Global effects of soil and climate on leaf photosynthetic traits and rates

Abstract Aim The influence of soil properties on photosynthetic traits in higher plants is poorly quantified in comparison with that of climate. We address this situation by quantifying the unique and joint contributions to global leaf‐trait variation from soils and climate. Location Terrestrial ecosystems world‐wide. Methods Using a trait dataset comprising 1509 species from 288 sites, with climate and soil data derived from global datasets, we quantified the effects of 20 soil and 26 climate variables on light‐saturated photosynthetic rate ( A area ), stomatal conductance ( g s ), leaf nitrogen and phosphorus ( N area and P area ) and specific leaf area ( SLA ) using mixed regression models and multivariate analyses. Results Soil variables were stronger predictors of leaf traits than climatic variables, except for SLA . On average, N area , P area and A area increased and SLA decreased with increasing soil pH and with increasing site aridity. g s declined and P area increased with soil available P ( P avail ). N area was unrelated to total soil N . Joint effects of soil and climate dominated over their unique effects on N area and P area , while unique effects of soils dominated for A area and g s . Path analysis indicated that variation in A area reflected the combined independent influences of N area and g s , the former promoted by high pH and aridity and the latter by low P avail . Main conclusions Three environmental variables were key for explaining variation in leaf traits: soil pH and P avail , and the climatic moisture index (the ratio of precipitation to potential evapotranspiration). Although the reliability of global soil datasets lags behind that of climate datasets, our results nonetheless provide compelling evidence that both can be jointly used in broad‐scale analyses, and that effects uniquely attributable to soil properties are important determinants of leaf photosynthetic traits and rates. A significant future challenge is to better disentangle the covarying physiological, ecological and evolutionary mechanisms that underpin trait–environment relationships.