Testing for Changes in Multivariate Dependent Observations with an Application to Temperature Changes

We develop procedures for testing for changes in the mean of multivariatem-dependent stationary processes. Several test statistics are considered and corresponding limit theorems are derived. These include functional and Darling-Erdos type limit theorems. The tests are shown to be consistent under alternatives of abrupt and gradual changes in the mean. Finite sample performance is examined by means of a simulation study, and the procedures are applied to the analysis of the average monthly temperatures in Prague.