Individual vital rates respond differently to local‐scale environmental variation and neighbour removal

Abstract Understanding how plant fitness varies along natural gradients is critical for predicting responses to environmental change. However, individual vital rates are often used as fitness proxies without knowing how other vital rates vary along the same gradients. We investigated how canopy cover, plant–plant interactions, water availability and soil properties influenced the emergence, survival, seed production and population growth rates of eight annual plant species in semi‐arid Western Australia. We sowed seeds into sun‐exposed and shaded blocks across a reserve, removed all neighbouring plants from half of the interaction neighbourhoods, and used rainout shelters to reduce and increase precipitation relative to ambient plots. Canopy cover had strong negative effects on emergence, but few direct impacts on other vital rates and population growth rates. Direct competitive effects on survival and seed production were rare, although evident for population growth rate for 3/8 species. Competition was stronger in open than shaded plots for half of the species. Canopy cover also interacted with the watering treatment to influence survival of half of the species, but watering alone had few direct impacts on species' vital and population growth rates. We found only positive significant correlations between pairs of rates, and survival and seed production were far more frequently correlated with population growth rate than emergence. Synthesis . Our study illustrates that vital rates can respond to the same local‐scale environmental variation in different ways that are likely not driven by life history trade‐offs. We caution against using emergence as a proxy for population growth rate and emphasise that no single vital rate was a reliable fitness proxy overall. Interactions among abiotic and biotic factors were important drivers of vital and population growth rates for some species, highlighting the need to account for plant–plant interactions when predicting population responses to environmental change.