Bayesian model selection for the glacial–interglacial cycle

A prevailing viewpoint in palaeoclimate science is that a single palaeoclimate record contains insufficient information to discriminate between most competing explanatory models. Results we present here suggest the contrary. Using SMC^2 combined with novel Brownian bridge type proposals for the state trajectories, we show that even with relatively short time series it is possible to estimate Bayes factors to sufficient accuracy to be able to select between competing models. The results show that Monte Carlo methodology and computer power have now advanced to the point where a full Bayesian analysis for a wide class of conceptual climate models is now possible. The results also highlight a problem with estimating the chronology of the climate record prior to further statistical analysis, a practice which is common in palaeoclimate science. Using two datasets based on the same record but with different estimated chronologies results in conflicting conclusions about the importance of the orbital forcing on the glacial cycle, and about the internal dynamics generating the glacial cycle, even though the difference between the two estimated chronologies is consistent with dating uncertainty. This highlights a need for chronology estimation and other inferential questions to be addressed in a joint statistical procedure.

[1]  A. Doucet,et al.  A Tutorial on Particle Filtering and Smoothing: Fifteen years later , 2008 .

[2]  Michel Crucifix,et al.  Oscillators and relaxation phenomena in Pleistocene climate theory , 2011, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[3]  M. Loutre,et al.  Astronomical theory of climate change , 2004 .

[4]  Eric Moulines,et al.  Comparison of resampling schemes for particle filtering , 2005, ISPA 2005. Proceedings of the 4th International Symposium on Image and Signal Processing and Analysis, 2005..

[5]  A. Abe‐Ouchi,et al.  Insolation-driven 100,000-year glacial cycles and hysteresis of ice-sheet volume , 2013, Nature.

[6]  A. Doucet,et al.  Particle Markov chain Monte Carlo methods , 2010 .

[7]  Jun S. Liu,et al.  Sequential Monte Carlo methods for dynamic systems , 1997 .

[8]  P. Moral,et al.  Sequential Monte Carlo samplers , 2002, cond-mat/0212648.

[9]  Jacques Laskar,et al.  A long-term numerical solution for the insolation quantities of the Earth , 2004 .

[10]  K. Aihara,et al.  Dynamics between order and chaos in conceptual models of glacial cycles , 2013, Climate Dynamics.

[11]  Aki Vehtari,et al.  Understanding predictive information criteria for Bayesian models , 2013, Statistics and Computing.

[12]  M. Crucifix.  How can a glacial inception be predicted? , 2011, 1112.3235.

[13]  Rudolph van der Merwe,et al.  The unscented Kalman filter for nonlinear estimation , 2000, Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (Cat. No.00EX373).

[14]  N. Shackleton.  Oxygen Isotope Analyses and Pleistocene Temperatures Re-assessed , 1967, Nature.

[15]  E. Tziperman,et al.  Are the 41 kyr glacial oscillations a linear response to Milankovitch forcing , 2004 .

[16]  P. Huybers.  Combined obliquity and precession pacing of late Pleistocene deglaciations , 2011, Nature.

[17]  H. Elderfield,et al.  Evolution of Ocean Temperature and Ice Volume Through the Mid-Pleistocene Climate Transition , 2012, Science.

[18]  V. Brovkin,et al.  Glacial CO2 cycle as a succession of key physical and biogeochemical processes , 2011 .

[19]  N. Gordon,et al.  Novel approach to nonlinear/non-Gaussian Bayesian state estimation , 1993 .

[20]  Tomohiro Ando,et al.  Predictive likelihood for Bayesian model selection and averaging , 2010 .

[21]  Carl Wunsch,et al.  Consequences of pacing the Pleistocene 100 kyr ice ages by nonlinear phase locking to Milankovitch forcing , 2006 .

[22]  J. Carson,et al.  Uncertainty quantification in palaeoclimate reconstruction , 2015 .

[23]  Barry Saltzman,et al.  A first-order global model of late Cenozoic climatic change II. Further analysis based on a simplification of CO2 dynamics , 1991, Transactions of the Royal Society of Edinburgh: Earth Sciences.

[24]  Nicolas Chopin,et al.  SMC2: an efficient algorithm for sequential analysis of state space models , 2011, 1101.1528.

[25]  L. M. M.-T..  Theory of Probability , 1929, Nature.

[26]  Fabo Feng,et al.  Obliquity and precession as pacemakers of Pleistocene deglaciations , 2015, 1505.02183.

[27]  C. Andrieu,et al.  The pseudo-marginal approach for efficient Monte Carlo computations , 2009, 0903.5480.

[28]  W. Ruddiman,et al.  Ice-driven CO 2 feedback on ice volume , 2006 .

[29]  M. Raymo,et al.  A Pliocene‐Pleistocene stack of 57 globally distributed benthic δ18O records , 2005 .

[30]  Michel Crucifix,et al.  Why could ice ages be unpredictable , 2013, 1302.1492.

[31]  Peter John Huybers,et al.  Glacial variability over the last two million years: an extended depth-derived agemodel, continuous obliquity pacing, and the Pleistocene progression , 2007 .

[32]  Frank Kwasniok,et al.  Analysis and modelling of glacial climate transitions using simple dynamical systems , 2013, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[33]  M. Crucifix,et al.  Effects of Additive Noise on the Stability of Glacial Cycles , 2016 .

[34]  Barry Saltzman,et al.  Modelling the slow climatic attractor , 1988 .

[35]  M. L. Eaton.  Multivariate statistics : a vector space approach , 1985 .

[36]  Lorraine E. Lisiecki,et al.  Links between eccentricity forcing and the 100,000-year glacial cycle , 2010 .

[37]  C. Wunsch,et al.  Obliquity pacing of the late Pleistocene glacial terminations , 2005, Nature.

[38]  J. Backman,et al.  Oxygen isotope calibration of the onset of ice-rafting and history of glaciation in the North Atlantic region , 1984, Nature.

[39]  Didier Paillard,et al.  The timing of Pleistocene glaciations from a simple multiple-state climate model , 1998, Nature.

[40]  André Berger,et al.  Long-term variations of daily insolation and Quaternary climatic changes , 1978 .

[41]  C. Emiliani,et al.  Pleistocene Temperatures , 1955, The Journal of Geology.

[42]  André Berger,et al.  An alternative astronomical calibration of the lower Pleistocene timescale based on ODP Site 677 , 1990, Transactions of the Royal Society of Edinburgh: Earth Sciences.

[43]  J. D. Hays,et al.  The orbital theory of Pleistocene climate : Support from a revised chronology of the marine δ^ O record. , 1984 .

[44]  Darren J. Wilkinson,et al.  Bayesian inference for nonlinear multivariate diffusion models observed with error , 2008, Comput. Stat. Data Anal..

[45]  Aki Vehtari,et al.  A survey of Bayesian predictive methods for model assessment, selection and comparison , 2012 .

[46]  Barry Saltzman,et al.  A first-order global model of late Cenozoic climatic change II. Further analysis based on a simplification of CO2 dynamics , 1991 .

[47]  John Z. Imbrie,et al.  Modeling the Climatic Response to Orbital Variations , 1980, Science.

[48]  F. Parrenin,et al.  Terminations VI and VIII (∼ 530 and ∼ 720 kyr BP) tell us the importance of obliquity and precession in the triggering of deglaciations , 2012 .

[49]  Annabel M. Imbrie-Moore,et al.  A phase-space model for Pleistocene ice volume , 2010, 1104.3610.

[50]  Myles R. Allen,et al.  A comparison of competing explanations for the 100,000‐yr Ice Age cycle , 1999 .

[51]  Maureen E. Raymo,et al.  The timing of major climate terminations , 1997 .