Radio-tracking Interval Effects on the Accuracy of Diel Scale Crayfish Movement Variables

Author A. Ryan, Katie, C. Ebner, Brendan, H. Norris, Richard

Published 2008

Journal Title Freshwater Crayfish

Copyright Statement © 2008 International Association of Astacology. The attached file is reproduced here in accordance with the copyright policy of the publisher. Please refer to the journal's website for access to the definitive, published version.

Downloaded from http://hdl.handle.net/10072/28106

Link to published version http://iz.carnegiemnh.org/crayfish/IAA/

Griffith Research Online https://research-repository.griffith.edu.au Freshwater Crayfish 16:87–92, 2008 Copyright ©2008 International Association of Astacology ISBN: 978-0-9581424-8-9 print / 978-0-9805452-0-3 online

AWARDED BEST STUDENT ORAL PRESENTATION Radio-tracking Interval Effects on the Accuracy of Diel Scale Crayfish Movement Variables

Ka t i e A. Ry a n ,1,† Br e n d a n C. Eb n e r 2 a n d Ri c h a r d H. No rr i s 1

1 University of Canberra, Institute for Applied Ecology, Building 15, ACT 2601 2 Parks, Conservation & Lands, Department of Territory & Municipal Services, ACT Government, GPO Box 158, Canberra ACT 2601, Australia. † Corresponding Author — E-mail: [email protected]

Abstract.— The effect of increasing the interval between radio-tracking fixes above 1 h was examined for estimates of the diel activity, diel mobility, and diel range of the Murray River crayfish, Euastacus armatus. Estimates of these diel variables were determined by radio- tracking six individuals at intervals of 1 h for three diel periods. These data were then sub-sampled to simulate radio-tracking at intervals of 2 h, 3 h, 4 h and 6 h. Our perception of diel activity was altered as a consequence of different sub-sampling intervals, and variation in activity over the diel period was lost as tracking interval was increased. Both mean diel mobility and mean diel range decreased linearly with less frequent radio-tracking. A shift from 1 h to 6 h tracking corresponded to 49% and 55% reduction in estimated diel mobility and diel range, respectively. However, we suggest that these trade-offs between information gain and radio-tracking effort be viewed in the context of a comprehensive understanding of home-range across multiple seasons with a view to conserving this species. [Keywords.— diel activity; diel mobility; diel range; Euastacus; Radio-telemetry; temporal resolution].

Introduction be considered carefully in relation to the objectives of future studies. Radio-telemetry is an established method for studying the movement of terrestrial and freshwater vertebrates, and Similar findings have arisen from studying spotted bass, hence is valuable in providing information required for managing Micropterus punctulatus (Rafinesque) (Horton et al. 2004). That threatened and endangered species. In aquatic environments, where study also demonstrated the importance of radio-tracking the target animals are generally difficult to observe, radio-telemetry provides species while it is most active. Similarly, the nocturnal Japanese an effective means of studying animals whilst causing minimal fluvial sculpin, pollux Günther was tracked intensively at disturbance. Despite the common practice of radio-tracking river night and it was observed that home-range estimates decreased as fauna, the appropriate interval at which to manually radio-track the radio-tracking interval was increased from 1 h to 2 h, 3 h, 4 h remains an important issue. Uncertainty about tradeoffs between and 6 h (Natsumeda 2007). Focusing effort at times when animals accuracy, cost and practicality has led to a lack of standardisation are most active may be an effective way to conduct a radio- in radio-telemetry methods (Horton et al. 2004), which reduces telemetry study, however, it requires that the basic activity pattern comparability between studies. of a species is already known. The effect of radio-tracking interval on records of Unlike in the study of fish, appropriate radio-tracking intervals movement has been the subject of several studies (Baras 1998; have not been examined for crayfish. Previous radio-tracking Ovidio et al. 2000; Horton et al. 2004; Natsumeda 2007). Radio- studies of crayfish have been conducted at intervals ranging tracking of the Barbel, Barbus barbus (Linnaeus), less often than from 3 h (Webb and Richardson 2004) to every one or two days daily has been reported to bias interpretations of home-range (Robinson et al. 2000; Bubb et al. 2004, 2006). Other research and mobility (Baras 1998). That study showed radio-tracking at on the movement of crayfish has used mark-recapture techniques an interval of weekly or greater produced unacceptable losses of (e.g., Gherardi et al. 2000; Byron and Wilson 2001) and direct accuracy, on average 19% and 35.3% for 14 and 28-day intervals, observation (Barbaresi and Gherardi 2001). There is limited respectively (Baras 1998). A similar study of brown trout, Salmo comparability of these studies as a consequence of the disparity in trutta Linnaeus (Ovidio et al. 2000) found that a reduction in the methods used in each case. frequency of radio-tracking from two to 14 days resulted in a The Murray River crayfish, Euastacus armatus (Clark), is loss of accuracy in estimates of mobility by 28.7% and 82.9%, the second largest freshwater crayfish in the world yet its ecology respectively. Furthermore, estimates of home-range decreased in remains poorly understood (as with most Australian crayfish accuracy by 13.3% and 48.7% from two to 14 days, respectively. species). It is increasingly important to understand ecological Ovidio et al. (2000) concluded that radio-tracking interval should aspects of this vulnerable species (Crandall 1996) to ensure its 87 88 Fr e s h w a t e r Cr a y f i s h Vol. 16

Figure 1. Diel activity of E. armatus when sampled at intervals of a) 1 h, b) 2 h, c) 3 h, d) 4 h and e) 6 h. Points repre- sent the mean activity of 5 crayfish (+/- SE). Shading denotes hours of darkness. survival. To this end it would be of benefit to understand the long- Only males were captured which is likely to be a result of them term home-range requirements of this species. Whilst this is beyond being more active at the beginning of the breeding season. A radio- the scope of the current study, we were interested in developing an tag (Advanced Telemetry Systems, Isanti, USA) less than 10% of initial understanding of the home-range and diel activity of the body weight, following Robinson et al. (2000), was attached to species based on investigation at a fine temporal and spatial scale. each of the six individuals. To reduce the possibility of radio-tags Specifically, we aimed to determine the effect of radio-tracking becoming snagged and detached or inhibiting crayfish movement, interval on estimates of the diel activity, diel mobility, and diel radio-tag aerials were tightly coiled around the body of the radio- range of E. armatus. tag and set in place with 5 minute epoxy adhesive (Araldite®, Switzerland). Radio-tags were then attached to the dorsal surface Methods of the carapace using the same adhesive. After ten minutes a second Radio-tracking application of resin was applied to the top of the radio-tag and was smoothed to fill gaps between the radio-tag and the carapace. After Radio-tracking was used to investigate the movement of a drying time of around 2 h, individuals were released at the same E. armatus in the Murrumbidgee River, Australian Capital Territory location from where they were collected. (ACT), in autumn 2005. Euastacus armatus are known to be most active in the cooler months (March – August) during their breeding The locations of crayfish were determined using the Primary and brooding season (Geddes 1990). Crayfish were caught in a Marker System developed by ACT Parks, Conservation and single pool (235 m long and 20 – 50 m wide) using baited hoop Lands (Broadhurst et al. 2005, unpublished data). An arrangement nets following Lintermans and Rutzou (1991) and Asmus (1999). of physical markers was established at the study site as a basis 2008 Ry a n e t a l . — Ra d i o –Tr a ck i n g In t e r v a l a n d Cr a y f i s h Mo v e m e n t Acc u r a c y 89

Table 1. Reduction in accuracy of diel mobility of E. armatus when data from radio-tracking at intervals of 1 h were sub-sampled at intervals of 2 h, 3 h, 4 h and 6 h.

Reduction in accuracy Mean diel OCL (%) if fixes obtained at Individual Sex mobility (m) (mm) intervals of at 1 h intervals 2 h 3 h 4 h 6 h B 99 Male 45 15 15 44 55 C 92 Male 92 20 37 52 52 D 110 Male 0 0 0 0 0 E 92 Male 190 23 46 53 58 F 95 Male 138 31 48 60 62 Mean 98 93 18 29 42 45

Table 2. Reduction in accuracy of diel range of E. armatus when data from radio-tracking at intervals of 1 h were sub-sampled at intervals of 2 h, 3 h, 4 h and 6 h.

Mean diel range Reduction in accuracy (%) if OCL Individual Sex (m2) at 1 h fixes obtained at intervals of (mm) intervals 2 h 3 h 4 h 6 h B 99 Male 40 19 35 70 72 C 92 Male 211 21 40 61 69 D 110 Male 0 0 0 0 0 E 92 Male 638 18 32 50 75 F 95 Male 485 20 47 55 57 Mean 98 275 16 31 47 55 for recording the spatial position of individual crayfish based on Data Analysis radio-tracking over several diel periods. Reference markers were The locations of crayfish were plotted using GIS (ArcGIS) positioned so that they could be viewed at approximately eye level using the distances recorded from reference markers to create the from a canoe. The spatial location of each reference marker was point of best fit. Distances between these locations were measured determined from 60 records on a differential GPS (Trimble Pro XR, and used as a basis for estimates of diel activity (distance moved Sunnyvale, USA), obtained over 10 min. The average of these 60 between intervals) and diel mobility (sum of distances moved records was used to plot the true location of the reference markers over 24 h). Diel range (area inhabited over 24 h) estimates were in a Geographical Information System (GIS) (Acris; Arcview GIS calculated as minimum convex polygons in the Arcview extension, 3.2, Redlands, USA). Movement Program (version 2.04) (Hooge and Eichenlaub Crayfish were radio-tracked at 1 h intervals for three diel 1997). (24 h) periods (10th May 2005; 18th May 2005; 1st June 2006). The To determine the effects of less frequent radio-tracking, hourly order in which crayfish were radio-tracked throughout a diel period data were sub-sampled to simulate radio-tracking at intervals of remained constant to ensure intervals of 1 h were maintained. Radio- 2 h, 3 h, 4 h and 6 h. These intervals were selected because of their tracking was primarily conducted from a canoe and the amount divisibility into 24 h. Distance moved per time interval was plotted of disturbance caused by paddling was minimised when in close and used to assess the relationship between sampling resolution and proximity to radio-tagged crayfish. Individuals were located with estimates of diel activity. Similarly, we tabulated the percentage a 150 – 152 MHz receiver unit (Australis 26k; Titley Electronics, reduction in accuracy of diel mobility and diel range estimates in Ballina, Australia) and a yagi antenna (Titley Electronics, Ballina, relation to temporal sampling resolution. Comparisons were based Australia) until within a few metres of radio-tagged crayfish and on sample means and individual-specific data. Regression analysis then the exact position (assumed to be ± 20 cm: Ebner 2005, was conducted to determine the relationship between tracking unpublished data) was obtained from a gap loop antenna (Titley interval and mean diel mobility and mean diel range. Electronics, Ballina, Australia). The position was then recorded by measuring the distance to three of the reference markers with Results a laser range finder (Yardage Pro Trophy; Bushnell, Heatherton, Radio-tagged crayfish remained within the pool in which Australia) at distances > 5 m and with a measuring pole (marked they were originally collected, however two individuals with 10 cm increments) at distances < 5 m. remained inactive throughout all three diel periods. Further radio- 90 Fr e s h w a t e r Cr a y f i s h Vol. 16

data). Diel mobility ranged from 0 – 220 m (mean = 94 m, sd = 75). Sub-sampling of data collected at intervals of 1 h resulted in a significant and linear decrease in mean diel mobility P( < 0.05) and mean diel range (P < 0.05) (Figure 2). Regression analysis found that the change in tracking interval explained 85% of variability in diel mobility and 93% of the variability in diel range (Figure 2). The mean reduction in accuracy was up to 45% for diel mobility (Table 1) and 55% for diel range (Table 2). The reduction in accuracy of diel mobility and diel range varied between individuals. With the exception of the crayfish that was not active for the entire three diel periods (Individual D) the reduction in accuracy from 1 h to 2 h intervals ranged from 15 – 48% for diel mobility (Table 1) and 18 – 45% for diel range (Table 2). At 6 h intervals the reduction in accuracy ranged from 52 – 70% for diel mobility (Table 1) and 57 – 75% for diel range (Table 2).

Discussion Although the mode of crayfish movement is different from that used by (i.e., walking versus swimming), the impact of radio-tracking interval on estimates of diel movement, remain similar for both taxa (in that the accuracy of estimates reduces as intervals increase) (Baras 1998; Ovidio et al. 2000; Horton et al. 2004; Natsumeda 2007; the current study – Figures 1 and 2). The dormancy of Individual D, provided an exception in the current study. Stationary phases have been reported for other crayfish, including the white-clawed crayfish, Austropotamobius pallipes (Lereboullet) (Robinson et al. 2000) and the giant Tasmanian crayfish, Astacopsis gouldi (Clark) (Webb and Richardson 2004), and may have an important biological basis (i.e., reproductive state, Figure 2. Reduction in estimates of a) diel mobility (P < 0.05) and b) diel moulting, loss of appetite or response to environmental conditions) range (P < 0.05) of E. armatus (n = 5) when data from radio-tracking at in- (Webb and Richardson 2004). In contrast, extended periods of tervals of 1 h were sub-sampled at intervals of 2 h, 3 h, 4 h and 6 h. Points represent mean diel mobility and diel range of five crayfish (± SE). immobility are uncommon for many fishes and have been used as an indication of mortality or radio-tag rejection in some studies tracking following the current study found one of these crayfish (e.g., Ebner et al. 2007). Visual confirmation of the fate of dormant (Individual D) to be active; however, no movements were recorded aquatic animals is clearly desirable as an independent check in for the other crayfish (Individual A) (Ryan 2005, unpublished data) radio-tracking studies. Innovative solutions to confirming the and hence its data are not reported in the current study. fate of crayfish may be required where radio-tags are located in inaccessible microhabitats and/or in low visibility waters. Crayfish activity occurred during both day and night, with a maximum movement of 49 m recorded in 1 h. The average hourly Our finding of reduced diel mobility estimates with less activity for all crayfish over the three diel periods showed a highly frequent radio-tracking of E. armatus (Table 1 and Figure 2a) are variable pattern with some peaks in activity around sunrise, mid- comparable with similar studies of fish. Radio-tracking at intervals morning, and nightfall (Figure 1a). At less frequent radio-tracking of 14 days produced 56.8% and 82.9% reduction in estimates of intervals, diel activity patterns become progressively concealed mobility from daily data for barbel, B. barbus (Baras 1998) and (Figure 1). The main peaks in activity observed from fixes at brown trout, S. trutta (Ovidio et al. 2000), respectively. On a finer intervals of 1 h are still evident to some extent when data are temporal scale Horton et al. (2004) found up to 64% reduction in sub-sampled to simulate intervals of 2 h and 3 h; however, the accuracy of mobility estimates for spotted bass, M. punctulatus, variability in activity becomes decreasingly pronounced (Figure 1a when reducing the frequency of radio-tracking intervals from versus Figure 1b and c). At intervals of 4 h and 6 h the variability 15 min to 2 h. Thus, accurate mobility estimates are based on in activity is obscured and the activity pattern suggests primarily intensive manual radio-tracking or remote telemetry (Hanson et al. diurnal activity rather than showing peaks in activity at certain 2007) since this variable is a continuous measure of movement times of the day (Figure 1a versus Figure 1d and e). (rather than a summary of total area occupied, as is the case with home-range estimates). Diel range areas varied from 0 – 939 m2 (mean = 275 m2, sd = 296) and often overlapped in the case of: a) each individual among Diel activity of E. armatus showed peak periods and high diel periods; and b) among individuals (Ryan 2005, unpublished variation throughout diel periods in the current study (Figure 1). 2008 Ry a n e t a l . — Ra d i o –Tr a ck i n g In t e r v a l a n d Cr a y f i s h Mo v e m e n t Acc u r a c y 91

While studies in the Northern hemisphere have found freshwater species at coarse time intervals over multiple seasons with a view crayfish to be primarily nocturnal (e.g., Procambarus clarkii to understanding its habitat requirements. (Girard): Gherardi et al. 2000; A. pallipes: Barbaresi and Gherardi 2001), Webb and Richardson (2004) found the giant Tasmanian Acknowledgments crayfish, A. gouldi to be neither predominantly nocturnal nor ACT Parks, Conservation and Lands staff particularly Ben diurnal. However, diel activity patterns of E. armatus appear Broadhurst, Luke Johnston and Jason Thiem assisted with animal diurnal as tracking intervals increase to 4 h and 6 h (Figure 1). collection and radio-tracking. Mark Dunford provided invaluable Therefore, we caution against making comparisons of diel activity GIS support. Rohan Rehwinkel, Chris Levings, Steve Hall, Sarah between species where different sampling intervals have been used. Uglade, Elliot Duncan and Taylor Booth all provided valuable The diel activity patterns shown by E. armatus when radio- assistance in the field. The ACT Government provided a licence to tracked at intervals of 1 h (Figure 1) suggest that primary periods conduct this research. The research was conducted in Ngunnawal of activity are variable for this species. Previous studies of species Country. We thank our two anonymous reviewers who provided with a predictable diel activity pattern have been monitored helpful comments and suggestions on an earlier version of this with manual radio-tracking effort concentrated on periods of manuscript. high activity. For instance, Natsumeda (2007) demonstrated the importance of tracking the Japanese fluvial sculpin, C. pollux at Literature Cited times when it is most active to ensure accuracy in estimates of As m u s M (1999). Effects of recreational fishing on populations home-range. For animals with an unpredictable diel activity pattern, of Murray Crayfish Euastacus armatus in the Murrumbidgee such a study design is not effective and randomised or uniform River. Honours Thesis, Charles Sturt University, Wagga temporal sampling rather than stratified sampling is necessary. The Wagga, Australia. use of continuous remote-telemetry loggers would be ideal under these circumstances since it can yield information at a resolution Ba r a s E (1998). Selection of optimal positioning intervals in of minutes (cf., David and Closs 2001; Hanson et al. 2007). fish tracking: An experimental study on Barbus barbus. Hydrobiologia 371/372:19–28. Our finding of reduction in diel range estimates of E. armatus Ba rb a r e s i S a n d Gh e r a r d i F (2001). Daily activity of the white- in relation to less frequent radio-tracking (Table 2 and Figure 2b) clawed crayfish, Austropotamobius pallipes (Lereboullet): A is a function of underestimating both lateral and longitudinal in- comparison between field and laboratory studies. Journal of stream movement. Accuracy of range estimates has been found Natural History 35(12):1861–1871. to be less dependent on the frequency of radio-tracking intervals for fish that make movements along a linear axis (Baras 1998; Br o a d h u r s t B, Eb n e r B a n d Jo h n s t o n L (2005). Parks, Ovidio et al. 2000). The giant Tasmanian crayfish A. gouldi is Conservation & Lands, Department of Territory & Municipal known to have a large linear range in small (narrow) streams Services, ACT Government, Canberra, Australian Capital (Webb and Richardson 2004) and therefore less intensive radio- Territory, Australia. Unpublished data. tracking may suffice to determine its diel range compared with Bu bb DH, Th o m TJ a n d L u c a s MC (2004). Movement and dispersal that of E. armatus in a large, wide river (e.g., the Murrumbidgee of the invasive signal crayfish Pacifastacus leniusculus in River) where both substantial longitudinal and lateral movement upland rivers. Freshwater Biology 49(3):357–368. is possible. The current study has identified the need for intensive Bu bb DH, Th o m TJ a n d Lu c a s MC (2006). Movement patterns radio-tracking of E. armatus to accurately estimate diel range. For of the invasive signal crayfish determined by PIT telemetry. conservation purposes, including designating closed areas, where Canadian Journal of Zoology 84(8):1202–1209. E. armatus are protected outside of the ACT, we recommend a next By r o n CJ a n d Wi l s o n KA (2001). Rusty crayfish Orconectes ( step of obtaining home-range estimates and understanding habitat rusticus) movement within and between habitats in Trout use of this species over an annual cycle. In the absence of remote Lake, Vilas County, Wisconsin. Journal of the North American telemetry capability (cf., Hanson et al. 2007) we recommend Benthological Society 20(4):606–614. that the home-range and habitat use of E. armatus be based on weekly tracking across all four seasons in a subset of upland and Cr a n d a l l KA (1996). Euastacus armatus. In: IUCN 2007. 2007 lowland rivers. Radio-tracking E. armatus at night appears to be IUCN Red List of Threatened Species. Publ. Internet. http:// unnecessary based on the findings of the current study. www.iucnredlist.org/ [Accessed 01 November 2007]. Da v i d BO a n d Cl o s s GP (2001). Continuous remote monitoring of Understanding the relationship between radio-tracking interval fish activity with restricted home ranges using radiotelemetry. and estimates of diel mobility, diel activity, and diel range provides Journal of Fish Biology 59(3):705–715. a basis for designing future studies of E. armatus behaviour and habitat use. Estimates of diel mobility and diel range of E. armatus Eb n e r B (2005). Parks, conservation and lands, Department of are dependent on how frequently radio-tracking is conducted. Territory & Municipal Services, ACT Government, Canberra, These two variables are likely to be useful for measuring short- Australian Capital Territory, Australia. Unpublished data. term changes in behaviour in response to short-term disturbances Eb n e r BC, Th i e m JD a n d Li n t e rm a n s M (2007). Fate of two-year including environmental flows, to achieve ecological outcomes. old, hatchery-reared Maccullochella macquariensis, stocked However, the conservation of E. armatus may be best served into two upland rivers. Journal of Fish Biology 71(1):182– in the immediate future by collecting spatial information on the 199. 92 Fr e s h w a t e r Cr a y f i s h Vol. 16

Ge d d e s MC (1990). Crayfish. In: The Murray. Mackay N and the Australian Capital Territory. ACT Parks and Conservation Eastburn D (eds.), pp 302–307. Murray Darling Basin Service, Canberra, Australia. Commission, Canberra, Australia. Na t s u m e d a T (2007). Estimates of nocturnal home-range size Gh e r a r d i F, Ba rb a r e s i S a n d Sa l v i G (2000). Spatial and temporal of the adult Japanese fluvial sculpin, Cottus pollux (Pisces: patterns in the movement of Procambarus clarkii, an invasive ), in relation to bottom topography and sampling crayfish. Aquatic Sciences 62(2):179–193. intervals. Journal of Ethology 25(1):87–93. Ha n s o n KC, Co o k e SJ, Su s k i CD, Ni e z g o d a G, Ph e l a n FJ, Ti n l i n e Ov i d i o M, Ph i l l i pp a r t J-C a n d Ba r a s E (2000). Methodological R a n d Ph i l l i pp DP (2007). Assessment of largemouth bass bias in home range and mobility estimates when locating (Micropterus salmoides) behaviour and activity at multiple radio-tagged trout, Salmo trutta, at different time intervals. spatial and temporal scales utilizing a whole-lake telemetry Aquatic Living Resources 13(6):449–454. array. Hydrobiologia 582(1):243–256. Ro b i n s o n CA, Th o m TJ a n d Lu c a s MC (2000). Ranging behaviour Ho o g e PN a n d Ei c h e n l a u b B (1997). Animal movement extension of a large freshwater invertebrate, the white-clawed crayfish to Arcview. Ver. 1.1. Alaska Biological Science Center U.S. Austropotamobius pallipes. Freshwater Biology 44(3):509– Geological Survey, Anchorage, Alaska, USA. 521. Ho r t o n TB, Gu y CS a n d Po n t i u s JS (2004). Influence of time Ry a n K (2005). University of Canberra, Canberra, ACT, Australia. interval on estimations of movement and habitat use. North Unpublished data. American Journal of Fisheries Management 24(2):690–696. We bb M a n d Ri c h a r d s o n A (2004). A radio telemetry study Li n t e rm a n s M a n d Ru t z o u T (1991). The status, distribution and on movement in the giant Tasmanian freshwater crayfish, management of the Murray Crayfish Euastacus armatus in Astacopsis gouldi. Freshwater Crayfish 14:197–204.