Trends in the potential spread of seasonal climate simulations over South Africa. (5th December 2014)

• Record Type: Journal Article
• Authors: Lawal, Kamoru A.; Stone, Dáithí A.; Aina, Tolu; Rye, Cameron; Abiodun, Babatunde J.
• Is Part Of: International journal of climatology. Volume 35:Number 9(2015)
• Page Start: 2193
• Page End: 2209
• Publication Date: 2014-12-05
• Subjects: Climatology -- Periodicals; Climat -- Périodiques; Climatologie -- Périodiques; 551.605
• DOI: 10.1002/joc.4234
• Languages: English
• ISSNs: 0899-8418

Abstract

<abstract abstract-type="main" id="joc4234-abs-0001"> <title>ABSTRACT</title> <p id="joc4234-para-0001">This study assesses the existence and importance of trends in the spread of South African climate simulations covering 50 years of a large initial‐condition ensemble from a dynamical atmospheric model. It quantifies ensemble spread using two contrasting measures – standard deviation and 10–90th percentile range. The study then evaluates and examines the characteristics of long‐term trends in the ensemble spread in relation to trends in the ensemble mean and in the observational record, by considering the skill of the monthly mean precipitation and near surface air temperature simulations. Results provide evidence that variations in ensemble spreads generated by the atmospheric model used in this study reflect fundamental properties of atmospheric variability in the real climate system. We find significant long‐term trends in the measures of spread, with a general coastal–inland gradient, suggesting the possibility of existence of interannual variations in the potential range of seasonal climate simulations over South Africa. We also find robust relationships between trends in the observational record, in the simulated ensemble means and in measures of the simulated ensemble spread. Irrespective of the direction of trends, the correspondence of higher model skill when trends in the ensemble spread are larger suggests that the skill produced by a dynamical modelling system may not be independent of the model ability to capture the real atmospheric trends in whatever the model is simulating or forecasting. Therefore, based on historical data, further understanding of how potential predictability is changing has the prospect to improve the interpretation of current estimates of simulation skill.</p> </abstract>