Statistical Modelling 12 (2012), 195–209

A short note on quantifying and visualizing yearly variation in online monitored temperature data

Göran Kauermann
Department of Statistics,
Ludwig-Maximilians-University Munich
Ludwigstrasse 33
D–80539 Munich
Germany
eMail: goeran.kauermann@stat.uni-muenchen.de, gkauermann@wiwi-uni-bielefeld.de

Thomas Mestekemper
University Bielefeld,
Center for Statistics
Bielefeld
Germany

Abstract:

The paper demonstrates how seasonal variation in sequentially arriving temperature data can be visualized by the specification of landmarks and subsequent time warping. We exemplify the idea with water temperature data from the river Wupper in northwestern Germany and with air temperature data from Berlin, Germany. Landmarks are thereby based on temperature thresholds. The method allows to assess whether the seasonal variation is running ahead or behind the average.

Keywords:

warping; landmark specification; monotonic smoothing; temperature curves

Downloads:


back