Background & Summary

Mobulid Rays (family Mobulidae) are pelagic elasmobranchs distributed globally in tropical and warm-temperate waters1. They have very low fecundity and are targeted in small scale fisheries as well as captured as bycatch in large scale industrial fisheries making them exceptionally vulnerable to overexploitation2. In addition, their large-scale and long-term movement patterns are generally unknown3. While numerous tagging studies of elasmobranchs exist, few are focused on Mobulid Rays, with even fewer focused on the smaller devil ray species3,4,5,6. In an effort to elucidate these patterns to inform fisheries managers and conservation efforts, we initiated a tagging program in June 2004 within the southern Gulf of California, Mexico. The Gulf of California is home to five of the nine Mobulid Ray species (Mobula birostris, M. mobular, M. munkiana, M. tarapacana, and M. thurstoni), with M. munkiana listed as Vulnerable by the IUCN Red List and the other four species listed as Endangered7. We focused our efforts on the three most abundant species in the region, M. mobular, M. munkiana, and M. thurstoni. We present here the dataset from the electronic tags applied to these three species. The tags recorded temperature, as well as pressure and light-level data that allow for depth and location to be calculated. All of these data and related metadata are now publicly available through the United States (US) Animal Telemetry Network (ATN) Data Assembly Center (DAC), part of the National Ocean Service (NOS) National Oceanic and Atmospheric Administration (NOAA) Integrated Ocean Observing System (IOOS).

Methods

Tagging deployments and study subjects

Between 2004 and 2014, 48 Mobula spp. were tagged near Isla El Pardito in the Gulf of California, Mexico (14 Mobula mobular, 23 M. thurstoni, and 11 M. munkiana) with Wildlife Computers, Inc. (Redmond, WA) pop-up archival transmitting tags (PAT tags; also known as PSATs – pop-up satellite archival transmitting tags). The tag deployment metadata file contains details on each of the 48 tag deployments including information on the type of tag deployed (model, Platform Transmitter Terminal identification), individual Mobula spp. tagged (species, unique identifying number, sex, size), and capture event details (time, date, location) (Tables 1, 2). Tag deployment and Mobulid Ray demographic details are summarized in Fig. 1.

Table 1 Metadata descriptions of the tag deployment and Mobula spp. details for all tags included in the database.
Full size table
Table 2 Continuation of metadata descriptions of the tag deployment and Mobula spp. details for all tags included in the database.
Full size table
Fig. 1
figure 1

Metadata and deployment summaries of the Mobulid Ray tagging program. (a) Types of PAT tags deployed on Mobulid Rays; two-thirds of tags deployed were Wildlife Computers MiniPAT tags. (b) Number of tagged individuals by Mobulid Ray species: 14 (29%) M. mobular, 11 (23%) M. munkiana, and 23 (48%) M. thurstoni. (c) Sex distribution of tagged individuals by Mobulid Ray species. More male M. mobular (71%) were tagged than M. munkiana (27% male) or M. thurstoni (43% male). (d) Deployments by year by Mobulid Ray species. Tag deployments occurred during the month of June in all years except 2014 when tag deployments occurred during March. There were no deployments in 2013. (e) Total body length (cm) binned in 5 cm increments for each Mobulid Ray species.

Full size image

Tagging methods are detailed in Croll et al.3, but are summarized here. Mobula spp. were encircled with a 150 m long, 15 m deep, 25 cm mesh braided nylon surface net (Fig. 2a,b). Once captured, Mobula spp. were held in the water alongside the skiff to allow for water to flow through their gills (Fig. 2c). Individuals were measured for length and width or half-width (depending on species) and sexed prior to tag attachment. Tags were attached to the dorsal surface along the pectoral fin margin with an aluminium pole and a medical-grade plastic umbrella dart (Fig. 2d). A secondary attachment loop was used to keep the tag flush with the surface of the animal (Fig. 2d).

Fig. 2
figure 2

Typical Mobula spp. tagging operation in the Gulf of California, Mexico. (a) Braided nylon surface net being deployed from a skiff. (b) The capture net floating at the surface as the skiff encircles the Mobula spp. (c) Research team measuring Mobula mobular while it is held in the net alongside the skiff. (d) Successfully applied Mk10 PAT tag with secondary loop at the base of the float. Size and shape of PAT4 tags are the same as Mk10 PAT tags. (e) Successfully applied MiniPAT tag with secondary loop at the base of the float. All participants consent to having their picture taken and used in this manuscript.

Full size image

PAT platform sensors and configuration

Pop-up archival transmitting (PAT) tags deployed on Mobula spp. included PAT4, Mk10, and MiniPAT tags from Wildlife Computers, Inc. (Redmond, WA.). All PAT tag models included wet/dry, light level, pressure, and temperature sensors. Tags were programmed to collect external temperature, depth, and light level data while deployed on Mobula spp., and set to release from their anchor approximately 6 months post deployment or after remaining at a constant depth range (within 4–8 m) for 48–96 hours (depending on deployment year). Once released from the Mobula spp. tags floated to the surface and transmitted a subset of data to the Argos satellite system. PAT tag post-release transmission details are described in detail in O’Sullivan et al.8. PAT tags were programmed to prioritize which data to transmit and in what format (Table 3). The full archival dataset was downloaded from any recovered tags providing fine scale temperature, depth, and light level data recorded by the tag.

Table 3 Programmed PAT tag metadata descriptions.
Full size table

Light level data was used to estimate the location of the Mobula spp. by producing two light-level curves each deployment day as described by O’Sullivan et al.8. The light level curves are used to produce geolocation estimates (latitude and longitude) and are dependent on the quality of the light curves8. Wildlife Computers proprietary geolocation algorithm, GPE3, was used to process the light-level data to further refine the geolocation estimates. The GPE3 model runs contain two types of uncertainty around the geolocation estimate, observation light level mean sum of squares and model score (more information can be found at www.wildlifecomputers.com). Users are required to enter the estimated swim speed of the tagged animal, as well as the deployment start and end locations. For each tagged Mobulid ray, we conducted multiple runs of the GPE3 software using 0.75 ms−1, 1.0 ms−1, 1.25 ms−1, 1.5 ms−1, and 2.0 ms−1 for animal swim speed. Overall, we found the model runs at a swim speed of 1 ms−1 to have the highest overall model scores, and thus we exported those model runs to the ATN DAC and DataONE repository.

Data transmission and processing

Data from successful PAT tag deployments (n = 45) were transmitted to Wildlife Computers through Argos Services and decoded using the Wildlife Computers data analysis program (DAP; Wildlife Computers, Inc.). Archival data from recovered tags (n = 7) were manually uploaded directly to the Wildlife Computers data portal and decoded using the DAP. Similar to O’Sullivan et al.8, decoded raw telemetry data and processed GPE3 files 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 identifier (Tables 1,2). Downloaded data were zipped and maintained as is.

Data Records

Unique identification numbers assigned by researchers (i.e., AnimalID) were used to label each zip file (see Table 2). Similar to O’Sullivan et al.8, the subset of files included within each deployment folder are contingent on tag model, programming selections and whether a tag was successfully recovered. Transmitted and if available, recovered data were merged prior to release to the ATN DAC. Individual data files, regardless of tag type, were labelled using the tag’s assigned Platform Transmitter Terminal (PTT) id and the specific WC file type. Processed GPE3 files are labelled using animal id and the number of the selected GPE3 file run. To assist with future merging and reuse of these data, unique deploy id (i.e., AnimalID), PTT ID and tag type were included within each individual data file.

Full data records and metadata from the 45 tags deployed on Mobula spp. from 2004–2014, as well as an ISO 19115 metadata record with geospatial data, are publicly available through the Research Workspace (RW) Data Observation Network for Earth (DataONE) member node (https://search.dataone.org/portals/RW) as well as the ATN data portal (https://tinyurl.com/3k44ca4e) where the location files (i.e. GPE3-X.csv) are also visualized. These data have a standalone, upstream Digital Objective Identifier (https://doi.org/10.24431/rw1k7du) specific to the dataset itself9 and a standard CC-BY license. These data and free to use without restriction, however, we request that future users acknowledge the ATN as well as cite this data manuscript in any representations of the data and/or future publications. The M. mobular data were previously published in Croll et al.3.

Technical Validation

Post-processing of raw data

Similar to O’Sullivan et al.8, raw data files were exported directly from the tag manufacturer (Wildlife Computers) by the ATN DAC and preserved as is. Files were reviewed by ATN for completeness and to ensure the correct labels were applied to files and folders, and proper ids were provided. We strongly encourage new users to read and fully comprehend associated metadata, and data files prior to use.

Usage Notes

Three of our tagged M. thurstoni (Mx07_29_MTh, Mx11_02_MTh, and Mx14_03_MTh) did not report any data and were thus excluded from the archived dataset. In addition, two of our tagged M. thurstoni (Mx10_19_MTh and Mx14_02_MTh) did not report enough data to run the GPE3 location process.