Developing physical surrogates for benthic biodiversity using co-located samples and regression tree models: a conceptual synthesis for a sandy temperate embayment

Abstract

Strong correlations between pairs of the abiotic variables (|ρ| >= 0.75) are highlighted in Table S1. The four Local Moran's I of bathymetry variables were highly correlated (ρ = 1.0); as were the four Local Moran's I of backscatter variables (ρ >= 0.97) and the four TPI variables (ρ >= 0.89). This indicates that over the extent of the study area scale was not a significant issue for these three measures. Strong correlations were also found amongst the morphological parameters (slope, topographic relief and surface area) (ρ >= 0.75), between sediment sorting and a seabed exposure variable (|ρ| = 0.82), and between Fe and water temperature (|ρ| = 0.79). The correlations between bathymetry and two seabed exposure variables (ρ = 0.79) were expected because bathymetry was an input to the SWAN model. The strong correlations between skewness and kurtosis (measures of asymmetry) (|ρ| = 0.93) and between mud and sand (|ρ| = 1.0) were also expected because they are interrelated descriptors of the grain size distribution for each sample.

Interlinking backscatter, grain size and benthic community structure

“…Hamilton 1971;Li and Taylor-Smith, 1969;Williams, 2001). Recent attention has returned to this area of research because of its potential to provide more accurate seabed properties than image-based methods which simply average the angular-response of a given dataset during processing (Fonseca et al, 2009;Hasan et al, 2012;Huang et al, 2012;Hughes-Clarke, 1994;Lamarche et al, 2011;Rzhanov et al, 2012;Schimel et al, 2010).…”

Environmental predictors of foraging and transit behaviour in flatback turtles Natator depressus

“…Owing to the interactions among selected predictors, partial species response curves could not be properly constructed. As an alternative, we used predicted probabilities of presence and pseudo-absence samples (Table 2) to construct the response curves (Huang et al 2012). Input values of key environmental variables that were numeric were binned into equal intervals.…”

Modelling the effect of wave forces on subtidal macroalgae: A spatial evaluation of predicted disturbance for two habitat-forming species

“…This represents one of the few examples (but see also England et al, 2008;Huang et al, 2012) where wave modelling has been used in an ecological context and links individual-based models of macroalgal biomechanics to the spatial distribution of wave forces to understand the implications of biomechanical thresholds on a geographic level. Although the data requirements of wave modelling are significant, requiring the joint availability of accurate bathymetry, weather and wave information, when this information is available, the development of SWAN models can provide information that would otherwise be missing from studies of subtidal species and communities.…”

“…SWAN can be applied at any scale relevant to coastal applications, and includes terms that model wind-generation, depth-induced wave breaking and wave-wave interactions that can enhance or dissipate wave energy Ris et al, 1999). SWAN models have previously been applied in oceanographic and coastal engineering scenarios (Warner et al, 2008), but have not been used extensively in marine ecology (however, see England et al, 2008;Huang et al, 2012).…”