Presenter:
Megan Kirchmeier-Young
When:
November 18, 2015 - 3:00pm to 4:00pm
Where:
Room 002, University House 1
2489 Sinclair Rd. Victoria , BC

Abstract: Many studying climate impacts rely on downscaling to extract high-resolution data from coarse-resolution climate models. A probabilistic downscaling approach, like that developed at the University of Wisconsin-Madison, predicts a probability density function (PDF) for each day that describes what could have occurred at the local scale that day, given the large-scale meteorological conditions. While probabilistic downscaling is shown to have many advantages, daily-varying PDFs present a challenge for verification, as only a single observed value is available each day for comparison. As such, there are two main goals for this study: (1) to develop a methodology for verification of a probabilistic climate dataset, and (2) to identify a set of metrics that describe climate characteristics important to the users of this downscaled dataset. These user-minded metrics emphasize the many ways to characterize "extreme" events in temperature and precipitation.

About the speaker: Megan Kirchmeier-Young is a new post doctoral fellow with the Canadian Sea Ice and Snow Evolution Network (CanSISE) project, jointly affiliated with CCCma and PCIC. She will be presenting work from her Ph.D. in Atmospheric and Oceanic Sciences at the University of Wisconsin-Madison. Megan's research focuses on using statistics to explore topics in regional climatology.