Estimating causal networks in biosphere–atmosphere interaction with the PCMCI approach

Krich, Christopher; Runge, Jakob; Miralles, Diego G.; Migliavacca, Mirco; Perez-Priego, Oscar; El-Madany, Tarek; Carrara, Arnaud; Mahecha, Miguel D.

doi:https://doi.org/10.5194/bg-17-1033-2020

Articles | Volume 17, issue 4

https://doi.org/10.5194/bg-17-1033-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/bg-17-1033-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 17, issue 4

Research article

|

26 Feb 2020

Research article |

| 26 Feb 2020

Estimating causal networks in biosphere–atmosphere interaction with the PCMCI approach

Christopher Krich, Jakob Runge, Diego G. Miralles, Mirco Migliavacca, Oscar Perez-Priego, Tarek El-Madany, Arnaud Carrara, and Miguel D. Mahecha

Data sets

TIGRAMITE-Causal discovery for time series datasets J. Runge https://github.com/jakobrunge/tigramite/

Climatic Research Unit (CRU): Time-series (TS) datasets of variations in climate with variations in other phenomena v3 University of East Anglia Climatic Research Unit, P. D. Jones, and I. C. Harris http://catalogue.ceda.ac.uk/uuid/3f8944800cc48e1cbc29a5ee12d8542d

A Non-Stationary 1981–2012 AVHRR NDVI3g Time Series (https://ecocast.arc.nasa.gov/data/pub/gimms/3g.v1/) J. E. Pinzon and C. J. Tucker https://doi.org/10.3390/rs6086929

Short summary

Causal inference promises new insight into biosphere–atmosphere interactions using time series only. To understand the behaviour of a specific method on such data, we used artificial and observation-based data. The observed structures are very interpretable and reveal certain ecosystem-specific behaviour, as only a few relevant links remain, in contrast to pure correlation techniques. Thus, causal inference allows to us gain well-constrained insights into processes and interactions.