© 2003 by ICES/CIEM International Council for the Exploration of the Sea/Conseil International pour l'Exploration de la Mer
Spatial dynamics of a nearshore, micronekton sound-scattering layer
a Department of Zoology, Hawaii Institute of Marine Biology, University of Hawaii PO Box 1106, Kailua, HI 96734, USA
b Hawaii Institute of Marine Biology PO Box 1106, Kailua, HI 96734, USA
*Correspondence to K. J. Benoit-Bird; tel: +1 808 247 5063; fax: +1 808 247 5831. e-mail: benoit{at}hawaii.edu.
In the Hawaiian Islands, there is a distinct resident community of micronekton, distributed along a narrow band where the upper underwater slopes of the islands meet the oceanic, mesopelagic environment. This mesopelagic boundary community serves as an important food resource to many animals. The goal of this work was to examine spatial heterogeneity of the Hawaiian mesopelagic boundary biomass at a range of scales, in the context of its diel vertical and horizontal migrations. A modified echosounder was used to sample the coasts of three Hawaiian Islands, permitting a range of scales from several meters to several kilometers to be assessed rapidly. The Hawaiian mesopelagic boundary community fits the hierarchical model of patch structure with patches within patches that are part of a larger-scale matrix of patches. Large differences in the overall distribution patterns of the mesopelagic boundary community exist along with a wide range of overall mesopelagic-animal densities. High animal-density locations have boundary-community layers with a large (kilometers) horizontal extent, and low animal-density locations have small (tens of meters), discrete patches. Higher animal-density locations are also more complex than low-density sites, with more levels of patchiness within the same range of spatial scales. Both time of day and distance from shore significantly affected the geometric and density characteristics as well as the distribution of aggregations within the boundary layer. Horizontal and vertical structures of the mesopelagic boundary community are also coupled. In high-density sites, there is strong vertical layering in acoustic-scattering strength while in low-density sites vertical acoustic structure is absent. The differences observed in the distribution of the mesopelagic boundary community at different levels of overall mesopelagic-animal density suggest biological forcing as the dominant mechanism. A description of heterogeneity in the mesopelagic boundary community in Hawaii is the first step in understanding its importance to both neritic and oceanic ecosystems and its potential for linking these two systems.
Keywords: mesopelagic boundary community, acoustic sampling, patchiness, heterogeneity, vertical structure, biological forcing, Hawaiian islands
Received 10 November 2002; accepted 5 March 2003.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussionReferences |
|---|
In ecological studies of the ocean, it has been found that its properties and populations are neither uniformly nor randomly distributed in space, but are in patterns that vary with time (Farquar and Holliday, 1977). All ecological systems exhibit heterogeneity at a broad range of scales, with patches (significant spatial variation; Downes, 1990) in oceanic biomass observed at scales of less than 1 m (Davis et al., 1991) to several kilometers (Mackas et al., 1985). The variability of an ecosystem in space and time is usually one of its most important features, influencing both practical problems of sampling and conceptual questions about its structure (Steele, 1976). The development and maintenance of spatial and temporal patterns and the consequences of those patterns for the dynamics of populations and ecosystems are fundamental themes in ecology (Levin, 1992). The consequences of these patterns of heterogeneity on biota are many, affecting population dynamics, trophic interactions, community organization and stability, and the cycling of elements (Levin, 1992).
Like most ecosystem components (Mangel and Adler, 1994), mesopelagic micronekton has been observed using net samples (Pearcy, 1971; Pearcy and Mesecar, 1971; Johnson, 1977) and submersible observations (Barham, 1971) to be clumped or patchy. Acoustic observations by Greenlaw and Pearcy (1985) showed that mesopelagic micronekton formed an extensive layer with an extent of 65 km or more. They concluded that inter-patch intervals on the level of the coarse patch found by Haury et al. (1978) are likely. For micronekton, patchiness is not merely a random variation in their environment, but is probably an essential requirement for obtaining adequate concentrations of food (Steele, 1976). However, while assessment of plankton patchiness at fine scales is common, specific information on the distribution of micronekton is limited primarily to the estimation of error.
Research on the causes of patchiness has focused on plankton, organisms that have been presumed to have little or no mobility, which have been modeled as passive particles. Numerous studies have demonstrated regulation of plankton by physical oceanographic processes (see reviews in Denman and Powell, 1984 and Legendre and Demers, 1984). The emphasis on passive particles has limited research on the biological causes of patchiness in the ocean, such as the causes of pattern categorized by Hutchinson (1961): reproductive, interactions between parents and offspring; social, intra-specific signaling between individuals; and coactive, intra-specific actions such as competition, predation, and parasitism. Because micronektonic animals are capable of large-scale movement, as shown by their large diel migrations, the behaviors of the animals are likely to be an important cause of patchiness in their distributions. Oceanographic forcing, i.e. currents, upwelling, tides, etc., may also affect micronekton patchiness indirectly by creating spatial variability in the richness and suitability of the environment, particularly of food resources.
In Hawaiian nearshore waters, there is an island-associated community of mesopelagic micronekton, termed the mesopelagic boundary community (Reid et al., 1991). This distinct resident community of micronekton is distributed along a narrow band where the upper slopes of the underwater portion of the islands meet the oceanic, mesopelagic environment. This community, comprising various species of fishes, shrimps, and squids, some of which undergo diel vertical and horizontal migrations (Benoit-Bird et al., 2001), has a composition unique to the boundary region (Reid et al., 1991). The mesopelagic boundary community is an important component of the coastal ecosystem in Hawaii, having been found to be important prey for tuna (He et al., 1997), billfish (Skillman, 1998), bottomfish (Haight et al., 1993), and spinner dolphins (Norris et al., 1994). Myctophids and other vertically migrating micronekton also appear to account for most of the consumption of zooplankton prey (Clarke, 1973). Patchiness in the mesopelagic community off Hawaii has been shown from the earliest studies. In 1969, Holton noted that lanternfishes, one of the primary components of the Hawaiian mesopelagic community have an extremely patchy distribution. More recent studies in Hawaii have also noted that large differences in abundance and distribution of mesopelagic biomass were observed over horizontal space (Reid, 1994). However, the inherent biases of the trawling studies used made it impossible to determine the scale of the patchiness.
Boundary communities, land-associated communities of specifically adapted, resident populations, are probably globally distributed (Reid et al., 1991). The spatial dynamics of boundary populations are important for understanding the effects of these communities on interactions between neritic and oceanic ecosystems. A description of spatial pattern is also critical in understanding the scope of intra- and inter-specific interactions and the reactions of individuals and communities to their physical environment (Haury et al., 1978). The goal of this work was to examine the spatial heterogeneity of the Hawaiian mesopelagic-boundary biomass at a range of scales, in the context of its vertical and horizontal migrations. Because trawling is not effective in assessing spatial and temporal heterogeneity of the boundary layer with fine resolution (Koslow et al., 1997), we used active acoustics to sample the distribution of the boundary layer off three Hawaiian Islands. Acoustic sampling permitted the rapid assessment of the distributional heterogeneity at a range of scales from several meters to several kilometers.
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Surveys
The coasts of three Hawaiian Islands were surveyed acoustically to determine the distribution of the mesopelagic-micronekton biomass at sites approximately 1 and 3 km from the shoreline. Off each island, transects parallel to shore were located 1.0–1.3 km from the shoreline (inshore) or 2.8–3.0 km from the shoreline (offshore)—these distances from the shore are called "distance classes" later. The survey vessel traveled 2.6 m s–1 (5 knots) over the bottom. Transects were surveyed for 1 h, at 1.5 h intervals between 1800 and 0400 h, local time. Sampling was conducted primarily during dark periods because the mesopelagic boundary community has a well-documented diel vertical migration, and would be too deep (400–1200 m) to be detected by the acoustic system during daylight hours (Struhsaker, 1973; Amesbury, 1975; Reid et al., 1991; Reid, 1994). Sampling took place during first and last light to measure the temporal boundaries of the community and to measure the background levels of the density of other organisms.
The Waianae coast of Oahu, Hawaii, was sampled acoustically between 1800 and 0100 h using a 10-m vessel, in the period May 1–8, 2001. One 9.3-km (5-nm) long transect was located at each of the two distance classes. The Kona coast of the Big Island of Hawaii was surveyed from the 50-m National Oceanic and Atmospheric Administration (NOAA) ship, Townsend Cromwell, during November 10–15, 1999 and June 1–5, 2001. Two offshore and three inshore transects off the Kona coast were sampled for 1 h at 3-h intervals during the first cruise and 90-min intervals during the second cruise between 2100 and 0400 h. The 68-m NOAA vessel, Ka'imimoana, was used to survey areas off the south and west shores of Lanai, between 15 and 18 January 2000. Two transects were 1.8 km (1 nm) and two were 5.6 km (3 nm) long. One of each transect length was located at each of the two distance classes. Each transect was surveyed for 1 h beginning at 2100, 2230, and 0000 h.
Off the Kona coast of Hawaii, transects oriented perpendicular to the shore were sampled between 2200 and 0000 h, and 0100 and 0300 h during the 1999 cruise. These transects began 1 km from the shoreline, and ended approximately 4 km from the shoreline, or at the outer edge of an encountered patch of mesopelagic animals. They covered the same 35 km along the shoreline as the parallel transects.
Acoustic data collection
Acoustic data were collected using a Computrol, Tournament Master Fishfinder NCC 5300 modified to read directly into a laptop computer (Benoit-Bird et al., 2001). The envelope of the echo was digitized at a sampling rate of 5 or 10 kHz using a Computer Board PC DAS16/12-AO or a Rapid System R1200. Data acquisition was initiated by the trigger of the echosounder and data were collected in blocks representing 156 m of the water column. The echosounder uses a 200 kHz outgoing signal with a pulse rate of 4 per second and pulse length of 130 µs. The transducer's signal is a downward pointing, 10-degree cone. The transducer was mounted on a towfish that "flew" 0.3 m beneath the surface of the water, regardless of the ship's speed over water. Beaufort Sea State was 0–2 during all sampling. See Benoit-Bird et al. (2001) and Benoit-Bird and Au (2001) for a description of the calibration procedures. The density of the mesopelagic animals was calculated with an echo-energy integration technique (MacLennan and Simmonds, 1992) as in Benoit-Bird et al. (2001).
Patches
The Webster method (Webster, 1973) was used to determine the edges of patches and gaps in the mesopelagic boundary community. The outermost boundary defined the patch off Hawaii and southern Oahu where the background density of mesopelagic animals was zero. Off northern Oahu and Lanai, where the background density was not zero, and for areas within primary patches off Hawaii and southern Oahu, internal boundaries determined the secondary, or internal heterogeneity, areas within the patch that are significantly patchy. A 2 m+2 m horizontal or vertical window was employed to determine discontinuities in the density (Figure 1). This technique averages the density observed in each 2-m window, reducing the statistical fluctuation in the scattered field, and then looks for differences between the windows with a t-test using the standard deviation of the entire sample as opposed to the standard deviation of each window to further stabilize the estimate (Legendre and Legendre, 1998). Areas of significantly higher density, equating to a difference of about 5 animals m–3 off each of the three islands, in either the horizontal (along-shore) or vertical (depth) were defined as boundaries. The boundaries were mapped in a geographic information system (GIS). This method resulted in sharp patch boundaries, allowing the patches to be easily determined using the centroid method (Legendre and Legendre, 1998). Unlike other spatial-averaging techniques, this "sliding window" method is relatively independent of the volume sampled by the transducer's beam. Rather than looking for the patch itself, this method looks only for the edges, making the scale of detectable patches determined by the size of the windows used and the intensity of the patch relative to the background, a factor which affects the statistical power. The choice of equal-size windows for both the vertical and horizontal directions provides different statistical power in the two dimensions, but it equates resolution for the determination of patch size. Because comparisons to determine horizontal edges were made between vertically adjacent areas, differences in statistical power due to the conical shape of the beam were negligible and were ignored. Vertical edges were determined within a single depth bin and no beam effects occur within a sample. However, the positions of vertical edges were corrected for beam effects as in Reid and Simmonds (1993). For both edge determinations, the use of the standard deviation of the entire sample helps reduce the differences in statistical power, caused by differences in sampling volume and the degree of autocorrelation, in edge determination between depths. Differences in the sampling intervals, however, make fluctuation in the horizontal edges more detectable than variation in the vertical edges (see Figure 1).
|
Patch characteristics for primary patches in the boundary community off Hawaii including mean density, maximum density, variance in density, maximum horizontal extent, maximum vertical extent, and distance to the nearest neighbor were measured with all horizontal measures presented being corrected for beam effects. Because of the consistent vertical extent of boundary-community patches at a given time and the distance from shoreline, the angle to the nearest neighbor for primary patches, which were all at the same depth was always zero. Also, because of the consistent vertical range of patches and patch shape, the area and circumference were simple mathematical permutations of horizontal and vertical extent and, because of co-linearity, were not tested. A multiple analysis of variance was used to test the effects of time, distance from shore, and island on patch-geometric and -density characteristics. Previous work had determined that spinner dolphins significantly affected both the geometric and density characteristics of boundary-community patches (Benoit-Bird and Au, 2003). Consequently, boundary-community patches containing spinner dolphins, as determined using the combination of passive acoustics, active sonar, and visual observations described in Benoit-Bird and Au (2003) were eliminated from the analysis.
Secondary heterogeneous regions (intra-patch heterogeneity) were characterized by their mean density, maximum density, variance in density, maximum horizontal extent, maximum vertical extent, area, perimeter, shortest distance to the nearest patch, and angle to the closest point in the nearest patch. Discontinuities within the boundary layer were characterized by their maximum horizontal extent, maximum vertical extent, distance to the nearest discontinuity, and angle to the closest point in the nearest discontinuity. Again, for gaps, area and perimeter were simple mathematical permutations of horizontal and vertical extent, and because of co-linearity, were not tested. Multiple analysis of variance was used to test the differences in patch characteristics between primary patches, areas of secondary heterogeneity, and gaps. Because animals within a patch and predators utilizing the patch experience all of its characteristics at once and may not distinguish between any one of them, all the characteristics were tested simultaneously using discriminant-function analyses. Within each level of heterogeneity, discriminant-function analysis was conducted first among all groups and then for only the effect of time and only the effect of distance from the shoreline, respectively.
The horizontal extent of patches and the secondary heterogeneity perpendicular from the shore were analyzed using single-factor analyses of covariance with time as a fixed factor and distance of the nearshore edge of the patch from the shoreline as a random factor. These results were compared with the along-shore horizontal sizes of the patches and sub-patches.
The distribution of patches
In order to look at the distribution of patches and secondary heterogeneous regions in the boundary community, distance to the nearest-neighbor methods was used with patch geometric centroid as the unit. Because acoustic sampling occurred beyond transects, a boundary strip of at least the expected distance to the first-nearest neighbor for a random distribution was included to limit bias. The Clark and Evans (1954) index of aggregation was calculated and the distribution was tested with Thompson's test (1956), a modified chi-squared test, at 
=0.05. This allows the distribution of patches to be tested as significantly greater than random, a uniform pattern, and significantly less than random, a clumped pattern. The distribution of the distances to the nearest-neighbors of primary patches off Hawaii and each level of secondary heterogeneity for all three islands for each group (time and distance from shore) were tested. The distribution between patch centroids and the second-nearest neighbor for secondary heterogeneity at all levels was tested in the same way. To look at the distribution only in the vertical and only in the horizontal dimension, the distribution of the centroids of patches in each direction was compared with the Poisson distribution using a coefficient of dispersion and the t-statistic. Bin size was 2 m vertically and 10 m horizontally. Primary patches off Hawaii could not be tested because each filled the entire vertical range of the boundary layer.
Vertical stratification
To look at vertical stratification in the scattering-layer strength, a single-factor ANOVA with 1-m depth bin as the factor was used. After a significant effect was found, Tukey's multiple-range statistic was used to determine which depth bins were homogeneous with neighboring depth bins (P>0.20) and which were significantly different (P<0.05). Adjacent 1-m bins that were not significantly different from each other, but were significantly different from depth bins surrounding the homogenous region, were considered to be a layer.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Spatial structure
Off northern Oahu and off Lanai, the boundary community extended in a layer over the surveyed area. There were discontinuities within the layer, but none penetrated through its entire vertical extent. Because the background density was not zero, patches were not distinct. Off the island of Hawaii and off southern Oahu, patches in the boundary community were distinct against a background density of zero (Figure 2).
|
Patches
Looking first at the primary patches in the boundary community found off the island of Hawaii and off southern Oahu, a multivariate ANOVA showed that all characteristics (mean density, maximum density, variance in density, maximum horizontal extent, maximum vertical extent, and distance to the nearest-neighbor) were significantly different (P<0.005 for all comparisons) as a function of distance from shoreline and time. There was also a significant interaction between time and the distance from shore for each boundary-community patch characteristic (P<0.001 for all comparisons). Only vertical extent was significantly different between southern Oahu and Hawaii (P<0.005). The changes in the characteristics of patches in the boundary community as a function of time and space are shown in Figure 3. Discriminant-function analyses correctly classified most patches in the boundary community (Table 1) with regard to first, all groups, second groups differentiated only by time, and third, groups differentiated only by distance from the shoreline.
|
|
Along-shore, horizontal extent of boundary-community patches changed significantly as a function of both time and distance from the shoreline (Figure 4). Boundary-community patches offshore were consistently about 25 m larger than inshore patches, though the size of both inshore and offshore boundary layer patches varied as a function of time. The horizontal extent of boundary-layer patches perpendicular from shore also varied significantly as a function of time and distance from shore (P<0.001, P<0.0001) with no significant interaction (P>0.05). The mean horizontal size of boundary-layer patches measured at the same time was approximately the same both along-shore and perpendicular to shore (Figure 5).
|
|
Gaps
Looking at the gaps found off the coasts of Lanai and northern Oahu, a multivariate ANOVA showed that there were significant differences in the vertical extent of the gaps in the boundary layer between the islands and as a function of distance from the shoreline and time (P<0.01 for all comparisons). The pattern of the changes in vertical extent of gaps matched the patterns in the vertical range of the entire layer. No other gap characteristics varied significantly (P>0.05). Discriminant-function analyses on gap characteristics as a function of time and distance from the shoreline, summarized in Table 1, reveal that gaps are misclassified more than 70% of the time. This is consistent with the hypothesis presented in Benoit-Bird and Au (2003) that gaps are the result of foraging by spinner dolphins that use the same foraging tactics regardless of time and distance from shore.
Secondary heterogeneity
Secondary, or internal, heterogeneity in the boundary community was found off all three Hawaiian islands surveyed. Off Hawaii and southern Oahu where the overall animal density was lower, two levels of boundary-community heterogeneity smaller than an individual patch were found. Off Lanai and northern Oahu where the density was high and the layer continuous, four levels of secondary heterogeneity were found within boundary-community layers. The spatial scales of primary and secondary patches in the boundary community are shown in Figure 6. Similar to primary patches, secondary heterogeneity 1 and 3 km from shore were not significantly different in their horizontal extent along-shore and perpendicular to shore (ANOVA, P>0.05, observed power=0.79). Along-shore, horizontal extent did not vary as a function of time or distance from shore although it did vary with island (Table 2).
|
|
The characteristics of secondary levels of heterogeneity in the boundary community were analyzed using ANOVAs and are summarized in Table 2. "Island" had a strong effect on the density characteristics of the first and second levels of boundary-layer heterogeneity as well as the horizontal and vertical size of these sub-patches. At the first level of boundary-community heterogeneity, similar differences were found between inshore- and offshore-sampling sites; density characteristics, vertical extent, and area varied significantly. At the second level, only "area" varies significantly between inshore- and offshore-sampling locations. At the third level of secondary heterogeneity, there are no characteristics that vary significantly between inshore and offshore sites. "Time" significantly affected the density characteristics and the vertical extent of the first and second levels of secondary heterogeneity since they varied significantly as a function of sampling time. At the third level, only maximum density varied significantly as a function of sampling time, and at the fourth level, maximum density and area varied significantly as a function of time. Discriminant-function analyses of each level of heterogeneity for each island are summarized in Table 1. Secondary heterogeneous regions in the boundary community at each island became categorized progressively less accurately with increasing levels of heterogeneity.
Distribution of patches
The distribution patterns of gaps, primary patches, and secondary heterogeneity of the boundary community are shown in Table 3. Gaps off northern Oahu and Lanai were significantly aggregated at all times, both inshore and offshore. Primary patches off Hawaii and southern Oahu were clumped, uniform, or random inshore, and clumped offshore. All levels of boundary-community secondary heterogeneity off Hawaii and southern Oahu were randomly or uniformly distributed inshore and clumped or random offshore. Inshore at both northern Oahu and Lanai, areas of secondary heterogeneity were either random or uniform. Offshore, level 1 secondary heterogeneity off southern Oahu and levels 1 and 2 off Lanai were significantly aggregated at all three sampling times. Offshore at Lanai and northern Oahu, areas of secondary heterogeneity at all other levels of heterogeneity were randomly distributed. When looking separately at vertical and horizontal distributions, aggregation or uniformity in the horizontal dimension accounted for all the patterns observed in both the distribution of primary patches and all levels of secondary heterogeneity of the boundary community. The distribution of distances to the second-nearest neighbor for all levels of secondary heterogeneity in the boundary community was significantly clumped (Thompson's test on Clark and Evans index of aggregation, P<0.05). In all cases, when vertical and horizontal distributions were assessed separately, only the horizontal, second-nearest neighbor distances were significantly clumped.
|
Vertical stratification
Vertical stratification in scattering-layer strength within the boundary community was not significant off the coasts of Hawaii and southern Oahu at any time either inshore or offshore. Off northern Oahu and Lanai, there was significant vertical stratification at all sampling times, both inshore and offshore (Figure 7). Strata ranged from 2 to 11 m high off Oahu and 2 to 15 m off Lanai. The number of significant strata varied from one to seven off Oahu, accounting for 30–62% of the entire layer. Off Lanai, the number of significant strata varied from five to seven and the stratified regions accounted for 40–52% of the entire depth range of the layer.
|
|
Discussion |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Acoustic surveys of the distribution of the Hawaiian mesopelagic boundary-community biomass show that the community fits the hierarchical model of patch structure proposed by Kotliar and Wiens (1990). The boundary community is found in a mosaic of patches within patches. A patch at a given scale has an internal structure of patchiness at a finer scale, and the mosaic containing that patch has a structure determined by patchiness at broader scales. Some experiments have been done to test the effects of hierarchical spatial distribution on pelagic animal behavior (see, for example, Thompson et al., 2001). However, measurements of animals living in three-dimensional environments distributed in these hierarchical-patch patterns are few (Ives et al., 1993; Fauchald et al., 2000). The information on the mesopelagic boundary community presented here will contribute to the understanding of hierarchical systems in the ocean, and because of the importance of this community to many predators, will be important to future development of applied foraging models.
The mesopelagic boundary community generally had high contrast with the background. Consequently, the outer boundaries of the boundary layer were discrete. Interestingly, the boundaries tended to be straight. This was true at the top and bottom edges of the layer that were at consistent depths within each sampling time. It was also true for the vertical edges which were found to vary in space by only a few meters in all cases. This is different from the characteristics of the edges of layers of mesopelagic micronekton observed by Greenlaw and Pearcy. (1985). The edges in volume-backscattering distributions that they observed tended to taper off relatively gradually, leaving indistinct edges that wavered on all sides of the layer.
The distance between patches in the boundary community ranged from 10 to 70 m (17–117 echoes) with no depth effect. In order for patches or other aggregations to be effective against predators, they must be farther apart than the perceptual distance of a predator (Brock and Riffenburgh, 1959). Predators of the boundary community are primarily large, pelagic animals with well-developed vision. Some of its predators are marine mammals that probably utilize echolocation in prey searching (Norris and Dohl, 1980). The distance between patches in the Hawaiian mesopelagic boundary community is much less than the sensory range of such predators. The distribution of boundary-community patches is probably beneficial to its predators. The costs of foraging in these closely spaced prey patches would be reduced because the intensity (density relative to the background density) of the patch represents the return to the predator, while the distance between patches represents the costs of obtaining the prey (Brock and Riffenburgh, 1959). At least one of the boundary-community's predators, Stenella longirostris, specifically exploits the characteristics of boundary-community patches (Benoit-Bird and Au, 2003).
Variation in the geometric and density characteristics of patches and sub-patches in the boundary community, as a function of sampling time and distance from the shoreline, were predictable at the largest scales and unpredictable at the smallest scales observed. Time affected the characteristics of spatial patterns more than distance from the shoreline. For both patches and sub-patches, geometric characteristics were more important than density characteristics for correctly classifying boundary-community distributions into inshore versus offshore groups. Offshore patches and sub-patches were consistently larger, both along and across the slope, than inshore patches. Density characteristics were more important for separating these aggregations by sampling time. The high densities found nearshore, probably due to packing as the mesopelagic animals avoid the bottom, were highly predictable. These differences in aggregation characteristics as a function of the migrations of the boundary community are consistent between islands and times of the year. This predictability probably affects the effectiveness of mesopelagic animals as predators and influences their population dynamics. The predictability also potentially allows predators to appropriately alter their foraging strategies to match the distributional patterns of their prey.
The overall distribution pattern of the mesopelagic boundary community at Lanai and northern Oahu, where the boundary community was distributed in a layer, was very different from the pattern off Hawaii and southern Oahu where it was distributed in discrete patches. Because these differences existed along a single shoreline off Oahu, they were not associated specifically with inter-island differences. Distributional changes were associated with changes in animal density—locations with high animal density have boundary-community layers, and locations with low animal density have patches. Off Oahu, there appeared to be a density threshold at which animal distribution changes. Instead of being spread out evenly with reduced overall density when animal abundance was lower, the animals in the boundary community were aggregated in patches of higher density. The reasons for this aggregation are probably related to the constraints of both prey capture and predator avoidance (Brock and Riffenburgh, 1959; Krause and Godin, 1995; Boyd, 1996).
Within the high-density layers found off Lanai and northern Oahu, narrow, vertical gaps were commonly found approximately 25 m apart (Benoit-Bird and Au, 2003). In most marine systems, gaps in the distribution of organisms usually arise from interactions between topography and a combination of physical and biological processes including physical advection, vertical-migration behavior, and predation (see, for example, Isaacs and Schwartlose, 1965; Genin et al., 1994). It has generally been concluded that physical rather than biological processes dominate gap formation in oceanic systems (Greene et al., 1994). However, the gaps in the Hawaiian mesopelagic boundary community in the present study are extremely small—less than a few meters wide—and much smaller than gaps formed primarily as a result of topography (Haury et al., 2000). Within the mesopelagic boundary layer in Hawaii, we have observed a strong association between gaps and dolphin predators. Spinner dolphins feed cooperatively and form distinct, high-density areas within the prey field (Benoit-Bird and Au, 2003). The observations of the effects of dolphin predation on the boundary layer, the lack of correlation of gaps with topographic features, and the ability of mesopelagic-boundary animals to move kilometers despite tides and currents (Benoit-Bird et al., 2001) suggest that predation, not physical forcing, dominates the formation of gaps.
Differences in overall mesopelagic-animal density affected not only the scales at which the boundary community was found (larger scales with greater density) but also the level of complexity. Northern Oahu and Lanai, the higher-density sites, had more complex spatial structure in the mesopelagic boundary community than Hawaii or southern Oahu, the lower-density sites. The high-density sites had more total levels of patchiness than the low-density sites and more levels of patchiness within the same range of spatial scales.
The horizontal and vertical structures of the Hawaiian mesopelagic boundary community are coupled. The large differences in distribution observed as a function of overall animal density are also correlated with differences in vertical-acoustic structure. In the lower-density sites of southern Oahu and Hawaii, vertical-acoustic structure was absent. In the higher-density sites of northern Oahu and Lanai, strong, vertical layering in acoustic-scattering strength was evident inshore and offshore, at all sampling times. Vertical layering in acoustic-scattering strength of midwater animals has also been observed in the Gulf of Mexico and Caribbean Sea (Baird and Wilson, 1977). While it is impossible to identify species from acoustic scattering alone, and density differences may confound differences in acoustic scattering, it appears that vertical layering in the mesopelagic layer is related to species composition and size differences. For example, video-work and trawling with the Multiple Opening and Closing Net Environmental Sensing System (MOCNESS) carried out simultaneously with acoustic surveys show that shrimp species are found more often in deeper layers (Benoit-Bird et al., 2001). The vertical structure of the boundary layer, or lack of it, probably affects the foraging tactics employed by predators and influences the population dynamics within the boundary community.
Differences in the overall mesopelagic-animal density of the boundary community also affected the spacing of aggregations. Both patches and sub-patches were more likely to be aggregated at lower densities and uniformly distributed at higher densities. Uniform spacing is usually associated with biological phenomena (Sandulli and Pinckney, 1999). Of course, the conclusions about uniformity are usually applied to individuals, not aggregations of individuals. However, the spacing of a cluster of animals potentially working in concert is probably related to causes similar to those underlying the distribution of individual animals. The aggregation of aggregations, rather than a random distribution of these patches, is probably related to the same factors as those affecting the overall distribution of animals at lower densities, predation efficiency and predator avoidance.
The patterns observed in the distribution of the mesopelagic boundary community as a function of overall mesopelagic-animal density suggest that biological forcing mechanisms are the dominant influence. The biological mechanisms probably act in concert with, and are indirectly affected by, physical forcing. The focus of study by oceanographers on spatial patterning of plankton in the ocean has led to an emphasis on physical forcing as the underlying process producing patchiness (but see Folt and Burns, 1999). For example, in a review of patchiness in the ocean and its causes, less than one page was devoted to the description of biological causes, while more than ten pages were dedicated to physical causes (Steele, 1976). Competition, predation, and parasitism, categorized as "coactive processes" by Hutchinson (1961), are still by far the least understood of the biological mechanisms of patch formation at any spatial or temporal scale (Haury et al., 1978). The boundary community is found in densities so high as to cause significant negative consequences in respiration and resource acquisition for individual animals. It is also a significant prey resource for many predators (Norris and Dohl, 1980; Haight et al., 1993; He et al., 1997; Skillman, 1998). Coactive forces are probably the primary forces acting within the Hawaiian mesopelagic boundary layer. It is clear that animals capable of movement over great distances despite the physical environment are capable of driving spatial patterning at both small and large spatial scales. Communities of mesopelagic micronekton, particularly those found in the accessible and relatively heterogeneous boundary region, are good systems in which to study the causes of patchiness in the ocean in organisms larger than plankton. Such study will probably clarify the importance of biological processes in structuring the distribution of biomass in the ocean. However, future work on underlying patch forcing will undoubtedly require information on species distributions. To study competition and resource utilization, we need to know how similar patches are with respect to species composition.
The variety and sheer number of predators that forage upon the mesopelagic boundary community in Hawaii (Norris and Dohl, 1980; Haight et al., 1993; He et al., 1997; Skillman, 1998) indicate the importance of this community to both nearshore and open-ocean predators. However, patchiness in the Hawaiian mesopelagic boundary community has probably affected previous estimates of its abundance and biomass (Clarke, 1973; Reid, 1994), impacting our views of the community's importance. Because aggregated spatial pattern is such a widespread phenomenon, there is little question of its ecological and evolutionary importance (Haury et al., 1978). An understanding of the spatial pattern in the mesopelagic boundary community has the potential to help us understand the adaptations these animals have undergone to live in the narrow zone between the oceanic and neritic systems. Ecological processes often depend on the pattern as well as the absolute density of resources (Mangel and Adler, 1994). Consequently, a description of heterogeneity in the mesopelagic boundary community in Hawaii is the first step in understanding its importance to both neritic and oceanic ecosystems and its potential for linking these two systems.
|
Acknowledgements |
|---|
We would like to thank the National Marine Fisheries Service's Honolulu Laboratory for providing ship-time aboard the Townsend Cromwell and the Sustainable Seas Expedition sponsored by the National Oceanic and Atmospheric Administration and National Geographic for providing ship-time aboard the Ka'imimoana. We thank Rusty Brainard for fostering our collaboration with the NMFS Laboratory, providing us the opportunity to be aboard the Cromwell, and Chief Scientist Robert Humphreys for working with us to maximize the accomplishment of both his and our cruise objectives. The officers and crew of the Townsend Cromwell and Ka'imimoana provided excellent scientific support, especially Phil. White. We would also like to thank Chris. Bird, Lisa Davis, and Marc Lammers, for assistance with field work. George Losey, Paul Nachtigall, Craig Smith, and James Parrish made helpful comments on earlier drafts of this manuscript. The OceanWide Science Institute provided funding for vessel-time off Oahu. This paper is funded by a grant from the National Oceanic and Atmospheric Administration, Project R/FM-7, which is sponsored by the University of Hawaii Sea Grant College Program, SOEST, under Institutional Grant No. NA86RG0041 from NOAA Office of Sea Grant, Department of Commerce. The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its sub-agencies. UNIHI-SEAGRANT-JC-02-15. This is HIMB contribution 1152.
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
-
Amesbury S. S. (1975) The vertical structure of the midwater fish community off leeward Oahu, Hawaii. PhD dissertation, University of Hawaii.
Baird R.C and Wilson D.F. (1977) Sound scattering and oceanic midwater fishes. In Anderson N.R and Zahuranec B.J (Eds.). Oceanic Sound Scattering Prediction(Plenum Press, New York) pp. 549–563.
Barham E.G. (1971) Deep sea fishes: lethargy and vertical orientation. In Farquar G.B (Ed.). Biological Sound Scattering in the OceanU.S. Government Printing Office pp. 100–118.
Benoit-Bird K.J and Au W.W.L. (2001) Target strength measurements of animals from the Hawaiian mesopelagic boundary community. Journal of the Acoustical Society of America 110:812–819.[CrossRef][Web of Science]
Benoit-Bird K.J and Au W.W.L. (2003) Prey dynamics affect foraging by a pelagic predator (Stenella longirostris) over a range of spatial and temporal scales. Behavioral Ecology and Sociobiology 53:364–373.[Web of Science]
Benoit-Bird K.J, Au W.W.L, Brainard R.E, Lammers M.O. (2001) Diel horizontal migration of the Hawaiian mesopelagic boundary community observed acoustically. Marine Ecology Progress Series 217:1–14.[Web of Science]
Boyd I.I. (1996) Temporal scales of foraging in a marine predator. Ecology 77:426–434.[CrossRef][Web of Science]
Brock V.E and Riffenburgh R.H. (1959) Fish schooling: a possible factor in reducing predation. Journal du Conseil-Conseil International pour L'exploration de la Mer 25:301–317.
Clarke T. (1973) Some aspects of the ecology of lanternfishes in the Pacific Ocean near Hawaii. Fishery Bulletin 71:401–434.[Web of Science]
Clark P.J and Evans F.C. (1954) Distance to nearest neighbor as a measure of spatial relationships in populations. Ecology 35:445–453.[Medline]
Davis C.S, Flierl G.R, Wiebe P.H. (1991) Micropatchiness, turbulence and recruitment in plankton. Journal of Marine Research 49:109–151.[Web of Science]
Denman K.L and Powell T.M. (1984) Effects of physical processes on planktonic ecosystems in the coastal ocean. Oceanography and Marine Biology Annual Review 22:125–168.
Downes B.J. (1990) Patch dynamics and mobility of fauna in streams and other habitats. Oikos 59:411–413.[CrossRef][Web of Science]
Farquar G.B and Holliday D.V. (1977) Recommendations of the working group on bioacoustics. In Anderson N.R and Zahuranec B.J (Eds.). Oceanic Sound Scattering Prediction(Plenum Press, New York) pp. 483–491.
Fauchald P, Erikstad K.E, Skarsfjord H. (2000) Scale dependent predator–prey interactions: the hierarchical spatial distribution of seabirds and prey. Ecology 81:773–783.[CrossRef][Web of Science]
Folt C.L and Burns C.W. (1999) Biological drivers of zooplankton patchiness. Trends in Ecology and Evolution 14:300–305.[CrossRef]
Genin A, Greene C.H, Haury L, Wiebe P, Gal G, Kaartvedt S, Meir E, Fey C, Dawson J. (1994) Zooplankton patch dynamics: daily gap formation over abrupt topography. Deep-Sea Research, Part A 41:941–951.
Greene C.H, Wiebe P.H, Zamon J.E. (1994) Acoustic visualization of patch dynamics in oceanic ecosystems. Oceanography 7:4–12.
Greenlaw C.F and Pearcy W.G. (1985) Acoustical patchiness of mesopelagic micronekton. Journal of Marine Research 43:163–178.[Web of Science]
Haight W.R, Parrish J.D, Hayes T.A. (1993) Feeding ecology of deepwater lutjanid snappers at Penguin Bank, Hawaii. Transactions of the American Fisheries Society 122:328–347.[CrossRef]
Haury L.R, McGowan J.A, Wiebe P.H. (1978) Patterns and processes in the time-space scales of plankton distributions. In Steele J.H (Ed.). Spatial Pattern in Plankton Communities(Plenum Press, New York) pp. 277–328.
Haury L, Fey C, Newland C, Genin A. (2000) Zooplankton distribution around four eastern North Pacific seamounts. Progress in Oceanography 45:69–105.[CrossRef][Web of Science]
He X, Bigelow K.A, Boggs C.H. (1997) Cluster analysis of longline sets and fishing strategies within the Hawaii-based fishery. Fisheries Research 31:147–158.[CrossRef][Web of Science]
Holton A.A. (1969) Feeding behavior of a vertically migrating lanternfish. Pacific Science 23:325–331.[Medline]
Hutchinson G.E. (1961) The paradox of plankton. American Naturalist 95:137–145.[CrossRef]
Isaacs J.D and Schwartlose R.A. (1965) Migrant sound scatterers: interactions with the seafloor. Science 150:1810–1813.
Ives A.R, Kareiva P, Perry R. (1993) Response of a predator to variation in prey density at three hierarchical scales: lady beetles feeding on aphids. Ecology 74:1929–1938.[CrossRef][Web of Science]
Johnson R.K. (1977) Acoustic estimation of scattering-layer composition. Journal of the Acoustical Society of America 61:1636–1639.[CrossRef][Web of Science]
Koslow J.A, Kloser R, Williams A. (1997) Pelagic biomass and community structure over the mid-continental slope off southeastern Australia based upon acoustic and midwater trawl sampling. Marine Ecology Progress Series 146:21–35.[Web of Science]
Kotliar N.B and Wiens J.A. (1990) Multiple scales of patchiness and patch structure: a hierarchical framework for the study of heterogeneity. Oikos 59:253–260.[CrossRef][Web of Science]
Krause J and Godin J.G. (1995) Predator preferences for attacking particular prey group sizes: consequences for predator hunting success and prey predation risk. Animal Behaviour 50:465–473.[CrossRef][Web of Science]
Legendre L and Demers S. (1984) Towards dynamic biological oceanography and limnology. Canadian Journal of Fisheries and Aquatic Sciences 41:655–672.
Legendre P and Legendre L. (1998) Numerical Ecology: Developments in Environmental Modeling(Elsevier, New York) 20: 853 pp.
Levin S.A. (1992) The problem of pattern and scale in ecology. Ecology 73:1943–1967.[CrossRef][Web of Science]
Mackas D.L, Denman K.L, Abbott M.R. (1985) Plankton patchiness: biology in the physical vernacular. Bulletin of Marine Science 37:652–674.[Web of Science]
MacLennan D.N and Simmonds E.J. (1992) Fisheries Acoustics(Chapman and Hall, New York).
Mangel M and Adler F.R. (1994) Construction of multi-dimensional clustered patterns. Ecology 75:1289–1298.[CrossRef][Web of Science]
Norris K.S and Dohl T.P. (1980) Behavior of the Hawaiian spinner dolphin, Stenella longirostris. Fishery Bulletin 77:821–849.[Web of Science]
Norris K.S, Würsig B, Wells R.S, Würsig M. (1994) The Hawaiian Spinner Dolphin(University of California Press, Berkeley, CA).
Pearcy W.G. (1971) Variations in the abundance of sound scattering animals off Oregon. In Farquar G.B (Ed.). Biological Sound Scattering in the OceanU.S. Government Printing Office pp. 647–666.
Pearcy W.G and Mesecar R.S. (1971) Scattering layers and vertical distribution of oceanic animals off Oregon. In Farquar G.B (Ed.). Biological Sound Scattering in the OceanU.S. Government Printing Office pp. 381–394.
Reid S. B. (1994) Spatial structure of the mesopelagic fish community in the Hawaiian boundary region. PhD dissertation, University of Hawaii.
Reid D.G and Simmonds E.J. (1993) Image analysis techniques for the study of fish school structure from acoustic survey data. Canadian Journal of Fisheries and Aquatic Sciences 50:886–893.
Reid S.B, Hirota J, Young R.E, Hallacher L.E. (1991) Mesopelagic boundary community in Hawaii: micronekton at the interface between neritic and oceanic ecosystems. Marine Biology 109:427–440.[CrossRef]
Sandulli R and Pinckney J. (1999) Patch sizes and spatial patterns of meiobenthic copepods and benthic microalgae in sandy sediments: a microscale approach. Journal of Sea Research 41:179–187.[CrossRef]
Skillman R. A. NMFS. (1998) . Central Pacific swordfish, Xiphias gladius, fishery development, biology, and research. 101–124 NOAA Technical Report 142.
Steele R. A. (1976) Patchiness. In Cushing D.H and Walsh J.J (Eds.). The Ecology of the Seas(W.B. Saunders, Philadelphia) pp. 98–115.
Struhsaker P. (1973) A contribution to the systematics and ecology of Hawaiian bathyal fishes. Unpublished PhD dissertation, University of Hawaii.
Thompson H.R. (1956) Distribution of distance to the nth neighbour in a population of randomly distributed individuals. Ecology 37:391–394.[Medline]
Thompson A.R, Petty J.T, Grossman G.D. (2001) Multi-scale effects of resource patchiness on foraging behaviour and habitat use by longnose dace, Rhinichthys cataractae. Freshwater Biology 46:145–160.[Medline]
Webster R. (1973) Automatic soil-boundary location from transect data. Mathematical Geology 5:27–37.[Medline]
Zar J.H. (1999) Biostatistical Analysis(Prentice Hall, Upper Saddle River, NJ) 663 pp.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. J. Benoit-Bird The effects of scattering-layer composition, animal size, and numerical density on the frequency response of volume backscatter ICES J. Mar. Sci., April 1, 2009; 66(3): 582 - 593. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||