Acceleration Data Loggers Research Articles

Data-processing artefacts in three-dimensional dive path reconstruction from geomagnetic and acceleration data

ABSTRACT: Tri-axis magnetism and acceleration data loggers have recently been used to obtaintime-series headings and, consequently, the 3-dimensional dive paths of aquatic animals. However,problems may arise in the resulting calculation process with multiple parameters. In this study, thedive paths of loggerhead turtles and emperor penguins were reconstructed. For both species, appar-ently unrealistic movements were found. Time-series heading data of turtles showed small regularfluctuations synchronous with stroking. In the dive paths of penguins, infrequent abrupt changes inheading were observed during stroke cycles. These were unlikely to represent true behavioursaccording to observations of underwater behaviour and tri-axis magnetism and acceleration data.Based on the relationship between sampling frequency and frequency of body posture change, wesuggest that (1) the changes in the animals’ posture concurrent with strokes and (2) the mismatchedtreatment (i.e. filtering and non-filtering) of the acceleration and magnetism data caused the arte-facts. These inferences are supported by the results of simulations. For data sets obtained at a givensampling frequency, the error pattern in calculated dive paths is likely to differ depending on the fre-quency and amplitude of body posture changes and in swim speed. In order to avoid misinterpreta-tion, it is necessary to understand the assumptions and inherent problems of the calculation methodsas well as the behavioural characteristics of the study animals.KEY WORDS: 3D dive path · Low-pass filter · Sampling frequency · Stroke activity · Data logger ·Dead-reckoning

Accelerating estimates of activity-specific metabolic rate in fishes: Testing the applicability of acceleration data-loggers

We report the results of a series of experiments to test the utility of acceleration data-loggers for determining the energy expenditure of juvenile hammerhead sharks ( Sphyrna lewini). In one experiment, three sharks were instrumented with miniature acceleration data-loggers and swum in a Brett-style respirometer at a range of speeds. For all three sharks, significant linear relationships were obtained between mean oxygen consumption (M ·o 2) and Partial Dynamic Body Acceleration in the lateral and dorso-ventral axes (PDBA y, z ) with high predictive power ( r 2 > 0.71). In a second experiment, PDBA was measured for sharks swimming freely in circular tanks. The free-swimming sharks exhibited wide ranges of PDBA y, z ; routine swimming was characterised by low PDBA y, z (0.01–0.12 g) whereas unsteady swimming, (especially fast-start swimming of > 1 g) was characterised by high PDBA y, z . Despite initial evidence of linearity in the oxygen consumption vs. PDBA relationship, incorporating previous estimates of standard metabolic rate of hammerhead sharks suggests a non-linear fit. Further work is needed to establish the exact shape of the relationship beyond the narrow range of speeds that hammerhead pups could be exercised in this study, particularly the low swimming speeds which are frequently observed in free-swimming animals.

Monitoring beak movements with an acceleration datalogger: a useful technique for assessing the feeding and breathing behaviors of sea turtles

This study was performed to determine whether the attachment of acceleration data- loggers to the lower beaks of loggerhead turtles Caretta caretta could be a useful technique for mon- itoring their feeding and breathing behaviors. Attaching acceleration dataloggers to the lower beak of turtles allows determination of the pitch of the head from the low frequency component of the acceleration data, and of dynamic movements (e.g. biting) from the high frequency component. In addition, to determine whether the acceleration datalogger could distinguish between different food sources and feeding locations based on acceleration characteristics, we fed the turtles with different types of food (squid rings, fins, and heads that include arms and tentacles) across different locations. Our results demonstrate that the acceleration datalogger was able to detect the lower beak move- ments of loggerhead turtles, which enabled detection of 99.6 ± 1.1 (SD) % of feeding and 100% of breathing behaviors, with respective false detection rates of 24.8 ± 12.4% and 2.4%. Furthermore, our results demonstrate that it is possible to (1) determine whether feeding on prey requires a strong biting force, and (2) differentiate between feeding at the sea floor and in the water column. Attach- ing an acceleration datalogger is thus established as a useful technique for monitoring the feeding and breathing behaviors of sea turtles. Future studies employing acceleration dataloggers should provide new insights into the biology of sea turtles and their feeding and diving strategies.

Validation of a device for accurate timing of feeding events in marine animals

Two acceleration data loggers, each measuring surging and heaving acceleration, were attached to the head and mandible of three captive hooded seals, Cystophora cristata, for detection of underwater feeding events. Three sizes of prey: Atlantic herring (large), capelin (medium), and half a capelin (small) were tested. A highpass frequency-filtering method at 3 Hz provided more distinct prey ingestion signals for both head and mandible acceleration. The surge-axis signals from head acceleration suggested that the seals ingested their prey not only by biting, but also by thrusting. Moreover, prey ingestion movements showed higher surging acceleration from mandible than from head (mean ± SD from head: 5.37 ± 4.45 m s−2, from mandible: 8.43 ± 5.15 m s−2, n = 153), indicating that the data from the head is not required for precise identification of feeding events. Thus, our mandible acceleration device provides a practical method for the timing of underwater feeding events in seals.

Thick-billed murres use different diving behaviors in mixed and stratified waters

Abstract Linking diving and foraging behavior of small seabirds with the fine-scale characteristics of water masses has been challenging largely due to sampling constraints. We examined the diving behavior of 12 chick-rearing thick-billed murres ( Uria lomvia ) at St. George Island, southeastern Bering Sea, in relation to sea-surface temperature (SST) and thermocline depth using ventrally attached depth–temperature–acceleration data loggers. Our results from summer 2004 showed that murres swam in water masses ranging from well-mixed (SST 7–9 ° C, estimated distance of 14 km from the breeding colony) to well-stratified (SST 9–12 °C, estimated distance of 30–50 km). Murres dove deeper (modal depth: 60–70 m) in the mixed water mass than in the stratified water, where most dives were to just below the thermocline depth (modal depth: 20–30 m). We suggest that the thermocline is important in shaping dive profiles of thick-billed murres, possibly through its effect on the vertical distribution of both zooplankton and fish prey.

Plunging Behaviour in Chick-rearing Brown Boobies

ABSTRACT Brown Boobies (Sula leucogaster) obtain prey by plunge dives. In order to investigate the foraging behaviour of Brown Boobies underwater, we attached acceleration data loggers to chick-rearing Brown Boobies at Nakanokamishima Island in Japan. We documented, for the first time, Brown Boobies performing many rapid and shallow V-shaped dives and some W-shaped dives during the daylight period. The average and maximum dive depth and duration were 1.03 m and 3.81 m and 1.83 s and 21 s, respectively. Furthermore, Brown Boobies used buoyancy to ascend to the water surface. Our data suggest that Brown Boobies mainly depend on shallow-plunging, contrary to the pursuit divers such as gannets, due to their ecological and/or physiological constraints.

Use of an acceleration data logger to measure diel activity patterns in captive whitetip reef sharks,Triaenodon obesus

Traditional telemetry methods have been used to quantify the horizontal and vertical displacement of marine species, but are unable to identify specific physical activities such as swimming or gliding, resting, foraging, or spawning. We tested the utility of an acceleration data logger to quantify activity patterns of three captive whitetip reef sharks (Triaenodon obesus ) in an enclosed lagoon using internal and external attachment methods. Data obtained using both attachment methods allowed swimming and resting behavior to be differentiated. All sharks showed constant swimming for 5–14 hours post-tagging before adopting a pattern of daytime rest and nocturnal activity throughout the 6–16 day deployments. Sharks showed a diel activity pattern, spending 10–24% of their time swimming during the day, and 42–67% swimming at night. Overall, sharks spent an average of 35 ± 11% (mean ± SD) of their time swimming. Mean tailbeat frequency was found to be 0.89 ± 0.03 beats s−1 in one shark for which it was measured. Respirometry experiments that measured the metabolic rate of two neonate whitetips showed significantly higher metabolic rates at night compared to the day. When taken in conjunction with the acceleration data, these results suggest that whitetips are nocturnally active and show diel circadian rhythms shortly after birth. Our study demonstrates that acceleration data loggers can be used to quantify activity patterns and offer promise for quantifying energy budgets of various reef sharks both in captivity and in the field.

Body density affects stroke patterns in Baikal seals

Buoyancy is one of the primary external forces acting on air-breathing divers and it can affect their swimming energetics. Because the body composition of marine mammals (i.e. the relative amounts of lower-density lipid and higher-density lean tissue) varies individually and seasonally, their buoyancy also fluctuates widely, and individuals would be expected to adjust their stroke patterns during dives accordingly. To test this prediction, we attached acceleration data loggers to four free-ranging Baikal seals Phoca sibirica in Lake Baikal and monitored flipper stroking activity as well as swimming speed, depth and inclination of the body axis (pitch). In addition to the logger, one seal (Individual 4) was equipped with a lead weight that was jettisoned after a predetermined time period so that we had a set of observations on the same individual with different body densities. These four data sets revealed the general diving patterns of Baikal seals and also provided direct insights into the influence of buoyancy on these patterns. Seals repeatedly performed dives of a mean duration of 7.0 min (max. 15.4 min), interrupted by a mean surface duration of 1.2 min. Dive depths were 66 m on average, but varied substantially, with a maximum depth of 324 m. The seals showed different stroke patterns among individuals; some seals stroked at lower rates during descent than ascent, while the others had higher stroke rates during descent than ascent. When the lead weight was detached from Individual 4, the seal increased its stroke rate in descent by shifting swimming mode from prolonged glides to more stroke-and-glide swimming, and decreased its stroke rate in ascent by shifting from continuous stroking to stroke-and-glide swimming. We conclude that seals adopt different stroke patterns according to their individual buoyancies. We also demonstrate that the terminal speed reached by Individual 4 during prolonged glide in descent depended on its total buoyancy and pitch, with higher speeds reached in the weighted condition and at steeper pitch. A simple physical model allowed us to estimate the body density of the seal from the speed and pitch (1,027-1,046 kg m(-3), roughly corresponding to 32-41% lipid content, for the weighted condition; 1,014-1,022 kg m(-3), 43-47% lipid content, for the unweighted condition).

A new technique for monitoring the detailed behaviour of terrestrial animals: A case study with the domestic cat

For many animal species that are difficult to access, the behaviour of free-ranging individuals cannot be assessed by direct observation. In order to remedy this, we developed a new technique using a motion detector (acceleration data-logger) for monitoring the activity and behaviour of free-ranging vertebrates and tested its efficiency on a domestic cat, Felis catus. A total of 3615 min of surging acceleration was measured along the longitudinal body axis of an adult male cat. The cat's behaviour was also filmed for 113 min, these video data being used to correlate the logger's signals with the cat's behaviour. Acceleration data-loggers attached on the cat's collar recorded acceleration signals which were influenced by both the gravitational acceleration resulting from the body posture and the dynamic acceleration resulting from the dynamic behaviour of the cat. By applying spectral analysis based on a fast Fourier Transform to acceleration signals, body postures and some of the dynamic behaviours of the cat such as drinking, eating, and several paces of travelling were efficiently determined. The present study shows that acceleration data-loggers represent a useful and reliable system for accurately recording the activities and detail behaviours of the terrestrial animals.

Why do macaroni penguins choose shallow body angles that result in longer descent and ascent durations?

It is generally assumed that air-breathing aquatic animals always choose the shortest route to minimize duration for transit between the surface and foraging depth in order to maximize the proportion of time spent foraging. However, empirical data indicate that the body angles of some diving animals are rarely vertical during descent and ascent. Why do they choose shallower body angles that result in longer descent and ascent durations? To investigate this question, we attached acceleration data loggers to eight female macaroni penguins, breeding on the Kerguelen Islands (48 degrees 45'-50 degrees 00' S, 68 degrees 45'-70 degrees 58' E; South Indian Ocean), to record depth, two-dimensional acceleration (stroke cycle frequency and body angle) and temperature. We investigated how they controlled body angle and allocated their submerged time. The instrumented females performed multiple dives (N=6952) with a mean dive depth for each bird ranging from 24.5+/-28.5 m to 56.4+/-75.1 m. Mean body angles during descent and ascent were not vertical. There was large variation in mean descent and ascent angles for a given dive depth, which, in turn, caused large variation in descent and ascent duration. Body angles were significantly correlated with time spent at the bottom-phase of the dive. Birds that spent long periods at the bottom exhibited steep body angles during ascent and subsequent descent. By contrast, they adopted shallow body angles after they had short or no bottom phases. Our results suggest that macaroni penguins stay at the bottom longer after encountering a good prey patch and then travel to the surface at steep body angles. If they do not encounter prey, they discontinue the dive, without staying at the bottom, ascend at shallow body angles and descend at shallow body angles in a subsequent dive. A shallow body angle can increase the horizontal distance covered during a dive, contributing to the move into a more profitable area in the following dive. During the ascent, in particular, macaroni penguins stopped beating their flippers. The buoyantly gliding penguins can move horizontally with minimum stroking effort before reaching the surface.

Simultaneous measurement of swimming speed and tail beat activity of free-swimming rainbow trout Oncorhynchus mykiss using an acceleration data-logger

A recently developed motion detector (acceleration data-logger), based on accelera- tion measurements, was used to monitor the swimming behavior of two free-swimming captive rainbow trout Oncorhynchus mykiss in an aquaculture net cage. Depth, swimming speeds and two- direction acceleration data were collected continuously for approximately 20 h per fish. To define relationships between swaying acceleration profiles and tail beat activity of rainbow trout, the tail beat activity of trout was monitored using a video camera in a small tank with the simultaneous use of the acceleration data-logger. During steady swimming, there were sharp, distinct peaks of swaying acceleration. When compared with the data from the video camera, the frequency of swaying acceleration was synchronized with the sequences of swimming activity. In addition, the tail beat frequency of the fish could be identified within the cycle of swaying acceleration during steady swimming phases. Mean tail beat frequency was 1.27 ± 0.40 and 1.40 ± 0.5 Hz (mean ± SD). Using the relationship between tail beat frequency and swimming speed, the 'preferred' swimming speed of trout was estimated to be between 0.48 and 0.58 body length (BL)/s, and trout rarely swam in excess of 2.0 BL/s. The present study shows that acceleration data-loggers used to record sponta- neous measurements of swimming speed and tail beat activity represent a useful and reliable system for accurately estimating both the rate of activities and movements of free-ranging fish.

A new device for monitoring the activity of freely swimming flatfish, Japanese flounder Paralichthys olivaceus

: The tail beat and activity behavior of four captive Japanese flounder Paralichthys olivaceus, were monitored with acceleration data-loggers while the fish swam in an aquarium. Depth, swimming speeds and two-axis acceleration data were collected continuously for approximately 20 h per fish. Simultaneously, the swimming behaviors of the fish were filmed at different angles. Using the specific characteristic of the acceleration profiles, in tandem with other types of data (e.g. speed and depth), four behavioral patterns could be distinguished: (i) ‘active’ swimming; (ii) burying patterns; (iii) ‘inactive’ gliding; and (iv) lying on the bottom. Tail beat frequency ranged from 1.65 ± 0.47 to 2.04 ± 0.25 Hz (mean ± SD; n = 4). Using the relationship between tail beat frequency and swimming speed, the ‘preferred’ swimming speed of the fish was estimated to be between 0.6 and 1.2 body lengths (BL)/s. Additionally, fish rarely swam faster than 1.2 BL/s. This study shows that the acceleration data-loggers represent a useful and reliable system for accurately recording the tail beat of free-ranging fish and estimating flatfish behavior.

On the Use of the Time Axis for Ecological Separation: Diel Rhythms as an Evolutionary Constraint

On the Use of the Time Axis for Ecological Separation: Diel Rhythms as an Evolutionary Constraint

Precise monitoring of porpoising behaviour of Adélie penguins determined using acceleration data loggers.

A new method using acceleration data loggers enabled us to measure the porpoising behaviour of Adélie penguins (Pygoscelis adeliae), defined as a continuous rapid swimming with rhythmic serial leaps. Previous hydrodynamic models suggested that leaping would be energetically cheaper when an animal swims continuously at depths of less than three maximum body diameters below the water surface. In the present study, free-ranging Adélie penguins leapt at a mean speed of 2.8 m s(-)(1) above the predicted threshold speed (0.18-1. 88 m s(-)(1)). Wild penguins reduced drag by swimming deeper (0.91 m) and did not swim continuously within the high-drag layer while submerged. This indicates that previous calculations may be incomplete. Moreover, leaps represented an average of only 3.8 % of the total distance travelled during the porpoising cycle, which would make energy savings marginal. Among the six penguins used in our study, two did not porpoise and three porpoised for less than 7 min, also indicating that this behaviour was not important during travel to and from foraging sites, as has been previously suggested. Birds mainly porpoised at the start and end of a trip. One explanation of porpoising might be an escape behaviour from predators.