spawned; top) and ‘ripe’ (mature gonads, but has not spawned; bottom) adult golden galaxias (Galaxias auratus) captured in Lake Sorell during July 2009. The definitions of ripe and spent correspond to developmental stages 4 and 5 of Hardie et al., (2007b).
In lakes Crescent and Sorell, water levels had been rising for 3-4 months (minima occurred in March in Sorell and April in Crescent) prior to the beginning of the egg sampling in July, and water levels continued to rise during all months when eggs were found in both lakes (Figure 3.52). In Lake Crescent, eggs were not found until September when the mean monthly water level had risen 609 mm above the approximate edge of the rocky shore spawning habitat (which is at a water level of 802.20 m AHD), and were most abundant in October when the mean water level was 999 mm above this critical minimum (Figure 3.52).
3.7 Golden galaxias larval emergence and abundance
In Lake Crescent, G. auratus larvae were present in pelagic habitats between September and January (inclusive), whereas in Lake Sorell they were present between August and January (inclusive) (Figure 3.53). No larvae were found in either lake in July. Larvae were abundant between October and January in both lakes, with monthly mean catches of 89-986 fish tow-1 during this period (Figure 3.53). Small numbers of juveniles (mean monthly catches of <8 fish tow-1) were also captured in pelagic habitats in Lake Crescent between July and December (Figure 3.53); however, juveniles were basically absent from the pelagic samples collected in Lake Sorell, with a mean monthly catch of <1 fish tow-1 in September (Figure 3.53).
83 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
1.8
804.2 1.6
804.0 1.4
1.2 803.8
1.0 803.6
0.8
+1) number of eggs of number +1) 803.4
x
(
10
Water level (m AHD) level (m Water
g o
L 0.6 803.2
0.4 803.0
0.2 802.8
0.0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month
Lake Crescent
2.0 803.5 1.8
1.6
803.0 1.4
1.2
1.0 802.5
+1) number of eggs of number +1) x
( 0.8
Edge of rocky shores
10
Water level (m AHD) level (m Water
g
o L 0.6 802.0
0.4
0.2
801.5 0.0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Figure 3.52 Mean monthly catch of Galaxias auratus eggs in sweep net samples (± SE) and water levels in lakes Sorell and Crescent between January and December 2009. The edge of the rocky shores in Lake Crescent is indicated.
84 Response of Lake Crescent and Lake Sorell to water level variability during 2009
During the 2000-2004, 2007 and 2009 breeding seasons, there was significant inter-annual variability in the mean seasonal abundance of G. auratus larval in Lake Crescent (ANOVA,
F6,127 = 6.15, P <0.0001) and Lake Sorell (ANOVA, F6,133 = 19.42, P <0.0001) (Figure 3.54). In both lakes, the mean abundances of larvae during the 2009 breeding season were greater than the 2001 breeding season (the 2001 season was used as the reference for inter-annual comparisons in both lakes) (Figure 3.54). In Lake Crescent, the mean (± SE) for the 2001 spawning season was 160 ± 37 fish tow-1, while the mean for 2009 was 332 ± 62 fish tow-1. In Lake Sorell, the mean (± SE) for the 2001 spawning season was 115 ± 20 fish tow-1, while the mean for 2009 was 463 ± 90 fish tow-1. The difference between the 2001 and 2009 breeding seasons was significant in Lake Sorell (P <0.01), but not in Lake Crescent (P = 0.098) (Figure 3.54). In Lake Crescent, only mean larval abundances in the 2000 (P <0.05) and 2002 (P <0.001) breeding seasons differed significantly from 2001, while in Lake Sorell, mean larval abundances in the 2000 (P <0.0001) and 2004 (P <0.01) spawning seasons were also significantly different from 2001 (Figure 3.54).
3.8 Golden galaxias larval abundance and water level fluctuations
Previously, using data collected in lakes Crescent and Sorell in the 2000-2004 and 2007 breeding seasons, inter-annual variation in larval abundances in Lake Crescent was found to have a strong relationship with lake hydrology; however, no such association was observed in Lake Sorell (DPIW, 2008). Inclusion of the 2009 spawning season data with the historical data set further supports this relationship (Figure 3.55). In Lake Crescent, the magnitude of the seasonal rise in water level (WL) between May and September, when G. auratus are known to spawn and most of the egg incubation typically period occurs, is very strongly and positively related to the mean abundance of larvae (AL) 2 in each spawning season (log10 AL = 2.177+ log10 (1.211 WL), R = 0.97) (Figure 3.55; see Appendix 14 for a plot of this relationship with untransformed water level data). Conversely, in Lake Sorell, during most years (except 2000), mean abundances of larvae were quite similar even though seasonal water level rises ranged from 0.221 to 1.363 m (Figure 3.55).
In Lake Crescent, the log-log relationship between water level rise and abundance of larvae is remarkably strong and has been shown to apply to a wide range of seasonal water level rises (0.160 to 1.506 m) (Figure 3.55), which encompass the range of seasonal water level fluctuations that have historically occurred in the lake (Figure 3.4). This relationship clearly indicates that production of the G. auratus population in this lake can be significantly limited by the availability of spawning substrate (especially during periods of low water levels), and provides the basis for using water level manipulation strategies to assist spawning of this population.
85 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.5 Larvae 3.5 Juveniles
3.0 3.0 Sorell
2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
3.5 3.5 +1) number of fish
x 3.0 3.0
( Crescent 10
g 2.5 2.5
o L 2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0 Jul Aug Sep Oct Nov Dec Jan Jul Aug Sep Oct Nov Dec Jan Month
Figure 3.53 Mean (± SE) catch of Galaxias auratus larvae (≤40 mm FL) and juveniles (>40 mm FL) in pelagic samples from lakes Sorell and Crescent between July 2009 and January 2010. No sampling was undertaken in either lake in August 2009.
86 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
** 2.5 **
2.0
1.5
+1) number +1) fish of x
( 1.0 10
g ***
o L
0.5
0.0
2000 2001 2002 2003 2004 2007 2009
Lake Crescent
2.5
2.0
* 1.5
***
+1) number +1) fish of x
( 1.0
10
g
o L
0.5
0.0
2000 2001 2002 2003 2004 2007 2009
Figure 3.54 Mean (± SE) catch of Galaxias auratus larvae (≤40 mm FL) in pelagic samples during seven spawning seasons (August-December) in lakes Sorell and Crescent between 2000 and 2009. Additional to the vertical bars, the mean catches for 2001 are also indicated by dashed lines. Catches which differ significantly from 2001 in each lake are indicated by symbols (*, P <0.05; **, P <0.01; ***, P <0.001).
87 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
2.5
2.0
1.5
1.0 Cohort 2000 2001 0.5 Lake Crescent 2002 2003 2.5 2004
+1) number of larvae of number +1) 2007 x
( 2009
10 g
o 2.0 L
1.5
1.0
0.5 -0.8 -0.6 -0.4 -0.2 0.0 0.2
Log10(Water level rise (m))
Figure 3.55 Rise in water level between May and September and the mean catch of Galaxias auratus larvae (≤40 mm FL) in pelagic samples during seven spawning seasons (August-December) in lakes Sorell and Crescent between 2000 and 2009. A linear model of the log-log relationship in Lake Crescent is indicated by a dashed line.
88 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.9 Abundance of golden galaxias young-of-the-year and water level fluctuations
In Lake Crescent, young-of-the-year (YOY) of the 2000-2009 cohorts accounted for 16-97% of the total catches of G. auratus at littoral sites during summer-autumn (Figure 3.56). In Lake Sorell, the proportion of YOY in the littoral catches was >63% in all years, except 2007 when YOY only accounted for 37% of the total catch (Figure 3.56). Interestingly, in Lake Crescent and Lake Sorell, the highest proportions of YOY in summer-autumn catches occurred in 2010 (i.e. the 2009 cohorts), with 97% and 93% of the catches in these lakes being YOY, respectively (Figure 3.56).
Despite there being some years in both lakes where larval abundance appears to correlate with the abundance of YOY between 2001 and 2010, no significant relationships were found in either population between inter-annual mean larval abundances and proportions of YOY in summer- autumn catches. However, during this period, in Lake Crescent, there was a weak, positive semilogarithmic relationship between the magnitude of the rise in water level (WL) between May and September, and the proportion of YOY (pYOY) in catches at littoral sites during the following 2 summer-autumn (pYOY = 78.02 + log10(54.00 WL), R = 0.44) (Figure 3.56). No such relationship occurred in Lake Sorell where in all years the proportion of YOY in summer-autumn catches was generally quite large even though there was considerable variation in annual water level rises during this period (Figure 3.56).
89 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell 100
80
60
40
20 Cohort 2000 2001 0 Lake Crescent 2003 2004 100 2006 2007 2009
80 Percent of YOY incatch YOY of Percent
60
40
20
0 -1.25 -1.00 -0.75 -0.50 -0.25 0.00 0.25 0.50
Log10(Water level rise (m))
Figure 3.56 Rise in water level between May and September in the natal year (i.e. calendar year of the breeding season when fish hatched) and the percent of Galaxias auratus young-of-the-year (YOY; ≤65 mm FL) in littoral fyke net catches during the following late summer – autumn (February-April) during seven years in lakes Sorell and Crescent between 2001 and 2010. A linear model of the semilogarithmic relationship in Lake Crescent is indicated by a dashed line.
90 Response of Lake Crescent and Lake Sorell to water level variability during 2009
3.10 Relative abundance of golden galaxias populations
It should be noted that the methods used to monitor the status of the G. auratus populations in lakes Crescent and Sorell (fyke netting in littoral habitats) are not quantitative; hence, do not provide density estimates of the populations (i.e. absolute numbers of individuals in the populations per unit area). The monitoring methods provide measures of the relative abundance of the populations: that is, abundance estimates from standardised quantities of fishing effort which are consistent across sites, lakes and years. These data provide information regarding trends in the relative abundance of populations between lakes and years, which can be used to examine the status of the populations.
Water levels in the lakes at the time of sampling are likely to influence the perceived relative abundances of the G. auratus populations in lakes Crescent and Sorell. The storage volumes and inundated surface areas of the lakes vary with water levels (Appendices 11 and 12), and this variation in the quantities of wetted habitat would directly influence the density (i.e. number per unit volume or area) of G. auratus populations by diluting or concentrating individuals in the available habitat. Low water levels in Lake Crescent are particularly likely to influence catches, as at levels ≤802.0 m AHD the volume and surface area of the lake begin to reduce substantially (Appendix 11).
Because of these issues, comparisons of mean catches of G. auratus between years need to be interpreted cautiously, especially for the Lake Crescent population. Furthermore, catch size structure data (section 3.11) should also be considered when interpreting catch data to draw conclusions about the status of the populations. Bearing these issues in mind, temporal differences in relatively abundance between and within the G. auratus populations in Lake Crescent and Lake Sorell are outlined below.
Mean catches of G. auratus (which include both juveniles and adults) during recent annual population monitoring (8 years between 2001 and 2010) indicate that there is a significant difference between the relative abundances of the populations in Lake Crescent and Lake Sorell (ANOVA,
F1,54 = 26.75, P <0.0001), with the Crescent population being on average 6.8 times more abundant than the Sorell population. Furthermore, there was significant inter-annual variability in mean catches of G. auratus in Lake Crescent (ANOVA, F7,20 = 13.57, P <0.0001) and Lake Sorell
(ANOVA, F7,20 = 3.72, P <0.01) (Figure 3.57).
In both lakes, the mean catches during autumn 2010 were greater than autumn 2002 (using 2002 as the reference for inter-annual comparisons in both lakes) (Figure 3.57). In Lake Crescent, the mean (± SE) for 2002 was 124 ± 23 fish site-1, while the mean for 2010 was 1801 ± 539 fish site-1. In Lake Sorell, the mean (± SE) for 2002 was 29 ± 9 fish site-1, while the mean for 2010 was 130 ± 39 fish site-1. The difference between the 2002 and 2010 catches was highly significant in Lake Crescent (P <0.0001), but not significant in Lake Sorell (P = 0.120) (Figure 3.57). In Lake Crescent, the mean catch in 2001 also differed significantly from 2002
91 Response of Lake Crescent and Lake Sorell to water level variability during 2009
(P <0.05); while the catch in 2005 was also close to being significantly different (P = 0.056). In Lake Sorell, none of the annual mean catches differed significantly from 2002, but catches in 2005 (P = 0.056) and 2009 (P = 0.067) were close to being significantly different (Figure 3.57).
As previously mentioned, water levels at the time of sampling are likely to have affected catches during annual population monitoring, especially in Lake Crescent. In Lake Crescent and Lake Sorell, water levels during annual population monitoring have varied considerably (Figure 3.58). In Lake Crescent, during years where water levels were ≤802.000 m AHD at the time of sampling (2001, 2008 and 2009), it is likely that catches have been influenced by the low water level conditions. At these times, G. auratus were probably concentrated by reductions in available habitat, thus, the perceived relatively abundances were inflated. Therefore, the relative abundance of the G. auratus population in Lake Crescent during 2001, 2008 and 2009 was likely to have been less than that which is indicated by comparing mean catches in these years to those from other years when the water levels were higher (Figure 3.58).
3.11 Size and age structure of golden galaxias populations
Length frequencies of catches of G. auratus in Lake Crescent and Lake Sorell during annual population monitoring between 2001 and 2010 (eight monitoring occasions) show that the size structure of both populations has varied temporally (Figure 3.59). Based on extensive monthly length frequency data sets for both populations (Hardie, 2003a; Hardie, 2007) and previously established length-at-age relationships (Hardie, 2007), temporal variability in the age structure of the populations and the strength of YOY cohorts can be confidently interpreted from these length frequency data. Distinguishable components of the annual length frequencies include: YOY cohorts (typically 30-70 mm FL; age = 0+), second year cohorts (typically 70-100 mm FL; age = 1+), and cohorts of fish in their third year and older (typically >100 mm; likely age range = 2+ to 9+).
The proportions of YOY cohorts in the annual catches in Lake Crescent have varied since 2001. These cohorts were quite abundant in 2004, 2008 and 2010, indicating strong recruitment at these times (Figure 3.59). Conversely, YOY cohorts were in low abundance in 2001, 2002, 2007 and 2009 which shows that recruitment strength in these years was relatively poor. These years of poor and strong recruitment in Lake Crescent are also supported by proportions of second year fish in the annual catches in the preceding years. For example, low abundance in the YOY cohort in 2007 was followed by few second year fish being captured 2008, and the abundant YOY cohort captured in 2008 resulted in an abundant second year cohort in 2009. This variability in recruitment of YOY fish also appears to have influenced the proportions of older fish in the Lake Crescent population since 2001, with, for example, older cohorts being more abundant in 2001 and 2002, and less abundant in 2010.
92 Response of Lake Crescent and Lake Sorell to water level variability during 2009
The length and age structure of the G. auratus population in Lake Sorell appears to be less variable than the Lake Crescent population (Figure 3.59). In this lake, YOY fish typically dominate summer- autumn catches, although they were much less abundant in 2009 than in the other years (Figure 3.59). Because of the dominance of YOY fish in Lake Sorell, similar patterns in recruitment and subsequent progression of cohorts to those found in Lake Crescent are less obvious; nevertheless, these trends are still evident in Sorell.
Recruitment patterns, which are illustrated by the length frequency data, show how susceptible these G. auratus populations are to short-term changes in population structure, with 1-2 consecutive years of poor recruitment having the potential to dramatically alter age structure. Interestingly, recruitment in both lakes from the 2007 breeding season, when the first water level manipulation was undertaken in the Crescent-Sorell system (DPIW, 2008), resulted in abundant YOY fish in both populations and these cohorts remained abundant into their second year (i.e. in 2009 catch data; Figure 3.59). Furthermore, following poor recruitment in 2009 in both Lake Crescent and Lake Sorell (as indicated by the small catches of YOY fish in both lakes; Figure 3.59), during 2010, YOY fish were very abundant in the littoral catches in both lakes (Figure 3.59), indicating strong recruitment occurred in both populations as a result of the 2009 spawning season (when larvae were also found to be abundant in both lakes (section 3.8) and the second water level manipulation was undertaken (section 3.1.1)). These findings show the value of undertaking water level manipulations in the Crescent-Sorell system to assist spawning of the G. auratus population in Lake Crescent, which can positively influence the abundance of cohorts in the short- and long-term.
93 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
3.5
3.0
2.5
2.0
+1) number of fish of number +1) x
( 1.5
10
g
o L 1.0
0.5
0.0
2001 2002 2004 2005 2007 2008 2009 2010
Year
Lake Crescent
3.5 ***
3.0 * 2.5
2.0
+1) number of fish of number +1) x
( 1.5
10
g
o L 1.0
0.5
0.0
2001 2002 2004 2005 2007 2008 2009 2010
Year
Figure 3.57 Mean (± SE) catches of Galaxias auratus in lakes Crescent and Sorell during annual population monitoring in late summer – autumn. Note catch data are not available for 2003 and 2006. Additional to the vertical bars, the mean catches for 2002 are also indicated by dashed lines. Catches which differ significantly from 2002 within each lake are indicated by symbols (*, P <0.05; ***, P <0.001).
94 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Sorell
3.5 803.8
3.0 803.6
2.5 803.4
2.0
803.2
+1) number of fish of number +1) x
( 1.5
10
g
Water level (m AHD) level (m Water
o L 803.0 1.0
0.5 802.8
0.0
2001 2002 2004 2005 2007 2008 2009 2010
Year
Lake Crescent
3.5 802.8
3.0 802.6
2.5 802.4
802.2 2.0
802.0
+1) number of fish of number +1) x
( 1.5
10
g Water level (m AHD) level (m Water
o 801.8 L
1.0 801.6
0.5 801.4
0.0
2001 2002 2004 2005 2007 2008 2009 2010
Year
Figure 3.58 Mean (± SE) catches of Galaxias auratus in lakes Crescent and Sorell during annual population monitoring in late summer – autumn and water levels in the lakes at the time of sampling, 2001-2010. Note catch data are not available for 2003 and 2006.
95 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Crescent Lake Sorell
30 20 n = 150 n = 87 Mar-01 10 0 30 n = 150 n = 68 20 Mar-02 10 0 30 n = 150 n = 119 20 Feb-04 10 0 30 n = 116 n = 140 20 Mar-05 10 0 30 n = 120 n = 73 20 Apr-07 Frequency (%) Frequency 10 0 30 n = 150 n = 142 20 Mar-08 10 0 30 n = 200 n = 50 20 Apr-09 10 0 30 n = 300 n = 250 20 Mar-10 10 0 40 60 80 100 120 140 160 40 60 80 100 120 140 160 Fork length (mm)
Figure 3.59 Cumulative length frequencies of catches of Galaxias auratus from lakes Crescent and Sorell during annual population monitoring in late summer – autumn between 2001 and 2010. Note length frequency data are not available for 2003 and 2006.
96 Response of Lake Crescent and Lake Sorell to water level variability during 2009
4. Discussion
4.1 Water level patterns in Lake Crescent and Lake Sorell
Lake Crescent and Lake Sorell are relatively shallow, have quite small lake area to catchment area ratios (1:1.4 and 1:1.8, respectively assuming no discharge from Lake Sorell into Lake Crescent), and are located in the upper region of their hydrological landscape (i.e. headwaters of the Clyde River). These factors are thought to affect their water level regimes, which are heavily influenced by local climatic variability. Patterns in water levels in Lake Crescent and Lake Sorell since 1970 illustrate this: during this period there have been substantial downward trends in the water levels in both lakes and this appears to have been primarily driven by reduced precipitation, unseasonal rainfall patterns and increased evaporation in their catchment. Furthermore, during this period, water levels in the lakes have also been significantly influenced by El Niño-induced drought events (e.g. 1982-84), which have caused record low levels in both lakes in recent years (i.e. minima during 2000 in Lake Sorell and 2009 in Lake Crescent). These water level patterns are not unique to lakes Crescent and Sorell, with other lakes in the eastern region of the Tasmanian Central Plateau (e.g. Arthurs Lake and Lagoon of Islands; Hydro Tasmania, unpubl. data) and greater eastern Tasmania (Lake Leake and Tooms Lake; DPIPWE, unpubl. data) experiencing similar reductions in water levels in recent years.
Additional to the long-term downward trend in water levels in Lake Crescent and Lake Sorell, inter- annual variability in water levels in both lakes has increased in recent years. This hydrological variability has caused inundating and dewatering patterns of littoral habitats in the Crescent-Sorell system (i.e. wetlands in both lakes and rocky shores in Lake Crescent) to become more irregular than historical patterns (i.e. those during 1970-1996). This alteration in the water level regimes of the lakes has had profound effects of their ecosystems (DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b; Heffer, 2003c; Uytendaal, 2003; Uytendaal, 2006); this is discussed further in sections 4.2-4.7.
The water levels in lakes Crescent and Sorell between 2006 and 2009 provide an example of this increased variability, with minimal seasonal rises in 2006-2008 followed by the largest rise in both lakes since 1970 during 2009 (the magnitudes of which were >2 times greater than the long-term mean seasonal rises in both lakes). Water levels during this 4-year period in Lake Crescent and Lake Sorell illustrate how rapidly these interconnected lakes can recover (hydrologically) from low water levels. However, the rate of recovery of components of the ecosystems of these lakes differs, and some appear to take considerable time to ‘recover’ from prolonged periods of low water levels (sections 4.2-4.7). It is also worth considering the direr environmental (and social and economic) consequences that would have occurred if the dry climatic conditions of 2006-2008 continued for
97 Response of Lake Crescent and Lake Sorell to water level variability during 2009 another 1-2 years. Examples of possible consequences of these circumstances include: (1) further reductions in water levels (with new record low levels since 1970 in both lakes) with Lake Crescent possibly being totally dewatered, (2) no ability to release water from Lake Sorell to Lake Crescent, nor downstream into the Clyde River, thus no available water for downstream users, and (3) total loss of aquatic fauna that require permanent water in Lake Crescent, including the G. auratus population.
Based on modelled predictions of future climatic conditions and surface water runoff in the Clyde River catchment (Ling et al., 2009; Graham et al., 2009), the recent dry conditions in the catchment may be indicative of future conditions. These conditions have negatively affected the condition of lakes Crescent and Sorell and their associated flora and fauna (DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Hardie, 2007; Uytendaal, 2003; Uytendaal, 2006; sections 4.2-4.7), and are likely to continue to do so in the future. Indeed, Graham et al., (2009) identified the Interlaken Lakeside Reserve Ramsar-listed wetland (an area of Lake Crescent) as a site which may be impacted by future moderate to dry climatic scenarios. Therefore, management strategies for water resources in the Clyde catchment (those relating to the environment and also consumptive water users), should include plans to deal with dry climatic conditions. The recent climatic patterns, and the associated impacts on the Crescent-Sorell ecosystem, may provide a good indication of what to expect in the future. Modelling long-term historical (i.e. last century or more) water level patterns in lakes Crescent and Sorell would also help managers understand historical conditions in these lakes, particularly the frequency of low water level events. Given knowledge gained from the many studies that have been undertaken recently on the ecology of lakes Crescent and Sorell, and the driving ecological and physical processes in these systems (e.g. DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Hardie, 2007; Heffer, 2003c; Stuart-Smith, 2007; Uytendaal, 2003; Uytendaal, 2006), water resource managers are well placed to develop strategies to conserve the environmental values associated with these lakes.
4.2 Effects of water levels on physico-chemical parameters
Previously, the effects of water levels on physico-chemical parameters in Lake Crescent and Lake Sorell have been investigated by Uytendaal (2003; 2006) using empirical field data, and coupled hydrodynamic and ecosystem modelling. According to Uytendaal (2003; 2006), in many ways these relatively shallow lakes are similar physically and chemically, and wind-induced resuspension of sediments is an influential process in both lakes; however, the shallower basin of Lake Crescent generally allows sediment resuspension to be a more influential process in this lake. The more profound effect of sediment resuspension in Lake Crescent causes this lake to typically be more turbid and to have a more abundant plankton community than Lake Sorell (Uytendaal, 2003; Uytendaal, 2006). Additionally, based on relationships between water levels, shear stress modelling
98 Response of Lake Crescent and Lake Sorell to water level variability during 2009 and turbidity levels in lakes Crescent and Sorell, Uytendaal (2003; 2006) determined important water level thresholds to conserve water quality (particularly turbidity) in the lakes.
The work undertaken in the present study enabled some of the findings of Uytendaal (2003; 2006) – and also DPIW (2008) which provided a more recent, albeit brief, examination of water quality in the lakes – to be re-examined. It also allowed long-term patterns in turbidity levels to be examined and comparisons of physico-chemical parameters in the lakes under very low water level conditions to be made. Furthermore, the present study documented turbidity levels in littoral wetlands of the lakes, while they were inundated from the lakes, for the first time.
Overall, since the late 1990s, total and colloidal turbidity levels in lakes Crescent and Sorell have increased substantially. As has been conclusively shown by previous studies in the Crescent-Sorell system (Uytendaal, 2003; Uytendaal, 2006), these increases have been caused by prolonged periods of low water levels which have provided suitable conditions for wind-induced resuspension of sediments to be a highly influential process in the lakes. The inclusion of turbidity data collected between 2007 and 2010 in long-term data sets (DPIW, 2008; Uytendaal, 2003; Uytendaal, 2006) did not significantly alter nor weaken previously established strong relationships between water levels and turbidity levels in lakes Crescent and Sorell. In fact, these longer data sets provide further evidence of the effects of wind-induced sediment resuspension on water quality in lakes Crescent and Sorell, and show the importance of the water level thresholds that were defined by Uytendaal (Uytendaal, 2003; Uytendaal, 2006). For example, significant increases in total and colloidal turbidity levels in lakes Crescent and Sorell occurred during episodes of very low water levels in both lakes in 2000-2001 and 2008-2009, during which water levels fell well below preferred and critical minimum thresholds in both lakes. Conversely, relatively high water levels, which were above the preferred and critical minimum thresholds in both lakes during 2004-2006 and late 2009 – early 2010, caused substantial decreases in turbidity levels. These observations show the importance of maintaining relatively high water levels in order to conserve reasonable water quality in the lakes.
There are strong linear relationships between water levels and total turbidity (which is associated with small-sized sediment particles) in Lake Crescent and Lake Sorell; however, there is marked hysteresis (i.e. lagged response) in relationships between water levels and colloidal turbidity (which is associated with small-sized sediment particles) in both lakes, especially in Lake Sorell. This has meant that, since 1991, fluctuations in water levels in lakes Crescent and Sorell have caused total turbidity levels to also fluctuate in direct response to water levels, with low water levels causing high levels of total turbidity. However, during this period, colloidal turbidity has increased incrementally in a step-like pattern during consecutive periods of low water levels. The main processes which reduces colloidal turbidity levels in the lakes are dilution (via inflows) and flushing (via releases of water into the Clyde River), and both of these processes only occur when there are moderate to large seasonal yields and relatively high water levels in the lakes. Therefore, infrequent low water
99 Response of Lake Crescent and Lake Sorell to water level variability during 2009 level events are a significant threat to water quality in the lakes in the short- and long-term, as they cause concentrations of colloidal material in the water columns of the lakes to increase, and it is likely to take considerable time (years to decades) to remove this material once it is in suspension (Uytendaal, 2003; Uytendaal, 2006).
The difference in the average depth of the main basins of Lake Crescent (approximately 0.6 m) and Lake Sorell (approximately 1.5 m) during late summer – autumn 2009 caused considerable differences in physico-chemical conditions in the two waterbodies. Based on data from continuously recording physico-chemical loggers which were deployed in Lake Crescent and Lake Sorell at this time, generally, water quality was poorer and much more dynamic in the shallower Lake Crescent. In Lake Sorell, the magnitude and/or degree of temporal variation in water temperature, dissolved oxygen concentration, electrical conductivity and total turbidity levels were all within the ranges that had previously occurred in this lake during periods of low water levels (Uytendaal, 2003; Uytendaal, 2006). However, in Lake Crescent, the magnitude and/or degree of temporal variation in these parameters was quite extreme in comparison to conditions which had previously been observed in this lake.
Notable characteristics of the physico-chemical conditions in Lake Crescent (cf. Lake Sorell) were the large diurnal fluctuations in temperature and dissolved oxygen, relatively high electrical conductivity levels, and very high total turbidity levels which also showed a large degree of temporal variation. These physico-chemical conditions were symbolic of a waterbody that was on a trajectory towards being totally dewatered: with the relatively shallow water column allowing substantial fluctuations in water temperature and dissolved oxygen diurnally and on an inter-daily basis due to variation in ambient weather conditions; concentrating of dissolved salts in the lake causing electrical conductivity to increase; and, significant wind-induced shear stress on large areas of the substrate of the lake causing resuspension of sediments to markedly increase turbidity levels. Interestingly, despite reasonably high water temperatures and significant sediment loads in the water column, dissolved oxygen did not reach critically low concentrations in Lake Crescent while water levels were very low in summer-autumn 2009. This appeared to be due to wind action aerating the water column of the lake.
There was considerable spatial variability in total and colloidal turbidity levels in the Interlaken Lakeside Reserve wetland (Lake Crescent) and Silver Plains Marsh (Lake Sorell) during spring 2009; however, the water in these wetlands was generally quite turbid (mean total turbidities of >35 NTU). As the catchments of these wetlands have reasonable coverage of terrestrial vegetation, and the substrate of the wetland themselves is mostly covered by aquatic vegetation, these areas are unlikely to provide sources for suspended sediment (including colloidal material) to enter the water columns of the wetlands. Furthermore, during spring 2009 when turbidity levels in the wetlands were monitored, the wetlands were well connected to the main basins of the lakes, thus,
100 Response of Lake Crescent and Lake Sorell to water level variability during 2009 allowing movement of water between these habitats. For these reasons, it is suspected that most of the suspended sediment in the water columns of the wetlands – especially colloidal material which accounted for, on average, ≥60% of the reduction in water clarity in these habitats – would have originated from the lake basins (i.e. lake water), not from runoff in the catchments of the wetlands nor resuspension of sediments in the wetlands themselves. Thus, additional to the impacts of low water levels on turbidity levels in the main basins of the lakes, the present study also indicates that water clarity in the littoral wetlands (including the Ramsar-listed Interlaken Lakeside Reserve wetland) can also be influenced by low water levels in the Crescent-Sorell system.
4.3 Effects of poor water quality on the Crescent-Sorell ecosystem
Low water levels in Lake Crescent and Lake Sorell reduce water quality in the lakes. Poor water quality is likely to impact negatively on many aspects of the Crescent-Sorell ecosystem. Components of ecosystem which are known to be affected include colonisation and growth of aquatic macrophytes (Heffer, 2003c), and condition, survival and feeding behaviour of fishes (Stuart-Smith, 2001; Stuart-Smith, 2007; Stuart-Smith et al., 2006; Stuart-Smith et al., 2004; Stuart- Smith et al., 2007b; Stuart-Smith et al., 2007a; Stuart-Smith et al., 2008). Many other components of the ecosystem of these inter-connected lakes are also likely to be influenced by poor water quality, but, to date, limited scientific work has been undertaken in this area. Furthermore, besides the direct effects of reduced water quality, changes to the availability (to aquatic flora and fauna) and condition of littoral habitats in the lakes that are associated with low water levels also impact on the Crescent-Sorell ecosystem (sections 4.4-4.7).
While water levels were very low in lakes Crescent and Sorell in February-April 2009, mean euphotic depths in both lakes were <0.15 m, which indicates that light penetration into the water column of the lakes was very limited. These light climate conditions, coupled with increased physical disturbance at the sediment surface from increased wave action that resulted from reduced water levels, would have provided unfavourable conditions for aquatic macrophytes to colonise in these areas (Uytendaal, 2003; Uytendaal, 2006). According to Uytendaal (2006), total turbidity levels need to be well below 50 NTU in lakes Crescent and Sorell before a significant improvement in light availability occurs, thus, conditions become favourable for colonisation by macrophytes. Additionally, turbid water from the main basins of lakes Crescent and Sorell, which entered the littoral wetlands during winter-spring 2009 when water levels were high in the lakes, may have impacted on germination, growth and survival of aquatic macrophytes in the wetlands (DPIPWE, 2010a; Scheffer, 2004). This threatening process has not been focused upon previously, but given the strong relationships between water levels and turbidity levels in the main basins of the lakes, persistent high turbidity levels in these areas obviously provide a direct threat to the unique littoral wetlands of lakes Crescent and Sorell.
101 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Currently, there is limited information regarding specific physico-chemical tolerance limits of flora and fauna in the Crescent-Sorell ecosystem. The work of Stuart-Smith and others (Stuart-Smith, 2001; Stuart-Smith, 2007; Stuart-Smith et al., 2006; Stuart-Smith et al., 2004; Stuart-Smith et al., 2007b; Stuart-Smith et al., 2007a; Stuart-Smith et al., 2008) has provided information about the effects of turbidity on: (1) predator-prey interactions between native G. auratus and alien S. trutta, (2) the feeding behaviour of these species, and (3) habitat use strategies of G. auratus. While the findings from this work have provided greater understanding in these areas, physico-chemical thresholds which cause detrimental or lethal physiological impacts on native flora and fauna species in these lakes have not been investigated.
In the case of G. auratus, fish kills of this species in Lake Sorell have previously been reported (Anon, 1994), and these were suspected to have been associated with a sudden increase in turbidity levels during turbulent conditions in the lake. When water levels are low in lakes Crescent and Sorell, strong wind events, which cause rapid and substantial increases in turbidity levels, are likely to cause some, possibly many, G. auratus die due to physiological stresses associated with extremely high sediment loads in the water column (especially respiration problems). However, locating dying or deceased fish in these large, turbid lakes is difficult; particularly given A. australis (a fish species which has an opportunistic feeding strategy) and waterbirds are likely to predate on the food resource provided by stressed or dead fish before they can be located. Additional to year- round impacts of high sediment loads, it has also been suggested that poor water quality, including high organic loadings, can also increase the chances of fungal infection for G. auratus during their winter-spring spawning season (Hardie et al., 2007a). Furthermore, increased sedimentation in littoral habitats associated with low water levels may also pose a threat to incubating eggs of this species (DPIW, 2008).
Besides impacts on the endemic G. auratus, populations of the introduced S. trutta are also negatively affected by poor water quality in Lake Crescent and Lake Sorell. Salmo trutta are able to survive during drought conditions in refugia in freshwater systems in Australia which have sub- optimal environmental conditions (i.e. water temperatures and dissolved oxygen concentrations); however, they typically require water temperatures less than c. 25°C and dissolved oxygen concentrations of greater than c. 2.5-5 mg L-1 (Elliott, 2000). Water temperatures and dissolved oxygen concentrations did not appear to reach the limits of this species during summer-autumn 2009. Nevertheless, the compounding impacts associated with elevated water temperatures and extremely high suspended sediment concentrations, which would disrupt respiration and negatively effect the ability of this species to locate prey, hence feed, are thought to have caused fish kills of S. trutta in recent years, especially in Lake Crescent. This is somewhat confirmed by findings during sampling to remove salmonids (both S. trutta and O. mykiss) from Lake Crescent during summer- autumn 2009. During this work, it was noted that the salmonid populations in the lake were in low
102 Response of Lake Crescent and Lake Sorell to water level variability during 2009 abundance, and the fish that were captured were in very poor condition and most had empty stomachs (IFS, unpubl. data).
Given the known and suggested impacts of elevated turbidity levels on the Crescent-Sorell ecosystem outlined above, if low water levels in the lakes persist for prolonged periods in the future, examining the physico-chemical tolerances of some of the key native flora and fauna (especially endemic species) in the system would be worthwhile. Knowledge in this area would allow managers of the lakes to implement mitigation strategies (such as refining water level management strategies and translocating taxa to refugia) before populations of species are lost or decline to low densities that prevent recovery.
4.4 Water levels, and the inundation and condition of littoral habitats
Lake Crescent and Lake Sorell contain only a few broad types of littoral habitat: temporary wetlands which are dominated by aquatic macrophytes when inundated, shores containing sandy substrate and shores containing rocky substrata. The substrate in the main basins of the lakes is mostly dominated by consolidated sediment and, in the case of Lake Sorell, rocky substrate is also present in some areas. Littoral areas in the Crescent-Sorell system are important, as they provide substrata with a range of structural complexity which provide refuge, feeding and breeding habitat for flora and fauna in the lakes. Examples of fauna – littoral habitat associations in these lakes include use of rocky substrata in littoral areas by G. auratus for spawning (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b; section 4.5), habitat-related differences in the composition of macroinvertebrate communties in the littoral zone of the lakes (see Hardie, 2003b), and use of aquatic macrophytes in wetlands by waterbirds for breeding (Heffer, 2003c; S. A. Hardie pers. obs.). Additionally, the extensive littoral wetlands in the lakes are also likely to be vital components of the mosaic of temporary wetlands that occur in the landscape surrounding the lakes (e.g. Hazelwoods Lagoon and Lagoon Plain marsh; Heffer, 2003c; Heffer, 2003a; Maxwell and Tyler, 2009), and may provide refugia for fauna (e.g. waterbirds) when other nearby wetlands are dry.
Previously, water level thresholds which approximate the in-lake edge of the littoral wetlands in lakes Crescent and Sorell (Heffer, 2003c), and littoral areas of rocky substrate in Lake Crescent (Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b) have been determined. Heffer (2003c) also derived thresholds which estimate the extent and depth of inundation in the littoral wetlands. This work was, however, undertaken during a period (2000-2003) when water levels in the lakes did not allow the thresholds to be validated with on-ground surveys. During this previous work, water levels in the lakes were too low to inundate the littoral wetlands from the lakes, and they were well above the in-lake edge of the rocky shores, such that locating the rock-sediment interface was difficult in some locations. During 2009, significant variation in water levels in lakes Crescent and Sorell provided opportunties to further examine relationships between water levels and the inundation of
103 Response of Lake Crescent and Lake Sorell to water level variability during 2009 littoral habitats using on-ground surveys, aerial photography and GIS modelling. The physical structure of littoral areas of rocky substrata in Lake Crescent and processes that influence the condition of these habitats were also examined. Furthermore, recent studies of aquatic vegetation in the Interlaken Lakeside Reserve wetland of Lake Crescent provide information regarding the condition of these habitats in Lake Crescent and Lake Sorell (DPIPWE, 2010a; Smith and Mendel, 2009).
There was close agreement between the observed extents and depths of inundation in the littoral wetlands of lakes Crescent and Sorell during winter-spring 2009 and previously established water level thresholds which relate to the inundation of the littoral wetlands in these lakes (Heffer, 2003c). Therefore, the findings of the present study largely confirm these thresholds: both those which relate to the in-lake edges of wetlands and those that inundate significant proportions of the wetlands. In the case of littoral areas of rocky substrata in Lake Crescent, the findings of the present study also confirmed many aspects of previous surveys of this habitat. However, in the present study, the rigorous assessment of the extent of this habitat, and relationships between water levels and its inundation and condition, also provided valuable new information for managing these areas.
GIS modelling can provide a rigorous method to examine critical habitats in lakes (Rowe et al., 2002; Douglas et al., 2009). The use of GIS modelling in this study is the first time this technique has been successfully used to determine the area of critical habitat for a freshwater fish species in Tasmania. Based on this modelling, littoral areas of rocky substrate in Lake Crescent were found to account for 1.3% of the surface area of the lake (total combined area of rocky shores = 296 610 m2), which clearly shows that the extent of this habitat is exceptionally limited in Lake Crescent. The extent of this habitat was found to be slightly greater than previously documented during the survey in March 2002 (Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b). Rocky substrata in littoral areas occurred between levels of 803.800 m AHD (full supply level) to as low c. 801.200 m AHD in some locations (cf. a lower limit of ~801.500 m AHD in the 2002 survey; Hardie, 2003a). However, as has previously been discussed (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b), the mere occurrence of rocky substrata is not the only factor that is important for aquatic fauna which use this habitat (e.g. G. auratus), the quality of the habitat (i.e. degree of sedimentation and habitat structural complexity) is equally important.
According to field observations, GIS modelling of the area of optimal spawning habitat for G. auratus (i.e. rocky shore at depths of 0.2-0.6 m), and a detailed conceptualisation on the processes and conditions that occur on littoral areas of rocky substrata in Lake Crescent, a water level of 802.200 m AHD is a critical threshold for the inundation of this habitat. At this water level, it is estimated that 38% of the total area of this substrate in Lake Crescent is inundated; however, the quality of the substrate below this level is poor because it is mostly smothered by sediment. Furthermore, below this level, the area of optimal spawning habitat is reduced due to the
104 Response of Lake Crescent and Lake Sorell to water level variability during 2009 bathymetry of the rocky shores. Additional to the critical importance of the water level of 802.200 m AHD in Lake Crescent, levels of >802.700 m AHD are thought to inundate reasonably large areas (range in surface area = 40 000-50 000 m2) of littoral rocky substrata of greater quality. At the water level of 802.700 m AHD, it is estimated that 55% of the total area of this substrate in Lake Crescent is inundated, and above this level, rocky shores contain substrata of greater structural complexity (i.e. more diverse range of particle sizes) which are influenced to a lesser degree by sedimentation than areas of rocky substrata at lower levels. The critical importance of the water level of 802.200 m AHD and benefits of levels >802.700 m AHD for spawning of the G. auratus population in Lake Crescent are supported by reproductive data for this population. During the 2000-2002, 2007 and 2009 breeding seasons, mass spawning of this population has only been found to occur once water levels rise above 802.200 m AHD, typically by >0.2 m, and higher levels provide even more successful spawning events (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b) (sections 4.5 and 4.6).
Episodes of extremely low water levels and prolonged periods of low water levels in Lake Crescent and Lake Sorell adversely effect the availability of littoral habitats to aquatic fauna, and the condition of these areas. However, water level fluctuations in the lakes, which include reduced water levels for limited periods, are important to their ecosystems. Many processes associated with dewatering and inundation of littoral habitats during water level fluctuations help maintain the integrity of these habitats. For example, low water levels which dewater rocky shores, sand shores and littoral wetlands, allow aeolian (wind-related) activity to abrade and ‘clean’ dewatered rocks, re-work material on sandy shores and in their associated lunette systems, and provide suitable conditions for aquatic vegetation in wetlands to undergo a dorminant growth phase and be browsed by native terrestrial fauna. Conversely, high water levels which inundate rocky shores, sandy shores and littoral wetlands, allow wave action to erode sediments on rocky shores, reduce and stabilise aeolian activity on sandy shores and in lunette systems, and provide suitable conditions for the germination, growth and reproduction of aquatic vegetation in wetlands.
Similar to variability in flows in rivers, water level fluctuations are critical to physical and biological processes in lakes, and fauna that inhabit lentic waters (Coops et al., 2003; Hofmann et al., 2008; Wantzen et al., 2008). Based on the reasons outlined above, it is obvious that water levels control patterns of inundation and dewatering (hence availability to aquatic fauna) and condition of littoral habitats in lakes Crescent and Sorell. Therefore, besides the deterimental impacts of low water levels on water quality in these lakes, reduced levels also influence the extent, diversity and condition of these areas. Clearly, maintaining natural variability in fluctuations in water levels in lakes Crescent and Sorell is important. In particular, management of water levels in these impounded lakes should aim to minimise alterations to the timing, frequency, duration and magnitude of seasonal water level rises and falls.
105 Response of Lake Crescent and Lake Sorell to water level variability during 2009
4.5 Spawning activity of golden galaxias: timing, habitat availability and water level effects
Generally, during the 2009 breeding season, G. auratus populations in Lake Crescent and Lake Sorell were found to follow similar patterns in spawning habitat use, spawning time, and associations between spawning and water levels to those observed previously (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b; S. A. Hardie, unpubl. data). However, some subtle differences were observed. Inclusion of data collected in 2009 in the long-term reproductive data sets for this species revealed new information about the spawning of G. auratus which is pertinent to the management of lakes Crescent and Sorell.
Habitats in which G. auratus eggs were collected in Lake Crescent and Lake Sorell during winter- spring 2009 (this study) were similar to those observed in the 2000, 2001, 2002 (Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b), 2005 (S. A. Hardie, unpubl. obs.), and 2007 (DPIW, 2008) breeding seasons in these lakes (i.e. cobble substrata in shallow areas on rocky shores). However, the relatively large rises in water levels during 2009 provided conditions that enabled further investigation of the depths at which this species spawns (i.e. deposits its eggs).
Collectively, during the 2007 (DPIW, 2008) and 2009 breeding seasons, when G. auratus eggs were semi-quantitatively sampled, eggs of this species were collected at a depth range of 150- 840 mm on rocky shores in lakes Crescent and Sorell. Although, as the developmental stages of the eggs collected in these breeding seasons were not determined, it was not possible to discriminate between recently spawned (i.e. deposited) eggs and those which were spawned some time ago, and thus were in the later stages of incubation. (This could be explored by analysing preserved egg samples from these breeding seasons). Therefore, the depth range of 150-840 mm encompasses locations of eggs of all developmental stages. Observations at G. auratus spawning sites during the six breeding seasons which have been sampled – particularly during 2009 when water levels rose markedly, and thus inundated substantial areas of spawning substrate over a lateral gradient of approximately 1 m – suggests that spawning takes place at a reasonably consistent depth range, and the rate and magnitude of the rise in water levels during the breeding season controls the depth at which eggs incubate. Based on these observations, the depth range at which G. auratus deposits its eggs is thought to be c. 150-400 mm.
The deposition of eggs in relatively shallow (≤0.4 m) areas by G. auratus and the reliance of this species on water level rises (possibly as a spawning cue) have obvious ramifications for management of water levels in Lake Crescent and Lake Sorell. In both lakes, natural patterns in rises of water levels during winter-spring should not be altered significantly by operation of the lakes (unless it is an appropriately timed and managed water level manipulation to assist spawning of G. auratus; section 4.8). Most importantly, drawdown of water levels in the lakes during the
106 Response of Lake Crescent and Lake Sorell to water level variability during 2009 breeding season should be avoided, as this may effectively remove spawning cues, dewater recently spawned and incubating eggs in the littoral zone, and decrease the surface area of good quality spawning habitat that is inundated during the spawning period. In Lake Crescent, drawdowns which result in water levels failing to reach the threshold of 802.200 m AHD, or falling below this level during the spawning season of G. auratus, would seriously threaten production of this species.
Multi-season reproductive data for the G. auratus populations in lakes Crescent and Sorell show that this species exhibits considerable inter-annual variability in its reproductive cycle, with spawning occurring between late April and late September (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b; this study). However, as occurred in 2009, the G. auratus population in Lake Sorell typically spawns in mid-winter, whereas the population in Lake Crescent spawns in late winter – early spring. Previously, rising water levels and the presence of suitable spawning substrate have been shown to be critical for G. auratus spawning (Hardie et al., 2007b), and variability in these factors has been found to largely determine the time at which spawning occurs (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b). This is particularly so in Lake Crescent where production of the G. auratus population is thought to be significantly constrained by limited availability of spawning habitat during periods of relatively low water levels (DPIW, 2008; Hardie et al., 2007b).
Similar to during the 2000, 2001 and 2007 breeding seasons when water levels rose from below the critical level of 802.200 m AHD to eventually inundate areas of good quality spawning habitat (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b), during 2009, habitat availability in Lake Crescent appeared to again influence the time at which the population in this lake spawned en masse. In Lake Crescent, even though a strong seasonal rise in water level had been occurring since April, eggs were not found in abundance until September-October when water levels had risen >600 mm above the water level of 802.200 m AHD. Furthermore, the exceptionally large rise in water levels in Lake Crescent during 2009 inundated substantial areas of spawning substrate during the breeding season, and this resulted in the most successful spawning season (based on larval abundance) recorded to date in the lake. The large seasonal rise in Lake Sorell also resulted in strong recruitment of larvae in that lake. These finding show the importance of rises in water levels in Lake Crescent and Lake Sorell during the breeding season of G. auratus, especially rises above the 802.200 m AHD threshold in Lake Crescent. The importance of the level of 802.200 m AHD in Lake Crescent is also supported by habitat mapping, and modelling of relationships between water levels and the inundation of littoral areas of rocky substrate in this lake (section 4.4).
107 Response of Lake Crescent and Lake Sorell to water level variability during 2009
4.6 Abundance of young-of-the-year golden galaxias and water level fluctuations
Quantitative relationships between environmental variables, and spawning and/or recruitment in lentic waters are known to occur for some freshwater fishes (Miranda et al., 1984; Maceina and Stimpert, 1998); however, the causative mechanism(s) are often poorly understood. Additionally, water level fluctuations are known to influence the availability of spawning habitats in littoral areas of lakes for some fish species (Gafny et al., 1992; Probst et al., 2009; Rowe et al., 2002). For G. auratus, the mechanisms associated with a relationship between hydrology and spawning success in the Lake Crescent population are clear: water levels control the availability and condition of spawning substrate, and rising water levels possibly provide stimuli for spawning (DPIW, 2008; Hardie, 2007; Hardie et al., 2007b). This study examined the abundance of the 2009 cohorts of G. auratus (at the larval and juvenile stages) in Lake Crescent and Lake Sorell, when there were large seasonal rises in water levels in both lakes. Its findings provide further support for this relationship in Lake Crescent.
Similar to previous breeding seasons (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b), during 2009, G. auratus larvae began to emerge during early spring in lakes Crescent and Sorell. Larvae were most abundant in spring, but, in comparison to previous breeding seasons, larvae/juveniles (<40 mm FL) remained quite abundant in the water columns of both lakes until early summer. Based on mean abundance of larvae during breeding seasons (which has previously been suggested as a robust measure of early recruitment strength; Hardie, 2007), larvae were very abundant in both lakes during 2009. In Lake Crescent, mean larval abundance during 2009 was the highest recorded in seven breeding seasons and in Lake Sorell it closely approximated the highest mean abundance which was recorded in 2004. Therefore, spawning was very successful in the G. auratus populations in lakes Crescent and Sorell during 2009. This shows that, in the short-term, these populations are somewhat resilient to the effects of a small number of consecutive years of low water levels; this is discussed further in section 4.7.
Data collected during 2009 allowed the strong, positive relationship between the magnitude of water level rise during spawning and mean seasonal abundance of G. auratus larvae in Lake Crescent which was initially found by Hardie (2007) and re-examined by DPIW (2008), to be examined once again. The response of the populations in both lakes during 2009 to the hydrological conditions that occurred at this time were consistent with these previous findings, and the inclusion of the 2009 data did not significantly alter the strength of the linear relationship for Lake Crescent (i.e. R2 changed from 0.95 to 0.97 with the inclusion of the 2009 data). This remarkably strong relationship has now been shown to apply to a wide range of seasonal water level rises in Lake Crescent (0.160 to 1.506 m), which encompass the range of seasonal water level fluctuations that have historically occurred in the lake.
108 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Besides spawning habitat availability/condition, early-stage recruitment of temperate freshwater fishes is influenced by several factors, including food resource availability and water temperature (Cushing, 1990). Quantifying the abundance of young-of-the-year (YOY) G. auratus once they have moved into littoral habitats is difficult due to field sampling limitations; however, water levels in Lake Crescent during the breeding season appear to influence recruitment of late-stage YOY of this species. In Lake Crescent, over seven breeding seasons, there was a weak relationship between the magnitude of the water level rise during the spawning and the proportion of YOY in catches in littoral habitats in the following summer-autumn. This finding indicates that, in Lake Crescent, abundance of YOY correlates with larval abundance, and both are influenced by water level fluctuations during the spawning period. Interestingly, in Lake Crescent and Lake Sorell, the highest proportions of YOY in summer-autumn catches occurred in 2010 (i.e. the 2009 cohorts; ≥93% of catches being YOY). Thus, while there were no relationships between hydrology and early-stage recruitment in the Lake Sorell population, recruitment in 2009 in this lake was also very strong at the larval and juvenile stages.
Historically, during periods of relatively high water levels, Lake Crescent and Lake Sorell have contained submerged macrophytes in bays of the main basins of the lakes and in their littoral wetlands. Galaxias auratus also use submerged aquatic vegetation as a spawning habitat when available (Hardie et al., 2007b). Therefore, when available, these areas are likely to provide additional areas of spawning habitat for G. auratus. Because of this, the observed hydrology- recruitment relationship in Lake Crescent may be altered during periods of persistent high water levels as large areas of suitable spawning habitat would be available, thus lessening the potential for production of the G. auratus population to be habitat-limited.
4.7 Status of golden galaxias populations
According to recent studies (Hardie, 2003a; Hardie, 2007; Hardie et al., 2005; Hardie et al., 2006a; Hardie et al., 2006b), qualitative sampling regimes that are used to monitor threatened galaxiid species in lentic waters in Tasmania need to be refined if they are to effectively determine the ‘status’ of populations. In particular, development of indices which indicate the status of populations would be useful for management of these species. As noted previously in this report (section 3.10), the monitoring regime that is used to monitor the G. auratus populations in lakes Crescent and Sorell has limitations, especially in regard to influences from water levels in the lakes. However, it provides a broad assessment of temporal changes in the relative abundance and size structure (hence age structure; Hardie, 2007) of these populations, which can be used to assess the status of these populations through time.
While relative abundances of the G. auratus populations in Lake Crescent and Lake Sorell have varied temporally between 2001 and 2010, on average, annual monitoring data indicate that the
109 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Crescent population has been 6.8 times more abundant than the Sorell population. This difference is similar to that found previously using monitoring data collected at monthly intervals over short periods (<2.5 years) (Hardie, 2003a; Hardie et al., 2005). Hardie et al., (2005) found the G. auratus population in Lake Crescent to be 7.7 times more abundant than the population in Lake Sorell after correcting for spawning-related effects on catches. They suggested that the difference in relative abundance between the two populations is likely to be due to differences in rates of predation by introduced salmonids and primary productivity between the lakes: with Lake Crescent being more productive and having less abundant salmonid populations (IFS, unpubl. data) than Lake Sorell, therefore the capacity to support a more abundant galaxiid population. This is still likely to be the case.
While the G. auratus population in Lake Crescent is generally more abundant than the population in Lake Sorell, inter-annual variability in the relative abundances and size structures of the populations in both lakes suggest that water level fluctuations in recent years have impacted on the Crescent population. In Lake Crescent, taking into account the likely concentrating effect of low water levels, there appears to have been a decline in the relative abundance of the G. auratus population during 2007-2009. Furthermore, during 2008-2010, there appears to have been a shift in the size and age structure of the population, with fewer larger-sized (>100 mm FL), thus older, fish being captured during annual monitoring. These changes in the late-stage juvenile and adult proportions of Lake Crescent population are supported by more rigorous data of the abundance of larvae and early- stage juveniles, which suffered declines due to low water levels during these periods (section 4.6). Therefore, lake hydrology appears to influence the overall abundance and size structure of the G. auratus population in Lake Crescent.
Interestingly, littoral catch data for G. auratus in Lake Crescent and Lake Sorell during 2010 indicated that the relative abundances of the populations increased markedly following substantial rises in water levels in both lakes in 2009. This was particularly the case in Lake Crescent where the mean catch was the highest recorded in the lake during eight years of monitoring between 2001 and 2010. Length frequency data show that this catch was mostly comprised of YOY fish. This increase in relative abundance in the Lake Crescent population in 2010 coincided with very high larval abundance during the previous (2009) breeding season (section 4.6). This positive response by the G. auratus population to the high water levels in late 2009 – early 2010 suggests that the population is somewhat resilient to short-term declines in abundance. However, given changes in the relative abundance and size structure of this population during the 2000s, reductions in water levels and/or seasonal water level rises in several consecutive years have the potential to be a significant threat to the long-term viability of this population.
The wild populations of G. auratus in lakes Crescent and Sorell provide the strong-hold for this species. Hydrological alteration (i.e. reduced water levels and changes to seasonal water level
110 Response of Lake Crescent and Lake Sorell to water level variability during 2009 fluctuations) remains the single greatest threat to this species. The findings of annual population monitoring (DPIW, 2008; this study) and recent studies of the reproduction and recruitment (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b; this study) of G. auratus in lakes Crescent and Sorell show that the population in Lake Sorell is less abundant than the population in Lake Crescent, but it’s relative abundance and size structure are more viable in the latter. This greater stability in the Lake Sorell population is probably due to the slightly deeper bathymetry of the lake (which provides more stable habitat conditions) and the ample spawning habitat for G. auratus in that lake. Given the long-term downward trend in water levels in both lakes, and the significant effects of reduced water levels on the G. auratus population in Lake Crescent, the population in Lake Sorell appears to be more secure. Management of these lakes and this species should take this into account where appropriate, with effort given towards providing favourable water level conditions in Lake Crescent for G. auratus to breed and maintain its population, while ensuring the population in Lake Sorell is protected from adverse impacts from substantial reductions in water levels and marked increases in populations of introduced fishes (i.e. C. carpio, O. mykiss and S. trutta).
4.8 Environmental benefits of water level manipulation during 2009
Due to increasing pressures on water resources, which invariably alter hydrological regimes of freshwater systems (e.g. regulated flows in rivers and altered water level regimes in lakes), provision of ‘environmental water’ to manage freshwater-dependent ecosystems is becoming a common practice in many regions. Similar to environmental flows in lotic systems (King et al., 2009; King et al., 2010; Rayner et al., 2009), associations between water levels and fish recruitment in lentic systems (Havens et al., 2005; Maceina and Stimpert, 1998; Sammons et al., 2002; Winfield et al., 2004) suggest that water level manipulations in waterbodies can provide important hydrological conditions for fish species which depend on certain components of water level regimes to complete their life cycles. However, for such water level manipulations to be effectively implemented, knowledge of hydrological-ecological relationships and critical hydrological thresholds for target species is required.
During 2007, knowledge of relationships between water levels, and spawning and early-stage recruitment of G. auratus populations in lakes Crescent and Sorell, and the minimum water level for this species to spawn in Lake Crescent (Hardie, 2003a; Hardie, 2007; Hardie et al., 2007b), was used to guide the first environmental water level manipulation in this system (DPIW, 2008). This manipulation involved a water release of c. 3000 ML from Lake Sorell to Lake Crescent during winter-spring and successfully assisted spawning on the G. auratus population in Lake Crescent (DPIW, 2008).
111 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Similar to 2007, water was released from Lake Sorell into Lake Crescent during 2009; however, the 2009 water level manipulation had several purposes: it aimed to assist spawning of G. auratus and inundation of littoral wetlands in Lake Crescent, and also help manage pest fish, water quality and water resources in both lakes. During 2009, the circumstances in the lakes were quite different to those in 2007, with very low water levels in autumn 2009 followed by substantial rises in levels during winter-spring due to relatively heavy rainfall in the catchment and strong baseflows in tributaries. These conditions provided a significant contrast to 2007, which was a period of reasonably low water levels and average natural inflows in both lakes, thus enabled the effectiveness of the water level manipulation strategy to be examined in a different situation.
Between July and October 2009, c. 10000 ML was released from Lake Sorell to Lake Crescent, which was >3 times greater than the release used in the 2007 (DPIW, 2008). The manipulation in 2009 had a minor effect on Lake Sorell, where it reduced the water level by 0.190 mm, and a significant influence of levels in Lake Crescent, where it increased the water level by 0.540 m (with the entire rise being above the critical minimum for G. auratus spawning (i.e. edge of the rocky shores) in this lake). The water level manipulation caused positive short-term effects on the ecosystem in Lake Crescent and, based on the lasting effects of the 2007 manipulation, the manipulation in 2009 will also have long-term benefits for the whole Crescent-Sorell ecosystem.
In Lake Crescent, the water level manipulation had significant positive effects on water quality, spawning of G. auratus and inundation of the littoral wetlands, and these parameters were not compromised in Lake Sorell. The additional rise of >0.5 m in Lake Crescent caused total and colloidal turbidity levels to decrease further than would have occurred under conditions provided by a natural rise in water levels in the lake. The large-scale G. auratus spawning event in Lake Crescent was assisted by the water release from Lake Sorell, with a significant increase in mean larval abundance. The strength of early-stage recruitment of G. auratus juveniles from the 2009 breeding season in Lake Crescent was also likely to have been greater, with a large influx of this extremely abundant cohort dominating catches of this species in littoral habitats in autumn 2010. The manipulation also had a positive influence on the inundation of the two large littoral wetlands in Lake Crescent (Clyde Marsh and the Ramsar-listed Interlaken Lakeside Reserve wetland). These wetlands were inundated by depths >300 mm for 44 days prior to the end of the water release (30 October 2009), whereas, without the release, they would not have been inundated to this depth at all during this period. Furthermore, the water level in Lake Crescent remained above the surveyed in-lake edge of these wetlands for 203 days between winter 2009 and autumn 2010. This was the largest inundation event in these wetlands since 1996 and it provided suitable conditions for aquatic vegetation in these areas to regenerate (see DPIPWE, 2010a) and allowed aquatic (e.g. macroinvertebrates, fish, waterbirds, platypus, etc.) and terrestrial (e.g. grazing marsupials) fauna to use these productive, wetted habitats.
112 Response of Lake Crescent and Lake Sorell to water level variability during 2009
As Lake Sorell is significantly larger than Lake Crescent (c. 2-fold greater in surface area and volume), releases on water from Sorell to Crescent have minimal effects on water levels in Sorell and substantial effects on levels in Crescent (see DPIW, 2008). During 2009, the effect of the 10000-ML water release on Lake Sorell was also somewhat mitigated by the large natural rise in water level that occurred in this lake (and Lake Crescent). In Lake Sorell, the relatively small (0.19 m) reduction of water levels caused by the release of water into Lake Crescent had minimal effects on water quality, the G. auratus population and littoral wetlands. In this lake, total and colloidal turbidity levels declined substantially during late 2009. Large-scale spawning in the G. auratus population took place and this resulted in a highly abundant cohort of larvae. The strength of early-stage recruitment of G. auratus juveniles from the 2009 breeding season in Lake Sorell was also high in comparison to other years, with the large influx of this abundant cohort dominating catches of this species in littoral habitats in autumn 2010. The manipulation also had a minimal effect on the inundation of the three large littoral wetlands in Lake Sorell (Kemps/Kermodes Marsh, Robertson Marsh and Silver Plains Marsh). The water release had little effect on the inundation of these wetlands during the release period, with the wetlands being inundated by depths >300 mm for 52 days prior to the end of the release (the release reduced the period of inundation to these depths by 4 days). Furthermore, up to 20 September 2010, the water level in Lake Sorell had remained above the surveyed in-lake edge of these wetlands for 400 days (and this was continuing at the time this report was being prepared). Similar to Lake Crescent, this was the largest inundation event in these wetlands since 1996, which has benefited flora and fauna that rely on, or opportunistically use, these habitats when they are wetted.
Additional to the short-term benefits of the water level manipulation, which were assessed by this study, they are also likely to be several long-term benefits to the Crescent-Sorell ecosystem from the water release. For example, improvements in water quality in Lake Crescent during late 2009- early 2010, which were assisted by the manipulation, will persist while water levels remain relatively high in that lake. The higher water levels in Lake Crescent will also reduce exposure of the substrate of the lake to shear stress associated with wind-induced turbulence, thus allow turbidity levels to continue to improve. Length frequency data from annual monitoring of the G. auratus populations in Lake Crescent and Lake Sorell also show the importance of strong recruitment years. In Lake Crescent, the 2007 G. auratus cohort, which was assisted by a water level manipulation in its breeding season, accounted for significant proportions of the population in 2008 and 2009, and provided most of the spawning stock for the 2009 breeding season. Similarly, the very abundant 2009 cohorts in Lake Crescent (and Lake Sorell) will provide some degree of security for populations of this species in these lakes in the next 2-3 years.
The temporary littoral wetlands of lakes Crescent and Sorell have historically been inundated more often than not during spring, and Heffer (2003c) recommended that they should be inundated at
113 Response of Lake Crescent and Lake Sorell to water level variability during 2009 frequencies of at least once every 5 years to maintain their floral communities. Prior to 2009, the wetlands in both lakes had not been inundated since 1996, thus assisting the inundation of the wetlands in Lake Crescent during 2009 has provided a timely and much needed watering event for those habitats. The benefits to the flora and fauna that inhabit the wetlands of the lakes will continue in the coming years, particularly for flora and fauna which reproduce and disperse when inundated and undergo dominant phases during dry periods.
The 2007 and 2009 water level manipulations in the Crescent-Sorell system are the first water level manipulations to be formally used in a lentic system in Tasmania (and probably Australia) to manage native fish populations and habitats. Clearly, the findings of the present study and those of DPIW (2008) have demonstrated that appropriately timed releases of water from Lake Sorell to Lake Crescent can be used to manage the Crescent-Sorell ecosystem. This innovative approach has been trialled twice, under quite different water level conditions in these lakes, and positive responses in environmental parameters in the system on both occasions show that it provides a powerful tool for managing these lakes. It is recommended that this type of strategy should be employed in the future to manage environmental values in lakes Crescent and Sorell, especially during periods of relatively low water levels.
114 Response of Lake Crescent and Lake Sorell to water level variability during 2009
5. Conclusion and Recommendations
5.1. Key findings
This study took advantage of a unique opportunity to examine several components of the Crescent- Sorell ecosystem during a period (i.e. 2009 – early 2010) of extreme water level variation, which included very low water levels in summer-autumn 2009, followed by rapid and substantial seasonal rises in levels during winter-spring 2009. These conditions also provided quite different circumstances to undertake the second environmental water level manipulation in this lake system; with the first being undertaken in 2007 when water level minima were higher and only moderate natural seasonal rises in levels occurred in Lake Crescent and Lake Sorell. Similar to during 2007, during 2009 the response of G. auratus populations (at several life stages) in Lake Crescent and Lake Sorell to the water level manipulation was monitored. Using long-term data sets, which included data collected during the present study, several previously established correlations and relationships between water levels and parameters of the Crescent-Sorell ecosystem were re- examined. Furthermore, the resistance/resilience of some environmental values in lakes Crescent and Sorell in relation to impacts of low water levels and recovery during a period of substantial rises in levels was investigated.
This study confirmed many findings from previous studies in lakes Crescent and Sorell (e.g. DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Hardie, 2007; Heffer, 2003c; Stuart-Smith, 2007; Uytendaal, 2003; Uytendaal, 2006), but also furthered understanding in some areas. The following is a summary of the key findings of the present study, many of which are particularly pertinent to management of water levels in Lake Crescent and Sorell in the short- and long-term:
Since 1970, there has been a downward trend in water levels in Lake Crescent and Lake Sorell, and this appears to have been primarily driven by changes in local climatic conditions (e.g. reductions in rainfall). Inter-annual variability in water levels in both lakes has also increased in recent years. These alterations to water level patterns in lakes Crescent and Sorell maybe indicative of future conditions under a drying climatic. Management of water resources in the Clyde River catchment should consider the impacts of this long-term drying trend on water resource use and environmental values in lakes Crescent and Sorell.
During 2009, in Lake Crescent, water levels rose approximately 1.78 m, with a minimum of 801.510 m AHD (April 3; which was the lowest recorded level since 1970) and maximum of 803.290 m AHD (October 28). In Lake Sorell, water levels rose approximately 1.38 m, with a minimum of 802.780 m AHD (April 20) and maximum of 804.160 m AHD (October 2). Based on mean monthly levels, the magnitudes of rises during 2009 in both lakes were >2 times greater than their respective long-term averages.
115 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Since the late 1990s, total and colloidal turbidity levels in Lake Crescent and Lake Sorell have increased substantially. These increases have been caused by prolonged periods of low water levels which have enabled wind-induced resuspension of sediments to be a highly influential process in the lakes. Long-term patterns in turbidity and water level data show the importance of previously established water level thresholds that relate to managing water quality in lakes Crescent and Sorell.
In Lake Crescent and Lake Sorell, colloidal turbidity has increased incrementally in a step- like pattern during consecutive periods of low water levels. It is likely to take considerable time (years to decades) to remove (via dilution and flushing) suspended colloidal material from the lakes; thus, low water level events are a serious short- and long-term threat to water quality in the lakes.
During late summer – autumn 2009, when water levels were very low in both lakes, generally, water quality was poorer and much more dynamic in Lake Crescent (cf. Lake Sorell). This difference was likely to have been caused by differences in the average depths of the lakes, with Lake Crescent being significantly shallower than Lake Sorell at this time.
Reasonably high total and colloidal turbidity levels were recorded in the Interlaken Lakeside Reserve wetland (Lake Crescent) and Silver Plains Marsh (Lake Sorell) during spring 2009. This finding indicates that water quality (i.e. water clarity) in the littoral wetlands of Lake Crescent and Lake Sorell can be significantly influenced by periods of low water levels in the main basins of these lakes.
Low water levels in summer-autumn 2009 substantially reduced the surface area of Lake Crescent (14.96 km2; 65% of total area at full supply level) and Lake Sorell (45.71 km2; 82% of total area at full supply level), and decreased the diversity of wetted habitats in the lakes. These conditions dewatered: (1) extensive littoral wetlands in both lakes, (2) basically all sandy shore habitat in both lakes, and (3) the majority of rocky shore habitat in Lake Crescent and a substantial proportion of this habitat in Lake Sorell.
The large seasonal rises in water levels in lakes Crescent and Sorell during winter-spring 2009 increased the surface area of Lake Crescent (22.62 km2; 98% of total area at full supply level) and Lake Sorell (53.13 km2; 96% of total area at full supply level), and increased the diversity of wetted habitats in the lakes.
During late 2009 – early 2010, the littoral wetlands in Lake Crescent and Lake Sorell were inundated to the greatest extent since 1996 (proportion inundated: Lake Crescent = 89%; Lake Sorell = 73%). In Lake Sorell, the water level reached the surveyed edge of the wetlands (803.600 m AHD) on 16 August 2009 and, up to 20 September 2010, it had
116 Response of Lake Crescent and Lake Sorell to water level variability during 2009 remained above edge of wetlands (period of 400 days which was continuing while this report was being prepared). In Lake Crescent, the water level reached the surveyed edge of the wetlands (802.700 m AHD) on 10 September 2009 and it remained above the edge of the wetlands until 1 April 2010 (period of 203 days).
There was close aggreement between observed extents and depths of inundation in littoral wetlands of lakes Crescent and Sorell during winter-spring 2009 and previously established water level thresholds which relate to the inundation of littoral wetlands in these lakes.
Littoral areas of rocky substrate in Lake Crescent were found to account for 1.3% of the surface area of the lake (total combined area of rocky shores = 296 610 m2), which clearly shows that the extent of this habitat is exceptionally limited in Lake Crescent. The extent of this habitat was found to be slightly greater than previously documented during a survey in March 2002, with rocky substrata in littoral areas occuring from levels of 803.800 m AHD (full supply level) to as low as c. 801.200 m AHD in some places.
There is a positive linear relationship between water levels and the proportion of inundated littoral rocky substrate in Lake Crescent; however, the relationship between water levels and the area of spawning habitat for G. auratus is non-linear. Water levels of >802.200 m AHD inundate relatively large areas of G. auratus spawning habitat of moderate quality and water levels >802.700 m AHD inundate similar sized areas of good quality spawning habitat. The importance of these water level thresholds is supported by: field observations, reproductive data for G. auratus populations in lakes Crescent and Sorell, GIS modelling of the area of optimal spawning habitat for G. auratus in Lake Crescent, and a detailed conceptualisation on the processes and conditions that occur on littoral areas of rocky substrata in Lake Crescent.
In Lake Crescent, littoral areas of rocky substrate were almost completely dewatered during autumn 2009, but the large magnitude of the water level rise in the lake during winter-spring almost completely inundated these areas (97% inundated) during spring.
Generally, during the 2009 breeding season, G. auratus populations in Lake Crescent and Lake Sorell were found to follow similar patterns in spawning habitat use, spawning time, and associations between spawning and water level conditions to those observed previously.
Collectively, during the 2007 and 2009 breeding seasons, G. auratus eggs were collected at a depth range of 150-840 mm on rocky shores in lakes Crescent and Sorell. However, the depth range at which G. auratus spawns, hence deposits its eggs, is thought to be c. 150-
117 Response of Lake Crescent and Lake Sorell to water level variability during 2009
400 mm. The relatively shallow depth at which G. auratus deposits its eggs leaves them susceptible to unseasonal water level drawdowns during the spawning period.
In Lake Crescent, mean larval abundance during 2009 was the highest recorded in seven breeding seasons and in Lake Sorell it closely approximated the highest mean abundance which was recorded in 2004. Therefore, spawning was very successful in the G. auratus populations in lakes Crescent and Sorell during 2009. This shows that, in the short-term, these populations are somewhat resilient to the effects of a small number of consecutive years of low water levels.
The strength of early-stage recruitment of G. auratus populations in Lake Crescent and Lake Sorell during 2009 was consistent with the findings of previous studies. There was a remarkably strong, positive relationship between the magnitude of water level rise during the G. auratus spawning period and mean seasonal abundance of larvae in Lake Crescent. This relationship has now been shown to apply to a wide range of seasonal water level rises in Lake Crescent (0.160 to 1.506 m), which encompass the range of seasonal water level fluctuations that have historically occurred in this lake.
While the relative abundances of the G. auratus populations in Lake Crescent and Lake Sorell have varied temporally between 2001 and 2010, on average, the Crescent population has been 6.8 times more abundant than the Sorell population.
During 2007-2009, there was a decline in the relative abundance of the G. auratus population in Lake Crescent, and in recent years there appears to have been a shift in the size and age structure of this population, with fewer larger-sized, older fish being present. These findings show that reductions in water levels and/or seasonal water level rises in several consecutive years have the potential to be a significant threat to the long-term viability of this population.
In the immediate term, the G. auratus population in Lake Sorell appears to be more secure than the population in Lake Crescent. This is because Lake Sorell is slightly deeper (which provides more stable/secure habitat conditions), contains ample spawning habitat for G. auratus, and reduced water levels do not significantly impact on the G. auratus population in that lake (cf. the significant effects of low water levels on the G. auratus population in Lake Crescent).
Between July and October 2009, c. 10000 ML of water was released from Lake Sorell to Lake Crescent, which was >3 times greater than the release used in the 2007. The manipulation in 2009 had a minor effect on Lake Sorell, where it reduced the water level by 0.190 mm, and a significant influence of levels in Lake Crescent, where it increased the level
118 Response of Lake Crescent and Lake Sorell to water level variability during 2009
by 0.540 m (with the entire rise being above the critical minimum for G. auratus spawning (i.e. edge of the rocky shores) in this lake).
The water level manipulation caused positive short-term effects on the ecosystem in Lake Crescent, particularly in regard to water quality, inundation of littoral wetlands, and spawning and recruitment of G. auratus. Based on the lasting effects of the 2007 manipulation, the whole Crescent-Sorell ecosystem will also benefit from the water level manipulation in 2009 over the next few years.
The findings of the present study, and those of the study of the water level manipulation in lakes Crescent and Sorell during 2007, demonstrate that appropriately timed releases of water from Lake Sorell to Lake Crescent can be used to manage the Crescent-Sorell ecosystem, particularly G. auratus populations.
5.2 Environmental management of water levels in lakes Crescent and Sorell
5.2.1 Management of littoral habitats, water quality and golden galaxias populations
The environmental provisions within the current water management plans (WMPs) for lakes Crescent and Sorell (DPIWE, 2005a) and the Clyde River (DPIWE, 2005b) are supported by the findings of a considerable volume of technical work and applied research which has been undertaken prior to, and after the adoption of the WMPs (Davies and Pinto, 2000; DPIW, 2008; Hardie, 2003a; Hardie, 2003b; Hardie, 2007; Hardie et al., 2004; Hardie et al., 2005; Hardie et al., 2006a; Hardie et al., 2007a; Hardie et al., 2007b; Smith and Mendel, 2009; Uytendaal, 2003; Uytendaal, 2006; Uytendaal et al., 2003; Heffer, 2003c; Heffer, 2003b). Collectively, these studies have provided extensive knowledge of the influences of water levels on the ecosystems of these lakes; more so than any other lake system in Tasmania.
The present study had several purposes, including documenting the effects of significant variability in water levels in lakes Crescent and Sorell during 2009, and re-examining previously established correlations and relationships between water levels and parameters of the Crescent-Sorell ecosystem using long-term data sets. This work, along with a recent study of the littoral wetlands in these lakes (DPIPWE, 2010a), has furthered understanding of the effects of water level fluctuations on lakes Crescent and Sorell. It has also provided findings which are highly relevant to the management of water levels in lakes Crescent and Sorell, and hence water resources in the Clyde River catchment.
Based on the findings of the present study, and previous studies, components of the water level regimes of Lake Crescent and Lake Sorell which are important to the management of their ecosystems are synthesized in Table 5.1. In summary, Table 5.1 identifies water level thresholds which related to: (1) the inundation of important littoral habitats (rocky shores and wetlands) in Lake
119 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Crescent and Lake Sorell, (2) maintaining water quality (particularly low turbidity levels) in Lake Crescent and Lake Sorell, and (3) allowing mass spawning of the G. auratus population in Lake Crescent. Furthermore, the importance of maintaining water level fluctuations in lakes Crescent and Sorell that follow historical seasonal patterns (which relate to climatic conditions in the area) is also specified. Note further information regarding management of littoral wetlands in the Crescent-Sorell system is provided in DPIPWE (2010a).
Based on the information presented in Table 5.1, to sustainably manage the unique Crescent-Sorell ecosystem it is recommended that the:
Critical minimal operating water level in Lake Crescent is 802.200 m AHD
Critical minimal operating water level in Lake Sorell is 803.200 m AHD
Preferred minimum operating water level in Lake Crescent is 802.700 m AHD
Preferred minimum operating water level in Lake Sorell is 803.500 m AHD
Preferred operating range for water levels in Lake Crescent is 802.200-803.800 m AHD
Preferred operating range for water levels in Lake Sorell is 803.500-804.360 m AHD.
The critical minimum operating water level in Lake Crescent is a threshold which is of utmost importance to maintaining the G. auratus population in this lake, whereas the critical minimum level in Lake Sorell is crucial for maintaining water quality (especially low colloidal turbidity) in this lake (see Table 5.1). Operating the water levels in Lake Crescent and Lake Sorell with these critical thresholds as seasonal minima is, however, a high risk strategy. Instead, in order to avoid risking the environmental values in lakes Crescent and Sorell, where possible the preferred operating water levels in lakes Crescent and Sorell should be used as seasonal minima (Table 5.1). The preferred minimum operating water levels in both lakes provide benefits to several components of their ecosystems. In Lake Crescent, this threshold provides the G. auratus population in that lake with greater access to good quality spawning habitat, maintains water quality (low total and colloidal turbidity), and increases the chances of seasonal inundation of littoral wetlands. Similarly, the preferred minimum level in Lake Sorell maintains water quality (low total and colloidal turbidity), and increases the chances of seasonal inundation of littoral wetlands (although the preferred minimum in this lake is ≥100 mm below the in-lake edge of the wetlands). The critical and preferred minimum operating water levels for Lake Crescent and Lake Sorell should be viewed in this context.
Additional to the minimum operating levels in lakes Crescent and Sorell, seasonal patterns in water level fluctuations (i.e. timing, rate, duration and magnitude of rises and falls in water levels) are also crucial to the Crescent-Sorell ecosystem (Table 5.1). For example, rises during winter-spring
120 Response of Lake Crescent and Lake Sorell to water level variability during 2009 provide spawning cues for G. auratus, the magnitude of water level rises between May and September in Lake Crescent control the strength of G. auratus recruitment, and drawdowns during winter-spring could dewater (and kill) incubating G. auratus eggs, with water level decreases of ≥150 mm being a serious threat. Furthermore, gradual rises during winter-spring provide suitable conditions for germination and growth of aquatic plants in littoral wetlands.
Management of the water levels in lakes Crescent and Sorell should also consider inter-annual variability in the important water level regime components identified in Table 5.1. The water level regimes of these lakes are significantly influenced by variability in climatic conditions, and this climate-induced variability in water levels is thought to be important to the ecosystems of the lakes. However, releases of water from Lake Crescent into the Clyde River have the potential to increase the magnitude of this variation. During periods of prolonged low water levels, this may exacerbate the impacts of low water levels on the ecosystem of these inter-connected lakes. Because of this, during periods of low water levels, management of water levels in the lakes should consider the short- and long-term environmental consequences of breaching critical and preferred minimum operating levels and altering seasonal patterns in water level fluctuations.
Water quality, littoral wetlands and G. auratus populations are affected by inter-annual variability in water levels and this should be considered in the management of the water levels in lakes Crescent and Sorell. Water levels below the preferred minimum level in Lake Crescent and below the critical minimum level Lake Sorell, respectively, increase colloidal turbidity in the lakes. It is likely to take considerable time (years to decades) to remove (via dilution and flushing) suspended colloidal material from the lakes; therefore, periods during which water levels are below these thresholds should be viewed cumulatively across years. The temporary littoral wetlands of Lake Crescent and Lake Sorell require dry and inundated conditions to maintain their aquatic vegetation communities. Based on historical (since 1970) patterns of inundation, these habitats are thought to require a minimum frequency of being dry and/or inundated of 1 in every 5 years. Given the relatively short life span of G. auratus (most fish live for ≤3 years), the minimum frequency for water levels in Lake Crescent to be above the critical minimum water level for that lake during winter-spring is 1 in every 3 years. Furthermore, water level rises of >0.4 m above this thresholds in Lake Crescent are required for a moderate level of G. auratus recruitment to occur.
121 Response of Lake Crescent and Lake Sorell to water level variability during 2009 Table 5.1 Synthesis of important water level regime components for managing the ecosystems of Lake Crescent and Lake Sorell. Water level regime components include water level thresholds and fluctuation patterns. Only key ecosystem parameters which are directly influenced by the water level regime components are included; in some instances water level regime components may indirectly influence several other important elements of the ecosystems of the lakes (e.g. water quality in Lake Crescent influences the G. auratus population in that lake). Note further information regarding management of the littoral wetlands of the Crescent-Sorell system is provided in DPIPWE (2010a).
Lake Water level regime component Ecosystem parameter Environmental/operational importance and critical times References¶ (m AHD)
Lake Crescent 801.700 - Sill of sluice gates. NA
802.200* Rocky shore habitat Water levels above this threshold inundate areas of moderate quality (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie G. auratus population rocky shore habitat. et al., 2007b; Stuart-Smith, Rises above this level in early winter – spring are especially important 2007; Stuart-Smith et al., for G. auratus spawning, as they provide sufficient habitat for mass 2007b; Stuart-Smith et al., spawning in the G. auratus population. 2007a; Stuart-Smith et al., Water levels above this threshold allow G. auratus to access rocky 2008; this study) shore habitat of moderate quality, which they use for refuge and feeding throughout the year.
802.700† Water quality (total and Water levels above this threshold maintain low to moderate total and (DPIPWE, 2010a; DPIW, colloidal turbidity) colloidal turbidity levels, by minimising wind-induced resuspension of 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., Rocky shore habitat sediments. 2007b; Heffer, 2003c; Water levels above this threshold inundate areas of good quality rocky Littoral wetlands Uytendaal, 2003; Uytendaal, shore habitat. G. auratus population 2006; this study) Water levels above this threshold begin to inundate Clyde Marsh. Note the in-lake edge of Ramsar-listed Interlaken Lakeside Reserve wetland is 100 mm above this threshold. Rises above this level in early winter – spring provide good conditions for large-scale spawning in the G. auratus population.
803.800 - Full supply level. NA
Seasonal water level rise‡ Rocky shore habitat Water level fluctuations maintain the condition of rocky shore habitat. (DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Littoral wetlands Seasonal rises provide suitable conditions for germination and growth Hardie, 2007; Hardie et al., of aquatic plants in littoral wetlands. G. auratus population 2007b; Heffer, 2003c; Smith Rises during winter-spring provide spawning cues for G. auratus. and Mendel, 2009; this Rises during winter-spring provide suitable conditions for incubation of study) G. auratus eggs in shallow, littoral areas. The magnitude of water level rises between May and September controls the strength of G. auratus recruitment.
Seasonal water level fall§ Rocky shore habitat Water level fluctuations maintain the condition of rocky shore habitat. (DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Littoral wetlands Gradual seasonal falls provide suitable conditions for growth and Hardie, 2007; Hardie et al., reproduction of aquatic plants in littoral wetlands. G. auratus population 2007b; Heffer, 2003c; Smith Natural patterns in water level fluctuations indicate that falls occur and Mendel, 2009; this
122 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Lake Water level regime component Ecosystem parameter Environmental/operational importance and critical times References¶ (m AHD) after incubation of eggs is completed. study)
Lake Sorell 803.000 - Sill of sluice gates. NA
803.200* Water quality (colloidal Water levels above this threshold maintain low to moderate colloidal (Uytendaal, 2003; turbidity) turbidity levels, by minimising wind-induced resuspension of Uytendaal, 2006; this study) sediments.
803.500† Water quality (total and Water levels above this threshold maintain low to moderate total and (Uytendaal, 2003; colloidal turbidity) colloidal turbidity levels, by minimising wind-induced resuspension of Uytendaal, 2006; this study) sediments.
803.600 Littoral wetlands Water levels above this threshold begin to inundate Kermodes Marsh (Heffer, 2003c; this study) and Silver Plains Marsh. Note the in-lake edges of Robertsons Marsh and Kemps Marsh are 100 mm and 300 mm above this threshold, respectively.
804.360 - Full supply level
Seasonal water level rise‡ Rocky shore habitat Water level fluctuations maintain the condition of rocky shore habitat. (DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Littoral wetlands Seasonal rises provide suitable conditions for germination and growth Hardie, 2007; Hardie et al., of aquatic plants in littoral wetlands. G. auratus population 2007b; Heffer, 2003c; Smith Rises during winter-spring provide spawning cues for G. auratus. and Mendel, 2009; this Rises during winter-spring provide suitable conditions for incubation of study) G. auratus eggs in shallow, littoral areas.
Seasonal water level fall§ Rocky shore habitat Water level fluctuations maintain the condition of rocky shore habitat. (DPIPWE, 2010a; DPIW, 2008; Hardie, 2003a; Littoral wetlands Gradual seasonal falls provide suitable conditions for growth and Hardie, 2007; Hardie et al., reproduction of aquatic plants in littoral wetlands. G. auratus population 2007b; Heffer, 2003c; Smith Natural patterns in water level fluctuations indicate that falls occur and Mendel, 2009; this after incubation of eggs is completed. study)
*Recommended as a critical minimum operating water level. †Recommended as a preferred minimum operating water level. ‡Seasonal rises in water levels in Lake Crescent and Lake Sorell have historically occurred approximately between March-April and October-November. §Seasonal falls in water levels in Lake Crescent and Lake Sorell have historically occurred approximately between November-December and February-March. ¶Only references relating to studies on lakes Crescent and Sorell and their fauna are included. See these references for supporting literature from other regions which focus on similar ecosystems, and flora and fauna.
123 Response of Lake Crescent and Lake Sorell to water level variability during 2009
5.2.2 Water level manipulation as a routine management strategy
The findings of the present study, and those of DPIW (2008), have demonstrated that appropriately timed releases of water from Lake Sorell to Lake Crescent can be used to manage the Crescent- Sorell ecosystem. This innovative approach has been trialled twice, under quite different water level conditions in these lakes, and positive responses in environmental parameters in the system on both occasions show that it provides a powerful tool for managing these lakes. Therefore, additional to the environmentally important components of the water level regimes of lakes Crescent and Sorell (Table 5.1), where appropriate circumstances occur, it is also recommended that water level manipulations be routinely employed to assist management of the ecosystems of these lakes. Given the recent, climate-driven downward trend in water levels in Lake Crescent and Lake Sorell, water level manipulations could provide a method to help mitigate the effects of low water levels on the ecosystems of these lakes in the future.
Similar to the present study, water level manipulations in the Crescent-Sorell system can: (1) assist spawning of the G. auratus population in Lake Crescent, (2) increase the likelihood of inundation of the littoral wetlands in Lake Crescent, and increase the extent, magnitude and duration of an inundation, (3) help manage water quality in both lakes, particularly by improving water quality in Lake Crescent, (4) aid management of pest fish (C. carpio) populations in both lakes, and (5) aid management of water resources in the lakes in regard to downstream releases from Lake Crescent. Water level manipulations are thought to be particularly beneficial during periods of relatively low water levels in both lakes.
Water level manipulations in lakes Crescent and Sorell should be viewed as an adaptive management approach, and as such, there are no strict guidelines for their implementation. Instead, extensive knowledge of the influences of water levels on the Crescent-Sorell ecosystem which has been gained through various studies in these lakes between 2000 and 2010 (see Table 5.1), and the findings of ongoing environmental monitoring in the lakes (section 5.2.3), should be used guide water level manipulations. Based on experience gained from the 2007 (DPIW, 2008) and 2009 (this study) manipulations, key factors which should be considered when making decisions regarding the use of water level manipulations include:
Water levels in Lake Crescent and Lake Sorell during previous 5 years.
Water levels in Lake Crescent and Lake Sorell during autumn of the year in which a manipulation is being considered.
Long-term climatic forecast for winter-spring of the year in which a manipulation is being considered.
124 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Discharge in Mountain Creek (the primary tributary of Lake Sorell) during autumn – early winter of the year in which the manipulation is being considered.
Water quality in Lake Crescent and Lake Sorell during the previous 5 years.
The last time littoral wetlands in Lake Crescent and Lake Sorell were inundated, and the extent of that inundation.
The relative abundance and size structure of the G. auratus populations in Lake Crescent and Lake Sorell during autumn of the year in which a manipulation is being considered.
The last time there was a large-scale breeding event of the G. auratus population in Lake Crescent.
Status of pest fish (C. carpio) populations in Lake Crescent and Lake Sorell.
To ensure decisions as to whether or not to undertake water level manipulations adequately consider all environmental components of lakes Crescent and Sorell (and the needs of water users), it is recommended that an advisory group is formalised. At a minimum, this group should include: water managers, scientists with expert knowledge of the Crescent-Sorell system and/or the habitats which these lakes contain, and scientists who manage freshwater fauna that inhabit these lakes. Once a decision to undertake a manipulation is made, this group (or subset of this group) would be responsible for governing real-time management of water releases from Lake Sorell to Lake Crescent, to ensure they achieve their aims. Real-time management of the water level manipulation is particularly important, as local climatic conditions and inflows in tributaries can have a significant influence on water levels in lakes Crescent and Sorell over relatively short periods.
5.2.3 Environmental monitoring
Research in Lake Crescent and Lake Sorell in 2000-2003 (Hardie, 2003a; Hardie, 2003b; Hardie, 2007; Hardie et al., 2005; Hardie et al., 2007b; Heffer, 2003c; Uytendaal, 2003; Uytendaal, 2006) provided baseline data on several parameters of the ecosystems of these lakes. Since the inception of the current WMPs for lakes Crescent and Sorell (DPIWE, 2005a) and the Clyde River (DPIWE, 2005b) in 2005, several parameters of the Crescent-Sorell ecosystem (particularly water levels, water quality, littoral wetlands and G. auratus populations) have continued to be monitored (DPIPWE, 2010a; DPIW, 2008; Hardie, 2007; Hardie et al., 2007b; Smith and Mendel, 2009; Uytendaal, 2006; this study; DPIPWE, unpubl. data). Therefore, long-term data sets for several important environmental parameters of lakes Crescent and Sorell exist; many of which have been examined in the present study.
125 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Recently, these data sets have been invaluable to the management of water levels in lakes Crescent and Sorell, particularly during periods of low water levels, as they provide information regarding the status of the ecosystems of these lakes as well as the influences of water levels on physical, chemical and biological components of these systems. Furthermore, knowledge of the status of water levels, water quality, littoral wetlands and G. auratus populations was crucial to formulating water level manipulation strategies in 2007 and 2009, and this information will also be important for future decisions regarding manipulations.
For these reasons, to ensure management of water resources in lakes Crescent and Sorell (and the Clyde River catchment) is environmentally sustainable, it is recommended that water levels, water quality, littoral wetlands and G. auratus populations continue to be monitored (Table 5.2). Table 5.2 provides the details of routine monitoring, which is recommended for inclusion in future WMPs for lakes Crescent and Sorell. Monitoring that is considered important in the event of prolonged periods of very low water levels in the lakes is also described (Table 5.2). Additionally, further applied research topics are proposed to address knowledge gaps that relate to the management of water levels in lakes Crescent and Sorell (Table 5.2). Note further information regarding monitoring in littoral wetlands in the Crescent-Sorell system is provided in DPIPWE (2010a).
126 Response of Lake Crescent and Lake Sorell to water level variability during 2009 Table 5.2 Environmental monitoring and further research in Lake Crescent and Lake Sorell to assist management of water resources in these lakes (and the Clyde River catchment). Where appropriate, it is proposed that each ecosystem component be monitored in both lakes. Note further information regarding monitoring in littoral wetlands of the Crescent-Sorell system is provided in DPIPWE (2010a).
Ecosystem component Parameters Frequency Sampling* Purpose Routine monitoring
Water levels Levels in m AHD Daily to fortnightly Continuously recording water level Record water levels in the lakes to assist loggers and/or manual readings of management of their water resources and surveyed gauges (this study; DPIPWE, ecosystems. unpubl. data).
Water quality Total turbidity Monthly to bimonthly In situ water sampling, and field and Measure turbidity levels to access water laboratory analyses (Uytendaal, 2003; quality in the lakes, and further examine Colloidal turbidity Uytendaal, 2006; this study). relationships between water levels and water quality. Aid decisions regarding water level manipulations in the lake system.
Littoral wetlands Aquatic vegetation Once every 5 years Quadrat-based sampling in at least one Assess long-term changes in floral diversity littoral wetland of Lake Crescent and one and community composition. Aid decisions of Lake Sorell during summer (DPIPWE, regarding water level manipulations in the 2010a; Heffer, 2003c; Smith and Mendel, lake system. 2009; this study).
Galaxias auratus populations Relative abundance Annual Fyke netting in littoral habitats during late Monitor the status of G. auratus populations summer – autumn (DPIW, 2008; Hardie, in the lakes, and their responses to variability Size structure 2003a; Hardie, 2007; Hardie et al., 2005; in water levels. Aid decisions regarding water Hardie et al., 2006a; Hardie et al., 2007b; level manipulations in the lake system. this study). Additional monitoring to be considered during periods of low water levels
Water quality Total turbidity As required according to Continuously recording water quality Conduct more intensive monitoring of water conditions in the lakes loggers and/or in situ water sampling, and quality in the lakes during periods of very low Colloidal turbidity field and laboratory analyses (Uytendaal, water levels to examine conditions in the Electrical conductivity 2003; Uytendaal, 2006; this study). lakes at these times. This will help determine Dissolved oxygen the physico-chemical tolerances of aquatic fauna (particularly G. auratus) in these lakes. Water temperature
Galaxias auratus populations Relative abundance As required according to Fyke, tow and sweep netting (DPIW, Conduct more intensive monitoring of the conditions in the lakes 2008; this study). Collect specimens for G. auratus populations in the lakes during Size structure histological and pathological analyses. periods of very low water levels to rigorously Spawning and recruitment Periodic surveys of shorelines during and assess their status. This sampling will help Fish health immediately after turbulent conditions in determine the resistant/resilience of this the lakes (DPIW, 2008; Hardie, 2003a; species to low water level conditions. Occurrence of fish kills Hardie, 2007; Hardie et al., 2005; Hardie et al., 2006a; Hardie et al., 2007b; this study). Further research
127 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Ecosystem component Parameters Frequency Sampling* Purpose
Water levels in wetlands Levels in m AHD Weekly to bimonthly (for at least Install water level gauges in at least one Examine seasonal water level patterns in 12 months) littoral wetland of Lake Crescent and one wetlands, especially when they are not of Lake Sorell to record water levels. inundated by water from the main basins of the lakes. This monitoring would help determine relationships between water levels in Lake Crescent and Lake Sorell and their respective wetlands.
In-lake macrophytes Plant cover Opportunistically during periods Transect-based surveys (DPIPWE, If in-lake macrophytes begin to re-colonize in of high water levels in the lakes 2010a; Heffer, 2003c; Smith and Mendel, the main basins of Lake Crescent or Lake Plant diversity 2009; this study). Sorell, then macrophyte beds should be surveyed to document their extent and species composition. Water level conditions that have allowed this to occur should also be examined.
Use of littoral wetlands by Relative abundance Opportunistically during periods Examine using various fish sampling Determine the importance of the wetlands to G. auratus of high water levels in the lakes methods when littoral wetlands are G. auratus populations, and how this species Size structure connected sufficiently to the main basins uses these areas. Spawning and recruitment of the lakes to allow galaxiids to Micro-habitat use access/inhabit these areas (DPIW, 2008; Hardie, 2003a; Hardie, 2007; Hardie et al., 2005; Hardie et al., 2006a; Hardie et al., 2007b; this study).
Model viability of G. auratus Use existing biological and - Statistical modelling (DPIW, 2008; Hardie, If low water levels persist in lakes Crescent populations population status data. 2003a; Hardie, 2007; Hardie et al., 2005; and Sorell, Population Viability Analysis (or a Hardie et al., 2006a; Hardie et al., 2007a; similar modelling approach) could be used to Hardie et al., 2007b; this study; S. A. assess the viability of the G. auratus Hardie, unpubl. data; DPIPWE, unpubl. populations in these lakes. data).
Lunette systems Extent - Un-determined. Examine the extent of the unique lunette systems of lakes Crescent and Sorell, and Influential processes determine the influence of lake water level Age regimes on these areas. Analysing sediments in these systems could allow historical water level fluctuation patterns in these lakes to be determined – knowledge in this area would further understanding of the ecology of lakes Crescent and Sorell.
*References provide information regarding sampling methods and relevant existing data sets.
128 Response of Lake Crescent and Lake Sorell to water level variability during 2009
6. References
Anderson, K. E., Paul, A. J., McCauley, E., Jackson, L. J., Post, J. R. and Nisbet, R. M. (2006). Instream flow needs in streams and rivers: the importance of understanding ecological dynamics. Frontiers in Ecology and the Environment 4, 309-318. Anon (1994). Lake Sorell fish deaths. Inland Fisheries Commission Newsletter, p. 7. Anon (1995). Carp in Lake Crescent. Inland Fisheries Commission Newsletter, (pp. 1-4. Arthington, A. H., Naiman, R. J., McClain, M. E. and Nilsson, C. (2010). Preserving the biodiversity and ecological services of rivers: new challenges and research opportunities. Freshwater Biology 55, 1-16. Arthington, A. H. and Pusey, B. J. (2003). Flow restoration and protection in Australian rivers. River Research and Applications 19, 377-395. Bêche, L. A., Connors, P. G., Resh, V. H. and Merenlender, A. M. (2009). Resilience of fishes and invertebrates to prolonged drought in two California streams. Ecography 32, 778-788. Beumer, J. P., Burbury, M. E. and Harrington, D. J. (1981). The capture of fauna other than fishes in eel and mesh nets. Australian Wildlife Research 8, 673-677. Bunn, S. E., Abal, E. G., Smith, M. J., Choy, S. C., Fellows, C. S., Harch, B. D., Kennard, M. J. and Sheldon, F. (2010). Integration of science and monitoring of river ecosystem health to guide investments in catchment protection and rehabilitation. Freshwater Biology 55, 223-240. CFEV (2005). Conservation of Freshwater Ecosystem Values Project Database, v. 1.0 (periodic updating). Water Resources Division, Department of Primary Industries and Water, Hobart, Tasmania. Cheng, D. M. H. and Tyler, P. A. (1973). Lakes Sorell and Crescent - a Tasmanian paradox. Internal Revue Gesamten Hydrobiologie 58, 307-343. Cheng, D. M. H. and Tyler, P. A. (1976a). Nutrient economies and trophic status of Lakes Sorell and Crescent, Tasmania. Australian Journal of Marine and Freshwater Research 27, 151-163. Cheng, D. M. H. and Tyler, P. A. (1976b). Primary productivity and trophic status of Lakes Sorell and Crescent, Tasmania. Hydrobiologia 48, 59-64. Cleveland, W. S., McGill, M. E. and McGill, R. (1988). The shape parameter of a two-variable graph. Journal of the American Statistical Association 83, 289-300. Coops, H., Beklioglu, M. and L., C. T. (2003). The role of water-level fluctuations in shallow lake ecosystems – workshop conclusions. Hydrobiologia 506-509, 23-27. Crook, D. A., Reich, P., Bond, N. R., McMaster, D., Koehn, J. D. and Lake, P. S. (2010). Using biological information to support proactive strategies for managing freshwater fish during drought. Marine and Freshwater Research 61, 379-387. Cushing, D. H. (1990). Plankton production and year-class strength in fish populations: an update of the match/mismatch hypothesis. Advances in Marine Biology 26, 249-293. Davies, P. and Pinto, R. (2000). Clyde River Environmental Flows: IFIM Evaluation of Minimum Flows. Report Series No. WRA 00/16. Freshwater Systems, and Department of Primary Industries, Water and Environment, Hobart, Tasmania. http://www.dpiw.tas.gov.au/inter.nsf/Attachments/JMUY-5VG7LA?open Diggle, J., Day, J. and Bax, N. (2004). Eradicating European carp from Tasmania and implications for national European carp eradication. Inland Fisheries Service, Hobart, Tasmania. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications
129 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Douglas, J., Hunt, T., Abery, N. and Allen, M. (2009). Application of GIS modelling to quantify fish habitats in lakes. Lakes & Reservoirs: Research & Management 14, 171-174. DPIPWE (2010a). Interlaken Lakeside Reserve Wetland Vegetation Survey. Water Assessment Aquatic Ecology Report Series, Report No. WA 10/04. Water and Marine Resources Division. Department of Primary Industries, Parks, Water and Environment, Hobart, Tasmania. DPIPWE (2010b). Use of GIS to analyse effects of water levels on spawning habitat of the golden galaxias (Galaxias auratus) in Lake Crescent, Tasmania. Report No. WA 10/08. Information and Land Services Division and Water and Marine Resources Division. Department of Primary Industries, Parks, Water and Environment, Hobart, Tasmania. DPIW (2008). Water level manipulation in lakes Crescent and Sorell, Tasmania to assist spawning of the threatened golden galaxias (Galaxias auratus): a strategy to mitigate impacts of drought-induced habitat alterations. Water Assessment Aquatic Ecology Report Series, Report No. WA 08/49. Department of Primary Industries and Water, Hobart, Tasmania. http://www.stors.tas.gov.au/au-7-0054-00303 DPIWE (2005a). Lakes Sorell and Crescent Water Management Plan. Department of Primary Industries, Water and Environment (DPIWE), Hobart, Tasmania. http://www.dpiw.tas.gov.au/inter.nsf/WebPages/JMUY-6JKUYK?open DPIWE (2005b). River Clyde Water Management Plan. Department of Primary Industries, Water and Environment (DPIWE), Hobart, Australia. http://www.dpiw.tas.gov.au/inter.nsf/WebPages/JMUY-6JKURX?open Elliott, J. M. (2000). Pools as refugia for brown trout during summer droughts: trout responses to thermal and oxygen stress. Journal of Fish Biology 56, 938-948. Ficetola, G. F. and Denoël, M. (2009). Ecological thresholds: an assessment of methods to identify abrupt changes in species-habitat relationships. Ecography 32, 1075-1084. Frijlink, S. (1999). The Early Life History of the Golden Galaxiid, Galaxias auratus in Lakes Crescent and Sorell, Tasmania. B.Sc. (Hons) thesis. University of Tasmania, Hobart, Tasmania. Furey, P. C., Nordin, R. N. and Mazumder, A. (2004). Water level drawdown affects physical and biogeochemical properties of littoral sediments of a reservoir and a natural lake. Lake and Reservoir Management 20, 280-295. Gafny, S., Gasith, A. and Goren, M. (1992). Effect of water level fluctuation on shore spawning of Mirogrex terraesanctae (Steinitz), (Cyprinidae) in Lake Kinneret, Israel. Journal of Fish Biology 41, 863-871. Graham, B., Hardie, S., Gooderham, J., Gurung, S., Hardie, D., Marvanek, S., Bobbi, C., Krasnicki, T. and Post, D. A. (2009). Ecological impacts of water availability for Tasmania. A report to the Australian Government from the CSIRO Tasmania Sustainable Yields Project, CSIRO Water for a Healthy Country Flagship, Australia. http://www.csiro.au/partnerships/TasSY.html Grant, T. R., Lowry, M. B., Pease, B., Walford, T. R. and Graham, K. (2004). Reducing the by-catch of platypus (Ornithorhynchus anatinus) in commercial and recreational fishing gear in New South Wales. Proceedings of the Linnean Society of New South Wales 125, 259-272. Hakanson, L. (1977). Influence of wind, fetch, and water depth on distribution of sediments in Lake Vanern, Sweden. Canadian Journal of Earth Sciences 14, 397-412. Hardie, S. A. (2003a). Current Status and Ecology of the Golden Galaxias (Galaxias auratus). Rehabilitation of Lakes Sorell and Crescent Report Series No. 7/1. Inland Fisheries Service, Hobart, Tasmania. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications
130 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Hardie, S. A. (2003b). Current Status of the Macroinvertebrate Communities in Lakes Crescent and Sorell. Rehabilitation of Lakes Sorell and Crescent Report Series No. 7/2. Inland Fisheries Service, Hobart, Tasmania. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications Hardie, S. A. (2007). Conservation Biology of the Golden Galaxias (Galaxias auratus Johnston) (Pisces: Galaxiidae). PhD thesis. University of Tasmania, Hobart, Tasmania. Hardie, S. A., Barmuta, L. A. and White, R. W. G. (2004). Threatened fishes of the world: Galaxias auratus Johnston, 1883 (Galaxiidae). Environmental Biology of Fishes 71, 126. Hardie, S. A., Barmuta, L. A. and White, R. W. G. (2005). Spawning related seasonal variation in fyke net catches of golden galaxias (Galaxias auratus): implications for monitoring lacustrine galaxiid populations. Fisheries Management and Ecology 12, 407-409. Hardie, S. A., Barmuta, L. A. and White, R. W. G. (2006a). Comparison of day and night fyke netting, electro-fishing and snorkelling for monitoring a population of the threatened golden galaxias (Galaxias auratus). Hydrobiologia 560, 145-158. Hardie, S. A., Jackson, J. E., Barmuta, L. A. and White, R. W. G. (2006b). Status of galaxiid fishes in Tasmania, Australia: conservation listings, threats and management issues. Aquatic Conservation: Marine and Freshwater Ecosystems 16, 235-250. Hardie, S. A., Pyecroft, S. B., Barmuta, L. A. and White, R. W. G. (2007a). Spawning-related fungal infection of golden galaxias Galaxias auratus (Galaxiidae). Journal of Fish Biology 70, 1287- 1294. Hardie, S. A., White, R. W. G. and Barmuta, L. A. (2007b). Reproductive biology of the threatened golden galaxias Galaxias auratus Johnston and the influence of lake hydrology. Journal of Fish Biology 71, 1820-1840. Havens, K. E., Fox, D., Gornak, S. and Hanlon, C. (2005). Aquatic vegetation and largemouth bass population responses to water-level variations in Lake Okeechobee, Florida (USA). Hydrobiologia 539, 225-237. Heffer, D. K. (2003a). Hazelwoods Lagoon: Aquatic Vegetation Survey. Rehabilitation of Lakes Sorell and Crescent Report Series No. 6/3. Inland Fisheries Service, Hobart. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications Heffer, D. K. (2003b). Interlaken Lakeside Reserve: Ramar Wetland Management Plan. Rehabilitation of Lakes Sorell and Crescent Report Series No. 6/2. Inland Fisheries Service, Hobart. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications Heffer, D. K. (2003c). Wetlands of Lakes Sorell and Crescent: Conservation and Management. Rehabilitation of Lakes Sorell and Crescent Report Series No. 6/1. Inland Fisheries Service, Hobart, Tasmania. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications Hofmann, H., Lorke, A. and Peeters, F. (2008). Temporal scales of water-level fluctuations in lakes and their ecological implications. Hydrobiologia 613, 85-96. Hydro Tasmania (2003). South Esk - Great Lake Water Management Review. Scientific Report on Lake Augusta. Hydro Tasmania, Hobart, Tasmania. IFS (2004). Carp Management Program Report Lakes Crescent and Sorell, 1995 - June 2004. Inland Fisheries Service, Hobart, Tasmania. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications IFS (2010). Carp Management Program Annual Report 2009-10. Inland Fisheries Service, New Norfolk, Tasmania. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications Imboden, D. M. (2004). The motion of lake waters. The Lakes Handbook (eds P. E. O'Sullivan and C. S. Reynolds), pp. 115-152. Blackwell Science, Oxford. Kiernan, K. (1990). The extent of late cenozoic glaciation in the Central Highlands of Tasmania, Australia. Arctic Alpine Research 22, 341-354.
131 Response of Lake Crescent and Lake Sorell to water level variability during 2009
King, A. J., Tonkin, Z. and Mahoney, J. (2009). Environmental flow enhances native fish spawning and recruitment in the Murray River, Australia. River Research and Applications 25, 1205- 1218. King, A. J., WARD, K. A., O'Connor, P., Green, D., Tonkin, Z. and Mahoney, J. (2010). Adaptive management of an environmental watering event to enhance native fish spawning and recruitment. Freshwater Biology 55, 17-31. Ling, F. L. N., Gupta, V., Willis, M., Bennett, J. C., Robinson, K. A., Paudel, K., Post, D. A. and Marvanek, S. (2009). River modelling for Tasmania. Volume 5: the Derwent-South East region. A report to the Australian Government from the CSIRO Tasmania Sustainable Yields Project, CSIRO Water for a Healthy Country Flagship, Australia. http://www.csiro.au/partnerships/TasSY.html Maceina, M. J. and Stimpert, M. R. (1998). Relations between reservoir hydrology and crappie recruitment in Alabama. North American Journal of Fisheries Management 18, 104-113. Mason-Cox, M. (1994). Lifeblood of a Colony: A History of Irrigation in Tasmania, Rivers and Water Supply Commission, Hobart, Tasmania. Maxwell, C. J. and Tyler, P. A. (2009). Lagoons with Islands, Tasmanian Highlands. Australasian Plant Conservation 17, 26-26. Miranda, L. E., Shelton, W. L. and Bryce, T. D. (1984). Effects of water level manipulation on abundance, mortality, and growth of young-of-year largemouth bass in West Point Reservoir, Alabama-Georgia. North American Journal of Fisheries Management 4, 314-320. Nankivell, P. B. and Hollick, A. J. (1986). Geological Atlas: Interlaken. Department of Mines, Hobart, Tasmania. Parsons, M., Thoms, M., Capon, T., Capon, S. and Reid, M. (2009). Resilience and thresholds in river ecosystems. Waterlines Report Series No. 21. National Water Commission, Canberra, ACT. Pemberton, M. (1986). Title Land Systems of Tasmania. Region 5 - Central Plateau. Department of Agriculture Tasmania, Hobart, Tasmania. Poff, N. L., Allan, J. D., Bain, M. B., Karr, J. R., Prestegaard, K. L., Richter, B. D., Sparks, R. E. and Stromberg, J. C. (1997). The natural flow regime. Bioscience 47, 769-784. Probst, W., Stoll, S., Peters, L., Fischer, P. and Eckmann, R. (2009). Lake water level increase during spring affects the breeding success of bream Abramis brama (L.). Hydrobiologia 632, 211-224. R Development Core Team (2009). R: A language and environment for statistical computing, Version 2.9.2. R Foundation for Statistical Computing, Vienna, Austria. http://www.R- project.org Ramsar (2002). The Ramsar Convention on Wetlands, Ramsar Convention Bureau. Rayner, T. S., Jenkins, K. M. and Kingsford, R. T. (2009). Small environmental flows, drought and the role of refugia for freshwater fish in the Macquarie Marshes, arid Australia. Ecohydrology 2, 440-453. Rowe, D. K., Shankar, U., James, M. and Waugh, B. (2002). Use of GIS to predict effects of water level on the spawning area for smelt, Retropinna retropinna, in Lake Taupo, New Zealand. Fisheries Management and Ecology 9, 205-216. Sammons, S. M., Bettoli, P. W., Isermann, D. A. and Churchill, T. N. (2002). Recruitment variation of crappies in response to hydrology of Tennessee reservoirs. North American Journal of Fisheries Management 22, 1393-1398.
132 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Sammons, S. M., Dorsey, L. G., Bettoli, P. W. and Fiss, F. C. (1999). Effects of reservoir hydrology on reproduction by largemouth bass and spotted bass in Normandy Reservoir, Tennessee. North American Journal of Fisheries Management 19, 78-88. Scheffer, M. (2004). Ecology of Shallow Lakes, Kluwer Academic Publishers, The Netherlands. Scheffer, M., Hosper, S. H., Meijer, M.-L., Moss, B. and Jeppesen, E. (1993). Alternative equilibria in shallow lakes. Trends in Ecology and Evolution 8, 275-279. Scheffer, M. and van Nes, E. (2007). Shallow lakes theory revisited: various alternative regimes driven by climate, nutrients, depth and lake size. Hydrobiologia 584, 455-466. Scheifhacken, N., Fiek, C. and Rothhaupt, K. O. (2007). Complex spatial and temporal patterns of littoral benthic communities interacting with water level fluctuations and wind exposure in the littoral zone of a large lake. Fundamental and Applied Limnology 169, 115-129. Sharples, C. (1997). A reconnaissance of landforms and geological sites of geoconservation significance in the Western Derwent Forest District. A report to Forestry Tasmania. Hobart, Tasmania. Smith, J. and Mendel, L. (2009). An Assessment of the Effectiveness of the Inundation Regime on Wetlands Plant Species at Lake Crescent. A report to the Department of Primary Industries and Water. Hobart, Tasmania. Souchon, Y., Sabaton, C., Deibel, R., Reiser, D., Kershner, J., Gard, M., Katopodis, C., Leonard, P., Poff, N. L., Miller, W. J. and Lamb, B. L. (2008). Detecting biological responses to flow management: missed opportunities; future directions. River Research and Applications 24, 506-518. Stuart-Smith, R. (2001). The Feeding Ecology of Brown Trout (Salmo trutta L.) in Lake Sorell, Tasmania. BSc (Hons) thesis. University of Tasmania, Hobart, Tasmania. Stuart-Smith, R. D. (2007). The Importance of Complex Habitats for the Predator-Prey Interaction Between a Threatened Galaxiid Fish and an Introduced Salmonid. PhD thesis. University of Tasmania, Hobart, Tasmania. Stuart-Smith, R. D., Barmuta, L. A. and White, R. W. G. (2006). Nocturnal and diurnal feeding by Galaxias auratus, a lentic galaxiid fish. Ecology of Freshwater Fish 15, 521-531. Stuart-Smith, R. D., Richardson, A. M. M. and White, R. W. G. (2004). Increasing turbidity significantly alters the diet of brown trout: a multiyear longitudinal study. Journal of Fish Biology 65, 376-388. Stuart-Smith, R. D., Stuart-Smith, J. F., White, R. W. G. and Barmuta, L. A. (2007a). The effects of turbidity and complex habitats on the feeding of a galaxiid fish are clear and simple. Marine and Freshwater Research 58, 429-435. Stuart-Smith, R. D., Stuart-Smith, J. F., White, R. W. G. and Barmuta, L. A. (2007b). The impact of an introduced predator on a threatened galaxiid fish is reduced by the availability of complex habitats. Freshwater Biology 52, 1555-1563. Stuart-Smith, R. D., White, R. W. G. and Barmuta, L. A. (2008). A shift in the habitat use pattern of a lentic galaxiid fish: an acute behavioural response to an introduced predator. Environmental Biology of Fishes 82, 93-100. Thompson, R. M. and Ryder, G. R. (2008). Effects of hydro-electrically induced water level fluctuations on benthic communities in Lake Hawea, New Zealand. New Zealand Journal of Marine and Freshwater Research 42, 197-206. Threatened Species Section (2006). Recovery Plan: Tasmanian Galaxiidae 2006-2010. Department of Primary Industries and Water, Hobart, Tasmania. http://www.dpiw.tas.gov.au/inter.nsf/Attachments/LJEM-6VM72L?open
133 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Uytendaal, A. (2003). Water Quality in Lakes Crescent and Sorell: Underlying Processes and Management Options. Rehabilitation of Lakes Sorell and Crescent Report Series No. 5. Inland Fisheries Service, Hobart, Tasmania. http://www.ifs.tas.gov.au/ifs/fisherymanagement/publications Uytendaal, A. (2006). Water Clarity in Two Shallow Lake Systems of the Central Plateau, Tasmania, Australia. PhD thesis. University of Tasmania, Hobart, Tasmania. Uytendaal, A., Heffer, D. and Hardie, S. (2003). Lakes Sorell and Crescent Rehabilitation Project: Ecosystem Dynamics and Ecological Modelling. Inland Fisheries Service, Hobart, Tasmania. Valdovinos, C., Moya, C., Olmos, V., Parra, O., Karrasch, B. and Buettner, O. (2007). The importance of water-level fluctuation for the conservation of shallow water benthic macroinvertebrates: an example in the Andean zone of Chile. Biodiversity and Conservation 16, 3095-3109. Wantzen, K. M., Junk, W. J. and Rothhaupt, K.-O. (2008). An extension of the floodpulse concept (FPC) for lakes. Hydrobiologia 613, 151-170. Winfield, I. J. (2004). Fish in the littoral zone: ecology, threats and management. Limnologica 34, 124-131. Winfield, I. J., Fletcher, J. M. and James, J. B. (2004). Modelling the impacts of water level fluctuations on the population dynamics of whitefish (Coregonus lavaretus (L.)) in Hawestwater, U.K. Ecohydrology and Hydrobiology 4, 409-416. Wisniewski, C. (2003). Lakes Sorell and Crescent Capital Works. Integrated and multi-dsciplinary approach to the rehabilitation of Lakes Sorell and Crescent. Inland Fisheries Service, Hobart, Tasmania.
134 Response of Lake Crescent and Lake Sorell to water level variability during 2009 7. Appendices
Appendix 1. Littoral wetlands (depicted by dark shading) of lakes Crescent and Sorell, Tasmania and the primary tributary (Mountain Creek) and outflow (Clyde River) of the lake system. Lake perimeters are at full supply levels. Figure taken from Hardie (2007).
135 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 2. Water levels in lakes Crescent and Sorell between January 1970 and December 2008, and habitat thresholds. Levels at which littoral wetlands connect to the lakes, rocky shores begin to be inundated in Lake Crescent are indicated. Maximum (full supply) and minimum (sill) operating levels of sluice gates are also shown. Water levels are measured in metres of elevation according to the Australian Height Datum (m AHD) and these data were provided by IFS (unpubl. data).
Lake Sorell
Full supply
804 Wetlands
Sill 803
802
Lake Crescent Water level (m AHD) level (m Water 804 Full supply
803 Wetlands
Rocks 802 Sill
Jan-70Jan-71Jan-72Jan-73Jan-74Jan-75Jan-76Jan-77Jan-78Jan-79Jan-80Jan-81Jan-82Jan-83Jan-84Jan-85Jan-86Jan-87Jan-88Jan-89Jan-90Jan-91Jan-92Jan-93Jan-94Jan-95Jan-96Jan-97Jan-98Jan-99Jan-00Jan-01Jan-02Jan-03Jan-04Jan-05Jan-06Jan-07Jan-08
136 Response of Lake Crescent and Lake Sorell to water level variability during 2009 Appendix 3. Littoral areas of rocky substrate (regions filled with diagonal lines) in Lake Crescent surveyed during March 2002. Lake perimeters are at full supply levels. Wetland areas are depicted by dark shading. Figure taken from Hardie et al., (2007b).
137 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 4. Mean monthly total and colloidal turbidities, and water levels in lakes Crescent and Sorell, November1991 – December 2008. Turbidity data are from Uytendaal (2003; 2006), and DPIPWE (unpubl. data) and IFS (unpubl. data). Note different y-axis scales.
Lake Sorell
350 Water level Total turbidity Colloidal turbidity
300 804.5
250 804.0
200
150 803.5
Turbidity(NTU) Water level (m AHD) level (m Water
100 803.0
50
0 802.5
1995 2000 2005
Year
Lake Crescent
Water level 804.5 700 Total turbidity Colloidal turbidity
600 804.0
500 803.5
400 803.0
300 Turbidity(NTU)
802.5 AHD) level (m Water
200
802.0 100
801.5 0
1995 2000 2005
Year
138 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 5. Water levels in lakes Crescent and Sorell between January 1970 and December 2008, and preferred and critical operating levels under the Water Management Plan for the lakes (DPIWE, 2005a). The preferred operating range of the water levels in both lakes is shaded grey. Maximum (full supply) and minimum (sill) operating levels of sluice gates are also shown. Water levels are measured in metres of elevation according to the Australian Height Datum (m AHD) and these data were provided by IFS (unpubl. data).
Lake Sorell
Full supply
804 Pref. min Crit. min 803 Sill
802
Lake Crescent Water level (m AHD) level (m Water 804 Full supply
803 Pref. min
Crit. min 802 Sill
Jan-70Jan-71Jan-72Jan-73Jan-74Jan-75Jan-76Jan-77Jan-78Jan-79Jan-80Jan-81Jan-82Jan-83Jan-84Jan-85Jan-86Jan-87Jan-88Jan-89Jan-90Jan-91Jan-92Jan-93Jan-94Jan-95Jan-96Jan-97Jan-98Jan-99Jan-00Jan-01Jan-02Jan-03Jan-04Jan-05Jan-06Jan-07Jan-08
139 Response of Lake Crescent and Lake Sorell to water level variability during 2009 Appendix 6. Mean (± SE) catches of Galaxias auratus in lakes Crescent and Sorell during annual population monitoring between 2001 and 2008, and water levels at the time of sampling. On each sampling occasion, fish were collected using fine-meshed fyke nets in littoral areas of sand, rock and sediment substrates in each lake. Catch data are from Hardie (2003a), and DPIPWE (unpubl. data) and IFS (unpubl. data). Water levels influence catches, especially in Lake Crescent, through dilution and concentration of individuals, and this needs to be considered when interpreting these plots; this is examined further in this report (sections 3.10 and 3.11).
Lake Sorell
500
803.8
400 803.6
300 803.4
803.2
200
Water level (m AHD) level (m Water Catch (number of fish) of (number Catch
803.0
100
802.8
0 Mar-01 Mar-02 Feb-04 Mar-05 Apr-07 Mar-08
Lake Crescent
500
802.8
400 802.6
802.4 300
802.2
200
802.0
Water level (m AHD) level (m Water Catch (number of fish) of (number Catch
801.8 100
801.6
0 Mar-01 Mar-02 Feb-04 Mar-05 Apr-07 Mar-08
140 Response of Lake Crescent and Lake Sorell to water level variability during 2009 Appendix 7. Cumulative length frequencies of catches of Galaxias auratus from lakes Crescent and Sorell during annual population monitoring between 2001 and 2008. Where possible, the lengths of 50 fish from each sample site were measured in the field. Length frequency data are from Hardie (2003a), and DPIPWE (unpubl. data) and IFS (unpubl. data).
Lake Crescent Lake Sorell
30 n = 150 n = 87 Mar-01 20 10 0
30 n = 150 n = 68 Mar-02 20 10 0
30 n = 150 n = 119 Feb-04 20 10
0
30 n = 116 n = 140 Mar-05 Frequency (%) Frequency 20 10 0
30 n = 120 n = 73 Apr-07 20 10 0
30 n = 150 n = 142 Mar-08 20 10 0 40 60 80 100 120 140 160 40 60 80 100 120 140 160 Fork length (mm)
141 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 8. Monthly rainfall deciles for Tasmania. Source: Bureau of Meteorology (http://reg.bom.gov.au/jsp/awap/rain/index.jsp, accessed 7 October 2010)
(a) 1 June 2006 to 31 May 2009.
(b) 1 June 2009 to 31 August 2009.
142 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 9. (a) annual rainfall totals for the Interlaken area between 1984 and 2009 with the mean annual total for this period (660.2 mm) indicated by a dashed line, and (b) mean monthly rainfall totals for the Interlaken area between 1984 and 2009, and monthly totals during 2009. Data are from the Bureau of Meteorology (Interlaken weather station) and the Interlaken Estate property.
(a) 1000
800
600
400 Total rainfallTotal(mm)
200
0 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008
Year
(b)
200
1984-2009 2009
150
100 Rainfall(mm)
50
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month
143 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 10. Stockings of salmonid (trout and salmon) fish species in Lake Crescent, 1983-2010 (IFS, unpubl. data). No fish were stocked in years that are not included in the table.
Date Species* Size/age class Stock Type Mean weight (g) Approximate no.
01/08/1983 S. trutta Fry Wild Diploid - 100 000 01/08/1984 S. trutta Fry Wild Diploid - 100 000 07/02/1985 S. trutta Fingerling Wild Diploid 3.3 2 000 01/08/1985 S. trutta Fry Wild Diploid - 30 000 13/01/1986 S. trutta Fingerling Wild Diploid 3.1 25 000 26/03/1986 S. trutta Fingerling Wild Diploid 3.5 8 500 11/04/1986 O. mykiss Fingerling Domestic Diploid 39 7 670 22/04/1986 O. mykiss Fingerling Domestic Diploid 33 5 700 01/08/1986 S. trutta Fry Wild Diploid - 100 000 15/12/1986 S. trutta Fingerling Wild Diploid 0.8 38 000 14/05/1987 S. trutta Fingerling Wild Triploid 12 6 000 23/06/1988 O. mykiss Fingerling Domestic Diploid 27 4 000 14/12/1988 O. mykiss Fingerling Wild Triploid 1.0 2 000 23/06/1989 O. mykiss Fingerling Domestic Diploid 80 2 100 10/07/1989 O. mykiss Fingerling Domestic Diploid 15 2 000 11/07/1989 O. mykiss Yearling Domestic Diploid 125 500 01/08/1989 S. trutta Fry Wild Diploid - 25 000 06/11/1989 O. mykiss Yearling Domestic Diploid 100 1 000 03/02/1990 O. mykiss Fingerling Domestic Diploid 2.5 16 000 06/07/1991 O. mykiss Fingerling Wild Diploid - 3 000 14/07/1991 O. mykiss Fingerling Wild Diploid - 1 000 17/05/1992 O. mykiss Fingerling Wild Diploid - 3 000 29/05/1993 S. trutta Yearling Wild Diploid 14 2 500 26/03/1994 S. trutta Fingerling Wild Diploid 4.0 4 000 26/03/1994 S. trutta and S. salar Fingerling Wild/Domestic Diploid 4.0 6 000 17/11/1994 S. trutta and S. salar Yearling Wild/Domestic Diploid 20 1 000 10/12/1994 O. mykiss Fingerling Domestic Diploid 3.7 3 000 20/12/1994 S. trutta Advanced fry Wild Diploid - 10 000 20/04/1999 O. mykiss Yearling Domestic Diploid 60 3 500 04/11/1999 S. trutta Advanced fry Wild Diploid - 45 000 29/11/1999 O. mykiss Advanced fry Wild Diploid - 18 000 06/12/2001 S. trutta Advanced fry Wild Diploid - 10 000 08/01/2002 O. mykiss Fingerling Domestic Triploid 7.0 10 000 24/04/2002 O. mykiss Fingerling Domestic Triploid 20 10 000 20/06/2002 S. trutta Adult Wild Diploid 1000 3 050 20/06/2002 S. trutta Fingerling Wild Diploid 6.0 10 000 04/06/2003 S. trutta Adult Wild Diploid 1000 1 000 18/02/2004 O. mykiss Fingerling Domestic Triploid 15 5 000 01/06/2004 S. trutta Adult Wild Diploid 1100 1 000 20/06/2005 S. trutta Adult Wild Diploid 1000 600 29/08/2005 O. mykiss Yearling Domestic Triploid 200 3 000 *Species are: Atlantic salmon (Salmo salar), brown trout (Salmo trutta) and rainbow trout (Oncorhynchus mykiss).
144 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 11. Relationships between (a) water level and storage volume, and (b) water level and inundated surface area in Lake Crescent (DPIPWE, unpubl. data).
(a)
50000
20000
10000
5000
2000
1000
500
Storage volume (ML) volume Storage 200
100
50
20
801 802 803 804 805
Water level (m AHD)
(b)
20.0
10.0
5.0
) 2
2.0
1.0 Surface area (km area Surface
0.5
0.2
801 802 803 804 805
Water level (m AHD)
145 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 12. Relationships between (a) water level and storage volume, and (b) water level and inundated surface area in Lake Sorell (DPIPWE, unpubl. data).
(a)
100000 50000
20000
10000 5000
2000
1000
Storage volume (ML) volume Storage 500
200
100
50
800 801 802 803 804 805
Water level (m AHD)
(b)
50.0
20.0
10.0 )
2 5.0
2.0
1.0 Surface area (km area Surface 0.5
0.2
0.1
800 801 802 803 804 805
Water level (m AHD)
146 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 13. Actual (observed) daily water levels in lakes Crescent and Sorell between January and November 2007, and levels that were estimated (modelled) to have occurred in both lakes without the release during the 2007 manipulation period. Levels at which littoral wetlands connect to the lakes and rocky shores begin to be inundated in Lake Crescent are indicated. Maximum (full supply) and minimum (sill) operating levels of sluice gates in both lakes are also shown. Data are plotted with the same y-axis scales as those used in Figure 3.3 to compare the magnitudes of the water releases in 2007 and 2009. For a clearer presentation of the 2007 data (i.e. more appropriate y-axis scales) see DPIW (2008).
Lake Sorell
Observed Modelled 804.5 Full supply
804.0
Wetlands
803.5 Water levelAHD)Water (m
Sill 803.0
802.5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Month
Lake Crescent
Full supply
Observed Modelled 803.5
803.0
Wetlands
802.5
Rocks Water levelAHD)Water (m
802.0
Sill
801.5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Month
147 Response of Lake Crescent and Lake Sorell to water level variability during 2009
Appendix 14. Rise in water level between May and September and the mean catch of Galaxias auratus larvae (≤40 mm FL) in pelagic samples during seven spawning seasons (August-December) in lakes Sorell and Crescent between 2000 and 2009. Seasonal water level rise values are untransformed (cf. Figure 3.55).
Lake Sorell
2.5
2.0
1.5
1.0 Cohort 2000 2001 0.5 Lake Crescent 2002 2003 2.5 2004
+1) number of larvae of number +1) 2007 x
( 2009
10 g
o 2.0 L
1.5
1.0
0.5 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 Water level rise (m)
148