ASSESSING THE MAGNITUDE OF RECENT COMPOSITIONAL CHANGES IN MARINE ECOSYSTEMS: APPLYING THE CONSERVATION PALEOBIOLOGY APPROACH TO THE PERSIAN GULF

Nearly every modern marine ecosystem has experienced major changes due to anthropogenic stressors such as habitat modification, pollution, overexploitation and climate change. However, our knowledge of ecosystem dynamics in a historical time-frame (decades to few centuries) is restricted by the lack of direct observations.

An under-exploited source of information is hidden in death assemblages (DAs), the taxonomically identifiable, dead or discarded skeletal remains on a seabed. DAs represent archives that accumulate information on species composition and community states over time and are inert to recent changes. Previous studies have shown that live-dead (LD) agreement (the extent to which DA community composition and structure match the LA ones) tends to be poorer in anthropogenically disturbed settings, because living assemblages (LAs) respond faster than DAs to pressures, thus increasing the LD disagreement in composition. Moreover, dating individual shells (using radiocarbon calibrated amino acid racemization) within a DA allows the timing of ecosystem change to be identified. These approaches help recognize community shifts in time, overcoming the lack of direct observation. We apply these techniques to assess the impacts of oil platforms on the benthic molluscan assemblages in the Persian (Arabian) Gulf.

Contaminants show a weak gradient within each of two representative oilfields, which does not explain the spatial variation in taxonomic composition and abundance of the LA. Therefore, we expect the LD agreement to be mainly determined by time-averaging, inducing an increase in species richness and in evenness in the DA. In contrast to this expectation, rarefied species richness and evenness were not higher in the DA. Importantly, this finding was largely determined by the dominance in the DA of a single bivalve, Ervilia purpurea, which represented 40% of the whole DA, but was totally absent from the LA. By removing E. purpurea from the dataset, the results are in accordance with expectations. The ages of E. purpurea showed considerable spatial and temporal heterogeneity in production over the last decades and centuries, implying boom-and-bust population dynamics; its absence in LAs is thus unlikely related to the onset of oil platform production in the 20^th century.