SEARCH

SEARCH BY CITATION

Keywords:

  • littoral habitat;
  • recruitment;
  • reservoir;
  • spawning;
  • vulnerable species;
  • water-level variation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References

This study assessed the spawning activity of the threatened Australian lungfish Neoceratodus forsteri by measuring egg densities within the artificial habitat of a large impoundment (Lake Wivenhoe, Australia). Eggs were sampled (August to November 2009) from multiple locations across the impoundment, but occurred at highest densities in water shallower than 40 cm along shorelines with a dense cover of submerged terrestrial vegetation. The numbers of eggs declined over the study period and all samples were dominated by early developmental stages and high proportions of unviable eggs. The quality of the littoral spawning habitats declined over the study as flooded terrestrial grasses decomposed and filamentous algae coverage increased. Water temperatures at the spawning site exhibited extreme variations, ranging over 20·4° C in water shallower than 5 cm. Dissolved oxygen concentrations regularly declined to <1 mg l−1 at 40 and 80 cm water depth. Spawning habitats utilised by N. forsteri within impoundments expose embryos to increased risk of desiccation or excessive submergence through water-level variations, and extremes in temperature and dissolved oxygen concentration that present numerous challenges for successful spawning and recruitment of N. forsteri in large impoundment environments.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References

The impoundment of rivers can impact negatively on aquatic ecosystems as a result of habitat alteration, modified hydrological patterns, unnatural fluctuations in water quality and changes in biotic communities, often leading to declines in species richness and abundance (Arthington & Pusey, 2003; Hoeinghaus et al., 2009; Kolding & van Zwieten, 2012; Lintermans, 2012). The Australian lungfish Neoceratodus forsteri (Krefft, 1870) is one of six extant lungfish species and is the only Australian representative of a once diverse assemblage prevalent in the Devonian period (c. 413–365 million years ago) (Denison, 1968). The current geographic distribution of N. forsteri is restricted to the Mary, Burnett, Brisbane and North Pine Rivers, with remnant translocated populations of uncertain status in other river systems in south-east Queensland (Kind, 2011). Neoceratodus forsteri faces a range of threats to long-term population viability, including pollution, habitat modification, introduced pests, flow modification, declining water quality and the construction of impoundments (Arthington, 2009; Kind, 2011). The international significance of N. forsteri and the recognized threats to population viability have resulted in their listing as a nationally threatened species under Australian legislation (Environmental Protection and Biodiversity Conservation Act of 1999) and protected by the CITES treaty (CITES, 1979).

In riverine habitats, N. forsteri select specific spawning habitat conditions utilizing shallow water (c.10–60 cm) typically in gently flowing water amongst submerged macrophyte beds (Kemp, 1984; Brooks & Kind, 2002). Spawning can occur in slack water, but at depths <20 cm. Neoceratodus forsteri utilize a range of microhabitats when preferred macrophyte habitats are not available, such as the submerged roots of Callistemon spp. trees and the floating water hyacinth Eichhornia crassipes (Kemp, 1984). Brooks & Kind (2002) and Espinoza et al. (2012) report that egg densities were positively correlated with increasing macrophyte cover, but were negatively influenced by excessively tall and dense stands or in the presence of filamentous algae. Water quality conditions associated with riverine spawning habitats of N. forsteri are characterized by high dissolved oxygen (DO) concentrations (between 2·0 and 15·5 mg l−1) and moderate temperature ranges (between 12 and 32° C) (Brooks & Kind, 2002). In addition to these specific spawning habitat preferences, N. forsteri embryos have long development times of between 23 and 30 days from oviposition to hatching (Kemp, 1986). Larvae remain largely immobile and sedentary for a further 14–21 days while the yolk sac is absorbed (Kemp, 1986). The combination of specific spawning habitat requirements and long development times to free swimming (maximum of 50 days) presents many challenges for N. forsteri spawning, egg viability and larval survival. These challenges are expected to increase within habitats that experience large water level and physicochemical variations over daily and monthly time frames, as is potentially the case within impoundment habitats (Wantzen et al., 2008).

Impoundments in sub-tropical climates exhibit unnatural hydrological regimes compared with free flowing rivers, particularly with respect to the timing and extent of water-level variations. The natural hydrological regime of rivers that N. forsteri inhabits are characterized by high interannual variability in summer wet season flows (December to March). During the peak N. forsteri spawning period (August to December) (Kind, 2011; Espinoza et al., 2012), however, flow is typically at the lowest and least variable (Pusey et al., 1993). In contrast, water levels in south-east Queensland impoundments rarely remain stable and typically decline during this period. This variability can be exacerbated by drawdown for consumptive use and increased evaporation rates. Another feature of sub-tropical impoundments is the often depauperate riparian vegetation (Peden et al., 2011; Zohary & Ostrovsky, 2011) and lack of discernible water currents (Littlejohn, 2004). Typically the shorelines are dominated by pasture grasses and low-growing perennial vegetation that colonize during periods of lowering water levels, and subsequently become inundated and decay following rises in water level. This combination of features are likely to result in greater diurnal fluctuations in temperature and DO concentrations in the littoral zone (Pusey & Arthington, 2003) than would occur in natural riverine habitats with intact riparian vegetation and constant water flow. These unnatural hydrological and physical features are likely to influence N. forsteri spawning and recruitment success within impoundment habitats.

This study investigated the spawning activity and the spawning habitat utilized by N. forsteri in the artificial impoundment habitat of Lake Wivenhoe, a large 10 750 ha sub-tropical impoundment on the Brisbane River in south-east Queensland.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References

Sampling locations and methods

The study was undertaken between August and November 2009, in the sub-tropical Lake Wivenhoe, Queensland, within an embayment of the lake (Logans Inlet, 27° 19′ 48″ S; 152° 32′ 20″ E) (Fig. 1). A qualitative survey aimed to establish the prevalence of spawning activity across the lake, the typical habitats utilized for spawning and the location for the quantitative component of the study. The qualitative survey involved sampling 21 random locations across the lower reaches of the lake for the presence of eggs using a 30 cm × 30 cm push net (3 mm mesh) pushed along the substratum for a distance of 10 m, in water depths of 20–40 cm, with three replicates per location. The contents of each net were inspected for the presence of eggs and their relative abundance noted.

image

Figure 1. Location of Lake Wivenhoe showing main study site at Logan's Inlet (image) and the locations of qualitative survey sites assessed for the presence of Neoceratodus forsteri eggs (image).

Download figure to PowerPoint

The quantitative survey used a specially constructed sampling device to enable non-destructive quantitative sampling of eggs and habitat. The device consisted of a 1 m2 box enclosure with an open top and bottom and one side comprised a 1 mm removable nylon mesh net. This net formed a collection codend for eggs and suspended detritus. A petrol-driven water pump was operated by one person to gently agitate the water and substratum in the enclosure, suspending the eggs for collection. At the same time, a 30 cm × 30 cm push net (3 mm mesh) was used by a second person to sweep the enclosure water column to collect suspended eggs. Sampling ceased when three consecutive sweeps of the quadrat with the push net yielded no further eggs. All material collected in the sweep net and the codend was placed into white sorting trays and hand-picked in the field for eggs and larvae. This facilitated the rapid counting and immediate return of eggs back to the collection site without damage. To evaluate the efficiency of hand picking in the field, all remaining organic matter was preserved in 70% ethanol and returned to the laboratory for re-sorting. Any additional eggs found were only included in the total egg count, as the preservation process prevented accurate determination of egg developmental stages.

The efficiency of the sampler in retrieving the demersal eggs was evaluated by quantifying the recovery rate of a known number of surrogate N. forsteri eggs randomly placed into the enclosure over typical habitat types encountered at the survey site. The surrogates chosen for this evaluation were boiled tapioca balls as these had similar, although slightly larger, diameter to N. forsteri eggs, and similar negative buoyancy properties.

The quantitative survey comprised five separate sampling events, undertaken every 14 days from mid-September until mid-November. At the commencement of the survey, a 100 m reach of the shoreline was marked out with 10 m transects perpendicular to the shoreline. On each sampling occasion, samples were collected at three water depths (20, 40 and 80 cm) along each transect. Neoceratodus forsteri eggs within each quadrat were then counted and the stage of development was determined. Quadrat locations were offset by 2 m on any transect sample location that was repeated to avoid pseudo-replication. The total number of replicate transects per sampling date varied between 4 and 7.

Habitat descriptor variables were measured within each quadrat prior to egg sampling. The percent total cover of each vegetation type was estimated and the mean vegetation height was calculated from 10 replicated height measurements from within each quadrat. For non-rooted vegetation types including Ceratophyllum demersum and filamentous algae, only the total percent cover was estimated. Water quality at the sampling location was measured throughout the study period with multiparameter data loggers (Eureka Manta2; www.meas-spec.com; D-OptoLogger; www.d-opto.com; YSI 6920; www.ysi.com) and temperature loggers (Hobo Pendant UA-002-08; www.onsetcomp.com). The multiparameter data loggers were deployed directly on the substratum surface at a depth of 40 and 80 cm ensuring that individual sensors were not in direct contact with the substratum, thus representing typical conditions of N. forsteri eggs resting among vegetation. Temperature loggers were deployed at c. 5 cm water depths and all loggers recorded data at 5 min intervals.

Developmental stages of eggs were determined in the field using a four-stage classification scheme that comprised an aggregate of the more detailed stage descriptions provided by Kemp (1981). Stage classifications included 1, 2 and 3 week-old embryos, and peri-hatch embryos (stage immediately before hatching is imminent). A newly hatched larva termed hatchlings by Kemp (1981), have been called a larva here and refers to the period of time over which the yolk sac is absorbed and reaches the free swimming stage. An additional category of unviable (dead) eggs (Table 1) was used to describe eggs distinguishable by the size of the unviable embryo, tending to be larger and having a distinctive uneven shape and a green and mottled discolouration due to bacterial growth, when compared with viable embryos that were more evenly shaped with an even brown colouration of a developing embryo. All eggs other than a small selection of dead eggs were returned to the sample quadrat from which they had been collected.

Table 1. Categories used to differentiate egg development stages and viability status of Neoceratodus forsteri eggs
CategoryDevelopment stages (from Kemp, 1986)Notes
Week 1Stages 0–24No visible embryo differentiation
Week 2Stages 25–34Head and tail noticeable
Week 3Stages 35–40Elongation of embryo
Peri-hatch>Stage 41Pigmentation/head developed
Dead (unviable) Embryo uneven and discoloured

Statistical analyses

Total egg density (T), which included all developmental stages, was used as the response variable for all statistical analyses concerning spatiotemporal variation. Histograms of T resembled the Poisson-like distribution typical of count data, generally displaying most observations at or near zero and relatively few observations of higher values. Transformation was not sufficient to normalize data distribution and non-parametric statistics were used for analysing T.

Two types of generalized linear model (GLM) were used based on the ‘glm’ (quasi-Poisson distribution) and ‘glm.nb’ (negative binomial distribution) functions in R, version 2.10.1 (www.r-project.org/), in order to analyse differences in T relative to sampling date and sampling depth while accounting for its overdispersed mean-to-variance relationship (MVR) (O'Hara & Kotze, 2010). Based on model residuals and predicted MVRs, the negative binomial model provided a better fit (Ver Hoef & Boveng, 2007). Initial models included transect location as a random factor. Because there was no significant difference among means of T using transect as a grouping factor across sampling dates and water depths (Kruskal–Wallis χ2 14·73, P > 0·05, d.f. = 13) and because GLM results remained virtually unaffected by its exclusion, however, it was excluded from the model. The Wilcoxon rank sum test (W), wilcox.test function in R, was used for post hoc pair-wise analyses of main effects, i.e. for comparing means of T using (1) pairs of sampling dates across sampling depths and (2) pairs of sampling depths across sampling dates. Transect data were pooled for all these tests such that there were no sub-groups (d.f. = 1) and a minimum of n = 12 data points per group.

Vegetation community structure at Logan's Inlet was analysed using the combined programs Primer, version 6.1.11, and PERMANOVA+, version 1.0.1 (Anderson, 2001, 2005; Clarke & Warwick, 2001). Missing data points led to exclusion of the vegetation categories Myriophyllum spp. and an unidentified small green leaved species, both of which represented only a very small percentage of the observed vegetation. Prior to analysis, all data were square-root transformed and distributions investigated using Draftsman plots. Following the computation of a Bray–Curtis resemblance matrix, permutational multivariate analysis of variance (PERMANOVA) was applied to analyse community shifts using sampling date and sampling depth as fixed factors. Initially, the transect location was included as a random factor, but because, here too, it did not affect outcomes, it was removed from the model. The BEST function of the Primer package was used to determine the types of vegetation best explaining observed community structure.

Vegetation types identified by BEST results to be the most important in structuring the community were analysed individually in more detail. As previously described for T, the Wilcoxon test was used to analyse differences in percent cover of these influential vegetation types among sampling dates (across sampling depths) and among sampling depths (across sampling dates). Additionally, percent cover of the influential vegetation types was correlated with T using the ‘cor’ and ‘cor.test’ functions (Kendall and Spearman methods) in R in order to identify trends in the relationship between egg density and vegetation. Principal findings with respect to these and other Primer and PERMANOVA outcomes were illustrated using multidimensional scaling (MDS) plots.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References

Spawning patterns

The qualitative surveys across Lake Wivenhoe (Fig. 1) revealed the presence of N. forsteri eggs at 76% of sites. In general, the abundance of eggs at these sites was low and a large proportion of eggs were classed as unviable. The survey indicated that the presence of N. forsteri eggs was associated with shorelines of relatively shallow water (<1 m deep), protected from prevailing winds (generally the south-western shorelines), with substrata covered by short, but dense, inundated terrestrial grass and sporadic macrophyte cover. Eggs were mostly absent from steep shorelines with either tall dead grass cover or predominantly bare ground. The qualitative surveys confirmed that Logan's Inlet had the highest egg densities of all sites surveyed and that the habitat types available were typical of other locations containing N. forsteri eggs.

Field sorting eggs from retained samples was 89·6% efficient. The mean ± s.d. recovery rate of surrogate eggs from 10 replicate samples of 50 surrogate eggs placed in the sampler was 91 ± 10%. Quantitative egg surveys at Logan's Inlet resulted in 701 N. forsteri eggs being collected in field-sorted samples from a 93 m2 of habitat sampled. This comprised 211 unviable eggs, 363 week 1, 38 week 2 and five week 3, and 10 peri-hatch stage eggs, with only one larva found. An additional 73 ethanol-preserved eggs from the field sorting efficiency trial were recovered. Overall, this represents a mean ± s.d. egg density per sample event of 9·6 ± 13·0 m2. The density of N. forsteri eggs varied among sampling events and decreased significantly from the first to last sample event (Fig. 2). There was no significant difference among egg densities at different sampling depths, nor was the interaction between the factors' sampling date and depth significant according to the GLM (Table 2). Post hoc tests principally confirmed these findings, but revealed that, across sampling dates, egg densities were significantly higher at a water depth of 40 cm than at 20 and at 80 cm (W ≥ 236, P < 0·001, n = 31). Pair-wise comparisons of egg densities among sampling dates statistically confirmed the trend illustrated in Figs 2 and 3, by showing that reduction in egg density was most substantial between the first and any of the following sampling dates (>80%, W ≥ 176, P < 0·001, n1 ≥ 12, n2 ≥ 18) (Fig. 3).

image

Figure 2. Total Neoceratodus forsteri egg density and proportional developmental stages (image, unviable; image, empty shell; image, week 1; image, week 2; image, week 3; image, peri-hatch; image, other) collected on five sampling dates in 2009. The other category refers to unstaged eggs collected from ethanol-preserved field detritus.

Download figure to PowerPoint

Table 2. Generalized linear model results for differences in Neoceratodus forsteri egg density at Lake Wivenhoe with respect to sampling date and water depth
SourceEstimated.f.s.e.t-valueP
Intercept4·097 0·7735·299<0·001
Date−0·64240·233−2·762<0·01
Depth−0·00120·015−0·065>0·05
Date × depth−0·00480·005−0·800>0·05
image

Figure 3. Boxplot (mean, quartiles and ranges) of total Neoceratodus forsteri egg density of all egg developmental categories combined, collected on five sampling dates (1 = 16 September; 2 = 2 October; 3 = 16 October; 4 = 2 November; 5 = 16 November 2009) at three water depths (20, 40 and 80 cm).

Download figure to PowerPoint

PERMANOVA results showed that the structure of the vegetation community changed significantly over the course of the sampling period and with increasing water depth. The interaction between the factors' sampling date and sampling depth was also significant (Table 3). Three variables were identified to best explain observed community structure: grass (i.e. unidentified terrestrial grass), filamentous algae and Nitella spp. Based on pair-wise comparisons across all sampling depths, it was found that percent cover of grass decreased from 70 to 28% (W = 215, n1 = 12, n2 = 21, P < 0·001). In the same manner, grass cover decreased steadily and significantly with increasing water depth (from 61 to 25%, W = 838, n1 = 31, n2 = 31, P < 0·001). Temporal variation in percent cover of filamentous algae was marked by a peak at 58% in the middle of the sampling period compared with 11% in the beginning (W = 0, n1 = 12, n2 = 21, P < 0·001) and 39% in the end (W = 306, n1 = 21, n2 = 21, P < 0·05). The decrease of filamentous algae cover with increasing water depth was similar to that observed for grass from 56 to 28% (W = 718, n1 = 31, n2 = 31, P < 0·001). Percent cover of Nitella spp. showed very different patterns, which unravelled the interactive force of community shifts over time and with water depth revealed by PERMANOVA. Mean percentage cover of Nitella spp. varied between 9 and 15% across all sampling depths. Analysing depth across all dates, in contrast, Nitella spp. cover showed a significant increase below 20 cm (4% at 20 cm compared with 15% at 40 cm and 17% at 80 cm; W ≤ 203, n1 = 31, n2 = 31, P < 0·001).

Table 3. PERMANOVA results for differences in vegetation community structure at Lake Wivenhoe with respect to sampling date and water depth. Number of permutations ≥ 9921
Sourced.f.SSMSPseudo-FP
Date4534313368·148<0·001
Depth23882194111·84<0·001
Date × depth83487435·82·659<0·01
Residuals 9835163·9  

Representation of community structure in MDS plots described shifts in percent cover of the three most influential vegetation types. An overlay of these MDS representations with observed egg densities visualized that highest numbers of eggs were generally found when and where cover of grass was greatest, and when and where cover of filamentous algae was lowest (Fig. 4). This trend was also evident by plotting values of T v. observed cover of grass and filamentous algae. Correlations were, however, weak although significant (Kendalls's τ = 0·159, n = 93, P < 0·05 for grass, and Kendall's τ = −0·154, n = 93, P < 0·05 for filamentous algae). Apparent threshold levels of ≥40% grass cover and ≤30% filamentous algae cover marked the conditions for which egg densities higher than the observed mid-range of 41·5 were found (Fig. 5). The interrelationship between egg density and cover of Nitella spp. was not clear.

image

Figure 4. (a) Multidimensional scaling (MDS) plots of vegetation community structure depicting shifts in structure in relation to Neoceratodus forsteri egg density at Lake Wivenhoe. Results are presented by sampling dates 1–5 (1 = 16 September; 2 = 2 October; 3 = 16 October; 4 = 2 November; 5 = 16 November 2009) and water depths (image, 20 cm; image, 40 cm; image, 80 cm). (b) A bubble plot of the MDS in (a), depicting observed N. forsteri egg abundance shifts over each sampling date (1–5). (c)–(e) Bubble plots of the MDS in (a), depicting the percentage cover of vegetation categories, (c) grass, (d) filamentous algae and (e) Nitella sp. The circled areas in plots (c)–(e) provide a visual reference to the sample occasions when egg densities were >20 m2 as is apparent in plot (b). Stress of the MDS was 0·16.

Download figure to PowerPoint

image

Figure 5. Correlations between the density of Neoceratodus forsteri eggs and percentage cover with two influential groups of vegetation, (a) grass and (b) algae, at Lake Wivenhoe. Both correlations were significant at the 95% level (n = 93). Shading highlights the key areas of coverage.

Download figure to PowerPoint

Water quality

Continuous data loggers deployed at 5, 40 and 80 cm depth at the study site showed considerable seasonal and diurnal variations in temperature and DO concentration (Table 4). The mean temperatures and DO concentrations were similar across depths and showed similar increases in temperature, and declines in DO concentration over the study period. The range (minimum–maximum) in temperature and DO concentration, however, was markedly different between depths and times. At 5 cm depth, the maximum range observed was 20·4° C at the start of the study and 11·8° C at the end. At 40 cm depth the range in temperature was 4·5° C at the commencement and 4·3° C at the end of the study, whereas at 80 cm depth, the range was 4·6° C at the start and 4·4° C at the end of the study. The range in DO concentrations remained relatively constant over the study period, albeit much greater at 80 cm depth. The minimum DO concentrations decreased substantially over the study period. At 40 cm depth the minimum DO concentration was 7 mg l−1 at the commencement and 0·8 mg l−1 at the end of the study, whereas at 80 cm depth, the minimums regularly reached 0 mg l−1 throughout the study period.

Table 4. Mean (minimum and maximum) water temperature at 5, 40 and 80 cm depths, and dissolved oxygen concentrations at 40 and 80 cm depths, for 14 days preceding each sampling date in 2009 from Logans Inlet, Lake Wivenhoe
Sample dateVariable
Temperature (°C)Dissolved oxygen (mg l−1)
5 cm40 cm80 cm40 cm80 cm
16 September23·2 (16·0–36·4)22·2 (20·7–25·2)No data10·1 (7·0–13·4)No data
2 October22·4 (12·7–35·6)21·0 (17·7–23·8)No data9·3 (4·9–15·1)No data
16 October22·9 (17·8–30·1)22·2 (20·6–24·6)22·6 (20·8–25·4)8·1 (3·7–12·7)8·6 (0–17·5)
2 November25·1 (18·0–36·0)25·2 (23·6–28·7)25·0 (21·9–28·2)7·0 (1·5–11·1)7·5 (0–19·4)
16 November26·0 (22·2–34·0)25·2 (23·8–28·1)25·9 (24·4–28·8)4·1 (0·8–8·8)3·4 (0–12·3)

Lake water levels leading up to the commencement of quantitative egg sampling steadily declined (Fig. 6). In the 50 days prior to the commencement of sampling (maximum development times for N. forsteri to free swimming), water levels in Lake Wivenhoe declined by 24 cm. Water-level changes over the 10 week period of the study were much less, ranging by c. 6 cm as a direct result of management interventions taken to moderate water levels by releasing water from an upstream impoundment.

image

Figure 6. Water-level variation in Lake Wivenhoe from July to December 2009.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References

Previous research on N. forsteri spawning has primarily focused on riverine habitats (Kemp, 1984; Brooks & Kind, 2002) or in smaller weir pools on river systems (Espinoza et al., 2012), with limited published records of N. forsteri spawning within larger impoundment habitats (Kemp, 1984). The results of this study highlight a range of issues influencing N. forsteri spawning success within impounded environments. The preferred spawning habitat utilized by N. forsteri in Lake Wivenhoe is consistent with the depth range (c. 40–50 cm) and habitat types (dense macrophyte cover) utilized in riverine environments (Brooks & Kind, 2002; Kind, 2011; Espinoza et al., 2012). In addition, riverine populations of N. forsteri will spawn in slack water conditions, but typically at shallower depths. In Lake Wivenhoe N. forsteri spawn more in the shallower (20–40 cm) margins of the lake when suitable habitat is available. In this study, egg densities in Lake Wivenhoe were high compared to surveys from riverine habitats (Brooks & Kind, 2002; Espinoza et al., 2012). It is uncertain if this is a methodological artefact or if egg densities were higher in Lake Wivenhoe compared with previously reported riverine studies. Higher egg densities may also be attributed to the concentration of spawning activity over fewer areas of suitable spawning habitat in Lake Wivenhoe.

The submerged vegetation utilized for spawning by N. forsteri in Lake Wivenhoe was not restricted to aquatic macrophytes. It was dominated by terrestrial grasses that had been recently inundated following a rise in lake levels in the months preceding the spawning season. The rapid qualitative egg surveys across multiple sites in Lake Wivenhoe revealed that the presence of eggs was heterogeneous and was only amongst areas covered in short vigorous grass shoots, and not amongst taller moribund grass. It is hypothesized that the submerged fresh grass shoots closely resembled stands of the aquatic macrophyte Vallisneria sp. Neoceratodus forsteri has previously been shown to exhibit spawning habitat plasticity, utilizing semi-aquatic and submerged littoral and terrestrial vegetation (Kind, 2011). It would appear possible that for limited periods, freshly flooded terrestrial grasses of a certain growth form can provide substitute spawning habitat for N. forsteri in impoundments.

The reduction in egg densities over the study period could be related to a number of factors including changes in spawning habitat conditions or reduced spawning activity. The habitat quality declined owing to the progressive decomposition of the submerged terrestrial grasses and the progressive increase in filamentous algae covering the substratum. Filamentous algae is considered unsuitable spawning habitat for N. forsteri (Brooks & Kind, 2002). Signs of spawning activity were first detected on 13 August, when eggs were abundant at the study site. These qualitative and subsequent quantitative samples indicated that the spawning season of N. forsteri in Lake Wivenhoe closely matched that previously found for riverine populations (Kind, 2011). As such, the reduction in egg densities observed may be a result of both declining spawning activity in the later stages of the spawning season, shown to be around the end of November or early December (Brooks & Kind, 2002; Espinoza et al., 2012), and declines in habitat quality.

The littoral zone of Lake Wivenhoe experiences large fluctuations in temperature and DO concentration throughout the spawning season. These large fluctuations may be due to the lack of appreciable water movement within large impoundment habitats, combined with a lack of riparian shading. Additionally, the littoral benthos often comprises soft organic rich sediments, which could lead to increased oxygen demand. Little is known about the physiological tolerances of N. forsteri eggs to extremes in temperature and DO concentration. Brooks & Kind (2002) and Espinoza et al. (2012) measured variations in temperature and DO concentration within spawning habitats of the Burnett River and found maximum temperatures of between 30 and 32° C, and minimum DO concentration of between 2·0 and 6·9 mg l−1. Under laboratory conditions, temperatures >30° C increase N. forsteri egg mortality (Kemp, 1981). Based on the limited temperature and DO concentration data reported in the literature from riverine N. forsteri spawning habitats, the ranges measured in the littoral zone of Lake Wivenhoe may present considerable physiological challenges for N. forsteri egg development and survival. These physiological challenges may be a contributing factor to the general low proportion of viable eggs in samples observed here, compared with riverine spawning studies (Kemp, 1984; Brooks & Kind, 2002) that report much higher proportions of viable eggs in samples.

Variation in water level is likely to be a critical factor during the spawning season for demersal egg-laying species (Clark et al., 1998; Ozen & Noble, 2005; Clark et al., 2008). Given the narrow depth range that N. forsteri utilize for spawning and the long development time of eggs to the free swimming juvenile stage, the desiccation of macrophytes or surrogate spawning habitats during declines in water level would affect recruitment success (Brooks & Kind, 2002; Arthington, 2009). Additionally, water-level changes will impede the establishment of macrophyte habitats following inundation or desiccation of littoral zones (Duivenvoorden, 2008). Within riverine habitats, average river flows during the N. forsteri spawning season (August to December) are at their lowest and most stable (Corfield, 2007). Stable water levels provide ideal conditions for the establishment and maintenance of critical shallow water habitats reducing the likelihood of desiccation of eggs and immobile larval stages. In contrast, water levels in Lake Wivenhoe during the N. forsteri spawning season typically experience sustained declines in water level (Fig. 6), with potential changes of 20–40 cm over the N. forsteri egg developmental period. This degree of water-level change has the potential to expose eggs laid in shallow water to greater extremes in temperature and risk of desiccation. Given these findings, the capacity to maintain stable water levels within artificial impoundment habitats over the spawning season may provide some management options to increase the potential spawning success of N. forsteri within impoundments.

Neoceratodus forsteri are a large bodied, long-lived fish, with a spawning and recruitment strategy that has relatively few eggs per fish (maximum of 500 eggs) (Kemp, 1984), long egg development times from oviposition to free swimming stage and no parental care or egg protection strategies. The infrequent occurrence of juvenile N. forsteri in riverine habitats (Kind, 2011) suggests that egg survival and juvenile recruitment may be naturally low or sporadic. It has been postulated that life-history adaptations to highly variable juvenile recruitment rates include an increase in age at maturity, reduced reproductive investment and adult longevity (Schaffer, 1974; Stearns, 1977), all attributes that N. forsteri possess.

This study has shown that N. forsteri will spawn within the artificial habitats of Lake Wivenhoe, but the microhabitats that exist appear unsuitable for high rates of egg survival and juvenile recruitment. Combined with highly variable water levels and extremes in water quality conditions, large impoundments such as Lake Wivenhoe appear unlikely to support reliable recruitment of N. forsteri. In heavily regulated rivers, such as those that N. forsteri inhabit, these findings place even greater emphasis on identifying, maintaining and improving the remaining riverine habitats upstream and downstream of impoundments in order to provide N. forsteri with suitable spawning habitats and opportunities for recruitment. Until a more complete understanding of the mechanisms affecting N. forsteri recruitment is achieved, a conservative approach to the management of remaining natural habitats would be a prudent strategy to ensure ongoing recruitment and maintenance of populations of this species.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References

The authors thank N. Christiansen, P. Bond and J. Batista for assistance in the laboratory and field, and A. Kemp (University of Queensland), P. Kind and S. Brooks (Fisheries Queensland) for their insights and support for the project and their founding research into N. forsteri ecology. All work conducted in this study was approved by the University of Queensland Animal Ethics Committee (permit number CMS/487/09/SEQ Water). This research was funded by the Queensland Bulk Water Supply Authority (trading as Seqwater).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References
  • Anderson, M. J. (2001). A new method for non-parametric multivariate analysis of variance. Austral Ecology 26, 3246.
  • Anderson, M. J. (2005). PERMANOVA: Permutational Multivariate Analysis of Variance, a Computer Program. Auckland: Department of Statistics, University of Auckland.
  • Arthington, A. H. (2009). Australian Lungfish, Neoceratodus forsteri, threatened by a new dam. Environmental Biology of Fishes 84, 211221.
  • Arthington, A. H. & Pusey, B. J. (2003). Flow restoration and protection in Australia rivers. River Research and Applications 19, 377395.
  • Clarke, K. & Warwick, R. (2001). Changes in Marine Communities: An Approach to Statistical Analysis and Interpretation, 2nd edn. Plymouth: PRIMER-E.
  • Clark, M., Rose, K., Chandler, J., Richter, T., Orth, D. & Van Winkle, W. (1998). Simulating smallmouth bass reproductive success in reservoirs. Environmental Biology of Fishes 51, 161174.
  • Clark, M., Rose, K., Chandler, J., Richter, T., Orth, D. & Van Winkle, W. (2008). Water-level fluctuation effects on centrarchid reproductive success in reservoirs: a modeling analysis. North American Journal of Fisheries Management 28, 11381156.
  • Denison, R. H. (1968). Early Devonian lungfishes from Wyoming, Utah and Idaho. Fieldiana Geology 17, 353413.
  • Duivenvoorden, L. J. (2008). Effects of water level fluctuations on Vallisneria nana in the Burnett River in southeast Queensland, Australia. River Research and Applications 24, 13621376.
  • Espinoza, T., Marshall, S. & McDougall, A. (2012). Spawning of the endangered Australian Lungfish (Neoceratodus forsteri) in a heavily regulated river: a pulse for life. River Research and Applications. doi: 10.1002/rra.2607
  • Hoeinghaus, D., Agostinho, A., Gomes, L., Pelicice, F., Okada, E., Latini, J., Kashiwaqui, E. & Winemiller, K. (2009). Effects of river impoundment of ecosystem services of large tropical rivers: embodied energy and market value of artisanal fisheries. Conservation Biology 23, 12221231.
  • Kemp, A. (1981). Rearing of larvae of the Australian lungfish, Neoceratodus forsteri, under laboratory conditions. Copeia 1981, 776784.
  • Kemp, A. (1984). Spawning of the Australian Lungfish, Neoceratodus forsteri (Krefft) in the Brisbane River and in Enoggera Reservoir, Queensland. Memoirs of the Queensland Museum 21, 391399.
  • Kemp, A. (1986). The biology of the Australian Lungfish, Neoceratodus forsteri (Krefft 1870). Journal of Morphology 190, 181198.
  • Kind, P. K. (2011). The natural history of the Australian Lungfish Neoceratodus forsteri (Krefft, 1870). In Biology of Lungfishes (Jorgensen, J. M. & Joss, J., eds), pp. 6196. Enfield, NH: CRC Press.
  • Kolding, J. & van Zwieten, P. (2012). Relative lake level fluctuations and their influence on productivity and resilience in tropical lakes and reservoirs. Fisheries Research 115-116, 99109.
  • Krefft, G. (1870). Description of a giant amphibian allied to the genus Lepidosiren from the Wide Bay district, Queensland. Proceedings of the Zoological Society of London 1870, 221224.
  • Lintermans, M. (2012). Managing potential impacts of reservoir enlargement on threatened Macquaria australasica and Gadopsis bispinosus in southeastern Australia. Endangered Species Research 16, 116.
  • Littlejohn, C.L. (2004). Influence of artificial destratification on limnological processes in Lake Samsonvale (North Pine Dam), Queensland, Australia. MPhil Thesis, Griffith University, Brisbane, Australia. Available at http://nla.gov.au/anbd.bib-an26738367
  • O'Hara, R. & Kotze, D. (2010). Do not log-transform count data. Methods in Ecology and Evolution 1, 118122.
  • Ozen, O. & Noble, R. (2005). Relationship between Largemouth Bass recruitment and water level dynamics in a Puerto Rico Reservoir. Lake and Reservoir Management 31, 8995.
  • Peden, L., Skinner, S., Johnston, L., Frawley, K., Grant, F., & Evans, L. (2011). Survey of vegetation and habitat in key riparian zones in tributaries of the Murrumbidgee River in the ACT: Cotter, Molonglo, Gudgenby, Naas and Paddys Rivers. Technical Report 23. Canberra: Environment and Sustainable Development Directorate.
  • Pusey, B. & Arthington, A. (2003). Importance of the riparian zone to the conservation and management of freshwater fish: a review. Marine and Freshwater Research 54, 116.
  • Pusey, B., Arthington, A. & Read, M. (1993). Spatial and temporal variation in fish assemblage structure in the Mary River, south-eastern Queensland: the influence of habitat structure. Environmental Biology of Fishes 37, 355380.
  • Schaffer, W. M. (1974). Optimal reproductive effort in fluctuating environments. The American Naturalist 108, 783790.
  • Stearns, S. C. (1977). The evolution of life history traits: a critique of the theory and a review of the data. Annual Review of Ecology and Systematics 8, 145171.
  • Ver Hoef, J. & Boveng, P. (2007). Quasi-Poisson vs. negative binomial regression: how should we model overdispersed count data? Ecology 88, 27662772.
  • Wantzen, K., Rothhaupt, K., Mörtl, M., Cantonati, M., G.-Tóth, L. & Fischer, P. (2008). Ecological effects of water-level fluctuations in lakes: an urgent issue. Hydrobiologia 613, 14.
  • Zohary, T. & Ostrovsky, I. (2011). Ecological impacts of excessive water level fluctuations in stratified freshwater lakes. Inland Waters 1, 4759.

Electronic References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Electronic References
  • Brooks, S. and Kind, P. (2002). Ecology and Demography of the Queensland Lungfish (Neoceratodus forsteri) in the Burnett River, Queensland (With Reference to the Impacts of Walla Weir and Future Water Infrastructure Development). Department of Primary Industries Final Report. Brisbane, QLD: The State of Queensland, DPI. Available at http://catalogue.nla.gov.au/Record/2114730 (last accessed 30 March 2011).
  • Corfield, J. (2007). Potential Effects of Water Abstraction on Aquatic MNES Species in the Mary River and Six Mile Creek. Brisbane, QLD: Hydrobiology Pty Ltd, Environmental Services. Available at http://projects.linkwater.com.au/news.php?newsid=21 (last accessed 30 March 2011).
  • CITES. (1979). Convention on International Trade in Endangered Species of Wild Fauna and Flora. Signed at Washington, D.C., 3 March 1973, Amended at Bonn, 22 June 1979. Available at http://www.cites.org/eng/disc/text.php (last accessed 3 May 2013).