Article · 2012

Robust estimates of extinction time in the geological record

C. J. A. Bradshaw, A. Cooper, C. S. M. Turney and B. W. Brook

Quaternary Science Reviews 33 14-19

An article that develops a method for estimating extinction dates from dated fossil records, tests it by simulation and applies it to Alaskan mammoths and horses and to Neanderthal remains from Gorham's Cave.

Bradshaw and colleagues develop the Gaussian-resampled, inverse-weighted McInerny method, abbreviated GRIWM, for estimating when a species died out from its dated fossil records. They test it by simulation and apply it to Pleistocene mammoth and horse records from Alaska and to Neanderthal artefacts from Gorham's Cave, Gibraltar.

An estimate built on the most recent records, they argue, has to allow for both the size of the sample and the uncertainty in radiometric dates. GRIWM weights records by their distance in time from the latest one and resamples dates from Gaussian distributions. They contend that this reduces bias across several plausible fossil-sampling patterns, while producing wider uncertainty intervals than some alternative methods (pp. 15–19).

In the simulations the simpler weighted method performs best for uniform and logarithmic distributions. The Alaskan analysis finds that mammoth and horse extinctions were not synchronous. For Gorham's Cave the tests use 16 calibrated radiocarbon dates for Mousterian artefacts, and GRIWM gives a later lower bound than the other estimates discussed (pp. 17–18).

Dates are given in years before present, with radiocarbon years kept apart from calendar years in the case studies. The Gorham's Cave dates use the Cariaco Basin calibration, and converting the radiocarbon age of the latest confidently associated Neanderthal level into a calendar age is discussed (p. 18).

What it records

  • The simulations use five underlying distributions and 1,000 samples of 50 specimens for each.p. 15
  • Bradshaw and colleagues report that GRIWM has the lowest average bias for linear, sigmoidal and exponential sampling distributions.p. 17
  • GRIWM estimates centre the Alaskan mammoth and horse extinctions on 12,200 and 11,150 14C years before present, at least 300 14C years apart.p. 18
  • Finlayson and Tzedakis report that the most recent level confidently associated with Neanderthal occupation at Gorham's Cave is dated to 28,170 ± 240 14C years before present.p. 18
  • The GRIWM lower bound for the Neanderthal extinction estimate is 29,950 calendar years before present.p. 18

Where to read

  • Simulation design and extinction-estimation methodspp. 15–17
  • Alaskan mammoth and horse extinction estimatespp. 17–18
  • Gorham's Cave Neanderthal datespp. 17–18

Contents

  • 1. Introduction14
  • 2. Materials and methods15
  • 2.1. Time series simulation15
  • 2.2. Estimating time of extinction15
  • 2.3. Real extinction data17
  • 3. Results17
  • 3.1 Simulations17
  • 3.2 Pleistocene mammoth and horse extinctions17
  • 3.3 European Neanderthal extinction18
  • 4. Discussion18

People, places and institutions

People

  • McInernyauthor of the method the authors modify
  • Solowauthor of extinction-estimation methods discussed
  • Robertsco-author of an extinction-estimation method
  • Guthriesource of the Alaskan fossil data
  • Finlaysonsource of the Gorham's Cave artefact data
  • Tzedakissource of the Gorham's Cave date analysis

Places

  • Gorham's Cavesource of Neanderthal artefact dates
  • Alaskasource region for mammoth and horse fossil dates

Also named: Gibraltar.

Extent pp. 14–19, figs 1–5

Chronology and DatingAnimals

Cited on Ancient Cyprus

1 page

Corrections and additions

Know this document?

Help us cite it fully and point readers to a copy.

We would like to hear about

  • The full citation, or a correction to ours
  • Where a copy can be read or downloaded
  • Missing parts, volumes or plates
  • Related documents

Tell us if you would like to be thanked by name on the page.