I get all the news I need on the weather report, October 2010: Presentation

Abstract

In order to produce a weather forecast, the initial state of the atmosphere needs to be entered into the forecast model taking into account factors such as temperature, wind and moisture. However, the scale of some atmospheric motions is far too small for proper observation and representation in the model. It therefore becomes inevitable that the initial state is determined with some level of error. The successive nature of forecasts causes minor errors in the initial data to increase and eventually have an effect on the accuracy of the forecast due to the nonlinear and unstable characteristics of the atmospheric flow. The author explains how the CSIR Climate Studies and Modelling (CSM) group uses common forecast systems for: high-resolution weather forecasting which is over one to ten days; long-range forecasting which ranges over a period of weeks or even months; as well as decadal and multi-decadal climate change projections. The CSIR then provides the South African Weather Service with information that contributes to computing a more accurate forecast on a daily basis. The author noted that noted that over recent years, the skill of Numerical Weather Prediction (NWP) in the Southern Hemisphere has reached the level of that in the Northern Hemisphere and that there is a strong international as well as national focus on improving our forecast and projection capabilities and skill. According to Landman, the accuracy of weather forecasts and climate projections is highly dependent on consistent observations of the atmosphere, ocean and land. However, the unavoidable level of error in the determination of the initial state, the chaotic behaviour of the atmosphere and uncertainties associated with future global ocean and land surface changes as well as anthropogenic forcings, require that weather - and climate-related forecast or statements be issued probabilistically. And as such, it is not always the weatherman's fault when he seems to get things wrong sometimes. The presentation was part of the 'Street smart; Science Talk' series hosted by the Knowledge Commons on 26 October 2010.