Identifying "skylites" for AUV operations under pack ice: Insights from ice-draft profiling by moored sonar

The past decade has seen a remarkable evolution of sea-floor-based ice-draft profiling capabilities. Efforts have progressed from an original Beaufort Sea deployment of a single upward-looking acoustic echo sounder to the almost routine present-day positionings of special-purpose profiler units which operate in conjunction with adjacent current profiling and ice drift measurement instruments. These units allow detailed specification of draft statistics and high resolution mapping of moving ice undersurfaces for both on-board storage and, in real time, via cable and VHF connections. The data acquired have been employed for a wide variety of purposes including: monitoring the effects of climate change; characterization of, pack ice properties relevant to offshore platform- and facility-design; studies of wave climates inside marginal ice zones; and provision of realtime assistance for navigation and ice management decision-making. Presently, special purpose ice profiling sonars are being incorporated into under-ice science-related missions using autonomous underwater vehicles (AUVs) and on manned submarines This presentation begins with a short outline of present profiling capabilities, identifying important characterizations of acoustic scattering by an ice undersurface and outlining the development of the high frequency sampling technique which is an essential element in providing the detail and accuracy required for most modern applications. Quantitative data are provided on key issues determining instrument performance and their implications for optimal use of similar instruments for identifying suitable "skylites" or patches of open water or thin ice suitable for bringing AUVs to the sea surface for recovery or to carry out operational tasks. Two fundamentally different identification approaches, based upon, respectively, echo amplitude and range measurements are discussed and related to typical AUV operating constraints and needs. It is concluded that neither approach will, in itself, meet user needs, necessitating future efforts toward development of a hybrid identification methodology in accord with suggested operating principles.