An index for climate change: a multivariate time series approach


Francisco Andrés Rincón, Car 18A No 101 42 Apt 501 Bogota, Colombia. E-mail:


The analysis of environmental data represents an opportunity to use statistical tools to provide a better understanding of changes over time, being necessary to combine nontraditional approaches looking for the best way to capture patterns or trends. Throughout this paper, multivariate time series methodologies are presented to build an index to measure climate change; the analysis was carried out over nine meteorological time series for the period 1978–2002. The index was built using a spectral envelope approach combined with the use of band-pass filters, allowing to identify changes over the period observed as well as the main meteorological variables. Copyright © 2012 John Wiley & Sons, Ltd.