Modelling range shifts for Swiss breeding birds

In recent years, evidence has accumulated about the impacts of climate change on different aspects of bird life. In the spatial dimension, climate change has been shown to induce pole ward and upward shifts of species’ distributional ranges. Switzerland is a small country situated in western Europe and is characterized by an important elevational gradient; main distributional changes are therefore expected in the vertical dimension. This aspect has been addressed by the first phase of the project ClimBird 1 recently developed at the Swiss Ornithological Institute. In particular, this phase aimed at assessing ongoing changes in the elevational distribution of common breeding birds in Switzerland (Maggini et al . unpubl. data ). The analysis was performed on data gathered by the Swiss national common breeding bird survey (MHB ; Schmid et al ., 2004), a monitoring program me started in 1999 and conducted on an annual basis. The survey encompasses 267 1 - km 2 squares that are distributed across Switzerland as a grid and hence are representative of different biogeographical zones, elevation bands and habitats. Each square is surveyed three times – twice for squares above the treeline – per breeding season (15 April to 15 July) by qualified volunteers who map all birds seen or heard along a square - specific route. At the end of the season, territories are identified from the mapped points. Shifts were assessed between a first (1999 – 2002) and a second period of monitoring (2004 – 07). For each period, the observed number of territories per km 2 was modelled as a function of elevation using generalize d additive models (GAMs). Significant changes in the shape of the regression curve were evaluated at five reference points. The comparison showed an upward shift for at least one - third of the 95 species considered. This was rarely a complete shift, but mainly a partial shift with either an expansion of the leading edge or a retraction of the trailing edge. Due to the short period covered by the analysis, the different patterns are assumed to represent successive steps in the process of an upward shift. However, similar patterns may result from factors other than climate such as the encountering of barriers, a change in habitat suitability, competition with other species or different warming rates along the gradien