The electronic components in the AquaticVID consist of an action camera with internal battery, miniSD card, type A male-to-male mini-USB cable, a keep-alive load (these components are the same in all versions of the AquaticVID) and a power bank (the power bank size and capacity varies depending on AquaticVID configuration) (Fig. 1, Table 1).

Camera: The Mobius action cam 1 is used as a base camera as it is cheap, reliable, takes good quality video footage and has a wider range of easily configurable settings than other action cameras (the camera can be swapped with alternatives with no modifications to the system needed; see S1). Spare parts, alternative lens types and add-ons are low cost and easily sourced. Technical camera details: Mobius 1080p HD action camera 1 (three resolution settings 848x480(WVGA)–1920x1080, frame rate ­5–60fps, MP4, MOV, AVI, video codec h.264) with 130-degree wide-angle C2 lens (170-degree diagonal FOV). Camera and lens 68 mm(L) x 35 mm(W) x 18 mm(H). Weight 41 grams (see MobiusActionCam (2017) for recording instructions). Mobius provides a windows application with a simple graphical user interface for adjusting the camera settings. Alternatively, users can use a text file to enter settings (MobiusActionCam 2017). Camera settings and setup are available via the Mobius manual. Note that, while the cameras are reported to have good sound recording capabilities, we disable the sound function primarily to maximise battery life when recording video and because we do not generally require sound data in our camera-based studies. As a result, we have not field-tested sound recording or sought to optimise sound recording in the AquaticVID systems.

Data Storage: Mobius recommends the use of 128 GB miniSD cards as a maximum; however, we tested up to 400 GB cards and some models (formatted to FAT32 or exFAT32 in camera) were compatible (Suppl. material 1). Today, power supply is often the limiting factor, particularly in the smaller system versions, so storage capacity should be based on power and settings used (see Suppl. material 1 for power and storage capacity test outcomes), so that power runs out before the data capacity fills (include a small buffer as power use and data size vary based on light conditions).

Keep-alive Load: Sota beams variable load 6–150 mA with adjustable current level. Most modern power banks have an automatic sleep mode (they switch off if the current being drawn stops or is too low after a short time period) that cannot be turned off. This means that when charging a camera with an internal battery, the power bank will charge the camera battery once and then turn off when the camera battery is full despite having the capacity to charge the camera battery multiple times. Adding a component that maintains a low constant current draw that keeps the battery back ‘alive’ prevents power bank low-current shut-down. The keep-alive load uses a USB type A female connector for incoming power supply.

Power supply: The Mobius standard battery is 520 mAh which gives roughly 90 minutes battery life on default 1920 x 1080p at 30fps (can be upgraded to 820mAh = 130 minutes). In the AquaticVID systems, additional battery supply is added: In the AquaticVID MICRO version, generally we use an additional off-the-shelf power bank of 2600­–5200 mAh and the MINI and MINI Air have an additional off-the-shelf power bank of 30,000 mAh (see Suppl. material 1 for detailed specifications). Note: Although we use a power bank in the example here, as they are easy to source in most locations and simple to ship via postage compared to batteries, a single battery or banked batteries can be used in the system.

Cables: To interconnect the Mobius camera with the keep-alive load, a type A male to type mini male USB cable is used. To interconnect the keep-alive load with the power bank, a USB cable with a type A male connection is used, while the other end is power bank specific.

Camera holder: a simple camera holder made of easily available material – we use plywood offcuts with foam padding points or foam holders (any lightweight wood or 3D printed plastic versions can be used). See Suppl. material 2 additional details.

GPS: When used above-water, an external portable GPS unit (sold as Renkforce gp102 in Sweden that is a house brand rebadged G-Porter CANMORE-gp-102) can be added to the system to collect position data (Table 1, Fig. 1) Output file is .FIT (Flexible and Interoperable data Transfer) format and contains position, altitude, distance and speed information.

The standard watertight enclosure used for the AquaticVID consists of a dome, a cast acrylic tube and acrylic end cap, two O ring-flanges (these come standard with six O-rings) all made by and distributed by BlueRobotics (Table 1, Fig. 1). In the Mini and Micro AquaticVID versions, a single enclosure vent with plug is inserted into the acrylic end cap (Table 1, Fig. 1). Closing the watertight enclosure compresses the air inside and the pressure can begin to push the enclosure open. The vent lets you release the pressure and reseal the enclosure before deployment. Double O-rings in the vent combined with screw threading prevent water entry. In the MiniAir used above-water, an additional one-way air valve is inserted into the acrylic end cap (Table 1, Fig. 1). Adding an air valve overcomes issues with air inside the housing warming when in sunlight and the air expanding and popping the endcaps off the enclosure and prevents overheating of the camera system when using power banks. Recently BlueRobotics released a new locking version of their watertight enclosure and the tubes in the 3 inch size increased in price significantly. In Table 1, we give the higher prices of the newer locking versions of the tubes and associated flanges and end caps. However, the cheaper tubes that we currently use are stock tubing that is available from a number of plastic manufacturers, for example, Eplastics and the design files for components that match these stock tubes are available open source at BlueRobotics to be utilised should the early versions be preferred for a project. For the Mini and Mini Air, we generally purchase an acrylic tube length of 400 mm (15' 6") and for the micro, an acrylic tube length of 300 mm (11.8"); however, if shorter powerbanks or battery packs are used, then the acrylic tube can be cut shorter to further streamline the system if needed. Where deep deployments are needed, the tubes can be directly swapped over with BlueRobotics aluminium components, which can be sent down to 950 m.

Optional additions:

It is easy to remove the endcap and turn the camera on and off; however, in some cases, for example, wave and splash zones, where users need to maintain the water seal, the above-water version can be modified to include an on/off switch (Fig. 1C).

Mounting options:

The AquaticVID system can be mounted or deployed in various ways to suit different applications. The small size and weight of the system make it well-suited for ad-hoc field deployments, such as using cables or tension ties to attach it to vessel fittings or fishing gear. The system is not built with a standardised deployment mount and we tend to adapt a mount to each research situation (see Suppl. material 3 for examples).

New fishing gear development requires in-situ studies evaluating the behaviour of target and non-target species around the gear, so that catch efficiency can be maximised and unwanted bycatch minimised (Thomsen and Humborstad 2010). Cameras are a widely used and effective non-intrusive means of undertaking these evaluations. For example, Ljungberg et al. (Ljungberg and Lunneryd 2016) in Sweden and Hedgärde et al. (Hedgärde and Berg 2016) in Denmark investigated the behaviour of Atlantic cod (Gadus morhua) entering fishing pots, with different entrance modifications designed to increase catch efficiency. Atlantic cod behaviour around fishing gear in the Baltic Sea has also been recorded to study which modification on cod pots can prevent seal raids (Stavenow and Ljungberg 2016). However, newer models of action cameras (for example GoPros, with their increasingly better quality images and faster frame-rates in each new model), are not always well suited to these types of studies, because they have greater data storage requirements, require more battery power and often cost more. In addition, lower frame-rates and quality settings have been phased out, restricting the ability to input custom settings that could overcome some of these issues. These issues are particularly problematic when wanting to deploy multiple systems and for extended periods, whilst minimising data.

The AquaticVID has been used for fishing gear-related behavioural studies (e.g. Fig. 2), primarily to overcome the cost and battery life issues associated with more expensive action cameras mentioned earlier. The key system features allowing the success of these studies have been the extended deployment times and the low frame-rate settings that minimise data storage requirements. In the Baltic Sea, the behaviour of cod entering pots was studied using an AquaticVID variant, with deployments lasting up to 76 hours using WVGA setting with 10 fps (Nyquist 2019). Similarly, when assessing the behaviour of Atlantic cod around a newly-developed pontoon trap (Ljungberg and Lunneryd 2018) and gear modification's ability to reduce seal interactions with catch (Lunneryd and Björklund 2018), an AquaticVID system was used and deployments times at 5 fps in WVGA resolution lasted up to 6.5 days. In the Bothnian Bay, Östman et al. (Östman and Sundblad 2023) used the system to observe salmonids moving through an open-ended pound-net as a way to corroborate catches in nets set by small-scale fishers. On the west coast of Sweden, the influence of competition between European lobster (Homarus gammarus) on catch in lobster traps (Haffling 2023; Fig. 2c) and the effectiveness of wide-scale fishing with pots for cod (Königson and Hedgärde 2023; Fig. 2b), was also demonstrated using AquaticVID. Since passive gears in some areas must be soaked several days before fish swim into the gear, the AquaticVID with its long recording time was particularly suited.

Battery power is often a limiting factor in deployments of baited remote underwater video systems (BRUVS) (Gore et al. 2020). BRUVS samples generally only last for 30–90 minutes (e.g. Sih and Cappo (2017), Fetterplace (2018), De Vos (2021), Becker and Taylor (2022), Cáceres and Kiszka (2022)). However, each system is usually deployed multiple times per day until battery power is exhausted (Fetterplace, pers. commun). In addition, in some cases, long deployment times are needed, requiring various modifications to cameras and housings (e.g. Stobart and Díaz (2015) (7 h), Harasti and Lee (2017)(5 h), Gore and Ormond (2020) (2 h+), Torres and Abril (2020) (24 h)). The cost of many commercially made BRUVS is a limiting factor in setting up BRUVS studies and the extent of sampling possible within studies. To maximise efficiency, BRUVS are almost always deployed simultaneously, in sets of 4–10, at a study site (e.g. Cáceres and Kiszka (2022), Coleman and Wood (2023)) and multiple sets may be required at different regional sampling sites (e.g. Knott and Williams (2021)) which saves labour, time and vessel costs, but means the initial outlay on camera equipment and repairs/replacements can be high. The AquaticVID system with its very long battery life, small size and low cost is a useful BRUVS option to overcome some of these issues. In Sweden, the AquaticVID system has been used in comparisons of bait types in the Baltic Sea (Königson and Strömberg 2018, Königson and Hedgärde 2023; Fig. 4) and in ongoing student thesis BRUVS projects. Although the use of BRUVS in Sweden only began relatively recently, there is increasing impetus to undertake BRUVS studies and we expect there will be increased use of the AquaticVID system in these types of studies.