A global compilation of coral sea-level benchmarks: Implications and new challenges

I present a quality-controlled compilation of sea-level data from U–Th dated corals, encompassing 30 studies of 13 locations around the world. The compilation contains relative sea level (RSL) data from each location based on both conventional and open-system U–Th ages. I have applied a commonly used age quality control criterion based on the initial 234U/238U activity ratios of corals in order to select reliable ages and to reconstruct sea level histories for the last 150,000 yr. This analysis reveals scatter of RSL estimates among coeval coral benchmarks both within individual locations and between locations, particularly during Marine Isotope Stage (MIS) 5a and the glacial inception following the last interglacial. The character of data scatter during these time intervals imply that uncertainties still exist regarding tectonics, glacio-isostacy, U-series dating, and/or coral position. To elucidate robust underlying patterns, with confidence limits, I performed a Monte Carlo-style statistical analysis of the compiled coral data considering appropriate age and sea-level uncertainties. By its nature, such an analysis has the tendency to smooth/obscure millennial-scale (and finer) details that may be important in individual datasets, and favour the major underlying patterns that are supported by all datasets. This statistical analysis is thus functional to illustrate major trends that are statistically robust (‘what we know’), trends that are suggested but still are supported by few data (‘what we might know, subject to addition of more supporting data and improved corrections’), and which patterns/data are clear outliers (‘unlikely to be realistic given the rest of the global data and possibly needing further adjustments’). Prior to the last glacial maximum and with the possible exception of the 130–120 ka period, available coral data generally have insufficient temporal resolution and unexplained scatter, which hinders identification of a well-defined pattern with usefully narrow confidence limits. This analysis thus provides a framework that objectively identifies critical targets for new data collection, improved corrections, and integration of coral data with independent, stratigraphically continuous methods of sea-level reconstruction.

► First global quality-controlled compilation of coral sea-level data.
► Coral U–Th and position, uplift rates and open-system age data are included.
► Coral based-sea level record spanning the last 150,000 yr is presented.
► Uncertainties remain in tectonics, isostacy, U-series dating, and coral position.
► Partial agreement is observed between corals and continuous sea level records.