Professor Myles Allen CBE FRS

Abstract:

The weather at middle latitudes is largely unpredictable more than a week or so in advance, whereas fluctuations in the ocean may be predictable over much longer timescales. If decadal fluctuations in North Atlantic sea surface temperature could be predicted, it might be possible to exploit their influence on the atmosphere to forecast decadal fluctuations in climate. Here we report analyses of shipboard observations that indicate significant decadal predictability of North Atlantic sea surface temperature, arising from the advective propagation of sea-surface-temperature anomalies and the existence of a regular period of 12-14 years in the propagating signals. The same timescale can be identified in a dipole-like pattern of North Atlantic sea-level pressure variability. We propose a mechanism which may connect these oceanic and atmospheric fluctuations, possibly as part of a coupled ocean-atmosphere mode of variability. Our results are encouraging for the prospects of forecasting natural fluctuations in the climate of the North Atlantic region several years in advance.

Abstract:

Recent work suggests a discernible human influence on climate. This finding is supported, with less restrictive assumptions than those used in earlier studies, by a 1961 through 1995 data set of radiosonde observations and by ensembles of coupled atmosphere-ocean simulations forced with changes in greenhouse gases, tropospheric sulfate aerosols, and stratospheric ozone. On balance, agreement between the simulations and observations is best for a combination of greenhouse gas, aerosol, and ozone forcing. The uncertainties remaining are due to imperfect knowledge of radiative forcing, natural climate variability, and errors in observations and model response.

Abstract:

Extended empirical orthogonal functions (EEOFs), alternatively known as multi-channel singular systems (or singular spectrum) analysis (MSSA), provide a natural method of extracting oscillatory modes of variability from multivariate data. The eigen-functions of some simple non-oscillatory noise processes are, however, also solutions to the wave equation, so the occurrence of stable, wave-like patterns in EEOF/MSSA is not sufficient grounds for concluding that data exhibits oscillations. We present a generalisation of the "Monte Carlo SSA" algorithm which allows an objective test for the presence of oscillations at low signal-to-noise ratios in multivariate data. The test is similar to those used in standard regression, examining directions in state-space to determine whether they contain more variance than would be expected if the noise null-hypothesis were valid. We demonstrate the application of the test to the analysis of interannual variability in tropical Pacific sea-surface temperatures.