Diagrammatic Representations of Uncertainty in Meteorological Forecasting

Abstract

Meteorological prediction is complex, but an important scientific practice to help keep people safe. In the past few decades, forecasts have become much more accurate. However, this accuracy has not removed uncertainty from forecasts. Indeed, it is widely acknowledged that weather forecasting is an inherently uncertain practice. Given that the aims of weather prediction are practical (to help keep people and property safe), it is important that weather forecasts be effectively communicated to the broader public. But, there is an interesting question here about how the uncertainty (itself an interesting epistemological concept) of the forecasts is communicated. One of the ways this is done by the National Weather Service in the United States is to use various forms of graphical and diagrammatic representations to identify and communicate the uncertainty that remains as part of a forecast. In this essay, I will explore some of these diagrammatic and graphical representations to identify the ways in which uncertainty is expressed. I argue that there is an interesting and important way in which the aims of meteorological practice constrain the representations—such that they do not aim at perfect (or true) representations, but rather at products which will help change behavior to save lives.