Tim Palmer

Abstract:

The authors examine the sensitivity of the Battisti coupled atmosphere-ocean model - considered as a forecast model for the E1 Niño-Southern Oscillation (ENSO) - to perturbations in the sea surface temperature (SST) field applied at the beginning of a model integration. The spatial structures of the fastest growing SST perturbations are determined by singular vector analysis of an approximation to the propagator for the linearized system. Perturbation growth about the following four reference trajectories is considered: (i) the annual cycle, (ii) a freely evolving model ENSO cycle with an annual cycle in the basic state, (iii) the annual mean basic state, and (iv) a freely evolving model ENSO cycle with an annual mean basic state. Singular vectors with optimal growth over periods of 3, 6, and 9 months are computed. The magnitude of maximum perturbation growth is highly dependent on both the phase of the seasonal cycle and the phase of the ENSO cycle at which the perturbation is applied and on the duration over which perturbations are allowed to evolve. However, the spatial structure of the optimal perturbation is remarkably insensitive to these factors. The structure of the optimal perturbation consists of an east-west dipole spanning the entire tropical Pacific basin superimposed on a north-south dipole in the eastern tropical Pacific. A simple physical interpretation for the optimal pattern is provided. In most cases investigated, there is only one structure that exhibits growth. Maximum perturbation growth takes place for integrations that include the period June-August, and the minimum growth for integrations that include the period January-April. Maxima in potential growth also occur for forecasts of ENSO onset and decay, while minima occur for forecasts initialized during the beginning of a warm event, after the transition from a warm to a cold event, and continuing through the cold event. The physical processes responsible for the large variation in the amplitude of the optimal perturbation growth are identified. The implications of these results for the predictability of short-term climate in the tropical Pacific are discussed.

Abstract:

Monsoon variability on intraseasonal and interannual time-scales is analysed using data from five 10-year European Centre for Medium-Range Weather Forecasts Atmospheric Model Intercomparison Project integrations, which differ only in their initial conditions. The results show that monsoon fluctuations within a season and within different years have a common dominant mode of variability. The spatial pattern of the common dominant mode in precipitation has a pronounced zonal structure, with one band of anomalous rainfall extending from 20°N to 5°N, covering most of the land areas, with the other band, of opposite sign, lying between 5°N and 10°S, mostly over the Indian Ocean. This mode therefore describes both the active/break monsoon spells associated with fluctuations of the Tropical Convergence Zone (TCZ) between the continental and the oceanic regime and the principal pattern of interannual variability of monsoon rainfall. In the observations the oscillations between active and break monsoon spells have similar behaviour, although the model is deficient in representing the rainfall variability over India. On the intraseasonal time-scale the transition between the two regimes seems to have a chaotic nature. In addition the probability density function of the principal mode is bimodal for the years in which this mode is particularly dominant. These two results indicate a possible similarity with the Lorenz 3-component chaotic model. Northward-propagating convective regions, simulated by the model, are not clearly associated with the phase transitions of the TCZ regime. The timing of the monsoon onset appears to be modulated by the phase of the El Niño/Southern Oscillation during the preceding season, consistent with observational studies. The results suggest that the dominant mode may also represent some components of the observed monsoon variability. The interannual fluctuations of the dominant mode exhibit only a weak level of reproducibility compared with the relatively large predictability of a broad-scale monsoon wind-shear index.

Abstract:

The dominant singular vectors of the tangent propagator of the European Centre for Medium-Range Weather Forecast (ECMWF) numerical weather-prediction model are an essential component of the ECMWF Ensemble Prediction System (EPS). These singular vectors describe the principal finite-time linear instabilities of the northern extratropical atmospheric circulation. The impact of increasing the horizontal resolution of the tangent model from T21 to T42 on three different types of initial perturbation, which make use of these singular vectors, is considered. The increase of resolution allows the possibility of describing more accurately instabilities with an upscale cascade of energy from subsynoptic to synoptic scales. Two of the perturbations are referred to as the pseudo-inverse and sensitivity vectors. These are both diagnostic, and involve estimating from the short-range forecast error the component of initial error in the unstable subspace. The third type of perturbation is used to construct the set of initial states for the ensemble prediction scheme. Linear and nonlinear integrations are described using these different types of perturbation. All the results point to the conclusion that the higher-resolution calculations lead to more accurate results. Specifically, it is shown that significant amounts of energy in the pseudo-inverse and sensitivity vectors reside in scales smaller than T21, and that the skill of the EPS is improved when generating the initial perturbations using T42 singular vectors. This is found to be especially true in a number of cases when the control forecast was particularly poor. As a consequence, the predictability of synoptic-scale disturbances in the atmosphere on time-scales of a few days is likely to be determined by errors in the initial state on subsynoptic scales. In addition to these conclusions, it is proved that the analysis of the relationship between the EPS and the pseudo-inverse perturbations can be used to define the initial amplitude of the EPS perturbations.