Simple additive effects are rare: a quantitative review of plant biomass and soil process responses to combined manipulations of CO2and temperature

In recent years, increased awareness of the potential interactions between rising atmospheric CO2 concentrations ([CO2]) and temperature has illustrated the importance of multifactorial ecosystem manipulation experiments for validating Earth System models. To address the urgent need for increased understanding of responses in multifactorial experiments, this article synthesizes how ecosystem productivity and soil processes respond to combined warming and [CO2] manipulation, and compares it with those obtained in single factor [CO2] and temperature manipulation experiments. Across all combined elevated [CO2] and warming experiments, biomass production and soil respiration were typically enhanced. Responses to the combined treatment were more similar to those in the [CO2]-only treatment than to those in the warming-only treatment. In contrast to warming-only experiments, both the combined and the [CO2]-only treatments elicited larger stimulation of fine root biomass than of aboveground biomass, consistently stimulated soil respiration, and decreased foliar nitrogen (N) concentration. Nonetheless, mineral N availability declined less in the combined treatment than in the [CO2]-only treatment, possibly due to the warming-induced acceleration of decomposition, implying that progressive nitrogen limitation (PNL) may not occur as commonly as anticipated from single factor [CO2] treatment studies. Responses of total plant biomass, especially of aboveground biomass, revealed antagonistic interactions between elevated [CO2] and warming, i.e. the response to the combined treatment was usually less-than-additive. This implies that productivity projections might be overestimated when models are parameterized based on single factor responses. Our results highlight the need for more (and especially more long-term) multifactor manipulation experiments. Because single factor CO2 responses often dominated over warming responses in the combined treatments, our results also suggest that projected responses to future global warming in Earth System models should not be parameterized using single factor warming experiments.