Competition contributes to quantitative mismatches between plant fitness and occurrence along environmental gradients

Species distributions depend on the complex interplay between the abiotic environment and biotic interactions. Empirical work in plant communities has increasingly revealed how competition mediates species' demographic responses to environmental variation, but understanding when and how the demographic consequences of competition manifest in observed occurrence patterns remains an important challenge. Here, we explore how competition contributes to quantitative mismatches between plant fitness and occurrence along spatial environmental gradients in an edaphically heterogeneous California annual grassland landscape. We experimentally quantified how the germination rate, fecundity and lifetime fitness of eight native annual plant species respond to spatial variation in the soil environment, either in the presence or in the absence of naturally occurring neighbours. We also surveyed the species' occurrence across the study landscape. Combining these fitness and occurrence data, we tested whether observed occurrence patterns are congruent with fitness responses to different soil factor gradients. We found that competition altered fitness responses to the primary soil gradient (soil texture and fertility) for some species. In the absence of competitors, fitness responses to this environmental gradient were often poorly aligned with estimated variation in the probability of occurrence. At least one species (Lasthenia californica) was more likely to occur in sandy, infertile soils, where its fitness was expected to be at its lowest. In contrast, we found that competition had relatively little effect on fitness responses and occurrence along a secondary soil gradient (soil nutrient composition, especially Ca:Mg). Synthesis. We demonstrate that competition can contribute to mismatches between how plant fitness and occurrence vary along spatial environmental gradients. Importantly, these quantitative mismatches depend on the species and environmental gradient in question. Our results caution against assuming that variation in occurrence implies variation in fitness, or vice versa, especially when the strength of competition co‐varies with the abiotic environment.