Single Changepoint Procedures
By: Robert Lund, Xueheng Shi
Potential Business Impact:
Finds when climate patterns change over time.
Single changepoint tests have become a staple check for homogeneity of a climate time series, suggesting how climate has changed should non-homogeneity be declared. This paper summarizes the most prominent single changepoint tests used in today's climate literature, relating them to one and other and unifying their presentations. Asymptotic quantiles for the individual tests are presented. Derivations of the quantiles are given, enabling the reader to tackle cases not considered within. Our work here studies both mean and trend shifts, covering the most common settings arising in climatology. SOI and global temperature series are analyzed within to illustrate the techniques.
Similar Papers
Monitoring Time Series for Relevant Changes
Methodology
Spots many changes in data over time.
Seasonal trend assessment of US extreme precipitation via changepoint segmentation
Applications
Finds rain changes differently each season.
The Asymptotic Distribution for a Single Joinpoint Changepoint Model
Methodology
Finds the exact moment when data changes.