Studying regime shifts in ice freeze-up and break-up dates as well as ice cover duration on lakes, which are sensitive to the air temperature, provides us with useful information about climate change and linkages to atmospheric teleconnection patterns. Such data, however, typically show a positive serial correlation, which implies that a positive/negative observation tends to be followed by a positive/negative observation in the future. The positive serial correlation among data is known to lead to deflation of p-values and the related inflation of significance levels of most statistical procedures and tests, making the obtained results unreliable.
In this talk, we present various approaches for detecting regime shifts using the Lake Baikal ice phenology data for the period 1896-1996. Since most regime shift tests are derived under the assumption of independence that is typically not justified for hydrological processes, we suggest employing the sieve-bootstrap method to account for serial correlation among observations and to yield more accurate and reliable estimates of regime shifts. Furthermore, we discuss possible connections between the estimated regime shifts and the Pacific Decadal Oscillation (PDO).
Keywords: Change-point analysis; Ice phenology; Sieve-bootstrap method
Biography: Kimihiro is a PhD candidate at the Department of Statistics, University of California, Davis.
He obtained his Master's degree in Statistics from the University of Waterloo in 2009 under supervision of Dr. Yulia Gel.