Correlation or Causality: Drivers of the Earth climate

Here, we analyze recent measured data on global mean surface air temperature anomalies (GMTA) and various external forcing’s covering the last 160 years using newly developed techniques that allow discrimination between correlation and causality. This evaluation is based on a new concept for calculating the information flow between time series. The strong correlation between the global CO2 concentration and GMTA of 0.889±0 is well known. But calculating the information flow in nat (natural unit of information) per unit time from the global CO2 concentration to GMTA we get 0.348±0.112 [nat/year] and -0.006±0.003 [nat/year] in the reverse direction. Causality is expressed by an information flow significantly different to 0.0, whereas an information flow close to 0.0 indicates that the two time series are not causally related. Our result demonstrates one-way causality in the sense that the CO2 increase is causing the temperature increase and not the other way around. The positive value of the information flow indicates further that CO2 has a positive feedback and therefore a destabilizing effect on GMTA; more CO2 would lead to a stronger increase in GMTA. The results of investigating the information flow between the major radiative forcings and the GMTA time series clearly show that total Green House Gases (GHG), dominated in particular by CO2 forcing, is the main driver of changing global surface air temperature. Radiative forcing caused by aerosols and clouds is still important, but significantly smaller. Neither forcing by solar irradiance nor volcanic forcing contributes in a significant manner to the GMTA development. However on paleological time scales (800,000 years) the climate system seems to have behaved different, as on those time scales temperature changes seem to be the cause for subsequent changes in CO2 concentration. Finally, we applied the same causality analysis to the globally-gridded GMTA product in order to assess regional “sensitivity” to anthropogenic forcings versus natural modes of variability. This analysis reveals a surprising causal pattern: the increased anthropogenic forcing mainly originated in the northern hemisphere has especially strong warming effects in the southern hemisphere.

STIPS Adolf;  MACIAS MOY Diego;  GARCIA GORRIZ Elisa; 

2016-01-18

University of Landau

JRC95414

https://publications.jrc.ec.europa.eu/repository/handle/JRC95414,