**3.3 Passive tracking studies**

Four experiments using VR2 receiver were performed in Kromme Bay during the November 2003–2006 squid fishery closed seasons. In addition to the VR2 receiver arrays, the VRAP system was deployed in November 2005 and 2006.

#### **3.3.1 VR2 study**

426 Modern Telemetry

database for playback and analysis at a later date (Figure 4). A number of studies have shown the VRAP system to calculate transmitter position with an accuracy of 1 to 3 m (Bégout Anras et al., 1999; Klimeley et al., 2001; Zamora & Moreno-Amich, 2002 as cited in Jadot et al., 2006; Aitken et al., 2005), within the buoy triangle, with accuracy decreasing

Fig. 3. One of the three VRAP buoys deployed in Kromme Bay

equilateral triangular formation

Fig. 4. A single animal track, recorded by the VRAP Buoys, and played back using VRAP software. The smaller triangles in the diagram denote the position of the buoys in the

outside of the array.

Each year researchers searched for an active spawning aggregation. Diver observations confirmed the presence of egg beds, the footprint of these aggregations. VR2 receivers were then deployed 500 m apart, in a hexagonal array, on and around these egg beds. Initial range tests showed the receiving range of the VR2 receivers to be <500 m in Kromme Bay. It was therefore decided to deploy receivers 500 m apart to allow for an overlap in receiving ranges. In 2004, an additional VR2 receiver was deployed on a spawning site off Cape St Francis. The position of these arrays can be seen in Figure 5. Depending on the thermal conditions of the water column (Singh et al., 2009) the hexagonal configuration allowed an area of up to 1.28 km2 to be monitored. Each receiver was deployed 5 m above the seabed using a hollow-core polypropylene rope tensioned with a subsurface buoy. The mooring was anchored to the seabed with a 50 kg weight. During each study temperature data were collected using an array of Star-oddi Starmon mini underwater temperature recorders deployed at depths of 9, 14, 18, 21, and 24 m. This thermistor array (Figure 5) recorded temperature hourly. Hourly wind data, recorded at Port Elizabeth (Figure 1) airport, for 2003-2006 were obtained from the South African Weather Services. Wind data were filtered using an UNH Lanczos filter (weighted 73), and stick vector plots generated.

Fig. 5. The positions of the hexagonal VR2 receiver arrays (2003–2006) and the thermistor array overlaid on the bathymetry (contour lines).

#### **3.3.2 VRAP study**

VRAP buoys were deployed in the centre of the VR2 receiver arrays (Figure 6) in a 300 m equilateral triangle. This configuration allowed for optimal buoy performance. Each buoy was anchored to the seabed with two 50 kg weights. The hydrophone cable was run down the hollow-core polypropylene rope used to attach the buoy to the weights. The omnidirectional hydrophone was positioned approximately 5 m above the seabed.

The Use of Acoustic Telemetry in South African Squid Research (2003-2010) 429

Fig. 7. A chokka squid, *Loligo reynaudii*, caught on a jig

needles

Fig. 8. Tagging instrumentation (taken directly from Downey et al. (2010)): (a) the

attachment of hypodermic needles to an acoustic transmitter, (b) the specially designed tag applicator used to tag *L. reynaudii*, and (c) the placement of the acoustic transmitter within the mantle of the squid, on the ventral side, to avoid piercing organs with the hypodermic

Fig. 6. The positions of the triangular VRAP arrays (2005 & 2006) within the VR2 receiver arrays.

#### **3.3.3 Transmitter attachment**

A total of 45 squid and eight predators were tagged over the four experiments. The predators tagged included three ragged tooth sharks (*Carcharias taurus*), three shorttail stingrays (*Dasyatis brevicaudata*) and two smooth hound sharks (*Mustelus mustelus*). Details of the acoustic transmitters used are given in Table 1. For those animals that were tagged with transmitters without pressure sensors, only presence-absence data were collected. Transmitters with pressure sensors provided both depth and presence-absence data.


Table 1. Details of acoustic transmitters used in the VR2 and VRAP studies

Squid were caught, using jigs (Figure 7), and tagged with V9 acoustic transmitters (Figure 8a). The modification of transmitters for attachment and the tagging process have been described in detail in Downey et al. (2010). Two-18-guage hypodermic needles were glued to the surface of each transmitter, to allow for attachment to the squid (Figure 8a). The length of the needles was dependent on the sex and size of the animal tagged. Hypodermic needles with a length of 17 mm were used for males and needles with a length of 14 mm for the smaller "sneaker" males and females. Each year squid were caught within the hexagonal array of VR2 receivers. Once the animals were removed from the water and their sex determined they were placed on a damp cloth (Figure 9a). Using an applicator specifically designed for this purpose (Figure 8b), a transmitter with the appropriate needles length was inserted into the mantle cavity (Figure 9a). A protective sheath covered the hypodermic needles during insertion (Figure 8b).

Fig. 6. The positions of the triangular VRAP arrays (2005 & 2006) within the VR2 receiver

A total of 45 squid and eight predators were tagged over the four experiments. The predators tagged included three ragged tooth sharks (*Carcharias taurus*), three shorttail stingrays (*Dasyatis brevicaudata*) and two smooth hound sharks (*Mustelus mustelus*). Details of the acoustic transmitters used are given in Table 1. For those animals that were tagged with transmitters without pressure sensors, only presence-absence data were collected.

> **Pressure sensor**

V9P-6L-S256 30 90 Yes 12 (*L. reynaudii*) 6 6 V16-5H-R04K 35 109 No 3 (*C. taurus*) Unknown V16-5H-R04K 35 109 No 1 (*D. brevicaudata*) Unknown V16-5H-R04K 35 109 No 1 (*M. mustelus*) Unknown

V9P-6L-S256 30 90 Yes 23 (*L. reynaudii*) 13 10 V9P-2H-S256 20 60 Yes 1 (*D. brevicaudata*) 1 V9P-2H-S256 20 60 Yes 1 (*M. mustelus*) 1

**Number of** 

**animals tagged Male Female** 

Transmitters with pressure sensors provided both depth and presence-absence data.

2003 V8SC-2H-R256 10 35 No 4 (*L. reynaudii*) 2 2

<sup>2006</sup>V9P-6L-S256 30 90 Yes 6 (*L. reynaudii*) 4 2 V9P-2H-S256 20 60 Yes 1 (*D. brevicaudata*) 1

Squid were caught, using jigs (Figure 7), and tagged with V9 acoustic transmitters (Figure 8a). The modification of transmitters for attachment and the tagging process have been described in detail in Downey et al. (2010). Two-18-guage hypodermic needles were glued to the surface of each transmitter, to allow for attachment to the squid (Figure 8a). The length of the needles was dependent on the sex and size of the animal tagged. Hypodermic needles with a length of 17 mm were used for males and needles with a length of 14 mm for the smaller "sneaker" males and females. Each year squid were caught within the hexagonal array of VR2 receivers. Once the animals were removed from the water and their sex determined they were placed on a damp cloth (Figure 9a). Using an applicator specifically designed for this purpose (Figure 8b), a transmitter with the appropriate needles length was inserted into the mantle cavity (Figure 9a). A protective sheath covered the hypodermic needles during insertion (Figure 8b).

Table 1. Details of acoustic transmitters used in the VR2 and VRAP studies

**Max offtime (s)** 

**Min offtime (s)**

arrays.

2004

2005

**3.3.3 Transmitter attachment** 

**Year Transmitter type** 

Fig. 7. A chokka squid, *Loligo reynaudii*, caught on a jig

Fig. 8. Tagging instrumentation (taken directly from Downey et al. (2010)): (a) the attachment of hypodermic needles to an acoustic transmitter, (b) the specially designed tag applicator used to tag *L. reynaudii*, and (c) the placement of the acoustic transmitter within the mantle of the squid, on the ventral side, to avoid piercing organs with the hypodermic needles

The Use of Acoustic Telemetry in South African Squid Research (2003-2010) 431

transmitter detections at the spawning site, bottom temperature, and wind data against date and time. To determine significant differences in mean depth by day vs. night for male, female, and all squid combined, as well as mean depth for males vs. females by day and night, duplicate data, i.e. single detections recorded by more than one VR2 receiver, were removed and the total number of successfully detected transmissions for each sex per day and night calculated. The data for each sex were separated into depth categories, and the percentage of detections recorded in each depth category by day and night plotted. Twosample, two-tailed t-tests were used to identify significant differences. To analyse diurnal patterns at the spawning sites, the percentage of transmissions successfully detected per hour in a typical 24-h period were plotted, separately for males and females, using the data from which duplicates had been removed. The plots generated and the results of this

Fig. 10. A V16 pinger with a stainless steel trace attached to allow for external attachment. The analysis of the VR2 data showed three general presence–absence behaviours to be found at chokka squid spawning sites (Downey et al., 2010). They are, as given in Downey et al., (2010): (i) arrival at dawn and departure after dusk, (ii) a continuous and uninterrupted presence for a number of days, and (iii) a presence interrupted by frequent but short periods of absence. These authors also concluded that , in contrast to the findings of earlier studies, a core aggregation of squid occasionally remains on active spawning sites at night. At dawn, more squid arrive at the spawning site and the size of the aggregation increases, resulting in a dense aggregation by day. Shortly after dusk, spawning pairs break apart, and some squid leave the spawning site. Those squid remaining at a spawning site at night search for prey throughout the water column and in the benthos, whereas lone females deposit egg strands. The authors also found that movement between the spawning sites continues at night. Their VR2 study confirmed previous observations that the initial formation of spawning aggregations, before the deposition of the first egg strand, is

To investigate presence-absence of predators on the monitored spawning sites, the VR2 data was analysed per year. Signal detections from all tagged squid (grouped), the tagged

analysis are given in Downey et al., (2010).

triggered by upwelling.

The applicator was initially held sideways and once inserted was turned 90° and the protective sheath removed (Figure 8b). After pushing the hypodermic needles through the mantle (Figure 9b), nylon washers were pushed onto the ends of the needles (Figures 8c and 9c) followed by copper crimps (Figures 8c and 9d and e). The tagged squid was then placed in a bin containing seawater or held alongside the boat (Figure 9f), depending on sea conditions, to recover. Once normal fin-beating had resumed, the animal was released within the array of VR2 receivers.

Fig. 9. Attaching a transmitter to a squid (taken directly from Downey et al. (2010)): (a) a transmitter is inserted beneath the mantle using the applicator; (b) the apparatus is turned through 90°, the protective applicator sheath removed, and the hypodermic needles pushed through the mantle. (c) Nylon washers are pushed onto the ends of the hypodermic needles and (d) a metal cylinder slipped over each hypodermic needle, (e) the metal cylinders are crimped using long-nose pliers, and (f) the squid are held submerged alongside the boat until strong swimming ability is displayed (fin beating). Only then is the animal released on the capture site

Predators were tagged with V16 pingers (2004) and V9 sensor acoustic transmitters (2005 & 2006). The transmitters were modified for attachment by gluing a stainless steel trace (Figure 10) to the surface of the transmitter. Predators were either tagged by divers who used a Hawaiian sling (modified spear), to embed the stainless steel trace into the muscle alongside the fin, by wrapping the transmitter in bait and feeding it to the predator, or by surgical implantation. By using the feeding technique, the likelihood of transmitter loss due to merely falling off was avoided, however transmitters can be regurgitated. Surgical implantation, although more invasive, removes the possibility of transmitter loss.

### **3.3.4 VR2 data analysis**

To correct time-drift of individual VR2 receiver clocks, VR2 data files were time-corrected using a program created by Dale Webber of Vemco. The VR2 data was analysed separately for each year. To measure spawning intensity the number of hours each squid was present on the spawning site, expressed as a percentage of the total number of hours of passive tracking, was plotted. The presence-absence of individual squid was determined by plotting

The applicator was initially held sideways and once inserted was turned 90° and the protective sheath removed (Figure 8b). After pushing the hypodermic needles through the mantle (Figure 9b), nylon washers were pushed onto the ends of the needles (Figures 8c and 9c) followed by copper crimps (Figures 8c and 9d and e). The tagged squid was then placed in a bin containing seawater or held alongside the boat (Figure 9f), depending on sea conditions, to recover. Once normal fin-beating had resumed, the animal was released within the array of

Fig. 9. Attaching a transmitter to a squid (taken directly from Downey et al. (2010)): (a) a transmitter is inserted beneath the mantle using the applicator; (b) the apparatus is turned through 90°, the protective applicator sheath removed, and the hypodermic needles pushed through the mantle. (c) Nylon washers are pushed onto the ends of the hypodermic needles and (d) a metal cylinder slipped over each hypodermic needle, (e) the metal cylinders are crimped using long-nose pliers, and (f) the squid are held submerged alongside the boat until strong swimming ability is displayed (fin beating). Only then is the animal released on

Predators were tagged with V16 pingers (2004) and V9 sensor acoustic transmitters (2005 & 2006). The transmitters were modified for attachment by gluing a stainless steel trace (Figure 10) to the surface of the transmitter. Predators were either tagged by divers who used a Hawaiian sling (modified spear), to embed the stainless steel trace into the muscle alongside the fin, by wrapping the transmitter in bait and feeding it to the predator, or by surgical implantation. By using the feeding technique, the likelihood of transmitter loss due to merely falling off was avoided, however transmitters can be regurgitated. Surgical

To correct time-drift of individual VR2 receiver clocks, VR2 data files were time-corrected using a program created by Dale Webber of Vemco. The VR2 data was analysed separately for each year. To measure spawning intensity the number of hours each squid was present on the spawning site, expressed as a percentage of the total number of hours of passive tracking, was plotted. The presence-absence of individual squid was determined by plotting

implantation, although more invasive, removes the possibility of transmitter loss.

VR2 receivers.

the capture site

**3.3.4 VR2 data analysis** 

transmitter detections at the spawning site, bottom temperature, and wind data against date and time. To determine significant differences in mean depth by day vs. night for male, female, and all squid combined, as well as mean depth for males vs. females by day and night, duplicate data, i.e. single detections recorded by more than one VR2 receiver, were removed and the total number of successfully detected transmissions for each sex per day and night calculated. The data for each sex were separated into depth categories, and the percentage of detections recorded in each depth category by day and night plotted. Twosample, two-tailed t-tests were used to identify significant differences. To analyse diurnal patterns at the spawning sites, the percentage of transmissions successfully detected per hour in a typical 24-h period were plotted, separately for males and females, using the data from which duplicates had been removed. The plots generated and the results of this analysis are given in Downey et al., (2010).

Fig. 10. A V16 pinger with a stainless steel trace attached to allow for external attachment.

The analysis of the VR2 data showed three general presence–absence behaviours to be found at chokka squid spawning sites (Downey et al., 2010). They are, as given in Downey et al., (2010): (i) arrival at dawn and departure after dusk, (ii) a continuous and uninterrupted presence for a number of days, and (iii) a presence interrupted by frequent but short periods of absence. These authors also concluded that , in contrast to the findings of earlier studies, a core aggregation of squid occasionally remains on active spawning sites at night. At dawn, more squid arrive at the spawning site and the size of the aggregation increases, resulting in a dense aggregation by day. Shortly after dusk, spawning pairs break apart, and some squid leave the spawning site. Those squid remaining at a spawning site at night search for prey throughout the water column and in the benthos, whereas lone females deposit egg strands. The authors also found that movement between the spawning sites continues at night. Their VR2 study confirmed previous observations that the initial formation of spawning aggregations, before the deposition of the first egg strand, is triggered by upwelling.

To investigate presence-absence of predators on the monitored spawning sites, the VR2 data was analysed per year. Signal detections from all tagged squid (grouped), the tagged

The Use of Acoustic Telemetry in South African Squid Research (2003-2010) 433

Active or manual tracking involves monitoring the movement of acoustically tagged animals from a vessel. South African researchers made use of the VR100 system for active

The manual tracking study discussed here made use of a VH110 directional hydrophone and a VR100 receiver. This general purpose, splash-resistant receiver is designed for tracking animals from vessels. The hydrophone is held in the water, either manually or by attachment to the side of the boat. The hydrophone detects transmitter signals and the VR100 records the ID Code, date, time, other received information (depth/temperature) and GPS location of the detections. This information can then be downloaded to a computer for

As part of a project investigating deep spawning (71-130 m) in *Loligo reynaudii*, a phenomenon researchers as yet know very little about, the movement of squid on the deep spawning grounds was monitored using the above-mentioned manual tracking system. As it is difficult to find and identify active spawning aggregations deeper than 60 m, using the two fixed telemetry systems previously described would not be feasible. This study was

Using the jigging fishing method (Figure 7), squid at depths >60 m can only be caught at night, using powerful lights to attract them to the surface. For the manual tracking study, squid were caught from an 8 m inflatable boat anchored next to a chokka boat. The two boats were close enough for the chokka boat lights to attract squid to the area around the smaller boat. Two squid were caught in this manner, on separate nights, and tagged with V9TP-6L continuous sensor transmitters. Details of the transmitters used are given in Table 2. Animals were tracked (Figure 11) from the time of tagging to shortly after sunrise. The tagging method and instrumentation used was the same as that described for the VR2 and

> **Pressure sensor**

V9TP-6L 450 1050 Yes Yes 63 Male V9TP-6L 450 1050 Yes Yes 75 Sneaker

The VR100 data was manually examined, using Microsoft Excel, for erroneous depth and/or temperature data. Erroneous data were identified by their large difference from previous and successive values, whereas these were similar. Those data entries containing errors

**Temperature sensor** 

**Frequency** 

**(kHz) Sex** 

male

conducted during the November 2010 squid fishery closed season.

**Max period (ms)** 

Table 2. Details of acoustic transmitters used in the VR100 tracking study

**Min period (ms)** 

**4. Active tracking telemetry system** 

tracking.

**4.1 VR100 receiver** 

viewing or analysis.

**4.2 Active tracking studies** 

**4.2.1 Tagging of animals** 

VRAP studies.

2010

**Year Transmitter type** 

**4.2.2 VR100 data analysis** 

predators (individually) and surface and bottom temperatures were plotted. The position of predators in the water column, in relation to squid, was analyzed by plotting all squid depth data (grouped), predator depth data (individually) and surface and bottom temperatures. Plots were generated only for those days predators were present.

The results of the predator study are as yet unpublished. This study, however, showed predators moved to and from the spawning sites a number of times, despite the continual presence of squid. The presence of predators on the spawning sites appeared to be strongly linked to surface temperature. When temperatures were stable at ~18 °C, predators remained on the spawning sites for long periods. When surface temperatures increased, predators either moved to the surface and left the spawning site shortly thereafter or immediately moved off.

#### **3.3.5 VRAP data analysis**

Invalid positional fixes were identified by their large distance from previous and successive fixes, whereas these were close in proximity. For each squid monitored by the VRAP system daily plots, separating day vs. night movement, were generated using Arcview GIS software. This allowed analysis of horizontal movement at the individual level as well as the identification of patterns in movement. Similarly depth over time was plotted for each individual. Depth data recorded by the VRAP system was not analyzed in great detail as the analysis of the VR2 receiver depth data was fairly comprehensive. The distance between two consecutive points, when the time between consecutive detections was less than 10 minutes, was used to calculate swimming speed. The distance (d) between two consecutive locations was calculated in Microsoft Excel using Equasion 1:

> d=acos(cos(radians(90-Latitude1)).cos(radians(90-Latitude2))+ sin(radians(90-Latitude1)).sin(radians(90-Latitude2)). (1) cos(radians(Longitude1-Longitude2))).R

The value 6371 km was used for the radius of the earth (R). This formulae made use of latitudes and longitudes in decimal degrees. Swimming speed was calculated by dividing the distance between two consecutive detections by the number of seconds taken to move between the two points (m.s-1). Average swimming speeds were then calculated. As these results are as yet unpublished and data is still being analysed, only the initial analysis and findings are reported here.

At night males appeared to move around the spawning site, covering a larger surface area, compared to females. This was possibly due to the males' main nocturnal activity being feeding, whereas females often continue to deposit eggs, using stored spermatophores for fertilization. On occasion however, males would also spend a number of hours in one specific area of the site, possibly resting. Both sexes spent time concentrated in one area for a number of hours during the day. Average swimming speed for males at night was calculated as 0.25 m.s-1, compared to 0.22 m.s-1 for females. These slight differences are possibly a result of the different nocturnal activities. Average swimming speed for males during the day (0.21 m.s-1) was slower than that calculated for females (0.24 m.s-1). The 1993/1994 telemetry studies (Sauer et al., 1997) also reported males to swim more slowly than females when part of a spawning aggregation. The swimming speeds reported by these authors were however, slower than those observed in this study (0.18 m.s-1 for females and 0.14 m.s-1 for males). No predators were detected by the VRAP system.
