Thesis (M. Sc.)--University of Adelaide, 1988. / Includes reprint of journal article. Includes bibliographical references (leaves 196-208).

Includes reprint of journal article. Bibliography: leaves 196-208.

The major question addressed by this project was to determine if the long-necked, freshwater turtle Chelodina oblonga, vocalise underwater and whether their vocal activity could be related to behavioural or ecological aspects of their lives. These turtles often live in wetlands where visibility is restricted due to habitat complexity or light limitation caused by factors such as tannin-staining, or turbidity. For many aquatic animals, sound is a useful means of communication over distances beyond their visual acuity. This thesis gives the first detailed account of the underwater vocal repertoire of C. oblonga. In total, over 230 days were spent in the field and more than 500 hours of tape recordings were made for this research. Initially, a number of recordings took place in three wetlands known to support turtle populations: Blue Gum Lake; Glen Brook Dam; and Lake Leschenaultia in Perth, Western Australia; in order to determine the nature of the freshwater sound field and place turtle vocalisations into the context in which they were vocalising. The wetlands differed in terms of degree of enrichment, substrate material, water depth and habitat complexity. Recordings were made over a four-week period in the last month of summer and the first week of autumn (Feb-Mar 2003). Invertebrate sweeps were also taken over a two-week period at each recording site to determine if invertebrate distributions were related to patterns of sonic activity. To determine the influence of wind on ambient noise; recordings were undertaken on winter mornings (June-August, 2003) at Blue Gum Lake and Glen Brook Dam at locations north, south, west and east for four different wind speeds – Beaufort Wind Scale (BWS) 0,1,2 & 3. There were seven distinctive calls recognised in the recordings. The frequency bandwidth most utilised by organisms was between 3 kHz up to around 14 kHz, with the exception of the ‘bird-like song’; which extended from 500 Hz up to around 10 kHz. Blue Gum Lake contained a more diverse and abundant assemblage of invertebrates than Lake Leschenaultia and Glen Brook Dam. Correspondingly, a greater diversity of calls was recorded at Blue Gum Lake, as well as the presence of chorus activity, which was not heard at the two less-enriched sites. The periods of greatest diversity and abundance of macroinvertebrates was synonymous with the increased sonic activity at dusk and midnight with noise levels greatest at dusk in particular, and to a lesser extent at midnight. There was no difference in ambient noise at Blue Gum Lake or Glen Brook Dam at wind speeds of Beaufort Wind Scale 0, 1 and 2. Turtles from three populations were recorded in artificial environments: consisting of round, plastic, above-ground ponds (1.8m dia. x 0.65m depth), which were set up to recreate small wetlands. Recordings occurred from September to October, 2003 and from February to December, 2004 as well as January, 2005. Seven hatchling and five juvenile turtles (CL <10cm) were also recorded in order to ascertain whether very young turtles vocalised. Hatchlings were recorded in a glass aquarium (35.5cm length x 20cm width x 22.0cm depth) and juveniles were placed into a below-ground outdoor pond (1m length x 0.5m width x 0.4m depth). Recordings occurred from as early as 4.30am (dawn recordings) to as late as 1.30am (evening recordings). The recordings revealed that turtles utilise an underwater acoustic communication system (calling at the water’s surface was also noted but these were not recorded or a part of this research) involving a repertoire of both complex and percussive sounds with short, medium and potentially long-range propagation characteristics. Complex structures included harmonically related elements (richly or sparsely) and different rates of frequency modulation. Frequency use extended beyond the in-air auditory sensitivity known for a single species of turtle studied from the family Chelidae; with calls ranging from around 100 Hz in some of the percussive displays, to as high as 3.5 kHz in some complex calls, with ‘clicks’ extending beyond the 20 kHz upper limit of the recording system. However, most of C. oblonga’s vocalisations had dominant frequencies below 1 kHz. Turtles were intermittent callers with an extensive vocal repertoire of seventeen (17) vocal categories - highly suggestive of complex social organisation. Vocalisations included: a) clacks; b) clicks; c) squawks; d) hoots; e) short chirps; f) high short chirps; g) medium chirps; h) long chirps; i) high calls; j) cries or wails; k) cat whines; l) grunts; m) growls; n) blow bursts; o) staccatos; p) a wild howl; and q) drum rolling. Also, two sustained ‘pulse-bouts’ were recorded during the breeding months, hypothesised to function as acoustic advertisement displays – possibly ‘calling songs’. Hatchling turtles were not heard to vocalise within the audible range. Only a single complex vocalisation was heard produced by the juvenile turtles, with a number of percussive calls. Preliminary playback trials were conducted under free-field conditions and within an artificial environment, which consisted of a below ground rectangular tank (2.4m length x 0.8m width x 0.6m deep). A number of turtle calls recorded in the artificial ponds were selected for playback. A UW 30 speaker was used for broadcast of calls. The free-field playbacks occurred at Mabel Talbot Lake and Blue Gum Lake during the months of April and May, 2005. Playback using 14 seconds of an artificially constructed sequence from the sustained ‘pulse-bout’ occurred in the artificial channels. This sequence consisted of some of the first phase pulses followed by a section of the ‘vibrato’. The preliminary free-field playback trials indicated that turtles had some interest in the calls being played by responding with an ‘alert posture’. Turtles were shown to remain in the alert posture for a significantly longer time than when no sound was played or when white noise was played. The extensive repertoire and initial responses to the free-field playbacks indicated that sound has some biological importance for C. oblonga, although results of playbacks under artificial conditions were inconclusive.

With 226-343 individual/ha, population density of Chelodina longicollis in the dams of the University of Western Sydney-Hawkesbury Richmond campus were in the range of other studies around Australia. Their size extremes (24.3 -223.3 mm) were within the range of previous studies, and the overall sex ratio was skewed toward males. The annual growth rates varied and were weakly correlated with animal size. Scute shedding occurred between September and April and peaked in December. Turtles were generally in excellent condition, indicating that sufficient food resources were available in local habitats. Only 3.4% of the population were in poor condition and few animals (8.8 %) carried signs of past injury. No gross abnormalities were recorded, however, there has been low levels of recruitment to the population compared with other Australian studies. Despite a maximum distance dams sampled of 2.8 km and ample evidence of interchange between dams, there was a great variation in animal size, cohort structure, sex ratio among dams. There are a range of factors which have the potential to bias sampling results. Turtles were not influenced by a dominance hierarchy or by the presence of eels, however, they appeared to be capable of avoidance behaviour when nets are set at a specific location. Different cohorts were caught differentially and this varied with month, season and year. In addition, catchability varied among cohorts. Juveniles were least likely, and sub-adult males were most likely, to be recaptured. In some dams there was evidence that animals moved at random while in others movement did not conform to this pattern. These results could not be accounted for in terms of dam size, physical structure of the dam or the distribution and abundance of vegetation. / Master of Science (Hons)

Thesis (Ph.D.)--Murdoch University, 2005. / Thesis submitted to the Division of Science and Engineering. Bibliography: leaves 203-217.

Most studies of wetlands tend to focus on the biotic and abiotic interactions within the aquatic habitat. Though wetlands and associated biota may appear to be somewhat isolated from the influence of the wider landscape, wetland habitats are critically linked with adjacent terrestrial habitats and other wetlands through the two-way flows of energy and nutrients and provision of structure. While an understanding of these inter-habitat linkages is breaking down the perceived boundaries between &quotaquatic" and &quotterrestrial" ecosystems, there is more limited knowledge on the ecology of wetland animals that must meet critical needs in both aquatic and terrestrial habitats at some time during their life or seasonal cycles. Here, I examine the terrestrial ecology of a freshwater turtle, the eastern long-necked turtle (Chelodina longicollis) in the temporally dynamic and heterogeneous landscape of Booderee national park in south-east Australia by 1) providing a description of terrestrial behaviours, 2) identifying the factors driving terrestrial behaviour and its functional significance, 3) examining factors that may limit or constrain terrestrial behaviour and 4) demonstrating how various terrestrial behaviours can factor prominently in the overall biology of a nominally aquatic animal. Chelodina longicollis used terrestrial habitats for reasons other than nesting, including aestivation and movements between wetlands. Radio-telemetry of 60 turtles revealed that nearly 25 % of all locations were in terrestrial habitats up to 505 m from the wetland, where turtles remained for extended periods (up to 480 consecutive days) buried under sand and leaf litter in the forest. Individuals also maintained an association with a permanent lake and at least one temporary wetland within 1470 m, though some inter-wetland dispersal movements were much longer (5248 m). As a result of their associations with several wetlands and terrestrial aestivation sites, C. longicollis traversed large areas and long distances (13.8 +/- 2.8 ha home range, 2608 +/- 305 m moved), indicating that this species is highly vagile. In fact, a three-year capture-mark-recapture study conducted in 25 wetlands revealed that 33% of the population moved overland between wetlands. After scaling this rate to the number of generations elapsed during the study, C. longicollis moved between discrete water bodies at a rate of 88-132% per generation. This rate is not only high for freshwater turtles, but is among the highest rates of inter-patch movement for any vertebrate or invertebrate. Chelodina longicollis demonstrated an impressive capacity for individual variation in nearly every aspect of its behaviour examined. Most of the variation in space use, movements, terrestrial aestivation and activity could be attributed to extrinsic local and landscape factors, seasonal influences and rainfall, whereas intrinsic attributes of the individual such as sex, body size, body condition and maturity status were less important. Turtles increased movement distance and home range size in regions where inter-wetland distances were farther and with increasing wetland size. Individuals spent more time in terrestrial habitats with decreasing wetland hydroperiod and increasing distance to the nearest permanent lake. Overland movements between wetlands were correlated with rainfall, but the directionality of these movements and the frequency with which they occurred varied according to the prevalent rainfall patterns; movements were to permanent lakes during drought, but turtles returned to temporary wetlands en masse upon the return of heavy rainfall. However, deteriorating conditions in drying wetlands forced turtles to move even in the absence of rainfall. Captures at a terrestrial drift fence revealed that immature turtles as small as 72.3 mm plastron length may move overland between wetlands with similar frequency as larger adults. Taken together, these results suggest that C. longicollis behaviour is in part conditional or state-dependent (i.e., plastic) and shaped by the spatiotemporal variation and heterogeneity of the landscape. Perhaps the most surprising aspect of individual variation was the alternate responses to wetland drying. Turtles either aestivated in terrestrial habitats (for variable lengths of time), or moved to other wetlands. Movement to other wetlands was the near universal strategy when only a short distance from permanent lakes, but the proportion of individuals that aestivated terrestrially increased with distance to the nearest permanent lake. When long distances must be travelled, both behaviours were employed by turtles in the same wetland, suggesting that individuals differentially weigh the costs and benefits of residing terrestrially versus those of long-distance movement. I propose that diversity in response to wetland drying in the population is maintained by stochastic fluctuations in resource quality. The quality of temporary wetlands relative to permanent wetlands at our study site varies considerably and unpredictably with annual rainfall and with it the cost-benefit ratio of each strategy or tactic. Residency in or near temporary wetlands is more successful during wet periods due to production benefits (high growth, reproduction and increased body condition), but movement to permanent wetlands is more successful, or least costly, during dry periods due to the fitness benefits of increased survival and body condition. I used the doubly-labelled water (DLW) method to provide the first estimates of water and energy costs of aestivation and overland movement for any freshwater turtle behaving naturally in the field. Chelodina longicollis remained hydrated while terrestrial with water flux rates (14.3-19.3 ml kg-1 d-1) on par with those of strictly terrestrial turtles, but field metabolic rate during aestivation (20.0-24.6 kJ kg-1 d-1) did not indicate substantial physiological specializations in metabolism during aestivation. Energy reserves, but not water, are predicted to limit survival in aestivation to an estimated 49-261 days, which is in close agreement with the durations of natural aestivation. The energy costs of overland movement were 46-99 kJ (kg d)-1, or 1.6-1.7 times more expensive than aestivation. When a wetland dries, a turtle that foregoes movement to other wetlands can free sufficient energy to fuel up to 134 days in aestivation. The increasing value of this energy &quottrade-off" with travel distance fits our behavioural observations of variance in response to wetland drying. Taken together, this evidence indicates that terrestrial habitats provide more than just organic and structural inputs and filtering services and that nearby wetlands are important for reasons other than potential sources of occasional colonists to a population. Terrestrial habitats are used for aestivation in response to wetland drying and different wetlands are diverse in their functions of meeting the annual or life-cycle requirements of C. longicollis in temporally dynamic wetland systems. As overland movements between these various habitat types are in response to spatiotemporal variation in habitat quality and associated shifts in the fitness gradient between them, I suggest that terrestrial and different aquatic habitats in Booderee offer complementary resources contributing to regional carrying capacity and population persistence of the turtle population. Thus, important ecological processes regulating C. longicollis in a focal wetland should not be viewed as operating independently of other nearby wetlands and their adjacent terrestrial habitats. Collectively, these findings highlight the complex and dynamic associations between a population of freshwater turtles and the wider terrestrial and aquatic landscape, demonstrating that turtle populations and the factors that impact them can extend well beyond the boundaries of a focal wetland.