Arctic EcosystemsJohn Horne and Silvana Gonzalez, University of Washington
Scope of Work:
Detecting and understanding the impacts of oil spills in the Arctic requires characterizing (i. e. describe and quantify) natural variability in the ecosystem. However, characterizing natural variability is not a trivial task. Marine ecosystems are variable across a wide range of spatial and temporal scales as a result of numerous physical and biological processes acting and/or interacting across an equally wide range of scales (e. g. Stommel, 1963; Haury et al., 1978). Current data and understanding of natural variability in Arctic ecosystems is poor relative to more southern ecosystems, and integrated solutions to acquire and analyze variability in animal densities and distributions in extreme environments are scarce. Therefore, there is a need to develop integrated data acquisition and analytic techniques to characterize and monitor the density and distributions of marine animals living in the water column (i. e. pelagic organisms) to enable the detection and measurement of anthropogenic impacts such as oil spills in Arctic ecosystems. Characterization of pelagic communities using traditional sampling methods (e. g. ship-based mobile surveys with net or acoustic sampling) is difficult at high-latitude environments due to ice cover during part or most of the year. The use of echo sounders attached to platforms or moorings (i. e. stationary acoustics) provides a non-invasive technology to characterize temporal variability in density and vertical distributions of pelagic organisms at high resolution over long temporal periods (i. e. high scope data) (e. g. Urmy et al., 2012; Jacques, 2014). Stationary acoustics can:
(1) characterize “natural” or baseline conditions of animal density dynamics as it provides a continuous sampling of pelagic fish and zooplankton year-round;
(2) quantify the amount of change relative to baseline variability to help determine thresholds of environmental effects; and
(3) monitor for episodic events or trends in animal densities or behaviors in response to natural or anthropogenic events.
The inclusion of active acoustics in instrumented platforms (i. e. marine observatories; see Godø et al., 2014) also allows the simultaneous and continuous collection of biological and physical data, which can be used to understand physical drivers of observed biological patterns, a prerequisite to predicting biological responses to environmental or anthropogenic change.
Despite covering a wide spectrum of temporal scales (e. g. seconds to months or years), point source measurements using stationary acoustics do not include a large range of spatial scales when characterizing variability in animal densities or behaviors. Fish and zooplankton have aggregated distributions at scales that range from meters to kilometers (Horwood and Cushing, 1978; George, 1981). This spatial heterogeneity is influenced by the environment, biological interactions (e. g. predation, competition, aggregation), and behaviors (e. g. species and organismal dispersal). Consequently, point source measurements of biological characteristics might be representative of only a portion of a site or region that needs to be characterized and monitored. Therefore, it is imperative to quantify the spatial area that is represented by a point source measurement (i. e. representative range). The area represented in measurements by a stationary echosounder depends on the spatial variability of animal distributions within the study area. In other words, measurement uncertainty will be related to the rate at which the quantity of interest changes with distance from a point source measurement (e. g. Ellis and Schneider, 1997). Beyond the representative range of measurement, meaningful inferences can no longer be derived as uncertainty and interpretation errors are expected to increase. In contrast, multiple-point source measurements within the representative range of a point source are sub-optimal and better use of sampling resources can be made (e. g. lower cost or expand the sampling domain).
A better understanding of the representative range of a site will ensure an appropriate characterization and monitoring of biological communities, and at the same time, optimize the cost-effectiveness of remote monitoring through the deployment of one or multiple platforms.