ABSTRACT

This is one of two verification chapters of the book. Whereas the verification of deterministic forecasts is the topic of this chapter, probabilistic forecast verification is topic of the next. The discussion here resolves around Murphy's conception on what constitutes a good forecast. In that, three types of goodness, namely, consistency, quality, and value are thoroughly examined. Consistency refers to the correspondence between forecasts and forecaster's judgment, whereas quality refers to the correspondence between forecasts and observations. Last but not least, value refers to the incremental benefits, which is often gauged in monetary terms, of using the forecast of interest from using some other forecasts. Indeed, forecast possesses no intrinsic value, but can only acquire value through its ability to influence decision-making. This chapter also exemplifies a complete workflow of deterministic forecast verification using state-of-the-art data.