摘要：Managing the impacts of vessel noise on marine fauna requires identifying vessel numbers, movement, behaviour, and acoustic signatures. However, coastal and inland waters are predominantly used by ‘small’ (<25 m-long) vessels, for which there is a paucity of data on acoustic output. We reviewed published literature to construct a dataset (1719 datapoints) of broadband source levels (SLs) from 17 studies, for 11 ‘Vessel Types’. After consolidating recordings that had associated information on factors that may affect SL estimates, data from seven studies remained (1355 datapoints) for statistical modelling. We applied a Generalized Additive Mixed Model to assess factors (six continuous and five categorical predictor variables) contributing to reported SLs for four Vessel Types. Estimated SLs increased through ‘Electric’, ‘Skiff’, ‘Sailing’, ‘Monohull’, ‘RHIB’, ‘Catamaran’, ‘Fishing’, ‘Landing Craft’,’ Tug’, ‘Military’ to ‘Cargo’ Vessel Types, ranging between 130 and 195 dB re 1µPa m across all Vessel Types and >29 dB range within individual Vessel Types. The most parsimonious model (22.7% deviance explained) included ‘Speed’ and ‘Closest Point of Approach’ (CPA) which displayed non-linear, though generally positive, relationships with SL. Similar to large vessels, regulation of speed can reduce SLs and vessel noise impacts (with consideration for additional exposure time from travelling at slower speeds). However, the relationship between speed and SLs in planing hull and semi-displacement vessels can be non-linear. The effect of CPA on estimated SL is likely a combination of propagation losses in the shallow study locations, often-neglected surface interactions, different methodologies, and that the louder Vessel Types were often recorded at greater CPAs. Significant effort is still required to fully understand SL variability, however, the International Standards Organisation’s highest reporting criteria for SLs requires water depths that often only occur offshore, beyond the safe operating range of small vessels. Additionally, accurate determination of monopole SLs in shallow water is complicated, requiring significant geophysical information along the signal path. We suggest the development of appropriate shallow-water criteria to complete these measurements using affected SLs and a comprehensive study including comparable deep- and shallow-water measures.
