Improvement Of Prediction Skill Research Articles

Decadal prediction of Sahel rainfall using dynamics-based indices

At decadal time scales, the capability of state-of-the-art atmosphere-ocean coupled climate models in predicting the precipitation in Sahel is assessed. A set of 14 models participating in the Coupled Model Intercomparison Project Phase 5 (CMIP5) is selected and two experiments are analysed, namely initialized decadal hindcasts and forced historical simulations. Considering the strong linkage of the atmospheric circulation signatures over West Africa with the rainfall variability, this study aims to investigate the potential of using wind fields for decadal predictions. Namely, a West African monsoon index (WAMI) is defined, based on the coherence of low (925 hPa) and high (200 hPa) troposphere wind fields, which accounts for the intensity of the monsoonal circulation. A combined empirical orthogonal functions analysis is applied to explore the wind fields’ covariance modes, and a set of indices is defined on the basis of the identified patterns. The WAMI predictive skill is assessed by comparing WAMI from coupled models with WAMI from reanalysis products and with a standardized precipitation index (SPI) from observations. Results suggest that the predictive skill is highly model dependent and it is strongly related to the WAMI definition. In addition, hindcasts are more skilful than historical simulations in both deterministic and probability forecasts, which suggests an added value of initialization for decadal predictability. Moreover, coupled models are more skilful in predicting the observed SPI than the WAMI obtained from reanalysis. WAMI performance is also compared with decadal predictions from CMIP5 models based on a Sahelian precipitation index, and an improvement in predictive skill is observed in some models when WAMI is used. Therefore, we conclude that dynamics-based indices are potentially more effective for decadal prediction of precipitation in Sahel than precipitation-based indices for those models in which Sahel rainfall variability is not well simulated. We thus recommend a two-fold approach when testing the performance of models in predicting Sahel rainfall, based not only on rainfall but also on the dynamics of the West African monsoon.

RE-ASSESSMENT AND UPDATE OF BULK LONGSHORE SEDIMENT TRANSPORT FORMULAS

Longshore sediment transport (LST) is one of the main drivers of beach morphology. Bulk LST formulas are routinely used in coastal management/engineering studies to assess LST rates and gradients. Over 50 years of research has resulted in several bulk LST formulas that have been tested with varying levels of rigor. In this study, the predictive skill of one of the most recent bulk LST formulas (Bayram et al., 2007) is evaluated. The calibration coefficients in the formula are improved using a least-squares optimization algorithm, resulting in a significant improvement in predictive skill. The generality of the improved formula is verified via the statistical methods of bootstrapping and cross-validation.

Toward Random Sampling of Model Error in the Canadian Ensemble Prediction System

Abstract An updated global ensemble prediction system became operational at the Meteorological Service of Canada in July 2007. The new elements of the system include the use of 20 members instead of 16, a single dynamical core [the Global Environmental Multiscale (GEM) model], stochastic physical tendency perturbations and a kinetic energy backscatter algorithm, an ensemble Kalman filter with four-dimensional data handling, and a decrease from 1.2° to 0.9° in horizontal grid spacing. This system is compared with the former operational one using a variety of probabilistic measures. For global upper-air dynamical fields, the improvement in predictive skill for equivalent forecast quality is from 9 to 16 h around day 6. Precipitation forecasts, verified over Canada, are also significantly improved. The impact of each of the abovementioned new elements of the ensemble prediction system is also evaluated separately in a series of sensitivity experiments for which one given element is removed from the system.

A new tide model for the Antarctic ice shelves and seas

AbstractWe describe a new tide model for the seas surrounding Antarctica, including the ocean cavities under the floating ice shelves. The model uses data assimilation to improve its fit to available data. Typical peak-to-peak tide ranges on ice shelves are 1–2 m but can exceed 3 m for the Filchner–Ronne and Larsen Ice Shelves in the Weddell Sea. Spring tidal ranges are about twice these values. Model performance is judged relative to the ~5–10 cm accuracy that is needed to fully utilize ice-shelf height data that will be collected with the Geoscience Laser Altimeter System, scheduled to be launched on the Ice, Cloud and land Elevation Satellite in late 2002. The model does not yet achieve this level of accuracy except very near the few high-quality tidal records that have been assimilated into the model. Some improvement in predictive skill is expected from increased sophistication of model physics, but we also require better definition of ice-shelf grounding lines and more accurate water-column thickness data in shelf seas and under the ice shelves. Long-duration tide measurements (bottom pressure gauge or global positioning system) in critical data-sparse areas, particularly near and on the Filchner–Ronne and Ross Ice Shelves and Pine Island Bay, are required to improve the performance of the data-assimilation model.

Medium-Range Forecasting—The Experience of ECMWF

The result of five years of operational medium-range forecasts at ECMWF is discussed. It is shown that a considerable improvement in predictive skill has taken place, resulting in more accurate forecasts and an extension in time of useful predictive skill. For the Northern Hemisphere extratropics, the improvement in skill during the winter season is about three days for the 500-mb geopotential compared to the very first experiments in the early 1970s. There are large geographical and temporal variations in the performance of the model, as well as large-scale errors of a systematic nature. Numerical experiments have indicated great sensitivity to the representation of orography and to tropical forcing.