Geoscientific Model Development (Dec 2019)

PALEO-PGEM v1.0: a statistical emulator of Pliocene–Pleistocene climate

  • P. B. Holden,
  • N. R. Edwards,
  • T. F. Rangel,
  • E. B. Pereira,
  • G. T. Tran,
  • R. D. Wilkinson

DOI
https://doi.org/10.5194/gmd-12-5137-2019
Journal volume & issue
Vol. 12
pp. 5137 – 5155

Abstract

Read online

We describe the development of the “Paleoclimate PLASIM-GENIE (Planet Simulator–Grid-Enabled Integrated Earth system model) emulator” PALEO-PGEM and its application to derive a downscaled high-resolution spatio-temporal description of the climate of the last 5×106 years. The 5×106-year time frame is interesting for a range of paleo-environmental questions, not least because it encompasses the evolution of humans. However, the choice of time frame was primarily pragmatic; tectonic changes can be neglected to first order, so that it is reasonable to consider climate forcing restricted to the Earth's orbital configuration, ice-sheet state, and the concentration of atmosphere CO2. The approach uses the Gaussian process emulation of the singular value decomposition of ensembles of the intermediate-complexity atmosphere–ocean GCM (general circulation model) PLASIM-GENIE. Spatial fields of bioclimatic variables of surface air temperature (warmest and coolest seasons) and precipitation (wettest and driest seasons) are emulated at 1000-year intervals, driven by time series of scalar boundary-condition forcing (CO2, orbit, and ice volume) and assuming the climate is in quasi-equilibrium. Paleoclimate anomalies at climate model resolution are interpolated onto the observed modern climatology to produce a high-resolution spatio-temporal paleoclimate reconstruction of the Pliocene–Pleistocene.