Understorey productivity in temperate grassy woodland responds to soil water availability but not to elevated [CO2]
Rising atmospheric [CO2] and associated climate change are expected to modify primary productivity across a range of ecosystems globally. Increasing aridity is predicted to reduce grassland productivity, although rising [CO2] and associated increases in plant water use efficiency may partially offset the effect of drying on growth. Difficulties arise in predicting the direction and magnitude of future changes in ecosystem productivity, due to limited field experimentation investigating climate and CO2 interactions. We use repeat near‐surface digital photography to quantify the effects of water availability and experimentally manipulated elevated [CO2] (eCO2) on understorey live foliage cover and biomass over three growing seasons in a temperate grassy woodland in south‐eastern Australia. We hypothesised that (i) understorey herbaceous productivity is dependent upon soil water availability, and (ii) that eCO2 will increase productivity, with greatest stimulation occurring under conditions of low water availability. Soil volumetric water content (VWC) determined foliage cover and growth rates over the length of the growing season (August to March), with low VWC (<0.1 m3 m−3) reducing productivity. However, eCO2 did not increase herbaceous cover and biomass over the duration of the experiment, or mitigate the effects of low water availability on understorey growth rates and cover. Our findings suggest that projected increases in aridity in temperate woodlands are likely to lead to reduced understorey productivity, with little scope for eCO2 to offset these changes. ; The EucFACE experiment is funded by the Australian Government, through the Education Investment Fund, the Department of Industry and Science and the Australian Research Council, and Western Sydney University. Funding for the cameras and data analysis was provided by Australian Research Council Discovery Grant number 130102576. The stereo camera and application development was funded by the Western Sydney University Internal Research Scheme. VRD acknowledges funding from a Ramón y Cajal fellowship (RYC‐2012‐10970).