Statistics and Model Diagnostics Research Section
The section is responsible for research and development of methodology for post-processing of NWP model output, and all aspects of research into verification methods. Much of the section's research activity involves statistical interpretation of model output, but we also are responsible for dependent models which using the output of NWP models.
We innovate by bringing into meteorological applications statistical methods that have been more widely used in other fields, such as multiple discriminant analysis, fuzzy inference systems, and classification and regression trees (CART). We also adapt strategies for statistical interpretation to suit the requirements of operational forecasting in Canada. Examples of new types of forecast products developed in the section are the stratospheric ozone forecasts used as the basis of the Canadian UV Index, special forecasts for the Calgary Winter Olympics, and updateable model output statistics (UMOS).
On the topic of forecast verification, the section has produced an international standard reference on verification methods, "A survey of common verification methods in Meteorology" (Stanski, Wilson and Burrows), published internally in 1989, then given worldwide distribution by WMO in 1990. More recent research activities have focused on verification methods for ensemble forecasts.
On the topic of dependent modelling, the section is responsible for the research in support of wind-wave modelling in Canada, including the adaptation and development of international community wave models. In recent years, we have been developing, extending and testing the European wave model WAM for use in Canadian waters, at high resolution on inland lakes and at lower resolution for the North Atlantic and North Pacific.
In the past twenty years, the section has produced over twenty separate forecast products that have been implemented and used in forecast operations, and more than 40 articles in refereed meteorological and oceanographic journals.
