Exploring taxonomic and phylogenetic relationships to predict radiocaesium transfer to marine biota

One potentially useful approach to fill data gaps for concentration ratios, CRs, is based upon the hypothesis that an underlying taxonomic and/or phylogenetic relationship exists for radionuclide transfer. The objective of this study was to explore whether these relationships could be used to explain variation in the transfer of radiocaesium to a wide range of marine organisms. CR data for ¹³⁷Cs were classified in relation to taxonomic family, order, class and phylum. A Residual Maximum Likelihood (REML) mixed-model regression modelling approach was adopted. The existence of any patterns were then explored using phylogenetic trees constructed with mitochondrial COI gene sequences from various biota groups and mapping the REML residual means onto these trees. A comparison of the predictions made using REML with blind datasets allowed the efficacy of the procedure to be tested. The only significant correlation between predicted and measured activity concentrations was revealed at the taxonomic level of order when comparing REML analysis output with data from the Barents Sea Region. For this single case a correlation 0.80 (Spearman rank) was derived which was significant at the 0.01 level (1-tailed test) although this was not the case once a (Bonferroni) correction was applied. The application of the REML approach to marine datasets has met with limited success, and the phylogenetic trees illustrate complications of using predictions based on values from different levels of taxonomic organization, where predicted values for the order level can mask the values at lower taxonomic levels. Any influence of taxonomy and phylogeny on transfer is not immediately conspicuous and categorizing marine organisms in this way is limited in providing a potentially robust prognostic extrapolation tool. Other factors may plausibly affect transfer to a much greater degree in marine systems, such as quite diverse life histories and different diets, which may confound any phylogenetic pattern.