Chapter 10 - Practice—evaluating forecast skill

Abstract

One recurrent theme throughout this book is about providing effective quantitative information from a decision-support perspective. Evaluating forecast skill is a critical component of hazard early warning systems. When we work with a weather, climate, or hydrologic prediction or estimation, how do we know how good is it? It turns out that there are a number of different ways to consider this question, each with specific strengths focusing on the specific questions being asked and the way that the data is being used. Here we describe some of the most commonly used forecast skill metrics. Forecast skill evaluation provides a rationale for the application of the forecasts for drought early warning and, hence, supports decision-making. Therefore forecast skill evaluation is an important component of implementing a drought early warning system (DEWS). This chapter describes commonly used forecast evaluation methods, their usages, and their potential strengths and weaknesses, as well as providing examples of how forecast skill evaluation guides the application of such methods in DEWSs.