Aim. This study investigates whether, and how the composition of riparian bird communities has been affected by climate warming and habitat change. Although these two forces act separately, their respective contributions are rarely examined. Moreover, while the response of a given community may be a function of latitude and altitude, most studies have focused on these gradients separately. Riparian ecosystems are an opportunity to investigate community change along latitudinal and elevational gradients.

Location. France, three major rivers (the Doubs, the Allier, the Loire)

Taxon. Birds.

Methods. Drawing upon bird community monitoring data over a period of 31 years (1982–2013, n = 1286 point counts), we assessed community adjustment to temperature increase with the Community Temperature Index (CTI), and the homogenisation pattern following habitat conversion with the Community Specialisation Index (CSI). We evaluated the spatial climatic debt accumulated by communities, and the interaction between CTI and CSI trends. Furthermore, we tested differences in trends for lowland and highland elevations.

Results. Over the study period the temperature increased by 1.17°C, and the CTI by 0.12°C, which corresponds to a spatial climatic debt of 4.9 km yr-1. Lowland, but not highland communities adjusted to warming, but there was a decline in species abundance. CSI trends revealed biotic homogenisation in both lowland and highland communities. This finding was uncorrelated with the CTI increase, and is thought to be due to pressure from land use change on community composition.

Main conclusions. Riparian breeding bird communities have been affected by a temperature increase and, potentially, habitat change. Highland communities are most vulnerable to climate warming. Both climate warming and habitat change appear to have rapidly affected the composition of local communities, with expectable concerns on their diversity and specificity in the long term.

Breeding birds were monitored in spring during a longitudinal long-term scientific study that recorded point counts of any species heard or seen (Blondel et al. 1981). Two sessions were run (20 minutes in April and 20 minutes between mid-May and mid-June), and the highest abundance for each species between the two sessions was retained. Point counts were regularly spaced along the riverbank (1–5 km depending on the river) to limit the risk of double counting birds moving across a large area, and were strictly identical for each session (Table S1).