healthmedicinet

ITAG design overview

The tag design leverages modularity to enable the ITAG to be deployed across taxa.
Specifically, the ITAG consists of two components (Fig. 1): (a) a buoyant waterproof eco-sensor package that records behavioral and environmental
data, and (b) a base that serves as the point of attachment to the animal. The tag
electronics are a universal component, while the base and attachment method are animal
specific. This modular format allows the same tag electronics to be used with a range
of animals with varying morphologies.

Fig. 1. Invertebrate tag (or ITAG) mounting locations and methods. a On large squid (using sutures) and b jellyfish (using suction cups). Positions of sensors, connectors, antennae, and bases
on the ITAG are also indicated. Dimensions are displayed in units of millimeters

The tag was designed with the shape, behavior, and propulsive modes of squid and jellyfish
in mind. Squids swim utilizing high-acceleration jet propulsion or by rowing of their
fins 29], and their hydrodynamically shaped bodies maximize propulsive efficiency by reducing
drag. Large jellyfish predominantly utilize rowing propulsion 30], which is characterized by lower acceleration rates when compared to jet propulsion
31]. In addition, large jellyfish possess bluff, less hydrodynamically streamlined body
shapes that are more optimized for efficient swimming 30], 31]. On squid, the ITAG is attached to the dorsal surface of the squid rostrum, and positioned
in line with the squid body to reduce drag. On jellyfish, the tag is attached to the
dorsal surface of the bell. In this configuration, the low height and minimal aspect
ratio of the tag reduces flow separation along the surface of the bell (Fig. 1a, b).

ITAG electronics module

The ITAG electronics are based on the marine mammal acoustic behavior DTAG 32]. The governing constraints for the ITAG design were that it should be small, lightweight,
neutrally buoyant in seawater, pressure tolerant, and capable of ca. 24 h deployments,
with the consideration that deployment durations could be increased in the future.
The result is a modular 108.4 mm × 64.0 mm × 28.7 mm package (Fig. 1), with a Li-ion rechargeable battery, external sensors, and a radio antenna embedded
in a syntactic foam frame that is sealed in urethane rubber. To avoid a bulky and
heavy pressure housing, the ITAG circuit boards are oil-filled with a flexible bladder.
The electronics are designed around a low-power programmable digital signal processor
(DSP), which combines data from the sensors and stores that information on a memory
array. Use of a DSP enables real-time filtering and loss-less compression of signal
streams when required. The ITAG currently has 3-axis accelerometer (KXSC7-1050, Kionix,
NY, USA), 3-axis magnetometer (HMC1043LMC1021 Honeywell, MA, USA), pressure (Series
1 TAB, Keller, VA, USA) temperature (USP10975, U.S. Sensor Corp, CA, USA), and light
sensors (PDV-P8104, Advanced Photonix, Inc, MI, USA). Orientation, which is parameterized
by the Euler angles pitch, roll, and heading, requires two sets of sensors; pitch
and roll are calculated using data from the accelerometers, and heading is calculated
from the magnetometer. To improve the heading estimate, the three magnetometer signals
are corrected for pitch and roll. This process, called gimballing, effectively transforms
the magnetic-field measurement to that which would be made on a horizontal surface
with the same heading. ITAG power consumption is approximately 50 mW while recording,
and the operating duration is currently limited by battery size. As configured, the
tag can collect high-resolution behavior and environmental data for deployments greater
than 30 h. An embedded VHF beacon is then used to locate the tag after release and
surfacing and can operate for over 7 days following the end of the recording period
(Table 1).

Table 1. Specifications of the ITAG and some comparison tools

ITAG base module

Squid musculature lends itself to relatively invasive tag attachment mechanisms using
needle pins to puncture the mantle or fin 33]. If a tag is mounted internally, there is high risk of damaging internal organs;
certain external attachment may impede swimming by minimizing fin motion or cause
abrasion to the skin 34], 35]. These mounting issues are exacerbated by a large tag package. Using the first generation
ITAG and VEMCO dummy tags (V16, 98 mm, 16 g, NS, Canada), the locations and attachment
mechanisms for captive squid Loligo forbesi were evaluated during preliminary trials conducted in March 2013 (VEMCO) and April
2014 (ITAG) at the University of the Azores in Horta, Portugal (Fig. 2). Mounting a hydrodynamic tag externally close to the posterior tip of the mantle
reduced internal-damage and swim-impedance issues. The presence of dense muscle and
cartilage at the mantle ensured secure suture attachment and reduced external tissue
damage. VEMCO tags were attached for up to 1 week on three large squid (mantle lengths
45, 39, and 47 cm) without any impairment of swimming movement or causing any obvious
serious tissue injury in the short term. The ITAG squid base utilizes a silicone pad,
ca. 2 mm in thickness, as a compliant interface between the tag electronics and the
animal to minimize tissue abrasion. A rigid plastic spine is embedded in the silicone
to provide the point of attachment to the tag electronics module. Additionally, the
silicone pad has four pairs of 1 mm diameter holes at each end that allow for easy
attachment via absorbable sutures (catgut, 3:0) to the dorsal surface of the animal
near the rostrum (Fig. 1a).

Fig. 2. Images of attachment and experimental setup. The invertebrate tag (or ITAG) was attached
to a captive jellyfish Aurelia aurita and b squid Loligo forbesi in laboratories in Woods Hole, MA and Horta, Azores, Portugal

The delicate nature of jellyfish tissue is particularly challenging for tag attachment.
Consequently, tags are limited in size and capability 36]. Jellyfish tagging has had significant success in very large jellyfish with oral
arms where tags can be attached using plastic cable-ties 25], 26]. Unfortunately, tag placement on the oral arms yields noisy acceleration data due
to fluid-induced motion of the tag when interacting with feeding currents and propulsive
jets, and cannot be used for many species where such morphological features are absent.
Attachment of tags on the dorsal surface of the jellyfish bell can facilitate acceleration
measurements that yield information about swimming cycles and energetics. However,
pins and sutures easily tear gelatinous tissue, and result in short attachment periods.
As such, the jellyfish ITAG base is a solid urethane platform with a rigid plastic
spine and four recessed suction cups (Fig. 1b). The gentle suction provided by the cups on the apex of the bell allows for secure
attachment, and the plastic spine connects the tag to the base.

In both the squid and jellyfish configurations, two corrodible, nickel–chromium wires
are used to couple the electronics module to the base module. The tag can be programmed
to send an electric current through these wires to activate fast corrosion and release
after a specified recording duration (e.g., 1.5, 8, 24 h). When the tag and base are
coupled, the ITAG is neutrally buoyant. Once the tag releases from the base, the positively
buoyant ITAG floats to the surface holding a vertical position, allowing the VHF radio
antenna to break the surface of the water and transmit its location.

Jellyfish trials

Laboratory investigations of attachment methods and deployments of the ITAG on jellyfish
were conducted between August and November 2013 at the Marine Biological Laboratory
and the Environmental Systems Laboratory in Woods Hole, MA. Initial tests of attachment
methods were conducted on captive Cyanea capillata and Chrysaora quinquecirrha (collected in Vineyard Sound). ITAG deployments were conducted in a 182.8 cm diameter
tank (1 m depth) on five Aurelia aurita (provided by the New England Aquarium) of approximately 17–22 cm bell diameter (Fig. 2a). Due to the limited bell size, neutrally buoyant ITAGs were affixed to five jellyfish
in water using two of the four suction cups on the dorsal surface of the jellyfish
bell near the apex, similar to an approach used in box jellyfish (Chironex fleckeri) 37]. Using the tags and simultaneous video recordings (Canon 7D, Japan) of attachment,
release, and swimming behavior, observations of jellyfish swimming behavior were correlated
with the accelerometer outputs. Data were later analyzed using custom Matlab algorithms.

Squid trials

Squid ITAG trials were conducted between March 15 and April 1, 2014 at the Porto Pim
Aquarium, a public facility run by Flying Sharks in Faial Island, Azores, Portugal (Fig. 2b). Squid (L. forbesi) were fished at the island slope (200 m) by hand jigging. Animals were delivered
to the facility via boat in coolers filled with running seawater. After capture and
transport, animals were placed in a 4 m × 8 m oval holding tank of 4 m depth containing
47 tons of filtered seawater, where they rested between 2 and 24 h before being recaptured
(using soft hand-nets) for tag attachment. Water in the facility was pumped in from
the adjacent Atlantic and maintained at 17.7 ± 0.9 °C. A series of initial mechanical
(centrifugal), pressurized sand filters, tower protein skimmers and UV filtration
were used to filter ca. both the input and tank water which aerated the water and
allowed for ca 77.7 % of water to be filtered per hour. Salinity and pH were maintained
at 35.55 ppt and 7.86, respectively. ITAGs were deployed on eight individual squid
with an average mantle length of 58 cm (52 cm min; 69 cm max). Animals were taken
out of the holding tank, placed on a v-shaped padded tagging table, and their gills
were ventilated with flowing seawater during tagging. The tag base was sutured near
(5–10 cm from the end) the posterior tip of the squid dorsal mantle with the tag affixed,
and animals were given a 15 min acclimatization period in a small soft and black plastic
raft before being released back into the holding tank. Animals were out of the water
for a mean 08:49 mm:ss (06:15 mm:ss min; to 11:00 ms:ss max). There were no obvious
changes in swimming behavior due to tag attachment.

After tag attachment and a subsequent acclimatization period, animals were released
back into the holding tank with conspecifics, tope sharks (Galeorhinus galeus), and smaller “prey” fish. Immediately after release into the holding tank, tagged
squid schooled with conspecifics, exhibiting normal coloration patterns and body positioning.
Other conspecifics did not show aggression or additional interest toward the tagged
animals. Mean durations of ITAG recordings were 20:43:51 (hh:mm:ss) but up to 24:28:49.
Tags separated from the base only as result of the timed release mechanism (i.e.,
no tags released prematurely) and thus attachments could have been extended.

In addition to tag data, three overhead, high-definition video cameras (GoPro, CA,
USA) and two sideways high-definition video cameras (Sony HDR-XR550, Tokyo, Japan)
located at the ends of the holding tank concurrently recorded specific behaviors of
tagged and untagged animals. Video cameras were synchronized with each other using
successive flashes of an external camera flash (Canon, NY, USA), and tag data were
synchronized by recording the arming of the tag with one of the five video cameras.
Using the tags and simultaneous video recordings of tag attachment, animal release
and swimming behavior, it was possible to coordinate camera-observed behaviors such
as forward and backward finning, direction reversal, jetting, and gradual, lateral
turns with ITAG sensor data. Finning was defined as movement where fins were predominantly
moving and strong mantle contractions were not obvious. During jetting, fins were
typically held near the ventral surface of the mantle or not extensively used. Camera
recordings were limited to ca. 90 min due to the memory limitations of the cameras.
Data were later analyzed using custom Matlab algorithms.

Additionally, we used the acceleration data to quantify the animal’s overall activity
level using one data set for jellyfish and one data set for squid to describe initial
outcomes and sensor analyses. Because the goal of the paper was describing the tag
design, development and initial outcomes, sensor analyses are presented primarily
from one animal. A biological assessment of behavioral trends of captive organisms
compiled from all eight tags deployments is currently underway and not included here.
Data were first separated into periods of light and dark using light sensor data.
Next root mean square (RMS) values for the absolute value of the acceleration magnitude
were calculated on an hourly basis. These hourly values of acceleration RMS were averaged
to produce a mean and standard deviation to represent the overall activity of the
squid during periods of light and dark. Only data when the squid was free of experimental
manipulation was used in this overall activity analysis. The mean, standard deviation,
and hourly rate of movements that resulted in peak acceleration rates exceeding 0.15 g
were also used to compare the relative activity of the animal during the tagging.
This threshold was selected after reviewing the tag and video data.