ICES Journal of Marine Science (2011), 68(9), 1954–1962. doi:10.1093/icesjms/fsr106
Utility of a spatial habitat classification system as a surrogate of marine benthic community structure for the Australian margin Rachel Przeslawski 1*, David R. Currie 2, Shirley J. Sorokin 2, Tim M. Ward 2, Franziska Althaus3, and Alan Williams 3 1
Marine and Coastal Environment Group, Geoscience Australia, GPO Box 378, Canberra, ACT 2601, Australia South Australian Research and Development Institute, Aquatic Sciences, PO Box 120, Henley Beach, South Australia 5022, Australia 3 CSIRO Wealth from Oceans Flagship, Marine Laboratories, GPO Box 1538, Hobart, Tasmania 7001, Australia 2
*Corresponding Author: tel: +61 2 6249 9101; fax: +61 2 6249 9920; e-mail:
[email protected]. Przeslawski, R., Currie, D. R., Sorokin, S. J., Ward. T. M., Althaus, F., and Williams, A. 2011. Utility of a spatial habitat classification system as a surrogate of marine benthic community structure for the Australian margin. – ICES Journal of Marine Science, 68: 1954 –1962. Received 14 October 2010; accepted 18 May 2011; advance access publication 13 July 2011.
Keywords: benthic invertebrates, continental margin, interpolation, seascapes, sponge, surrogacy.
Introduction One of the greatest challenges facing marine researchers and managers is the quantification of marine biodiversity to identify and conserve potentially unique, representative, or vulnerable habitats and species (Edyvane, 1999; Currie et al., 2008; UNESCO, 2009). In addition, the quantification of biodiversity is of interest to offshore industries in terms of their compliance with environmental regulation (Heap et al., 2010). However, even in relatively wellstudied regions, the distribution and community composition of certain taxa (e.g. infauna, meiofauna) or local habitats (e.g. deep sea) remain unknown (Poore, 1995). Although improved technology and protocols have facilitated biological sampling, it is not possible to catalogue all marine life using direct sampling techniques. To that end, surrogacy research, in which environmental variables are used to predict the patterns of marine biodiversity, provides a promising avenue (reviewed by McArthur et al., 2010). The best combined datasets for environmental data often reside in national data repositories. For example, global information on primary productivity and sea surface temperature are available through satellite imagery from NOAA’s (National Oceanic and Atmospheric Administration) Office of Satellite Data Processing and Distribution (www.osdpd.noaa.gov), and publicly available information on the geomorphology, bathymetry, and sediment Crown Copyright
#
2011. Published by Oxford University Press.
composition for the Australian Exclusive Economic Zone is available through Geoscience Australia’s Marine Sediments Database (www.ga.gov.au/oracle/mars). Such national data repositories are very useful in developing national habitat maps, e.g. the British Isles (Connor et al., 2004), and habitat classification systems derived on smaller regional or local scales (Greene et al., 2007). Using national physical datasets, Geoscience Australia has developed a continental-scale approach to represent Australia’s seabed habitats, termed seascapes. In this approach, environmental data are interpolated and combined to form discrete seascapes, each of which corresponds to benthic marine habitats with similar environmental characteristics (Harris, 2007; Whiteway et al., 2007; Heap et al., 2011). These environmental factors are known to affect some biological communities (McArthur et al., 2010), but the usefulness of the seascapes approach for representing marine benthic biodiversity remains poorly understood. Recent results indicate that the utility of seascapes to differentiate marine benthic communities varies among regions and may be greater in more-homogenous ecosystems than highly heterogeneous ecosystems (Heap et al., 2011). Here, we use biological data collected from two surveys including the Great Australian Bight (GAB) at the local (tens of
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
This study tests whether a continental-scale classification of Australian benthic habitats (termed “seascapes”) and the interpolated environmental data from which they are derived are useful as abiotic surrogates of biodiversity at a local [tens of kilometres, Great Australian Bight (GAB)] and regional scale [hundreds of kilometres, Western Australian (WA) margin]. Benthic invertebrate community structure is moderately associated with specific seascapes in both the GAB (R ¼ 0.418) and WA margin (excluding hard substrata, R ¼ 0.375; all substrata, R ¼ 0.313). Mud content, seafloor slope, and seafloor temperature are significantly correlated with invertebrate communities at both scales, with disturbance and primary production correlated with GAB communities. Seascapes are not consistently useful surrogates because the strength and significance of relationships between seascapes and community structure differs among seascapes, regions, and spatial scales. Nevertheless, a national system of seascapes is an appropriate surrogate for broad-scale benthic invertebrate community patterns when biological data are limited, provided the uncertainty is acknowledged and, where possible, an assessment made of each seascape’s ability to differentiate biological communities. Further refinement of seascape derivations may include updated and additional environmental data (particularly for hard vs. soft substrata) and validation among biological datasets from a range of habitats and scales.
A spatial habitat classification system as a surrogate of marine benthic community structure
1955
kilometres) scale and the Western Australian margin (WA margin) at the regional (hundreds of kilometres) scale. These datasets are suited to detecting biophysical patterns because they are based on benthic invertebrate megafauna, which are sessile or have limited mobility. In addition, sampling methods have been standardized within each survey, there was rigorous species-level identification of collections, and both sampling areas have been identified as regions of relatively high biodiversity (Williams et al., 2001, 2010; Ward et al., 2006; Currie et al., 2009). The GAB supports one of the world’s most diverse softsediment epibenthic systems (Ward et al., 2006), although its infaunal assemblages do not seem to be as diverse as its epifauna (Currie et al., 2009). The GAB also has an extraordinary level of endemism (Womersley, 1990; Poore, 1995; Sorokin et al., 2007) and a high species richness of some groups (Edyvane, 1999). The WA margin extends 1500 km through subtropical and temperate latitudes and supports diverse faunas of megabenthic invertebrates (Williams et al., 2010) and fish (Williams et al., 2001). New continental-scale environmental datasets were used to build on the findings of Ward et al. (2006), Currie et al. (2009), and Williams et al. (2010), specifically to investigate whether patterns of benthic community structure identified by the earlier studies are reflected in the seascapes defined for the study regions. Ward et al. (2006) and Currie et al. (2009) focused on
the western GAB, with intensive sampling across the continental shelf (,200 m; Figure 1). Species composition of benthic invertebrates in the GAB varied among sediment types, with species richness and biomass negatively correlated with depth and mud content (Ward et al., 2006). Infaunal assemblages in the area were also correlated with depth, oxygen saturation, and chlorophyll concentration, although mud content, seafloor temperature, turbidity, sediment grain size, and sorting were all unrelated (Currie et al., 2009). The study area of Williams et al. (2010) extended over most of Australia’s western margin with sparse sampling across depth at 18 latitude intervals. Williams et al. (2010) also found that the distribution of benthic invertebrates was affected by several environmental variables, but the magnitude of the relationships varied according to the spatial scale examined. At large spatial scales (more than hundreds of kilometres), seafloor temperature, oxygen concentration, and latitude were the driving abiotic factors, but at smaller spatial scales (tens of kilometres), the substratum type was the main factor influencing community structure (Williams et al., 2010). We tested whether the national-scale seascapes approach and the interpolated environmental data from which they are derived are useful abiotic surrogates for marine benthic biodiversity, defined as the community structure of epibenthic invertebrates at regional and local scales. The results of the study should facilitate future methodological improvement, including the
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
Figure 1. Location of the study areas overlaid on national seascape classifications. Yellow dots show stations on the WA margin, and green dots stations from the GAB. Seascapes are differentiated with distinct colours and numbers, as defined in Whiteway et al. (2007).
1956
R. Przeslawski et al.
identification of additional environmental data layers for deriving the seascapes. The work will also complement previous research in which dynamic variables related to ecological processes were used as surrogates (e.g. biological productivity in McArthur et al., 2010), rather than the static variables used here. In addition, the results will contribute to a greater understanding of national biodiversity patterns and the utility of national habitat classification systems based on the seascapes approach, as well as helping inform management of Australia’s marine jurisdiction.
Methods Biological data
Statistics Abiotic data Seascapes were available for Australia’s EEZ from the following two derivations, both based on cell sizes of 0.05 decimal degrees: on-shelf (0–200 m water depth) and off-shelf (.200 m depth). Derivations were based on national interpolations of the following
Multivariate analyses of the biomass of each species were undertaken using non-metric multidimensional scaling plots (n-MDSs) and analyses of similarities (ANOSIMs) to investigate differences in community structure among the seascapes. Biological data were square-root transformed to reduce the effect
Table 1. National on- and off-shelf seascapes for which survey stations overlapped, with characterizations relative to other on- or off-shelf seascapes at a national scale. Number of samples Seascape system On-shelf
Seascape number 1
On-shelf On-shelf On-shelf
4 6 7
On-shelf Off-shelf Off-shelf Off-shelf Off-shelf
8 1 3 5 9
Characterized by Moderate depth and productivity, flat seafloor, slightly gravelly mud Moderate depth, steep slope, gravel, low productivity Very deep water, very steep slope, muddy gravel, low productivity Deep, moderate temperature, very low disturbance and productivity Moderate depth, gravelly, moderate temperature Shallow, steep, cold, very low productivity Shallow, steep, moderate temperature Moderate depth, steep, very cold, very high productivity Moderate depth, steep, very cold, moderate productivity
GAB 30
WA (excluding hard) 0
WA (all) 0
6 4 0
0 17 0
5 21 5
0 0 0 0 0
4 0 43 0 10
7 4 55 4 10
GAB, Great Australian Bight study area; WA, Western Australian margin study area. Seascape numbers and characterizations are from Geoscience Australia (Whiteway et al., 2007; Heap et al., 2011).
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
For the GAB, specimens were collected from standard 500 m epibenthic sled tows at 40 stations, according to the methods described in Currie et al. (2008; Figure 1). The sled was based on the CSIRO Seamount, Epibenthic Sled (SEBS) design (Lewis, 1999), weighed 160 kg, had dimensions of 0.6 × 1.0 × 2.0 m, and was fitted with a collection bag of 50 mm mesh. Samples from each station were weighed, bagged, and frozen on board, then later sorted and identified to species or putative taxon. Voucher specimens were lodged at the South Australian Museum, Adelaide. For the WA margin, specimens were collected from an epibenthic beam trawl with a mouth width of 4 m or a CSIRO SEBS with a mouth width of 1.2 m (http://www.cmar. csiro.au/research/seamounts/epibenthic.htm); earlier analyses showed that combining standardized data from these two gear types did not influence the results (Williams et al., 2010). Samples were sorted to operational taxonomic units, counted, weighed, and preserved on board. In the laboratory, the samples from seven major groups (sponges, octocorals, echinoderms, molluscs, ascidians, sea spiders, and decapods) were identified to putative species by expert taxonomists. The counts and weights were standardized to area sampled using the mouth width of the gear and the distance of tow (Williams et al., 2010). At least four stations in the entire WA margin survey had to fall within each seascape for the data to be included.
environmental parameters: depth, mud content, gravel content, seafloor temperature, slope, primary production, and effective disturbance (not available for off-shelf seascapes; see Heap et al., 2011, for a full description of the methods). Effective disturbance is a univariate index derived from the magnitude and frequency of bed-shear stress (Hemer, 2006); such indices provide a measure of the overall natural disturbance a biological community may experience over time (Hughes et al., 2011). The GAB study area overlapped three on-shelf seascapes, and the WA margin study area overlapped four on- and four off-shelf seascapes (Table 1). To investigate the underlying environmental factors that may be driving the biological application of seascapes, biological data were analysed with all seven individual abiotic factors from which on-shelf seascapes are derived. For each location at which biological data were collected, the Hawths Tool in ArcGIS v. 9.3 was used to extract environmental data from interpolated spatial layers of the national on-shelf seascapes analysis, as described in Whiteway et al. (2007). As such, the resulting environmental data did not necessarily reflect the real environmental conditions at a location, but rather the interpolated values defined by methods used to derive the seascapes. These interpolated values were used to identify the individual factors in seascapes that were important drivers of biological assemblages in the GAB (and conversely, those that were not), to inform future refinements of seascapes and other national habitat approaches (Huang et al., 2011). Interpolated values were related to the measurements of several factors recorded during biological sampling of the GAB, as confirmed by linear regression (depth, r 2 ¼ 0.9260; mud content, r 2 ¼ 0.9468; gravel content, r 2 ¼ 0.8472), hence supporting the utility of these interpolations at scales of tens to hundreds of kilometres. In addition to the environmental variables from which the seascapes were derived, three categorical substratum classes (hard, soft, mixed) were assigned to the WA margin samples based on visual identification from multibeam sonar backscatter maps (Williams et al., 2010).
1957
A spatial habitat classification system as a surrogate of marine benthic community structure
Results Benthic invertebrate community structure showed a moderate association with certain seascapes in both the GAB (ANOSIM, r ¼ 0.418, p , 0.001) and the WA margin (excluding the hard substratum, r ¼ 0.375, p , 0.001; all substrata, r ¼ 0.318; Figure 2, Table 3). In the GAB, communities in seascape 1 were significantly different from communities in both seascapes 4 and 6 (Table 3), with the n-MDS confirming a distinct grouping of points from seascape 1 (Figure 2a), which generally showed greater species richness and higher biomass than the other two seascapes (Figure 3). Visually, seascapes 4 and 6 grouped together (Figure 2a), but the ANOSIM revealed no significant pairwise
differences (Table 3), nor were any obvious differences detected in species richness or biomass (Figure 3), possibly because of the small number of samples in these seascapes and a potential outlier from station 45 in seascape 4 (Figure 2a). Considering the WA margin samples from soft and mixed substrata, most pairs of seascapes were significantly different from each other, with r-values .0.4 for the three pairwise comparisons between on- and off-shelf seascapes (Table 3). This reflected differences in species richness and biomass, especially between on- and off-shelf seascapes (Figure 4a and b). This on- and off-shelf separation was also visible in the MDS (Figure 2b). When the WA margin stations encompassing a hard substratum were included in multivariate analyses, the r-values for the overall test were lower, although still significant. The relationships between the seascapes and the community structure were less clear (Figure 2c), in particular for the on-shelf seascapes, where most of the hard substratum samples were taken (Table 1). Most significant pairwise comparisons of the community structure were again between on- and off-shelf seascapes (Table 3). Among on-shelf seascapes, only seascape 4 was significantly different from seascapes 7 and 8 (the latter with low r ¼ 0.335; Table 3). Among offshelf seascapes, only seascape 9 was significantly different from seascapes 1 and 5, and seascape 3 was significantly different from seascapes 5 and 9, but with very low values of r (,2.7; Table 3). No significant difference was observed between on-shelf seascape 6 and the others, despite the difference in biomass recorded there compared with the other seascapes (Figure 4d). Qualitative assessments based on the sum of average biomass for major taxa indicated that sponges dominated all on-shelf seascapes in both the GAB and the WA margin. In the WA margin dataset excluding the hard substratum, seascape 6 was the most diverse in that it contained the most taxonomic groups (cnidarians, molluscs, echinoderms, decapods, ascidians), but insufficient stations were available to determine whether this seascape showed the same trend in the GAB. Off-shelf seascapes were characterized by different dominant taxa, with seascape 1 dominated by holothurians and ophiuroids, seascape 3 by sponges, seascape 5 by echinoids, and seascape 9 by holothurians, molluscs, and decapods. Of all possible combinations of environmental factors from which on-shelf seascapes were derived, the BIO-ENV procedure revealed that the strongest and most significant correlation between GAB benthic invertebrate communities was from a combination of two variables, depth and slope [Spearman’s rank correlation (r) ¼ 0.401, p , 0.01]. We were unable to conduct similar analyses on the WA margin dataset that excluded the hard
Table 2. Results of the BIO-ENV procedure that shows the contribution of each environmental factor to the overall differences in biological community structure, including transformations used on each environmental variable. GAB Factor Slope Depth Seafloor temperature Mud content Primary production Effective disturbance Gravel content
Transformation performed Log Log Rank Rank Rank Rank Log
WA margin
r-value 0.397 0.366 0.299 0.278 0.309 0.306 –0.012
p-value ,0.01 ,0.01 ,0.01 ,0.01 ,0.01 ,0.01 0.51
Transformation performed Rank Log Rank Square root Rank Log Square root
r-value 0.235 0.271 0.418 0.124 0.053 0.134 –0.001
p-value ,0.01 ,0.01 ,0.01 ,0.01 0.15 0.07 0.47
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
of large or abundant species (Zar, 1998). As the seascapes were derived according to data that did not include hard substrata, we analysed the WA margin data in a two-tiered approach, including (i) a subset of the samples excluding stations with a hard substratum, and (ii) the complete dataset of samples on all three substratum types (hard, mixed, soft), to investigate the influence of the substratum on the biological utility of the seascapes. Multivariate analyses were performed on species composition using the BIO-ENV procedure, in which each biological matrix is compared with the environmental matrix to determine the main abiotic factors driving the biological patterns (Clarke and Warwick, 2001). Abiotic data were not normally distributed, so were normalized before analysis (Clarke and Warwick, 2001; Table 2). These normalized variables were also used to calculate correlations between depth and all other environmental variables, to identify potential depth-related factors. The GAB and WA margin data were analysed separately because differences in time of collection and sampling gear and lack of standardized taxonomic nomenclature between the surveys precluded standardization of biomass data. Only stations overlapping on-shelf seascapes were included in the BIO-ENV and correlation analyses owing to the exclusion of effective disturbance in off-shelf seascapes (Whiteway et al., 2007). All multivariate analyses were performed with PRIMER (v. 6). Station 7 from the GAB and stations 72 and 52 from the WA margin were outliers that disproportionately affected the results, so were excluded from multivariate analyses as per the statistical recommendations of Chatterjee and Price (2000). Specifically, station 7 had no species present that occurred elsewhere, station 72 had a high biomass of an ophiuroid species that was found only at that station, and station 52 had extremely low biomass and species richness compared with all other stations. Correlation analyses were performed in the R statistical platform (v. 2.7.2).
1958
R. Przeslawski et al.
Discussion
Figure 2. n-MDS of benthic invertebrate community structure, with seascape as a factor at (a) the GAB (stress ¼ 0.09), (b) the WA margin, excluding stations with a hard substratum (stress ¼ 0.13), and (c) the WA margin, including stations with a hard substratum (stress ¼ 0.13). Each point represents the community structure (biomass of each species) at a given station. The distance between points indicates the similarity among communities, with closer points denoting more-similar communities than distant points. The circle denotes a potential outlier from station 45 in the GAB. Solid points indicate locations in on-shelf seascapes, and hollow points indicate locations in off-shelf seascapes. substratum because of the scarcity of stations in on-shelf seascapes. However, analysis of all WA margin stations including those with a hard substratum revealed that the strongest correlation between
Seascapes are not consistently useful surrogates at regional and local scales because the strength and significance of their relationships with community structure varied according to individual classifications. In other words, some pairs of seascapes differentiated communities, whereas others showed no biological differences. Nevertheless, a continental-scale system of seascapes does have some utility as a surrogate for broad-scale patterns of benthic invertebrate communities when both environmental and biological data are limited, as long as it is understood that uncertainty increases in data-poor areas and that these seascapes remain unvalidated. If at all possible, an assessment should be made of each seascape’s utility. For example, in the current study, off-shelf seascape 9 had significantly different communities from all other seascapes on the WA margin, indicating that biological communities located in that seascape are likely to be consistently different from others in the entire region. In contrast, on-shelf seascapes 4 and 6 showed no biological differentiation in either the GAB or the WA margin, indicating that those seascapes are not useful surrogates at either regional or local scales. Although seascapes were not consistent surrogates in the current study, the measure of biodiversity used should be considered in the evaluation. Biodiversity here is defined as community structure based on biomass, which, even following root-transformation, remains a measure that emphasizes large animals. Future research investigating other measures of biodiversity (e.g. numerical abundance, which is driven by small animals) may yield improved differentiation of communities among seascapes.
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
WA benthic communities was from depth and temperature (r ¼ 0.423, p , 0.01). Individually, most environmental factors exhibited weak or moderate correlations between benthic invertebrate communities in both the GAB and the WA margins: slope, depth, seafloor temperature, and mud content (Table 2). Primary production and effective disturbance were only significantly correlated with benthic communities in the GAB (Table 2). Gravel content had no discernible relationship with biological data in either region (Table 2). In the GAB, clear relationships could be seen when each standardized environmental factor was plotted against depth (Figure 5a). Mud content and slope both showed strong positive relationships with depth (mud, adjusted r 2 ¼ 0.761, p , 0.001; slope, adjusted r 2 ¼ 0.914, p , 0.001), whereas seafloor temperature, primary production, and effective disturbance showed negative relationships (temperature, adjusted r 2 ¼ 0.923, p , 0.001; primary production, adjusted r 2 ¼ 0.592, p , 0.001; effective disturbance, adjusted r 2 ¼ 0.392, p , 0.001; Figure 5a). Gravel content was the only environmental factor in the GAB that did not have any discernible relationship with depth (adjusted r 2 ¼ 0.015, p ¼ 0.217; Figure 5a). In contrast, the relationships between depth and most environmental factors from WA margin locations overlapping with on-shelf seascapes were weak (Figure 5b; effective disturbance, adjusted r 2 ¼ 0.058, p ¼ 0.079; gravel content, adjusted r 2 ¼ 0.089, p ¼ 0.038; mud content, adjusted r 2 ¼ 0.119, p ¼ 0.019; primary production, adjusted r 2 ¼ – 0.004, p ¼ 0.364), except seafloor temperature, which showed a moderate negative relationship with depth (adjusted r 2 ¼ 0.455, p , 0.0001), and slope, which showed a moderately strong positive relationship with depth (adjusted r 2 ¼ 0.667, p , 0.0001; Figure 5b).
1959
A spatial habitat classification system as a surrogate of marine benthic community structure
Table 3. Results of the ANOSIMs including overall and pairwise results from (top) GAB, (centre) WA margin excluding hard substrata, and (bottom) WA margin including all substrata. GAB: Overall r ¼ 0.418 (p , 0.001) On-shelf 1 On-shelf 4 0.416* On-shelf 6 0.543*
On-shelf 4 n.a. –0.159
WA margin (excluding hard substrata): Overall r ¼ 0.375 (p , 0.001) On-shelf 6 On-shelf 8 Off-shelf 3 On-shelf 8 0.053 n.a. Off-shelf 3 0.354* 0.466* n.a. Off-shelf 9 0.467* 0.906* 0.314*
On-shelf 7
On-shelf 8
Off-shelf 1
Off-shelf 3
Off-shelf 5
n.a. 0.139 0.588* 0.379* 0.625* 0.897*
n.a. 0.683* 0.370* 0.704* 0.879*
n.a. 0.251* –0.229 0.552*
n.a. 0.268* 0.251*
n.a. 0.727*
r-values are presented in normal text; asterisks indicate significant relationships (a ¼ 0.05); n.a., not available.
Figure 3. Univariate measures of biological diversity from the GAB: (a) species richness, and (b) biomass. Error bars are standard errors of the means. The utility of seascapes to predict broad differences in biodiversity varies among regions, and specific seascapes may not be similarly associated with biological communities in other regions. For example, seascapes may not be as useful as a proxy for biodiversity in more heterogeneous regions. Indeed, recent research comparing biological communities in the Great Barrier Reef, an environment of high heterogeneity (Pitcher and CRC Reef Research Centre, 2007), indicates that seascapes are not strong predictors of biological communities there at a local scale (Heap et al., 2011). One
obvious explanation is that in such heterogeneous regions, the substratum type, an environmental factor not included in seascape derivations, is overriding the effects from the other environmental factors incorporated in seascape derivations. However, results from the current study reveal that some seascapes are associated with distinct biological communities along the WA margin, an area of mixed hard and soft substrata. If the exclusion of substratum data in seascape derivations hinders the utility of seascapes as a surrogate for biodiversity, then a stronger association of WA margin seascapes with biological communities when the hard substratum is excluded would be expected. Instead the association between seascapes and biological communities is broadly the same regardless of whether substrata are mixed or homogenous, indicating that at the regional scale (hundreds of kilometres), the substratum may not play a strong role in terms of the biological application of seascapes, likely because that variable is more dependent on spatial scale than other abiotic variables. Future refinements to seascape derivations may improve their performance and utility to predict biodiversity across a range of spatial scales. Additional abiotic layers relevant to seabed characterization may be incorporated into future seascape derivations (e.g. oxygen availability), and a layer differentiating hard and soft substrata is important for characterizing biota at a local scale (Beaman et al., 2005). In addition, future refined seascape derivations should include updated layers of existing biophysical data. For example, an updated disturbance layer should incorporate recently derived national datasets that reflect more robust modelling and estimation of seabed exposure and disturbance over time as they relate to biota (e.g. GEOMACS, in Hughes et al., 2011). The current study identified depth as a significant factor underpinning biological patterns in both the GAB and the WA margin. Previous studies revealed depth as important in structuring benthic invertebrate communities from a variety of habitats, including coral reefs (Cleary et al., 2005), submarine canyons (Vetter and Dayton, 1998), polar regions (Jones et al., 2007), temperate rocky reefs (Williams and Leach, 1999), and the deep sea (Rex et al., 2006). Using the same biological dataset from which
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
WA (all substrata): Overall r ¼ 0.318 (p , 0.001) On-shelf 4 On-shelf 6 On-shelf 6 –0.18 n.a. On-shelf 7 0.422* 0.13 On-shelf 8 0.335* 0.071 Off-shelf 1 0.622* 0.318* Off-shelf 3 0.409* 0.291* Off-shelf 5 0.581* 0.382* Off-shelf 9 0.902* 0.488*
1960
R. Przeslawski et al.
the current study draws, Ward et al. (2006) also found that both depth and mud content were significant drivers of epifaunal invertebrate biodiversity. Currie et al. (2009) similarly found that depth was related to infaunal communities at the same stations. In contrast, previous research using an expanded dataset from the WA margin revealed that depth per se was not one of the main drivers of biological community structure, rather that depth-correlated variables (temperature and oxygen availability) were most influential (Williams et al., 2010). Depth itself does not affect species distribution directly, but is instead associated with other factors known to influence physiology, such as pressure, temperature, and nutrients (Carney, 2005), and the relationship between depth and these factors can vary among regions. Such variation is obvious in the current study, in which most individual environmental factors from which seascapes are derived are significantly correlated with on-shelf benthic invertebrate communities. These results can be explained in the current study by the relationship between GAB environmental factors and depth, with all environmental factors significantly related to biological communities also related to depth, whereas gravel content was not related to either depth or biological communities. However, no such strong relationships were detected in the WA margin. Therefore, depth is a key driver of community structure regardless of whether it is closely linked with other environmental factors underpinning seascapes (e.g. GAB), or only weakly linked (e.g. WA margin), demonstrating that within the on- or off-shelf classification system, depth is one of the key abiotic factors from which seascapes are derived. However, for habitat classification systems such as the seascape approach to reach their full potential, basic surrogacy research quantifying the relationships between depth and benthic community structure across a range of habitats, regions, and spatial scales is crucial (e.g. the Surrogates Programme at www. marinehub.org; McArthur et al., 2010).
The continental-scale seascape approach tested in the current study is one of several habitat-mapping approaches that currently exist (Huang et al., 2011). For example, the spatial hierarchical framework, in which nested classifications are used to describe biological systems, has recently been applied in Australian waters and provides ten scales at which to classify benthic biodiversity. The system of continental-scale seascapes may prove quite useful at the larger scales of the hierarchical framework (e.g. provinces and bathomes in Last et al., 2010), because the scale at which physical data were interpolated would be similar to the scale at which biological data are quantified. In addition, the method of seascape derivation used in the current study (Whiteway et al., 2007) can be effective when a biological data layer is incorporated into seascapes at smaller scales (Heap et al., 2011), e.g. the scale of geomorphological units in Last et al. (2010). Seascapes and other habitat-classification methods based on environmental variables are useful in providing a broad-scale indication of habitat diversity at various scales (Kostylev et al., 2001; Beaman et al., 2005; Whiteway et al., 2007; Huang et al., 2011), but they cannot entirely replace the need for biological sampling. Nevertheless, they may provide an indication of differences in biological communities at regional and local scales, and they can be applied to marine zone management. For example, if one of the criteria for a marine protected area was representativeness and biological data were limited, boundaries could be determined such that the maximum number of seascapes was encompassed. Ideally, the derivation of seascapes would be tailored to the scale of management (Stevens, 2002). Hence, the national system of seascapes used here would be relevant to a national system of marine protected areas, whereas new seascape derivations at smaller scales such as those presented in Heap et al. (2011) would be suitable for planning at the local scale (tens of kilometres). The collection of biological data can be logistically
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
Figure 4. Univariate measures of biological diversity from the WA margin: (a) species richness excluding samples from the hard substratum, (b) biomass excluding samples from the hard substratum, (c) species richness including samples from the hard substratum, and (d) biomass including samples from the hard substratum. Shaded bars indicate on-shelf seascapes and hatched bars off-shelf seascapes. Error bars are standard errors of the means.
A spatial habitat classification system as a surrogate of marine benthic community structure
1961
difficult in certain areas, and in those cases, habitat maps and abiotic surrogates from which they are derived have application to marine planning and environmental assessments (Stevens, 2002; Last et al., 2010) through the identification of representative, unique, and potentially vulnerable habitats.
Acknowledgements We thank Tanya Whiteway (Geoscience Australia) for her assistance in retrieving the data associated with seascapes. Andrew Heap (Geoscience Australia), H. G. Greene, and an anonymous reviewer provided useful comments that improved the draft manuscript. The paper is published with permission of the Chief Executive Officer, Geoscience Australia.
References Beaman, R. J., Daniell, J. J., and Harris, P. T. 2005. Geology – benthos relationships on a temperate rocky bank, eastern Bass Strait, Australia. Marine and Freshwater Research, 56: 943– 958.
Carney, R. S. 2005. Zonation of deep biota on continental margins. Oceanography and Marine Biology: an Annual Review, 43: 211– 278. Chatterjee, S., and Price, B. 2000. Regression Analysis by Example, 3rd edn. John Wiley, New York. 360 pp. Clarke, K. R., and Warwick, R. M. 2001. Change in Marine Communities: an Approach to Statistical Analysis and Interpretation, 2nd edn. PRIMER-E, Plymouth, UK. Cleary, D. F. R., Becking, L. E., de Voogd, N. J., Renema, W., de Beer, M., van Soest, R. W. M., and Hoeksema, B. W. 2005. Variation in the diversity and composition of benthic taxa as a function of distance offshore, depth and exposure in the Spermonde Archipelago, Indonesia. Estuarine, Coastal and Shelf Science, 65: 557 – 570. Connor, D. W., Allen, J. H., Golding, N., Howell, K. L., Lieverknecht, L. M., Northen, K. O., and Reker, J. B. 2004. The Marine Habitat Classification for Britain and Ireland Version 04.05. Joint Nature Conservation Committee, Peterborough. ISBN 1 861 07561 8 (internet version). Currie, D. R., Sorokin, S. J., and Ward, T. M. 2008. Performance assessment of the benthic protection zone of the Great Australian Bight Marine Park: Epifauna. Final Report to the South
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
Figure 5. Relationship between interpolated depth and all other environmental variables examined in (a) the GAB and (b) the WA margin. The y-axis represents the percentage of the maximum value of a given environmental factor based on interpolations from which the seascapes were derived. Maximum values are GAB: mud content ¼ 27.07%, seafloor temperature ¼ 17.088C, slope ¼ 73.218, primary production ¼ 385.50 mg C m22 d21, gravel content ¼ 17.59%, and effective disturbance ¼ 0.0439; WA margin: mud content ¼ 66.29%, seafloor temperature ¼ 22.758C, slope ¼ 86.408, primary production ¼ 294.42 mg C m22 d21, gravel content ¼ 15.57%, and effective disturbance ¼ 0.0648.
1962
Lewis, M. 1999. CSIRO-SEBS (seamount, epibenthic sampler), a new epibenthic sled for sampling seamounts and other rough terrain. Deep Sea Research, 46: 1101 –1107. McArthur, M., Brooke, B., Przeslawski, R., Ryan, D. A., Lucieer, V., Nichol, S., McCallum, A. W., et al. 2010. On the use of abiotic surrogates to describe marine benthic biodiversity. Estuarine, Coastal and Shelf Science, 88: 21 – 32. Pitcher, C. R., and CRC Reef Research Centre. 2007. Seabed biodiversity on the continental shelf of the Great Barrier Reef World Heritage Area. CSIRO Marine and Atmospheric Research, Cleveland. 300 pp. Poore, G. C. B. 1995. Biogeography and diversity of Australia’s marine biota. In The State of the Marine Environment Report for Australia. Annex 1: the Marine Environment, pp. 75– 84. Ed. by L. Sann, and P. Kailola. Department of Environment, Sport, and Territories, Canberra. Rex, M. A., Etter, R. J., Morris, J. S., Crouse, J., McClain, C. R., Johnson, N. A., Stuart, C. T., et al. 2006. Global bathymetric patterns of standing stock and body size in the deep-sea benthos. Marine Ecology Progress Series, 317: 1 – 8. Sorokin, S. J., Fromont, J., and Currie, D. 2007. Demosponge biodiversity in the benthic protection zone of the Great Australian Bight. Transactions of the Royal Society of South Australia, 132: 192 – 204. Stevens, R. 2002. Rigor and representativeness in marine protected area design. Coastal Management, 30: 237– 248. UNESCO. 2009. Global Open Oceans and Deep Seabed (GOODS) Biogeographic Classification. IOC Technical Series, 84. 87 pp. Vetter, E. W., and Dayton, P. K. 1998. Macrofaunal communities within and adjacent to a detritus-rich submarine canyon system. Deep Sea Research II: Topical Studies in Oceanography, 45: 25– 54. Ward, T. M., Sorokin, S. J., Currie, D. R., Rogers, P. J., and McLeay, L. J. 2006. Epifaunal assemblages of the eastern Great Australian Bight: effectiveness of a benthic protection zone in representing regional biodiversity. Continental Shelf Research, 26: 25 – 40. Whiteway, T., Heap, A., Lucieer, V., Hinde, A., Ruddick, R., and Harris, P. T. 2007. Seascapes of the Australian margin and adjacent sea floor: methodology and results. GA Record 2007/11. Geoscience Australia, Canberra. 133 pp. Williams, A., Althaus, F., Dunstan, P., Poore, G. C. B., Bax, N. J., Kloser, R. J., and McEnnulty, F. R. 2010. Scales of habitat heterogeneity and megabenthos biodiversity on an extensive Australian continental margin (100 – 1100 m depths). Marine Ecology: an Evolutionary Perspective, 31: 222– 236. Williams, A., Koslow, J. A., and Last, P. R. 2001. Diversity, density and community structure of the demersal fish fauna of the continental slope off Western Australia (20 to 35 degrees S). Marine Ecology Progress Series, 212: 247 –263. Williams, I. M., and Leach, J. H. J. 1999. The relationship between depth, substrate and ecology: a drop video study from the southeastern Australian coast. Oceanologica Acta, 22: 651– 661. Womersley, H. B. S. 1990. Biogeography of Australasian marine macroalgae. In Biology of Marine Plants, pp. 266– 295. Ed. by M. N. Clayton, and R. J. King. Longman Cheshire, Melbourne. 501 pp. Zar, J. H. 1998. Biostatistical Analysis, 3rd edn. Prentice-Hall, Upper Saddle River, NJ. 662 pp.
Downloaded from http://icesjms.oxfordjournals.org/ by guest on December 30, 2015
Australian Department for Environment and Heritage and the Commonwealth Department of the Environment and Water Resources, F2008/000647-1. 113 pp. Currie, D. R., Sorokin, S. J., and Ward, T. M. 2009. Infaunal macroinvertebrate assemblages of the eastern Great Australian Bight: effectiveness of a marine protected area in representing the region’s benthic biodiversity. Marine and Freshwater Research, 60: 459– 474. Edyvane, K. 1999. Conserving Marine Biodiversity in South Australia. 1. Background, Status, and Review of Approach to Marine Biodiversity in South Australia. South Australian Research and Development Institute, Adelaide. 173 pp. Greene, H. G., Bizzarro, J. J., O’Connell, V. M., and Brylinsky, C. K. 2007. Construction of digital potential marine benthic habitat maps using a coded classification scheme and its application. In Mapping the Seafloor for Habitat Characterization, pp. 93– 109. Ed. by B. J. Todd, and H. G. Greene. Geological Association of Canada, Toronto. 327 pp. Harris, P. T. 2007. Applications of geophysical information to the design of a representative system of marine protected areas in southeastern Australia. In Mapping the Seafloor for Habitat Characterization, pp. 463 – 482. Ed. by B. J. Todd, and H. G. Greene. Geological Association of Canada, Toronto. 327 pp. Heap, A. D., Anderson, T., Falkner, I., Przeslawski, R., Whiteway, T., and Harris, P. T. 2011. Seascapes for the Australian margin and adjacent seabed. GA Record 2011/06. Geoscience Australia, Canberra. 91 pp. Heap, A. D., Przeslawski, R., Radke, L. C., Trafford, J., and Battershill, C. 2010. Seabed environments of the Eastern Joseph Bonaparte Gulf, Northern Australia. GA Record 2010/09. Geoscience Australia, Canberra. 78 pp. Hemer, M. A. 2006. The magnitude and frequency of combined flow bed shear stress as a measure of exposure on the Australian continental shelf. Continental Shelf Research, 26: 1258– 1280. Huang, Z., Brooke, B., and Harris, P. T. 2011. A new approach to mapping marine benthic habitats using physical environmental data. Continental Shelf Research, 31: S4– S16. Hughes, M. G., Harris, P. T., and Brooke, B. P. 2011. Seabed exposure and ecological disturbance on Australia’s continental shelf: potential surrogates for marine biodiversity. GA Record 2010/43. Geoscience Australia, Canberra. 78 pp. Jones, D. O. B., Bett, B. J., and Tyler, P. A. 2007. Depth-related changes to density, diversity and structure of benthic megafaunal assemblages in the Fimbul ice shelf region, Weddell Sea, Antarctica. Polar Biology, 30: 1579– 1592. Kostylev, V. E., Todd, B. J., Fader, G. B. J., Courtney, R. C., Cameron, G. D. M., and Pickrill, R. A. 2001. Benthic habitat mapping on the Scotian Shelf based on multibeam bathymetry, surficial geology and sea floor photographs. Marine Ecology Progress Series, 219: 121– 137. Last, P. R., Lyne, V. D., Williams, A., Davies, C. R., Butler, A. J., and Yearsley, G. K. 2010. A hierarchical framework for classifying seabed biodiversity with application to planning and managing Australia’s marine biological resources. Biological Conservation, 143: 1675 – 1686.
R. Przeslawski et al.
