Oxygen depletion events and anoxia are a key threat to shallow marine coastal seas worldwide. The mortalities they trigger, however, are difficult to document in full. We developed an underwater device to experimentally induce hypoxia and anoxia on the seafloor. The EAGU (Experimental Anoxia Generating Unit) combines a time-lapse camera and flashes with an array of sensors and a datalogger. The unit was successfully deployed in 24 m depth in the Northern Adriatic Sea for 3 to 5 d and yielded detailed information on the behavior and sequence of mortality of macrobenthic organisms – both epiand infauna – under decreasing oxygen and increasing H2S concentrations. This unit, designed as a chamber with an instrument lid, also can be deployed in an open configuration to document low dissolved oxygen (DO) events. The equipment can provide data for a catalog of behavioral patterns, define indicator species, help reconstruct past mortalities, and better gauge the stability and status of benthic communities. *Present address: Department of Marine Biology, Faculty of Life Sciences, University of Vienna, Althanstrasse 14, 1090 Vienna, Austria. stachom5@univie.ac.at Acknowledgments Funding was provided by the Austrian Science Fund (FWF; project P17655-B03). The concept behind this work was discussed from the onset with Robert Diaz, Rutger Rosenberg, Heye Rumohr, and Kay Vopel, all of whom provided important advice and encouragement, especially Robert Diaz, whose own “dome of death” idea has come to fruition in our EAGU instrument. We would like to thank Valentin Perlinger (workshop), Gregor Eder (photography), Sabine Maringer (chemistry), Philipp Steiner, Alexandra Haselmair, and Ivo Gallmetzer (technical and diving support). Frank Bruder modified the camera housing, Markus Moll and Wolfgang Tick from Subtronic adapted the flashes. Lars Damgaard from Unisense answered our many questions about the sensors and the datalogger. Toshiba sponsored a field-notebook for the duration of the project. Special thanks to Joerg Ott for organizing and helping re-equip our boat, to Lado Celestina for refitting and maintaining the boat, and to the director (Alenka Malej) and staff (Gaspar Polajnar, Branko Cermelj, Janez Forte, Tihomir Makovec, Mira Avcin) of the Marine Biology Station (MBS) Piran, Slovenia, for their hospitality and support in matters large and small. Finally, an insightful anonymous reviewer provided a wealth of ‘spot on’ suggestions that we were happy to incorporate. Limnol. Oceanogr.: Methods 5, 2007, 344–352 © 2007, by the American Society of Limnology and Oceanography, Inc. LIMNOLOGY and OCEANOGRAPHY: METHODS Stachowitsch et al. Continuous documentation of anoxia 345 team of sport divers to record the extent of an ongoing oxygen depletion event. Additional benthic mortality events were also discovered in 1980, 1983, and 1989 during routine fieldwork (Stachowitsch 1991). Although seasonal anoxia in the northern hemisphere occurs mostly in late summer/fall (Pearson and Rosenberg 1978; Stachowitsch and Avcin 1988; Druon et al. 2004), its actual timing is related to local weather conditions. The onset and extent of such disturbances are difficult to predict and tend to elude investigation in the field. Finally, mortality events often run their course within a few days (Stachowitsch 1984), further hindering their full documentation. Laboratory chamber/aquarium experiments on respiration and responses to decreasing oxygen concentrations typically involve individual specimens or species (Renaud 1986; de Zwaan 2001; Miller et al. 2002; Matozzo et al 2005; Shimps et al. 2005). Their results, while physiologically accurate, do not combine all the relevant information about actual behavioral responses, intraand interspecific interactions, mortality sequences, and community-level processes in the natural environment. We addressed this dilemma by developing a device that can create and fully document small-scale experimental anoxia, in situ, as well as document the sequence of benthic mortalities. This instrument combines photo-documentation with detailed chemo-physical analyses and allows the behaviors and mortalities of benthic organisms to be analyzed during an oxygen depletion event from the onset. The focus is on the macrofauna because macroepiand infauna are widely used to detect and monitor community responses to environmental change. Here, as in the past, we refer to the macrofauna as those organisms that are visible in situ to the naked eye and to the camera, although in certain other habitats, e.g., the deep-sea benthos, such organisms may be referred to as megafauna. Many benthic organisms are sedentary and longlived, and the community structure therefore reflects environmental conditions integrated over extended periods (Bilyard 1987; Gray et al. 1988; Bourget et al. 2003; Ragua-Gil et al. 2004). Moreover, the benthos in the Northern Adriatic – via re-colonization and succession – can store information on prior disturbances over years or even decades and, therefore. can be regarded as a long-term memory of the overall system (Stachowitsch 1992). Materials and procedures Design of the experimental anoxia generating unit (EAGU)— The EAGU (Fig. 1) creates anoxia by sealing a 50 × 50 × 50 cm volume of water off from the surrounding environment. The instrument lid is positioned atop two different bases. The first is the “open” configuration (hereafter referred to as “frame”), a 2 cm aluminum-profile frame, (L × W × H = 50 × 50 × 50 cm) that is positioned over selected benthic organisms on the sediment surface. This configuration permits full water exchange and does not disrupt normal bottom-water currents. We observed no sediment accumulation or scouring of the seabed adjacent to the frame. This configuration is used to document animal behavior under normoxic conditions (as a control before reconfiguring to generate anoxia) or to record oxygen depletion events. The second, “closed” configuration (hereafter referred to as “chamber”) also consists of an aluminum-profile frame of the same size, but with 6-mm-thick plexiglass plates on its four vertical sides. This cube-like chamber is open above and below. The lower plexiglass edges are strengthened with sharpened aluminum elements. This chamber is pushed approximately 2 cm into the sediment to hinder water exchange through the substrate (Fig. 2). The watertight lid (simple rubber seal around upper edge of chamber) prevents exchange with the water column. This configuration is used to document behavioral responses to decreasing oxygen concentrations. The four lower corners of both configurations are equipped with removable 7-cm-long tapered metal tips that help stabilize the device in the sediment. The chamber is also equipped with two 50-cm-long handles to facilitate transportation and manipulations. The lid consists of a 12-mm-thick plexiglass plate measuring 51 × 70 cm and bears the equipment described below. Fig. 1. Experimental Anoxia Generating Unit (EAGU) with instrument lid positioned on top of plexiglass chamber. Here, only one sensor is connected to the datalogger and inserted through a sensor port. ch: camera housing, dl: datalogger, eb: external battery, fl: flashes, mb: metal brackets, os: oxygen sensor, pc: plexiglass chamber, sp: sensor port. Stachowitsch et al. Continuous documentation of anoxia 346 Camera equipment: A digital camera (Canon EOS 30D) with a zoom lens (Canon EFS 10-22mm, f/3.5-4.5 USM), mounted in an underwater carbon-fiber housing (Fig. 1) with a dome port (both Bruder). The camera’s number of effective pixels is 8.2 MP. The time-lapse function is effected by a Canon Timer Remote Controller (TC-80N3), and a 1 GB flashcard is used. The lens and its setting (14 mm) were chosen to provide an optimal combination of distortion-free images, a view of the entire 50 × 50 cm sediment area along with a portion of the vertical plexiglass walls, and to position the camera as close to the bottom as possible. This provided clearer images in turbid conditions (frame) and reduced the water volume in the chamber. Two underwater flashes (“midi analog,” series 11897; Subtronic). The flashes are modified to be adjusted manually (we used the 1/16 setting) and are attached to the lid by PVCswivel arms on two adjoining sides (Fig. 1). Two external battery packs power both the camera and the flashes (akku-safe 9Ah Panasonic; Werner light power Unterwassertechnik). The camera housing is positioned such that it lies centrally over the frame or chamber. The camera housing port fits snugly into an O-ring-equipped opening, with the dome projecting below the lid. The housing is further attached to the lid with an L-shaped aluminum bracket. The housing has four sockets: two for the flashes and two for the battery packs. Available power is usually the limiting factor in stand-alone long-term measurements. A special electronic control circuit (Fig. 3) was developed in order to run the equipment for at least 72 h with sufficiently small and light external batteries in combination with a commercially available camera and flash. The circuit was built on a small board (12 × 3 cm) using standard CMOS integrated circuits for logic functions and transistors for switching. The following functions were implemented: (1) A monitoring circuit (ICL7665 + Power Transistor) interrupts the 12 V supply power when the voltage falls below 10.2 V to prevent damage to batteries and electronics; (2) A stabilizing circuit (LM 317) provides a constant 7.5 V to the Canon camera. The camera automatically switches itself off 1 min after each shot; (3) A charging circuit (resistor + diode) constantly recharges the internal batteries in the flashes.
