Tagging deployments and study subjects
Table 1 contains an overview of the fields in the metadata file (JWS_metadata.xlsx) providing extensive background details on each of the 79 tag deployments and 63 study subjects. The data in this file give essential contextual information needed to understand the methodological, environmental, and demographic factors surrounding the deployments, which are critical for further examination and hypothesis testing of the sensor data. These metadata fall into several specific categories, but are not limited to, (i) information on the deployed electronic devices (platform, model, Platform Transmitter Terminal identifications), (ii) sharks (unique identifying numbers, sex, length), (iii) capture event (date, location, duration, methodology, interaction type), and (iv) the reporting period (duration, linear surface travel distance).
Figure 1 illustrates a typical C. carcharias tagging operation. This involves a contracted commercial fishing vessel with purpose-built gears to capture sharks (Fig. 1a) and a research crew to handle animals, monitor health (Fig. 1b) and attach electronic tags (Fig. 1c). More details on the tagging program and its methodologies are provided elsewhere14,19,20. Figure. 2 provides summaries of the deployment schedule, geographic locations, devices, and capture operations. Of note, 39.7% (25/64) of all tagging operations involved collaborations with commercial fishery operators (Fig. 2f–h), whose engagement was temporarily impacted (Fig. 2a) during the scientific review process when the population was under consideration for US Endangered Species Act listing. Figure 3 displays the demographic focus on small juvenile C. carcharias, with modest deployment durations and travel distances.
Depiction of a typical research operation for capturing and tagging juvenile White Sharks in the Southern California Bight. (a) Aquarium research vessel (RV Lucile) with crew approaching a contracted purse seine vessel containing a captured juvenile white shark. (b) Research crew on the RV Lucile leading the shark into a sling, where it is subsequently transferred to the vessel’s deck for tagging. (c) Successfully applied PAT and acoustic tags each positioned lateral of the dorsal fin, anchored via leaders, and affixed with titanium darts (yellow arrows). All images taken by Steve McNicholas (Great White Shark 3D) for the Monterey Bay Aquarium and used with permission.
Metadata summaries of the field program that deployed biologging tags on juvenile white sharks in the southern California Current. (a) Deployment schedule for 72 electronic tags released on 64 White Sharks from 2001–2020 (b) Tagging activity peaked in the late summer months when the population is most locally abundant. Field operations decreased from 2011–2013 when the population was being considered for listing under the U.S. Endangered Species Act (ESA). (c) Deployments focused on opportunities in the Southern California Bight coastline and included deployments in the nursery area of Bahía Sebastian Vizcaíno, Mexico and releases after exhibition at the Monterey Bay Aquarium. (d) Researchers released a variety of pop-up archival transmitting (PAT, 58 sharks), acoustic (21 sharks), and smart position and temperature (SPOT, 20 sharks) tags. This manuscript only reports the geolocation, temperature and depth data from the PAT and SPOT platforms. (e) Half (35 of 64, 54.7%) of all sharks received multiple tags, primarily to compare their relative performance. (f) Most tags (38 of 64, 60.3%) were deployed during focused scientific research operations. (g) The remainder were joint operations resulting from opportunistic bycatch in commercial fisheries using various gears and (h) Targeting various species. “Jab” gear refers to research operations that uses pole extensions to apply tags to sharks without capturing and handling.
Demographic and deployment summaries from the juvenile white shark tagging program. (a) Total body length (TL) histogram indicates that most individuals tagged were either neonates (<1.50 m total body length “TBL”), young-of-year (YOY, <1.75 m TBL), or other juveniles (<2.5 m TBL) which typically inhabit the geographic region of focus18,20,21,25. This plot excludes one 396 cm TBL female that was opportunistically tagged off Santa Rosa Island. (b) Over half of the individuals (33 of 64, 51.6%) were females, with the sex undetermined for 4 individuals (6.3%) that were jab-tagged and not landed. (c) A majority of tags had a short deployment length which is the duration when the tag is logging data on the shark’s activity. (d) For 50 of 64 sharks (78.1%) this was less than 6 months, and less than 3 months for half of all sharks averaged for each deployment year represented as boxplots with the raw observations. (e) Histogram of straight linear distance between the release site and the location of first tag reporting. While many juvenile White Sharks did not travel far during tag deployments, there are notable exceptions. (f) Two juvenile sharks, for example, swam a linear distance of nearly 2,000 km, each in under 200 days. Solid orange line is a locally weighted regression (shaded area is standard error) which is influenced by location of release and annual migration cycles.
SPOT5 platform sensors and configuration
Smart position and temperature transmitting (SPOT) tag data included in this dataset were obtained from SPOT5 tags (Wildlife computers, Redmond, WA) deployed on juvenile White Sharks between 2006 and 2009. These tags generated the locations of tagged sharks when the fin-mounted tag broke the surface of the water and there were Argos satellites overhead (http://www.argos-system.org). Several factors will influence the frequency of Argos locations from SPOT tags (hourly to weekly), as well as the accuracy of the positions (<250 m to >10 km), including: the sea state, the shark’s surface-oriented behavior, and the satellite coverage23. These factors determine how many messages from the tag reach the Argos satellite system, and therefore the location quality class. For instance, location class Z does not allow for a valid location to be estimated, classes B and A are assigned when only 2 and 3 messages are received, respectively and cannot estimate the accuracy of the location. Class 0, 1, 2 and 3 all require 4 messages to be received and have estimated errors of, respectively, over 1500 m (sometimes much greater), between 500 m and 1500 m, between 250 m and 500 m, and below 500 m. For additional technical details on location classes and position accuracy, see the CLS Argos User Manual (available at https://bit.ly/3uuMbzt).
The SPOT tags were programmed to only transmit location (and not time at temperature histograms or haul out statistics). They were also programmed to check if the wet/dry sensor is dry (and therefore the tag is out of the water and able to transmit) every 0.25 seconds. Messages sent by the tag were received by the Argos satellite system and transferred to the Wildlife Computers data portal from which data files were downloaded. The file formats included the proprietary .DIAG and .PRV file formats as well as a series of .CSV files. Location data are available within the locations.csv file for each tag.
PAT platform sensors and configuration
Pop-up archival transmitting (PAT) tags deployed on juvenile White Sharks used to collect data for this dataset included MK10 (deployed 2003–2016), MiniPAT (deployed 2010–2020), PAT2 (deployed 2001–2003) or PAT4 (deployed 2004–2005) tags from Wildlife Computers (Redmond, WA). All PAT tag models included wet/dry, light level, pressure, and temperature sensors. These tags were programmed to collect light level, depth and temperature data while deployed on White Sharks, and at a pre-determined date, release from their anchor and float to the surface where they could transmit a subset of their data to the Argos satellite system.
Once at the surface, a final (pop-up) location of the tag is calculated by the Argos satellite system by measuring the “Doppler shift” of repeated transmissions received by the satellite as it moves over the tag. Multiple Argos locations are calculated in this way, and just as with SPOT tags described above, Argos quality classes are associated with each location. If tags reported on schedule, the first high quality location (class 1, 2 or 3) of the tag as determined by the Argos satellite is considered the final known location of the tagged shark.
While the battery remains sufficiently charged, the tag transmits packets of information to the Argos satellite system with the archived data from the light, pressure, and temperature sensors. Due to the combination of deployment length, battery life and satellite availability over the location of the tag, only a small percentage of the archived data will be able to be transmitted. For this reason, PAT tags can be programmed by the user to prioritize which data to transmit and in what format (e.g., full time series of depth or temperature vs. binned histograms of time spent at depth and temperature). Table 2 contains an overview of the fields in the metadata file (PAT_programming.xlsx) that reveals how the PAT tags used in this study were programmed. When a tag that has popped up could be recovered, the full archived data set was downloaded. This provided fine scale data on the depth, temperature and light level experienced by the tag, which were then uploaded to the Wildlife Computers data portal.
One of the most useful outputs of the PAT tag platforms are estimates of the location of the tagged shark while at liberty (between the known position at the time of release, and the known position at pop-up directly through the Argos satellite system). To estimate position while on the shark, the tag records light levels and produces two light-level curves each day of deployment. Using the onboard clock set to UTC time, the times of the local dawn and dusk are compared to UTC, which provides an estimate of the longitude of the tag on that day. The time between dawn and dusk (i.e., the length of the day) is used to estimate the latitude of the tag based on the day of the year. Both the latitude and longitude estimates are very much dependent on the quality of the light curves, which in turn are very dependent on the environmental conditions experienced by the tag (cloud cover, depth water turbidity, etc.). Higher quality light curves will produce more accurate geolocation estimates. This approach has been used for decades19 and has been independently validated20.
To further refine these geolocation estimates, the light-level data are processed through a proprietary geolocation algorithm on the Wildlife Computer portal called GPE3. The user provides an estimate of the average swimming speed of the tagged animal, and the GPE3 process employs a discretized Hidden Markov model that uses light levels, sea surface temperatures from satellites to compare with the onboard temperature recordings, and any known locations (such as the deployment and pop-up locations) to reduce the uncertainty around each daily geolocation estimate. More information about the GPE3 can be obtained from Wildlife Computers (www.wildlifecomputers.com).
Data transmission and processing
Data from successful SPOT (n = 19) and PAT (n = 51) tag deployments were transmitted through Argos Services directly to the manufacturer and then decoded using their data analysis program (DAP; Wildlife Computers). Data from recovered archival tags (n = 26) were manually uploaded directly to the Wildlife Computers (WC) data portal by participating researchers and then decoded using DAP. Decoded raw telemetry data and when applicable processed GPE3 files (PAT tags only, see above) were then downloaded from the Wildlife Computers data portal to the ATN DAC via the Wildlife Computers API as .CSV files and in some cases in the proprietary WC file format using the unique manufacturer assigned deployment ids (Table 1). Downloaded data were zipped and maintained as is.
Source: Ecology - nature.com
