Seasonal forecasting of synoptic type variability: potential intraseasonal predictability relevant to the Cape south coast of South Africa

Abstract

An ensemble of 12 sea-level pressure (SLP) simulations from the United Kingdom Meteorological Office (UKMO) Global Seasonal Forecast System 5 (GloSea5) is used to investigate the potential predictability of synoptic types within 14 austral spring seasons (September-November, (SON)) for the 14 yr period from 1996 to 2009. Daily SLP model fields for the 14 SON seasons are mapped to the corresponding observed synoptic types, using self-organising maps (SOMs). Predictability of intraseasonal synoptic type characteristics is evaluated by comparing the frequency of synoptic types as simulated by GloSea5 against the observed synoptic type occurrence. Intraseasonal circulation variability for the Cape south coast of South Africa at interannual time scales is found to be predictable, although poorly. In this paper, the authors aim to assess the within-SON-season predictability by considering the predicted frequencies of different synoptic types affecting seasonal climate variations over the Cape south coast. If it can be shown that this intraseasonal variation in circulation over the Cape south coast can be predicted skillfully at interannual time scales, it will imply that skillful rainfall forecasts can be generated through the use of predicted daily circulation statistics.