What do we mean by consistency in weather forecasts?

Many sectors rely upon weather information, yet raw numerical weather prediction (NWP) output alone is rarely sufficient to support real‐world decisions. Post‐processing has therefore been developed to make NWP data more usable. However, different post‐processing approaches produce data with different properties, preserving some kinds of consistency while discarding others. At present, it is difficult to discuss these techniques clearly because there is no shared understanding of what is meant by consistency. This paper explores different types of consistency in forecast products and proposes a set of definitions to support clearer communication and the design of post‐processing systems.