Copy this text
Assessment of spatial distribution of organic contaminants and metallic compounds on a tropical island’ coral reef fish communities
The New Caledonian archipelago is an important hotspot of marine biodiversity. Due to mining activities, urbanization, and industrialization, significant amounts of contaminants are discharged into the lagoon. This study analysed the concentrations, spatial distribution, and potential drivers of 14 metallic compounds and trace elements (MTEs) and 22 persistent organic pollutants (POPs) in ~400 coral reef fish sampled from various sites around New Caledonia, across a gradient from mining centers to remote, uninhabited locations. Boosted regression trees modelling explained between 61 and 86 % of the global variation in MTEs and POPs concentration. Fish body size emerged as the most important correlate of MTEs and POPs concentrations in coral reef fish. Monthly rainfalls were the second most important variable for POPs, whereas the reef area was the second variable explaining MTE concentrations. Our modelling approach allowed us to predict and map the distribution of concentrations at the fish community level for 17 contaminants (9 MTEs and 8 POPs). Predicted concentrations ranged from ~1.5 ng.g−1 (β-endosulfan) to ~11.5 μg.g−1 (Ni), and revealed a widespread contamination throughout the lagoon, from the coast to the barrier reef. Contamination by mining-related elements (Ni, Cr…) were clearly influenced by the surface area of mining registry and to lithology to a lesser extent, whereas Hg contamination strongly depended on biological variables. Our study is the largest of its kind at the archipelago scale, combining data on 36 contaminants in ~400 fish samples with a modelling framework offering insights into underlying processes and spatial data for policy use.
Keyword(s)
Pollutants, Mining activities, Anthropogenic impacts, Modelling, Pacific Ocean
Full Text
File | Pages | Size | Access | |
---|---|---|---|---|
Publisher's official version | 13 | 5 Mo | ||
Fig. S1. Summary of the average relative contributions (%) of the fourteen variables used in boosted regression trees (BRTs) explaining their relative influence on coral reef fish contamination. | - | 180 Ko | ||
Fig. S2. Summary of the average relative contributions (%) of the nine variables used in boosted regression trees (BRT) explaining their relative influence on contamination of coral reef fish communit | - | 163 Ko | ||
Fig. S3. Spatial patterns of modelled concentrations of trace elements (Ag, As, Cd, Cr, Cu, Hg, Mn, and Se) in coral reef fish, using environmental, biogeographic, ... | - | 1 Mo | ||
Fig. S4. Spatial patterns of modelled concentrations of POPs (α-Endosulfan, Atrazine, Dieldrin, Heptachlor Epoxide B, pp’-DDD, pp’-DDT, and ∑PCBs) in reef fish, using a BRT model that incorporates... | - | 1 Mo | ||
Supplementary tables | - | 57 Ko |