Karner Blue Butterfly Emergence Model

Thumbnail links to larger image
Weather station locations. [PDF, 670KB]

The DNR has partnered with the University of Wisconsin - Department of Forestry and Wildlife Ecology to develop a useful tool for HCP partners and DNR property managers. The Karner Blue Butterfly emergence model uses climate software to track weather patterns and calculate degree days to predict when Karner blues will begin to fly, and when the peak flight is occurring. At each station, degree days are accumulated using near real-time data of minimum and maximum temperature for the current year up to the date indicated. Beyond this date, the model uses mean minimum and maximum temperatures calculated from historic weather data for the climate normal period of 1971-2000. When real-time data is missing for a given day, the calculated mean is also used. The accuracy of the model has not been validated, and will be tested throughout this year. A similar methodology has shown predictions to be within an average of 5 days of the actual dates.

Flight Predictions using current weather data through 6/27/2011

Station	Emergence First Flight	Peak First Flight	Emergence Second Flight	Peak Second Flight	% Days With Missing Data
Berlin	06/04/2011	06/19/2011	07/17/2011	08/03/2011	16%
Eau Claire	05/29/2011	06/08/2011	07/11/2011	07/28/2011	16%
Friendship	05/31/2011	06/09/2011	07/14/2011	07/31/2011	13%
Grantsburg	05/29/2011	06/15/2011	07/18/2011	08/07/2011
Hancock	05/30/2011	06/12/2011	07/13/2011	07/30/2011	68%
Hinckley	06/05/2011	06/22/2011	07/26/2011	08/18/2011	16%
Mather	06/01/2011	06/15/2011	07/18/2011	08/06/2011	36%
Neillsville	06/01/2011	06/15/2011	07/19/2011	08/09/2011	15%
Sparta	05/28/2011	06/08/2011	07/10/2011	07/26/2011	15%
Stevens Point	06/04/2011	06/18/2011	07/19/2011	08/06/2011	15%
Waupaca	06/02/2011	06/14/2011	07/15/2011	08/01/2011	16%
Wild River	05/28/2011	06/10/2011	07/16/2011	08/04/2011	13%
Wisconsin Rapids	05/31/2011	06/12/2011	07/15/2011	08/01/2011	13%

*Percent of days since Aprl 1 with missing data. Missing data is filled in using historical means from that station. The effect of the missing data on the model predictions depends on how much the actual temperature differ from historical means.

We will be testing and improving this tool throughout the year. If you have any suggestions, observations, or questions that would help us validate this system please let us know by sending an e-mail. Thank You!

Last Revised: Tuesday July 05 2011