Comments on Statistical Results
The model outputs are volume averages on a horizontal 5 km × 5 km grid, whereas the observations with which they are compared are point values that may in reality differ considerably from the averages. The observation sites were far from homogeneous on the grid scale, with differing terrain heights, slope inclinations, land usage, and soil conditions all of particular importance under the conditions of strong thermal forcing in phase I.
The rapidly changing synoptic conditions of phase II, including a frontal passage, were not adequately imposed, through either the lateral boundary conditions or internally, in any of the models. The RAMS model, using grid nesting, made the most ambitious attempt to represent these conditions. Even in this case, however, the verification region, with its high terrain, seemed not to be able to adjust sufficiently rapidly to the imposed larger-scale changes.
The modelers were free to decide their own initialization and data assimilation procedures, so that, inevitably, the results represent differing levels of sophistication of external forcing.
Some modelers were able to carry out a series of experiments to determine the most appropriate values of some assigned model parameters—for example, surface albedo and ground wetness—before generating the results published here. The results of such experiments are, in these cases, discussed in the text. They throw useful light on the physics of the modeled phenomena, particularly for phase I; only in the case of the RAMS model was such an approach adopted for phase II. Those results reported here that have not been preceded by such experiments must therefore not be regarded as reflecting the full potential of the models used.