Title: Yet Another Method for Compiling Climate Statistics.

Abstract: I'll talk about a tool I've created for partitioning a gridded field into areas associated with a target phenomenon and those that are not. The target phenomena I'll talk about are extra-tropical low pressure systems (a.k.a. cyclones). For this to work a method is needed for finding cyclone centers and track them in high time resolution SLP fields (6 hours or less). These centers are then used as seeds for a search that delimits the boundaries of each system. As will be discussed this process is less than straightforward as cyclones lack an absolute boundary and are often complex and even nested structures. The resulting database can be combined with other fields to investigate cyclone properties such as how cyclone related cloud varies with season, location and cyclone intensity or life-cycle. It can also be used as a sort of GIS (geographic information system) to extract specific cases such as just the cyclones that eventually pass near a certain location(s) (e.g., an ARM site or into the Arctic basin) or that have a combination of properties (e.g., only DJF high intensity systems that never pass over the ocean). This provides a powerful and specific way to inter-compare GCMs, compare them with reanalysis results or observations, or aid in model development. The non-cyclone part of a field can be equally interesting. For example, contrasting the cyclone and non-cyclone partitions can isolate cases where a quantity (say precipitation or dust load) is over or under determined by cyclone activity in a model. Examples and pretty pictures will be shown.