Antarctic subglacial lakes are thought to be extreme habitats for microbial life and may contain important records of ice sheet history and climate change within their lake floor sediments. To find whether or not this is true, and to answer the science questions that would follow, direct measurement and sampling of these environments are required. Ever since the water depth of Vostok Subglacial Lake was shown to be >500 m, attention has been given to how these unique, ancient, and pristine environments may be entered without contamination and adverse disturbance. Several organizations have offered guidelines on the desirable cleanliness and sterility requirements for direct sampling experiments, including the U.S. National Academy of Sciences and the Scientific Committee on Antarctic Research. Here we summarize the scientific protocols and methods being developed for the exploration of Ellsworth Subglacial Lake in West Antarctica, planned for 2012–2013, which we offer as a guide to future subglacial environment research missions. The proposed exploration involves accessing the lake using a hot‐water drill and deploying a sampling probe and sediment corer to allow sample collection. We focus here on how this can be undertaken with minimal environmental impact while maximizing scientific return without compromising the environment for future experiments.

Autonomous Underwater Vehicles (AUVs) are effective platforms for science research and 9 monitoring, and for military and commercial data-gathering purposes. However, there is an inevitable risk of 10 loss during any mission. Quantifying the risk of loss is complex, due to the combination of vehicle reliability 11 and environmental factors, and cannot be determined through analytical means alone. An alternative 12 approach -formal expert judgment -is a time-consuming process; consequently a method is needed to 13 broaden the applicability of judgments beyond the narrow confines of an elicitation for a defined 14 environment. We propose and explore a solution founded on a Bayesian Belief Network (BBN), where the 15 results of the expert judgment elicitation are taken as the initial prior probability of loss due to failure. The 16 network topology captures the causal effects of the environment separately on the vehicle and on the 17 support platform, and combines these to produce an updated probability of loss due to failure. An extended 18 version of the Kaplan Meier estimator is then used to update the mission risk profile with travelled distance. 19Sensitivity analysis of the BBN is presented and a case study of Autosub3 AUV deployment in the Amundsen 20Sea is discussed in detail. 22Keywords: Bayesian networks, survival statistics, expert judgment elicitation, autonomous vehicles. Autonomous Underwater Vehicles (AUVs) have a future as effective platforms for science research and 2 monitoring, and for military and commercial data-gathering purposes. Increasingly they are being used in 3 environments that are not benign [1,2]. Environments such as under sea ice [3], under shelf ice [4], or along 4 rocky coasts [5] intuitively give rise to a higher risk of loss should the vehicle malfunction. The risk of loss is 5 real; for example, Australian and British AUVs have been lost under ice sheets [6] and one team maintained a 6 lightweight tether to an AUV when operating under sea ice. The problem of predicting risk of loss is not only 7 one of predicting the reliability of the vehicle as a whole, its sub-systems and its components, but also of how 8 the operating environment, together with reliability, sets the probability of losing the vehicle. It is not obvious 9 that an approach based on separate statistical analyses of vehicle reliability and the affects of the 10 environment on probability of loss is either feasible or meaningful. Such an approach, when reduced to 11 summary statistics such as mean time to failure, would ignore the interaction between individual faults or 12 incidents and the environment, which we postulate to be at the centre of this problem. 13One alternative would be to assess the probability of loss in various environments directly, by counting the 14 frequency of occurrence. This frequentist approach is certainly appropriate for assessing the reliability of 15 identical engineered systems, where probability of failure is derived from a long-run frequency of occurrence, 16usually fr...

Autonomous underwater vehicles (AUVs) are used increasingly to explore hazardous marine environments. Risk assessment for such complex systems is based on subjective judgment and expert knowledge as much as on hard statistics. Here we describe the use of a risk management process tailored to AUV operations, the implementation of which requires the elicitation of expert judgment. We conducted a formal judgment elicitation process where eight world experts in AUV design and operation were asked to assign a probability of AUV loss given the emergence of each fault or incident from the vehicle's life history of 63 faults and incidents.After discussing methods of aggregation and analysis, we show how the aggregated risk estimates obtained from the expert judgments were used to create a risk model. To estimate AUV survival with mission distance, we adopted a statistical survival function based on the nonparametric Kaplan-Meier estimator. We present theoretical formulations for the estimator, its variance and confidence limits. We also present a numerical example where the approach is applied to estimate the probability that the Autosub3 AUV would survive a set of missions under

. 2012 A Behavioral probabilistic risk assessment framework for managing autonomous underwater vehicle deployments. Journal of Atmospheric and Oceanic Technology, 29 (11). 1689-1703. 10 ABSTRACTThe deployment of a deep-diving long-range autonomous underwater vehicle (AUV) is a complex operation that requires the use of a risk-informed decision-making process. Operational risk assessment is heavily dependent on expert subjective judgment. Expert judgments can be elicited either mathematically or behaviorally. During mathematical elicitation experts are kept separate and provide their assessment individually. These are then mathematically combined to create a judgment that represents the group view. The limitation with this approach is that experts do not have the opportunity to discuss different views and thus remove bias from their assessment. In this paper, a Bayesian behavioral approach to estimate and manage AUV operational risk is proposed. At an initial workshop, behavioral aggregation, that is, reaching agreement on the distributions of risks for faults or incidents, is followed by an agreed upon initial estimate of the likelihood of success of the proposed risk mitigation methods. Postexpedition, a second workshop assesses the new data and compares observed to predicted risk, thus updating the prior estimate using Bayes' rule. This feedback further educates the experts and assesses the actual effectiveness of the mitigation measures. Applying this approach to an AUV campaign in ice-covered waters in the Arctic showed that the maximum error between the predicted and the actual risk was 9% and that the experts' assessments of the effectiveness of risk mitigation led to a maximum of 24% in risk reduction.