A large array of species distribution model (SDM) approaches has been developed for explaining and predicting the occurrences of individual species or species assemblages. Given the wealth of existing models, it is unclear which models perform best for interpolation or extrapolation of existing data sets, particularly when one is concerned with species assemblages. We compared the predictive performance of 33 variants of 15 widely applied and recently emerged SDMs in the context of multispecies data, including both joint SDMs that model multiple species together, and stacked SDMs that model each species individually combining the predictions afterward. We offer a comprehensive evaluation of these SDM approaches by examining their performance in predicting withheld empirical validation data of different sizes representing five different taxonomic groups, and for prediction tasks related to both interpolation and extrapolation. We measure predictive performance by 12 measures of accuracy, discrimination power, calibration, and precision of predictions, for the biological levels of species occurrence, species richness, and community composition. Our results show large variation among the models in their predictive performance, especially for communities comprising many species that are rare. The results do not reveal any major trade‐offs among measures of model performance; the same models performed generally well in terms of accuracy, discrimination, and calibration, and for the biological levels of individual species, species richness, and community composition. In contrast, the models that gave the most precise predictions were not well calibrated, suggesting that poorly performing models can make overconfident predictions. However, none of the models performed well for all prediction tasks. As a general strategy, we therefore propose that researchers fit a small set of models showing complementary performance, and then apply a cross‐validation procedure involving separate data to establish which of these models performs best for the goal of the study.

Using data from the field and the literature on 67 species of birds, we analyzed intraclutch variation in egg size, especially the deviation of the last egg from the clutch mean (D). Values of D are closer to zero in precocial than in altricial species; D is negatively correlated with body size in interspecific comparisons, i.e. large birds, including precocial species, lay small final eggs; and D is higher in open-nesting passerines (on average D = +3.56%, 17 species) than in hole-nesting species (on average D = -0.05%, 13 species). Within populations of birds, a negative relationship exists between D and clutch size, particularly in species that have a generally low value of D. The results support the view that intraclutch variation in egg size has an ultimate, adaptive value. We suggest that birds adopting the "brood-reduction strategy" have a small final egg, particularly those birds with large clutches, whereas birds adopting the "brood-survival strategy" have a relatively large final egg, particularly those birds with large clutches.

db1022.html, respectively. Yearly d 13 C values for individual stations and global annual averages are presented in tables S3 and S4.

Large-scale multi-species data on population changes of alpine or arctic species are largely lacking. At the same time, climate change has been argued to cause poleward and uphill range shifts and the concomitant predicted loss of habitat may have drastic eff ects on alpine and arctic species. Here we present a multi-national bird indicator for the Fennoscandian mountain range in northern Europe (Finland, Sweden and Norway), based on 14 common species of montane tundra and subalpine birch forest. Th e data were collected at 262 alpine survey plots, mainly as a part of geographically representative national breeding bird monitoring schemes. Th e area sampled covers around 1/4 million km 2 , spanning 10 degrees of latitude and 1600 km in a northeast -southwest direction. During 2002 -2012, nine of the 14 bird species declined signifi cantly in numbers, in parallel to higher summer temperatures and precipitation during this period compared to the preceding 40 yr. Th e population trends were largely parallel in the three countries and similar among montane tundra and subalpine birch forest species. Long-distance migrants declined less on average than residents and short-distance migrants. Some potential causes of the current decline of alpine birds are discussed, but since montane bird population sizes may show strong natural annual variation due to several factors, longer time series are needed to verify the observed population trends. Th e present Fennoscandian monitoring systems, which from 2010 onwards include more than 400 montane survey plots, have the capacity to deliver a robust bird indicator in the climate-sensitive mountainous regions of northernmost Europe for conservation purposes.