Abstract Conservation palaeobiology informs conservation and restoration of ecosystems by using the fossil record to discriminate between baseline and novel states and to assess ecosystem response to perturbations. Variability in the time-scale of palaeobiological data can generate patterns that either exaggerate or mute the magnitude of biotic changes. We identify two approaches that remedy the challenges associated with the mixing of baseline and post-impact states and with the transformation of the stratigraphic depth to time. First, combining surface death assemblages with both (1) fossil assemblages preserved in the subsurface historical layers and (2) living assemblages can better resolve the nature of ecosystem shifts than within-core surveys or live–dead analyses alone. Second, post-mortem age distributions of skeletal particles and their preservation states are not only informative about stratigraphic resolution and time averaging of death assemblages but also about the timing of changes in abundance of skeletal producers. High abundance of the youngest age cohorts in surface death assemblages is a null expectation of disintegration and burial dynamic. When this dynamic is accounted for, age distributions of benthic invertebrates from Holocene sediments often reveal high volatility, prolonged turn-offs in production or pervasive regime shifts that are obscured in the raw stratigraphic record.

Abstract Death assemblages (DAs) are increasingly recognized as a valuable source to reconstruct past ecological baselines, due to the accumulation of skeletal material of non-contemporaneous cohorts. We here quantify the age and time-averaging of DAs on shallow subtidal (5–25 m) rocky substrates and in meadows of Posidonia oceanica in the eastern Mediterranean. We show that such DAs are very young – median ages 9–56 years – with limited time-averaging, one to two orders of magnitude less than on even nearby soft substrates. On rocky substrates, out-of-habitat transport is likely the main cause of loss of older shells. In Posidonia oceanica meadows, the root and rhizome system creates a dense structure – the matte – that quickly entangles and buries shells and limits the potential for bioturbation. The matte is, however, a peculiar feature of Posidonia oceanica , and age and time-averaging in meadows of other seagrass species may be different. The young age of DAs in these habitats requires a careful consideration of their appropriateness as baselines. The large difference in DA age between soft substrates, subject to numerous studies, and hard and seagrass substrates, rarely inspected with geochronological techniques, implies that DA dating is important for studies aiming at using DAs as baselines.

Abstract Inferring the composition of pre-Anthropocene baseline communities on the basis of death assemblages (DAs) preserved in a surface mixed layer requires discriminating among recently-dead shells sourced by living populations and older shells from extirpated populations. Here, we assess the distribution of postmortem ages in the DA formed by the brachiopod Gryphus vitreus at 580 m depth in the Bari Canyon (Adriatic Sea), with no individuals collected alive. The Gryphus DA exhibits millennial time averaging (inter-quartile range = 1250 years) and two modes in abundance at 500 and 1750 years BP. As high abundance of species in time-averaged DAs can reflect passive accumulation of shells sourced by populations with low standing population density, we reconstruct changes in annual density on the basis of the abundance maxima detected in the distribution of postmortem ages and on the basis of estimates of per-specimen disintegration rate. We find that adults (>20 mm) achieved densities of at least 10–20 individuals/m 2 (assuming lifespan is 10 years), and the pulses in abundance were thus associated with a high population density in the past, followed by the decline over the last few centuries. We infer that bathyal populations were volatile during the Late Holocene, with brachiopods sensitive to siltation that was induced by temporal changes in sediment dispersal into the Bari Canyon due to deforestation and climatic changes.

Abstract Evaluating the history of human impacts on marine ecosystems based on sediment cores is challenging on shelves characterized by very slow sedimentation. To assess the stratigraphic expression of such impacts in the condensed deposits of an epicontinental sea, we analysed a 3 m-long core collected at 31 m water depth off the Po prodelta in the Northern Adriatic Sea by integrating geochronological ( 14 C and 210 Pb), sedimentological, geochemical and palaeontological proxies. A depositional history of the last 10 000 years is expressed in four different facies: (1) alluvial floodplain, (2) transitional, shell-poor silts, (3) a condensed 30 cm-thick shell lag, and (4) a 10 cm-thick layer of distal prodelta silts comprising the last c. 500 years. 10 000-year-old shells of Lentidium mediterraneum spread over the shell lag and prodelta sediments document onshore transport during the early Holocene sea-level rise. Varicorbula gibba shells are age-homogeneous within the subsurface shell lag, documenting decimetre-scale mixing by bioturbation in the past. However, in spite of low sedimentation rates, the organic and heavy metal enrichment, the increase in proportional abundance of benthic foraminifers preferring organic-rich sediments ( Nonionella sp.), and the increase in size of molluscs ( V. gibba ) in the upper 10 cm formed by prodelta silts still detect the eutrophication in this region during the twentieth century. These eutrophication proxies are preserved in the stratigraphic record owing to temporarily increasing sedimentation rate and decreasing mixing depth.

Abstract Line Intercept Transects (LIT), Point Intercept Transects (PIT) and Photoquadrats (PQ) are the most common quantitative sampling techniques in modern and fossil coral reefs. Data from coral reefs obtained by the different methods are generally compared between various reef ages and localities. Quaternary reefs from warmer interglacial periods, which represent climate scenarios projected for the future, are particularly interesting for comparisons with modern reefs. Importantly, fossil reefs differ from modern reefs because they are diagenetically altered and time-averaged. While several studies have compared different quantitative methods in modern reefs, very few have dealt with the comparability among fossil and between fossil and modern reefs. Here, we compare LIT, PIT at 10, 20 and 50 cm intervals and PQ in two Pleistocene reef localities in Egypt. We find that alpha diversity, reef cover and community composition are dependent on the method. Results gained with plotless methods (LIT, PIT) differ strongly from results gained with plot methods (PQ). However, coral cover results are similar with LIT and PIT, and community composition is indistinguishable between the two, but alpha diversity depends on the interval used for PIT. We discuss the implications of our findings for comparing coral reefs of various ages and localities. We recommend surveying Pleistocene reefs with PIT at 20 cm intervals. This is because (1) alpha diversity is well captured, (2) the amount of time-averaging recorded by PIT is reduced compared to PQ, (3) the PIT results can be directly compared to reefs analysed by LIT and (4) the method is less time-consuming than LIT and PQ.