Weirathmueller Michelle J, Wilcock William S D, Hilmo Rose S
School of Oceanography, University of Washington, Seattle, Washington 98195, USA.
J Acoust Soc Am. 2017 Oct;142(4):2101. doi: 10.1121/1.5005494.
A semi-automated method is described to range to vocalizing fin whales using the timing and amplitude of multipath arrivals measured on seafloor receivers. Calls are detected and multipath arrivals identified with a matched filter. Multipath times and relative amplitudes are predicted as a function of range by ray tracing. Because the direct and first water-column multiple arrivals are not always observed, different hypotheses for the observed arrival paths must be considered. For two arrivals, an amplitude threshold is used to determine if the first arrival is the direct path and if not, the call is disregarded as distant. When three or more arrivals are detected, three hypotheses for the paths of arrivals are considered; the solution is the hypothesis and range that minimizes the timing and optionally, amplitude ratio or absolute amplitude residual. The method is tested with data from two ocean bottom seismometers, one on the Juan de Fuca Ridge and the other in the Cascadia Basin. Solutions obtained by minimizing a combined residual from timing and an empirical absolute amplitude model extracted from the data yield reliable ranges up to 5 km at both sites, and are sufficient to estimate call density using point transect distance sampling.
本文描述了一种半自动方法,该方法利用在海底接收器上测量的多径到达的时间和幅度来对发声的长须鲸进行测距。通过匹配滤波器检测叫声并识别多径到达。通过射线追踪预测多径时间和相对幅度作为距离的函数。由于并非总能观测到直达波和首次水柱多次波,因此必须考虑观测到的到达路径的不同假设。对于两次到达,使用幅度阈值来确定首次到达是否为直达路径,如果不是,则将该叫声视为远距离而不予考虑。当检测到三次或更多次到达时,考虑到达路径的三种假设;解决方案是使时间以及可选的幅度比或绝对幅度残差最小化的假设和距离。该方法使用来自两个海底地震仪的数据进行了测试,一个位于胡安德富卡海岭,另一个位于卡斯卡迪亚盆地。通过最小化时间和从数据中提取的经验绝对幅度模型的组合残差获得的解决方案,在两个地点都能产生高达5公里的可靠距离,并且足以使用点样线距离抽样来估计叫声密度。