Abstract : 3P.2
Subjective verification of high resolution and ensemble models during MAP D PHASE

Marco Stoll, Lionel Fontannaz, Paolo Ambrosetti
marco.stoll@meteoswiss.ch
MeteoSwiss

Within the framework of MAP D-PHASE large amount of high resolution model data as well as automated, model generated alerts will be available to forecasters to support their operational warning procedures of intense precipitation.

Model output can be verified against observations in an objective way, but the outcome of the verification may not always reflect the subjective feeling of forecasters. In warning cases, the forecasters are often in direct contact with end users and play the important role of an interface between the numerical model on the one hand and the decision-maker on the other. In this situation the value of the high-resolution and ensemble model output available should be evaluated from the forecasters point of view in order to find an optimal way in using the latest generation of atmospheric models.

Therefore we suggest a subjective evaluation protocol to be completed in "real time" by the forecasters. It will tackle the following questions to identify benefits and problems regarding additional model information:

Is there any quality gain in the precipitation forecast descending from the coarse to the high resolution local models? Comparison between ensemble and high resolution models, depending on lead time? Are synoptic an climatologic considerations more important than pure model output in quantifying precipitation amounts? Are forecasters in a better position assessing end users by having additional information?

In order to illuminate the role and benefit of the forecaster's work, we try to evaluate also the cognitive aspects in the forecaster's warning process. Knowledge about these aspects can be used later for training and development of decision tools.

Preliminary samples of MAP D-PHASE model data will be shown as well as an overview of the evaluation protocol, how it is applied in operational forecast shifts.