Bootstrap inference for environmental trends

In this thesis, statistical methods were developed to analyse time trends when a considerable part of the data series is missing. Missing observations complicate the analysis because many existing methods are not directly applicable in this situation. When studying climate time series, however, missing data are a frequently encountered problem due to, for example, maintenance of equipment or adverse weather conditions that prevent measurements from being taken. In addition, this research investigated how to model (economic) relationships that are changing over time. In both approaches, the focus lies on finding an accurate measure of uncertainty around the estimates produced with the help of the model. The methods developed in this thesis are used to study atmospheric ethane which is an indirect greenhouse gas contributing to global warming. The dissertation shows that atmospheric ethane has experienced a recent upward trend when the observations were obtained in the Northern Hemisphere. Time series coming from measurement stations located in the Southern Hemisphere do not show the same pattern.