Appendix A. Multivariate analysis of cross-shelf and vertical distributions of 34 species of crustacean larvae.
The cross-shelf and vertical distributions of the 34 most abundant species of crustacean larvae were analyzed for the four cruises using nonparametric analysis of similarity (ANOSIM) followed by hierarchical cluster analysis and ordination (Clarke 1993). Analyses were conducted using PRIMER software (version 6.1.10, Plymouth Routines in Multivariate Ecological Research). Data were square-root transformed to deemphasize the contribution of very abundant species and all analyses were conducted on a Bray-Curtis similarity matrix. All two-way crossed ANOSIM tests were conducted without replication with 999 maximum permutations using Spearman Rank correlation (Clarke and Warwick 1994).We first tested for cross-shelf differences in the larval assemblage among sampling dates after depths were averaged for each station. The larval distributions did not differ significantly among sampling dates (ANOSIM R = 0.013, p = 0.39) so subsequent analyses on cross-shelf distributions and depth were conducted on date-averaged data.
Agglomerative hierarchical clustering with group-average linking followed by a similarity profile test (SIMPROF) using 10,000 permutations tested the null hypothesis that the resultant groups in the dendrogram did not differ from one another in multivariate structure at the 1% level. Similarity percentages (SIMPER) revealed the percent contribution of each species-stage combination to the significant clusters. Nonmetric Multidimensional Scaling (nMDS) explored the relationship between the distance offshore and depth among larval assemblages (Kruskal and Wish 1978). The percent contribution of each species and stage to the significant clusters from the SIMPROF analysis objectively classified species-stage combinations by their cross-shelf and depth distributions.
The dendrogram from the cluster analysis revealed a high degree of spatial structure among groups that corresponded to differences in distance offshore and depth (Fig. 4A). Four distinct species assemblages occurred at a similarity of 18% relative to the distance offshore and depth (nearshore, midshelf A and B, outer shelf). The nearshore group was further split into two groups with a similarity of 28% that corresponded to different depth distributions. The five clusters were superimposed on a nMDS plot (2D stress: 0.14, 3D stress: 0.09) to examine the distance offshore and depth separation among the larval assemblages (Fig. 4B). The nearshore samples were more closely clustered than the outer-shelf group indicating that there is less variability in the depth distributions of larvae nearshore. The nearshore species assemblage also was more closely related to the midshelf species assemblage than to the outer-shelf assemblage. Further, the larval communities at the shallow depths of the nearshore and midshelf stations (nearshore A and midshelf A respectively) were more closely related than the deep larval communities at those sites.
Clarke, K. R. 1993. Nonparametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18:117–143
Clarke, K. R. and R. M. Warwick 1994. Similarity based testing for community pattern: the two-way layout with no replication. Marine Biology 118:167–176.
Kruskal, J. B., and M. Wish 1978. Multidimensional scaling. Sage Publications, Beverly Hills, California, USA.