Global average temperatures were relatively stable until the recent upward climb. This trend was described by Mann et al. (1999) using a hockey-stick model, which consists of two line segments that meet at a changepoint. Left of the changepoint is characterized by a constant flat trend and right of the changepoint follows a positive or negative trend. As the long-term average temperature can be described by the hockey-stick model, variables that respond to temperature may also be modeled in the same manner. This yields two important parameters to consider, the changepoint indicating when the system began to respond, and the right line segment slope illustrating how fast the system is responding. We aim to develop a suite of hockey-stick models and apply them to different biological and physical systems that respond to temperature. Our systems of interest are 1) North American lilac first bloom dates, 2) North American peak breeding bird season, and 3) Great Lakes ice coverage duration. Using this model, we can estimate the year that the system started to respond to the change in temperature and the magnitude of this response.
Results/Conclusions
After applying the hockey-stick model to the systems of interest, North American lilac first bloom dates had a changepoint of 1971 (1967/1975) and a right linear regression slope of -0.34 (-0.27/-0.42), meaning a shift of 1/3 of a day earlier each year since 1971. In Great Lakes ice coverage duration the changepoint was 1981 (1974/1995), with a slope after the changepoint of -0.8 (-0.3/-1.3). Finally, North America bird breeding season began to respond in 1992 (1985/2000) with a shift of -1.6(-0.8/-2.4) days earlier each year. It is important to understand the slope after the changepoint because each system is part of the ecosystem in that region and thus impacts other populations. For example, various fish stocks in the Great Lakes region have greater recruitment when there is more ice coverage for the year. Additionally, changes in North American lilac first bloom dates affect the peak food supply for populations that consume them. Understanding the slope after the changepoint is also important for managers and economic investments, such as the impact that changes in peak bird breeding season has on the birding industry.