A change point in a time series can be viewed as a time point at which the parameters of a statistical distribution or a statistical model change. Change points can be observed in a wide variety of fields (e.g. economy, social sciences, climate and hydrology). Most change point approaches were designed to detect a specific type of shift: in the mean, in the variance, in both the mean and the variance or in the parameters of a regression model, but rarely to discriminate between several types of changes. Furthermore, most change point methods make the hypothesis that the residuals are independent, but the presence of autocorrelation is a common feature of climate time series (especially at short time scales). If not taken into account in the analysis, the presence of positive autocorrelation can lead to the detection of false shifts. In this talk, methodologies allowing to identify the timing of a change point and to discriminate between no changes, a gradual change (trend) or an abrupt change (shift) are presented and compared. The autocorrelation structure is identified in each models (not restricted only to a first order autoregressive model) and integrated in the analysis. Applications of this approach to detect change points in the carbon cycle are presented.
Keywords: Change point; Autocorrelation; Carbon cycle
Biography: Dr. Beaulieu is a postdoctoral fellow in Atmospheric and Oceanic Sciences at Princeton University where she applies change point detection techniques to study structural changes in the carbon cycle. She received her Ph.D. in water sciences from the University of Quebec in 2009 that dealt with the homogenization of precipitation series. She was awarded the best Ph.D. thesis in science and engineering in the Quebec province for this year. She received a bachelor degree in statistics from Universite Laval in Quebec City.