Responses of ecosystems to environmental changes vary greatly across habitats, organisms and observational scales. The Quaternary fossil record of the Po Basin demonstrates that marine communities of the northern Adriatic reemerged unchanged following the most recent glaciation, which lasted approximately 100 000 years. The Late Pleistocene and Holocene interglacial ecosystems were both dominated by the same species, species turnover rates approximated predictions of resampling models of a homogeneous system, and comparable bathymetric gradients in species composition, sample-level diversity, dominance and specimen abundance were observed in both time intervals. The interglacial Adriatic ecosystems appear to have been impervious to natural climate change either owing to their persistence during those long-term perturbations or their resilient recovery during interglacial phases of climate oscillations. By contrast, present-day communities of the northern Adriatic differ notably from their Holocene counterparts. The recent ecosystem shift stands in contrast to the long-term endurance of interglacial communities in face of climate-driven environmental changes.

In marine settings, quantitative bathymetric models can be developed using various water depth proxies, including epibiont distribution, sedimentologic features, and the distribution of benthic taxa in time and space. Here, the late Quaternary bathymetric history of the Po coastal plain (Italy) has been reconstructed using mollusk samples from a network of 16 cores. Multiple analytical approaches have been applied in a comparative fashion. A direct ordination approach was used to estimate sample bathymetry using weighted averaging of genera with known preferred depth. Weighted averaging carries an advantage of analytical simplicity and produces direct ordination models expressed in environmentally meaningful units. Indirect ordination methods, based on depth estimates developed using posterioricalibrated ordination strategies (correspondence and detrended correspondence analysis calibrated against present-day bathymetric data), yielded results consistent with weighted averaging. Regardless of the choice of analytical methods, mollusk assemblages yielded bathymetric proxies congruent with independent sequence stratigraphic interpretations derived previously for both Late Pleistocene and Holocene transgressive-regressive cycles. The mollusk-derived proxies quantify spatial bathymetric gradients across the basin and local trends in absolute water depth in response to relative changes in sea level. However, for cores located in the most proximal part of the basin, mollusk-based ordinations failed to provide viable estimates due to inclusion of mixed marine and nonmarine mollusk faunas and scarcity of fossiliferous horizons necessary for adequate quantitative sampling. The multiple analytical approaches cross evaluated in this study consistently suggest that high-resolution quantitative bathymetric estimates can be derived for mollusk samples independent of stratigraphy for fully marine settings. When applied simultaneously to both samples and taxa, these approaches provide a viable strategy for quantifying stratigraphic and paleontological patterns and enhancing interpretations of basin-scale depositional systems.

Stratigraphic patterns of last occurrences (LOs) of fossil taxa potentially fingerprint mass extinctions and delineate rates and geometries of those events. Although empirical studies of mass extinctions recognize that random sampling causes LOs to occur earlier than the time of extinction (Signor–Lipps effect), sequence stratigraphic controls on the position of LOs are rarely considered. By tracing stratigraphic ranges of extant mollusc species preserved in the Holocene succession of the Po coastal plain (Italy), we demonstrated that, if mass extinction took place today, complex but entirely false extinction patterns would be recorded regionally due to shifts in local community composition and non-random variation in the abundance of skeletal remains, both controlled by relative sea-level changes. Consequently, rather than following an apparent gradual pattern expected from the Signor–Lipps effect, LOs concentrated within intervals of stratigraphic condensation and strong facies shifts mimicking sudden extinction pulses. Methods assuming uniform recovery potential of fossils falsely supported stepwise extinction patterns among studied species and systematically underestimated their stratigraphic ranges. Such effects of stratigraphic architecture, co-produced by ecological, sedimentary and taphonomic processes, can easily confound interpretations of the timing, duration and selectivity of mass extinction events. Our results highlight the necessity of accounting for palaeoenvironmental and sequence stratigraphic context when inferring extinction dynamics from the fossil record.

2016-12-23T18:47:23