Tree growth responses to temporal variation in rainfall differ across a continental-scale climatic gradient.

Abstract

Globally, many biomes are being impacted by significant shifts in total annual rainfall as well as increasing variability of rainfall within and among years. Such changes can have potentially large impacts on plant productivity and growth, but remain largely unknown, particularly for much of the Southern Hemisphere. We investigate how growth of the widespread conifer, Callitris columellaris varied with inter-annual variation in the amount, intensity and frequency of rainfall events over the last century and between semi-arid (800 mm mean annual rainfall) biomes in Australia. We used linear and polynomial regression models to investigate the strength and shape of the relationships between growth (ring width) and rainfall. At semi-arid sites, growth was strongly and linearly related to rainfall amount, regardless of differences in the seasonality and intensity of rainfall. The linear shape of the relationship indicates that predicted future declines in mean rainfall will have proportional negative impacts on long-term tree growth in semi-arid biomes. In contrast, growth in the tropics showed a weak and asymmetrical ('concave-down') response to rainfall amount, where growth was less responsive to changes in rainfall amount at the higher end of the rainfall range (>1250 mm annual rainfall) than at the lower end (<1000 mm annual rainfall). The asymmetric relationship indicates that long-term growth rates of Callitris in the tropics are more sensitive to increased inter-annual variability of rainfall than to changes in the mean amount of rainfall. Our findings are consistent with observations that the responses of vegetation to changes in the mean or variability of rainfall differ between mesic and semi-arid biomes. These results highlight how contrasting growth responses of a widespread species across a hydroclimatic gradient can inform understanding of potential sensitivity of different biomes to climatic variability and change.