Fig 1.
Conceptual diagram summarising the three main steps of the analyses.
(1) identifying the climate variables that affect each species and over which time periods (i.e. climate windows), (2) quantifying species and population responses to climate, and (3) investigate inter and intraspecific variation in climate responses. Step 3a investigates the relative amounts of intra- and inter-specific variation to ask how consistent responses are and whether there is a species signal. While in step 3b, comparative analysis is used to test for any species or site traits (e.g. phylogenetic relatedness or habitat type) that explain differences among species or sites sensitivities and future projections. Steps 1 and 2 are carried out on each species individually, while in Step 3 all species are combined.
Fig 2.
Illustration of hypothetical intra-specific and inter-specific variation in sensitivities to climate.
Plot (a) shows an example where a species signal is present. Here, the correlation between the climate sensitivity of two populations of the same species is much higher than the correlation between two populations of different species. The percentage of among-species variation explained is 64%, which suggests that population variation is low compared to among-species variation. Plot (b) shows an example where a species signal is absent. Here, the correlation between the climate sensitivity of two populations of the same species is lower than the correlation between two populations of different species (i.e. population sensitivities within a single species vary just as much as much as population sensitivities among different species). As such, the percentage of among-species variation explained is a much smaller 3%. The sensitivity estimates for each species is shown by the red points, while the black boxplots show the distribution of population sensitivity estimates (intraspecific or among-site variation in climate sensitivity). The two red vertical lines show the minimum and maximum of the species’ sensitivity estimates (i.e. the range of the red points).
Fig 3.
The time periods (or climate windows) over which climatic variables affected body condition.
Windows can potentially start from the 15th of August back 365 days before. Red and black lines show whether the relationship is quadratic or linear, respectively. Specific species names can be seen in Table B in S2 Appendix.
Fig 4.
Interspecific and intraspecific variation in future projections (a) and sensitivity (b-g) of 29 bird species at 80 sites.
The overall species’ sensitivity (or future projections) are shown by the red points and lines (with standard error bars). The two red vertical lines show the minimum and maximum of the species’ sensitivity estimates (i.e. the range of the red points). For each species, the intraspecific (among-site) variation in climate sensitivity (or future projections) is described by the black boxplots. Sample sizes are shown in brackets along the left side of the figure, with the first number showing how many sites were present followed by the average number of individuals per site per year.
Fig 5.
Boxplot of the projected change in percent body condition by 2050 (total future projections) and the contribution of each climate variable for all species for 39 passerine species.
Total future projections is the sum of all climate projections.
Fig 6.
The effect of phylogenetic relatedness on the dissimilarity in future projections and sensitivity of avian body condition.
The top figure illustrates what the slopes in the lower figure (the y-axis) represent. In the top figure, we specifically show the relationship between dissimilarity and phylogenetic distance for future projections (grey line), where the slope of 0 reflects that each species had the same projections. A positive slope would indicate that more closely related species have more similar responses of body condition to climate. The bottom figure summarises the slopes obtained from the linear regressions (slope±SE) of phylogenetic relatedness on the dissimilarity in climate sensitivity to each of the six specific climate variables and to future projections (where the grey dot relates to the grey slope in the top figure).
Fig 7.
The relationships between site habitat characteristics and sensitivities (±SE) for each climate variable.
The shaded bars indicate wet habitats and the clear bars represent dry habitats. Note that the units for each climate variable are not comparable as the units differ.
Fig 8.
The distribution across species of relationships between distance (km) and dissimilarity in sites sensitivities and future projections.
The top figure illustrates how the slopes were estimated in the lower figure (the y-axis). In the top figure, we specifically show the relationship between dissimilarity and distance (km) for future projections, where each line represents a different species. A slope of 0 would mean that projections did not differ with distance, while a positive slope would indicate that closer sites were more similar. The red slopes indicate when a slope was positive and their 95% CI did not cross zero. The bottom figure summarises the slope estimates for each species for future projections and climate sensitivity for each of the six specific climate variables. There were 7 species that showed a significant increase for future projections, 3 species were sensitive to rain, and 2 were sensitive to all other climate variables.