Workpackage 1: Causes of temperature change, and estimate of transient climate sensitivity

Workpackage 1 applies fingerprint methods to interpret the observed record of climate change.

The challenge in interpreting observed records of climate change is that observations reflect a combination of climate variations generated within the climate system ("internal climate variability") with changes associated with external forcing ("signals" or "fingerprints"; see Hegerl and Zwiers, 2011).

In this WP, fingerprint methods (Hasselmann, 1997, Allen and Tett, 1999) are applied to interpret the observed record. The observed climate change y is considered to be a linear combination of externally forced signals X, scaled to match the observations with an amplitude vector a, and residual internal climate variability. Fingerprints for external forcing represent the expected climate response to greenhouse gas increases, solar radiation changes, volcanism, and other anthropogenic influences. Fingerprints are derived from CMIP5 simulations of the last millennium, combined with those of the 20th century. The observations are analyzed to determine if they show evidence for the presence of the response to individual forcings, and if fingerprints for combinations of forcing can be detected and separated from each other. Fingerprints are sampled in the same way as observations.

A measure for climate variability is obtained from climate model simulations and, through the regression residual, from the observations. A scaling vector a accounts for possible errors in the amplitudes of the external forcing and the climate model’s response. It also provides the basis on which to decide whether a particular fingerprint is present in the observations.

The analysis is performed across different temporal and spatial foci, from inter-decadal to secular, covering sub-periods, the early anthropocene, and the entire record, in order to identify robust results. Instrumental data is used alone, and combined with proxy–based records, e.g. large-scale spatial averages (Tropics, Europe, Wilson et al., 2007; Luterbacher et al., 2004).

TITAN makes much more complete use of available data and provides a robust identification of the climate response to different external drivers. The challenge is to combine multiple sources of evidence to fully account for changing coverage and uncertainty. The analysis also results in an estimate, from observations, of the evolving greenhouse warming. This estimate is interpreted as an estimate of the transient climate response (TCR, Stott et al., 2006). The uncertainties propagated in the analysis yield the most rigorous estimate available so far, and the longer time horizon constrains it more than recent changes alone.

This WP addresses question i) and vi), and contributes to addressing question ii).