Professor Myles Allen CBE FRS

Abstract:

Realistic simulation of the internal variability of the climate system is important both for climate change detection and as an indicator of whether the physics of the climate system is well-represented in a climate model. In this work zonal mean atmospheric temperatures from a control run of the second Hadley Centre coupled GCM are compared with gridded radiosonde observations for the past 38 years to examine how well modelled and observed variability agree. On time scales of between six months and twenty years, simulated and observed variability of global mean temperatures agree well for the troposphere, but in the equatorial stratosphere variability is lower in the model than in the observations, particularly at periods of two years and seven to twenty years. We find good agreement between modelled and observed variability in the mass-weighted amplitude of a forcing-response pattern, as used for climate change detection, but variability in a signal-to-noise optimised fingerprint pattern is significantly greater in the observations than in a model control run. This discrepancy is marginally consistent with anthropogenic forcing, but more clearly explained by a combination of solar and volcanic forcing, suggesting these should be considered in future 'vertical detection' studies. When the relationship between tropical lapse rate and mean temperature was examined, it was found that these quantities are unrealistically coherent in the model at periods above three years. However, there is a clear negative lapse rate feedback in both model and observations: as the tropical troposphere warms, the mid-tropospheric lapse rate decreases on all the time scales considered.

Abstract:

This paper addresses some fundamental methodological issues concerning the sensitivity analysis of chaotic geophysical systems. We show, using the Lorenz system as an example, that a naive approach to variational ('adjoint') sensitivity analysis is of limited utility. Applied to trajectories which are long relative to the predictability time scales of the system, cumulative error growth means that adjoint results diverge exponentially from the 'macroscopic climate sensitivity' (that is, the sensitivity of time-averaged properties of the system to finite-amplitude perturbations). This problem occurs even for time-averaged quantities and given infinite computing resources. Alternatively, applied to very short trajectories, the adjoint provides an incorrect estimate of the sensitivity, even if averaged over large numbers of initial conditions, because a finite time scale is required for the model climate to respond fully to certain perturbations. In the Lorenz (1963) system, an intermediate time scale is found on which an ensemble of adjoint gradients can give a reasonably accurate (O(10%)) estimate of the macroscopic climate sensitivity. While this ensemble-adjoint approach is unlikely to be reliable for more complex systems, it may provide useful guidance in identifying important parameter-combinations to be explored further through direct finite-amplitude perturbations.