The Southern Hemisphere Dynamics of Forecast Uncertainty in the THORPEX Interactive Grand Global Ensemble (TIGGE)

Global ensemble models have been used for years to help account for forecast uncertainty. For ensemble models in TIGGE, it has been found that at earlier forecast times, uncertainties initially grow fastest at synoptic scales in areas that are most sensitive to baroclinic instability. As forecast times increase, the uncertainty grows exponentially and saturates. In the Northern Hemisphere extratropics, the areas of higher uncertainty form pockets that form a wave-like pattern determined by the large scale flow. In the Southern Hemisphere extratropics, the large scale flow is primarily zonal, so the waves in the large scale flow are not as prominent.

This project used local linear diagnostics, along with a Lorenz-curve based technique, to investigate the forecast uncertainty dynamics in the SH extratropics for model runs from January 1st, 2012 to February 29th, 2012. While results varied depending on the ensemble model, the areas of larger uncertainty tended to form areas of higher uncertainty where high frequency (synoptic scale) and low frequency transients were present. Spatial analysis showed that while the ensembles effectively captured the general characteristics of the uncertainty, the spread of the ensembles generally underestimated the magnitude and shape of the areas of larger uncertainty as forecast times increased. Also, the bias of these ensembles indicated that there may be systematic errors that grew as forecast times increased due to errors in predicting low frequency changes in the large scale flow. The results were similar to what has been found in the NH extratropics, despite the SH extratropics having primarily zonal flow during the time period investigated.