Predicting the Composition of Polychaete Assemblages in the Aegean Coast of Turkey

Abstract

Benthic infaunal species and communities have been extensively used to evaluate quality of the marine environment. Within the MSFD, community composition is addressed most commonly through Descriptor 6 (Seafloor integrity), criterion 6.2 (Condition of benthic communities). At the same time, the Directive has stipulations for addressing and assessing indicators linked with pressures in an explicitly spatial manner. At larger scales, achieving this through point sampling may be impractical or unfeasible; hence predictive methods are being increasingly employed to produce the large scale spatial data that are often required for marine spatial planning and management. The aim of the current work was to develop statistical and spatial modeling tools that can predict the distribution of soft-sediment benthic polychaetes in the Aegean coast of Turkey. To do that, we employed Species Archetype Models (SAMs), a novel analytical and modeling framework which uses mixture models to cluster species responses to the environment, producing a number of "archetypal" responses assumed to represent species with similar ecological/physiological tolerances. Polychaete presence/absence data were obtained from the literature and modeling was performed against environmental variables reflecting the main natural and anthropogenic gradients in the region. The resulting models are interpreted in light of the sensitivity/tolerance classification scheme for benthic invertebrates. Three Species Archetypes were identified through the analysis. In brief, Species Archetype 1 consists of the most prevalent species in the dataset and primarily follows the salinity and temperature gradients. Species Archetype 2, present in the central and southern Aegean, is dominated by sensitive and indifferent species and responds negatively to chlorophyll a, whereas Species Archetype 3 represents mostly tolerant and opportunistic polychaetes with increased probability of occurrence in eutrophic, shallow, inshore areas throughout the region. Predictive performance was constrained by the information contained in our data. These results from a limited data set show promise that SAMs as a modeling tool can offer valuable insights into patterns of benthic species distribution and coexistence and increase our capacity to provide predictive advice.

Sponge assemblages and predicted archetypes in the eastern Canadian Arctic

“…Astrophorids, such as the geodiids, Stelletta or Stryphnus, are massive sponges adapted to high-energy environments where they reach high densities (Knudby et al 2013, Beazley et al 2015) in contrast to hexactinellids or other demosponges that can have branching or tubular structures likely adapted to lower-energy regimes. The structureforming Geodia species in our study are found in areas with high primary production and fast currents that would provide high food supply needed to reach their large biomasses (up to 2 t found in a trawl set), as seen elsewhere in the Atlantic Ocean (Klitgaard & Tendal 2004, Fuller 2011, Murillo et al 2012, Knudby et al 2013.…”

“…A full model including all covariate and quadratic terms was used to select the number of species archetypes. The covariates used were previously standardized to a mean of 0 and variance of 1 (Dunstan et al 2013, Galanidi et al 2016. The most parsimonious model was determined by comparing models with different species archetypes (G) using the Bayesian information criterion (BIC) (Schwarz 1978).…”

“…Hui et al 2015, Grimm et al 2017. Often researchers remove covariates with high standard errors (Woolley et al 2013, Galanidi et al 2016 or fit a series of models comparing the fit statistics to determine the optimal model (Leaper et al 2014). In the present study, once the number of archetypes was selected, models with different covariates were evaluated in a stepwise approach to minimize the BIC.…”

“…In the northwest Atlantic, along the continental slopes, sponges can make up the majority of the ben-thic biomass (Fuller 2011, Murillo et al 2012, Knudby et al 2013, and in the eastern Canadian Arctic, sponges and sponge grounds occur in both Davis Strait and Baffin Bay (Knudby et al 2013, which have been identified as Significant Benthic Areas (Kenchington et al 2016a,b). Sponge communities from the northeast Atlantic (Klitgaard & Tendal 2004) and the northwest Atlantic (Fuller 2011, Murillo et al 2012 have been described, and include arctic and boreal astrophorid grounds with similar species compositions. However, there has not been a recent systematic review of the sponge species of the eastern Canadian Arctic, and the major sponge communities present in these areas remain unknown.…”

Projecting marine species range shifts from only temperature can mask climate vulnerability

“…However, examining predictions of species' co‐occurrence patterns in concert with dispersal estimates using joint species distribution modeling could provide a way forward (Anderson, ; Kissling & Schleuning, ; Ovaskainen et al, ; Warton et al, ). Lastly, we do not explore how our model projections translate to assemblage‐level outcomes; however, novel multivariate techniques (e.g., archetype modeling) could enable exploration of this topic in future (Dunstan, Foster, & Darnell, ; Galanidi, Kaboglu, & Bizsel, ).…”