Background/Question/Methods Understanding the processes that shape species’ distributions is an important goal of ecological research. Niche theory argues that species’ distributions may reflect their demographic responses to ecological gradients. However, disentangling the roles of abiotic versus biotic factors in shaping species’ distributions remains challenging in practice, thereby limiting our empirical understanding of when and how biotic interactions are expected to impact species’ realized distributions. This challenge would be further compounded if the strength of relevant biotic interactions varies by abiotic context. In order to explore how competitive interactions shape the performance and thus the distributions of plant species in a natural landscape, we grew eight annual plant species, either in the absence of competitors or within the resident plant community, at seven environmentally heterogeneous sites across an annual grassland in California. These study sites ranged from serpentine ‘hummocks’ with shallow, rocky soils to the non-hummock ‘matrix’ with deeper, finer-textured soils. We quantified for each species in each habitat type the effect of competitors on key demographic rates (germination rate, fecundity) and the resulting population growth rates. We then evaluated the extent to which these competitive effects can explain observed mismatches between the expected and realized distributions of our focal species. Results/Conclusions Preliminary analyses with a subset of species found that both the germination rates and fecundities of our focal species varied significantly by habitat type. Further, a significant interaction between habitat type and competitor treatment was found for both demographic rates, lending support to the importance of spatially-variable competitive interactions in this system. These patterns were especially striking when analyzed in terms of the overall per capita population growth rates. In the absence of competitors, species generally performed similarly well in either habitat type, if not better in the matrix. However, in the presence of resident competitors, species performed substantially worse in the matrix. Together, this translated into a larger effect of competitors in the matrix than on the hummocks for most focal species. Notably, several of these species have realized distributions that are largely confined to the hummocks, despite achieving a higher population growth rate in the matrix if competitors were absent. These results suggest a significant role of spatially-variable competitive interactions in shaping species’ distributions at the landscape-scale. Our work reinforces the need to take biotic interactions into consideration in the study of species’ distributions, especially at the landscape-scale.
