Approaches to defining a priori performance expectations in model benchmarking

While benchmarking within the land surface community is transitioning to more quantitative evaluation techniques, there is little discussion of the level of performance we should expect from a model. Expectations are clearly dependent on the nature of a particular experimental setup and the relative complexity of the model being used. This talk will show how empirical models offer one approach to defining performance expectations a priori and argue that understanding expectations should be a crucial part of any benchmarking exercise. An automated implementation of empirical model based benchmarking will be shown from the Protocol for the Analysis of Land Surface Models (PALS), a free online land surface model benchmarking tool.