On the Effect of Model Parameters on Forecast Objects

Many physics-based numerical models produce a gridded, spatial field of forecasts, e.g., a temperature "map". However, the field for some quantities such as precipitation generally consists of spatially coherent and disconnected "objects". Certain features of these objects (e.g., number, size, and intensity) are generally of interest. Here, a methodology is developed for assessing the impact of model parameters on features of forecast objects. Although, in principle, the objects can be defined by any means, here they are identified via clustering algorithms. The methodology is demonstrated on precipitation forecasts from a mesoscale numerical weather prediction model.