1 A sensor network platform to study impact of ocean acidification in deep water environments. Tom O’Reilly † , Kent Headley † , Bob Herlien † , Karen A Salamy † , Sameer Tilak ‡ , Duane Edgington † , Tony Fountain ‡ , Peter Brewer † , Bill Kirkwood † † Monterey Bay Aquarium Research Institute Email: {oreilly,headley,bobh,salamy,duane,brpe,kiwi}@mbari.org ‡ University of California, San Diego Email: {stilak,tfoundain}@ucsd.edu Scientists and engineers at the Monterey Bay Aquarium Research Institute (MBARI) are developing sensor network technology to better understand anthropogenic ocean acidifi- cation. Burning of fossil fuels has produced cumulative CO2 emissions on the order of one trillion tons since the beginning of the Industrial Revolution. While approximately half of this CO2 remains in the atmosphere today, the ocean is the main repository for the balance. Experiments and observations sug- gest that CO2 ocean enrichment and the resulting acidification will inhibit formation of calcium carbonate exoskeletons, and may have major impacts on marine ecosystems. To study these effects on marine ecosystems in situ, MBARI has developed the Free Ocean Carbon dioxide Enrichment (FOCE) sys- tem [1]. In contrast to laboratory experiments, FOCE directly measures effects in the ocean environment, and is designed for deployment in deep locations such as the Monterey submarine canyon. FOCE consists of a flume that is open to the ocean on the bottom and ends (see Figure 1b). CO2-enriched seawater is injected into either end of the flume (depending on the direction of ambient currents) and flows into a central chamber, which holds biological specimens, cameras, and sensors. The flow may be driven entirely by natural ambient currents or regulated using fans and baffles. The control loop will have several modes of operation, but typically will be used to maintain a constant pH offset from the surrounding envi- ronment. In this way, the conditions in the chamber simu- late ocean acidification conditions corresponding to projected future atmospheric CO2 levels. FOCE sensors include pH, temperature, salinity, and current sensors inside and outside the flume as well as lights and cameras. These sensors are connected to a PC-104 style embedded computer with 500 Mhz CPU, 1 GB memory, and 160 GB notebook drive for persistent storage. The FOCE system is now full scale and installed in deep waters in the Monterey Bay Canyon. FOCE is connected to Monterey Accelerated Research System (MARS: http://www.mbari.org/mars) deep-sea cable-to-shore observatory node in Monterey Canyon in 900 meters depth at the end of the 52 kilometer cable (Fig. 1a). The MARS cable provides Gigabit Ethernet communications over optical fiber, and hence supports sensors with very high data rates. To the best of our knowledge, this is the first deep water sensor network deployment aimed at understanding impact of ocean acidification on marine life. The FOCE equipment is designed with a goal of deployment cycle of 1-2 years duration. Designing and deploying the FOCE testbed presented sev- eral research and engineering challenges including provision- ing of electric power and networking capabilities, engineering of the FOCE tube, integration of a broad range of sensors, and design of development of software infrastructure for the data plane (for data acquisition) and control plane (to remotely command and control sensors), and sustainability of the testbed (e.g., upon deployment, FOCE testbed will be visited by an ROV on a periodic basis for CO2 replenishment). FOCE is a technology concept with potentially many different implementations ranging in size and complexity. For example, a version of of FOCE developed in collaboration with Uni- versity of Queensland is designed for use in a few meters of water on a coral reef. It requires much less infrastructure, and may be installed and operated without a vessel. The software framework is designed and developed in a modular fashion such that it is applicable to other projects at MBARI. Our goal is to make the hardware and software technology freely available to the oceanographic community world-wide. MBARI team members have collaborated with University of California (UCSD) team to design and develop data plane and control plane software framework for FOCE experiments. MBARI has developed network middleware known as SIAM to operate sensors deployed on TCP-IP network-based ocean observatories [2]. Each physical sensor in an observatory can be accessed through a corresponding SIAM instrument service. The instrument service interface defines generic methods to configure, control, and acquire data, while the service im- plementation maps between the generic interface and specific sensor protocols. Thus diverse non-standard sensor protocols are hidden by the uniform SIAM instrument service interface, resulting in an instrument network that can be accessed and managed in a straightforward way. This architecture also pro- vides a foundation for more sophisticated science applications with remote control and autonomous operation requirements. SIAM is designed for use on oceanographic observatories that are connected to shore through low-bandwidth satellite links, as well as systems that are connected to shore via power/communication cable. SIAM service network interfaces are implemented with Java RMI, which provides object-oriented remote procedure