A simple test for causality in complex systems

We provide a new solution to the long-standing problem of inferring causality from observations without modeling the unknown mechanisms. We show that the evolution of any dynamical system is related to a predictive asymmetry that quantifies causal connections from limited observations. A built-in significance criterion obviates surrogate testing and drastically improves computational efficiency. We validate our test on numerous synthetic systems exhibiting behavior commonly occurring in nature, from linear and nonlinear stochastic processes to systems exhibiting nonlinear deterministic chaos, and on real-world data with known ground truths. Applied to the controversial problem of glacial-interglacial sea level and CO$_{2}$ evolving in lock-step, our test uncovers empirical evidence for CO$_{2}$ as a driver of sea level over the last 800 thousand years. Our findings are relevant to any discipline where time series are used to study natural systems.

[1]  B. Hannisdal Non-parametric inference of causal interactions from geological records , 2011, American Journal of Science.

[2]  James M. McCracken,et al.  Nonparametric causal inference for bivariate time series. , 2015, Physical review. E.

[3]  Daniel Chicharro,et al.  Reliable detection of directional couplings using rank statistics. , 2009, Physical review. E, Statistical, nonlinear, and soft matter physics.

[4]  Ute Feldmann,et al.  Predictability Improvement as an Asymmetrical Measure of Interdependence in bivariate Time Series , 2004, Int. J. Bifurc. Chaos.

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

[6]  James Theiler,et al.  Testing for nonlinearity in time series: the method of surrogate data , 1992 .

[7]  A. Krakovská,et al.  Use of False Nearest Neighbours for Selecting Variables and Embedding Parameters for State Space Reconstruction , 2015 .

[8]  M. Paluš,et al.  Comparison of six methods for the detection of causality in a bivariate time series. , 2018, Physical review. E.

[9]  N. Shackleton,et al.  The 100,000-year ice-Age cycle identified and found to lag temperature, carbon dioxide, and orbital eccentricity , 2000, Science.

[10]  A. Kraskov,et al.  Estimating mutual information. , 2003, Physical review. E, Statistical, nonlinear, and soft matter physics.

[11]  P. Clark,et al.  Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation , 2012, Nature.

[12]  Kristian Haaga UncertainData.jl: a Julia package for working with measurements and datasets with uncertainties , 2019, J. Open Source Softw..

[13]  G. Rangarajan,et al.  Multiple Nonlinear Time Series with Extended Granger Causality , 2004 .

[14]  Davide Chicco,et al.  Ten quick tips for machine learning in computational biology , 2017, BioData Mining.

[15]  B. Matthews Comparison of the predicted and observed secondary structure of T4 phage lysozyme. , 1975, Biochimica et biophysica acta.

[16]  B. Hannisdal,et al.  Transfer entropy computation using the Perron-Frobenius operator. , 2018, Physical review. E.

[17]  P. Grassberger,et al.  A robust method for detecting interdependences: application to intracranially recorded EEG , 1999, chao-dyn/9907013.

[18]  J. Jouzel,et al.  Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica , 1999, Nature.

[19]  R. Burke,et al.  Detecting dynamical interdependence and generalized synchrony through mutual prediction in a neural ensemble. , 1996, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics.

[20]  P. c. Tzedakis,et al.  A simple rule to determine which insolation cycles lead to interglacials , 2017, Nature.

[21]  C. Ramsey,et al.  Sea-level variability over five glacial cycles , 2014, Nature Communications.

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

[23]  B. Hannisdal,et al.  Forcing of late Pleistocene ice volume by spatially variable summer energy , 2018, Scientific Reports.

[24]  Schreiber,et al.  Measuring information transfer , 2000, Physical review letters.

[25]  Joseph T. Lizier,et al.  An Introduction to Transfer Entropy , 2016, Springer International Publishing.

[26]  M. Paluš,et al.  Directionality of coupling from bivariate time series: how to avoid false causalities and missed connections. , 2007, Physical review. E, Statistical, nonlinear, and soft matter physics.

[27]  Thomas F. Stocker,et al.  Revision of the EPICA Dome C CO2 record from 800 to 600 kyr before present , 2015 .

[28]  J. D. Hays,et al.  Variations in the Earth ' s Orbit : Pacemaker of the Ice Ages Author ( s ) : , 2022 .

[29]  George Sugihara,et al.  Distinguishing time-delayed causal interactions using convergent cross mapping , 2015, Scientific Reports.

[30]  C. Granger Investigating causal relations by econometric models and cross-spectral methods , 1969 .

[31]  Shawn D. Pethel,et al.  Transfer Entropy for Coupled Autoregressive Processes , 2013, Entropy.

[32]  L. Tsimring,et al.  Generalized synchronization of chaos in directionally coupled chaotic systems. , 1995, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics.

[33]  James P. Crutchfield,et al.  Information Flows? A Critique of Transfer Entropies , 2015, Physical review letters.

[34]  N. Pisias,et al.  THE EVOLUTION OF PLEISTOCENE CLIMATE: A TIME SERIES APPROACH , 1981 .

[35]  R. Röthlisberger,et al.  Glacial terminations as southern warmings without northern control , 2009 .

[36]  R. L. Edwards,et al.  The Last Glacial Termination , 2010, Science.

[37]  Bernhard Schölkopf,et al.  Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks , 2014, J. Mach. Learn. Res..

[38]  Gemma Lancaster,et al.  Surrogate data for hypothesis testing of physical systems , 2018, Physics Reports.

[39]  L. Lisiecki,et al.  A Late Pleistocene sea level stack , 2014 .

[40]  George Sugihara,et al.  Detecting Causality in Complex Ecosystems , 2012, Science.

[41]  J. Martinerie,et al.  Statistical assessment of nonlinear causality: application to epileptic EEG signals , 2003, Journal of Neuroscience Methods.

[42]  R Quian Quiroga,et al.  Performance of different synchronization measures in real data: a case study on electroencephalographic signals. , 2002, Physical review. E, Statistical, nonlinear, and soft matter physics.

[43]  Daniele Marinazzo,et al.  Kernel method for nonlinear granger causality. , 2007, Physical review letters.

[44]  F. Parrenin,et al.  The Antarctic ice core chronology (AICC2012): an optimized multi-parameter and multi-site dating approach for the last 120 thousand years , 2012 .

[45]  J. Croll “Climate and Time” , 1875, Nature.

[46]  Dmitry A Smirnov,et al.  Spurious causalities with transfer entropy. , 2013, Physical review. E, Statistical, nonlinear, and soft matter physics.

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

[48]  Ulrich Parlitz,et al.  Mixed State Analysis of multivariate Time Series , 2001, Int. J. Bifurc. Chaos.

[49]  Filip Hanzely,et al.  Testing for causality in reconstructed state spaces by an optimized mixed prediction method. , 2016, Physical review. E.

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

[51]  M. Loutre,et al.  An optimized multi-proxy, multi-site Antarctic ice and gas orbital chronology (AICC2012): 120--800 ka , 2012 .

[52]  A. Seth,et al.  Granger causality and transfer entropy are equivalent for Gaussian variables. , 2009, Physical review letters.

[53]  T. Teräsvirta,et al.  A general framework for testing the Granger noncausality hypothesis , 1999 .

[54]  G. Sugihara,et al.  Generalized Theorems for Nonlinear State Space Reconstruction , 2011, PloS one.