Estimating received level in behavioral response studies through the use of ancillary data.

Abstract

Marine mammals are known to respond to various human noises, including and in certain cases, strongly, to military active sonar. Responses include small and short-term changes in diving behavior, horizontal avoidance of an ensonified area, and mass strandings. Considerable research has been conducted using short-term biologging tags to understand these responses. Yet researchers and managers want a better understanding of responses to sound over longer periods of time in a variety of contexts. The Atlantic Behavioral Response Study examines responses across multiple spatial and temporal scales using vessel-based focal follows, short-term biologging tags, and medium-term satellite transmitting tags. Since the latter do not record sound, we must intersect positions with a sound propagation model to estimate received sound pressure levels. We use all available information from (1) the observed x,y positions from the tag(s) and from focal follow vessels; (2) the discrete depth bin data (z) from the tag; (3) ocean bathymetry; and (4) outputs from sound propagation models. All these disparate streams of data contain varying levels of error in x, y, or z. We account for as much uncertainty as possible and include here a refined approach to better estimate the range of sound levels received by animals.