Our study showed that increases in seasonal productivity drive earlier autumn senescence of temperate trees. Norby argues that this finding is contradicted by observations from free-air CO2 enrichment (FACE) experiments, where elevated CO2 has been found to delay senescence in some cases. We provide a detailed answer showing that the results from FACE studies are in agreement with our conclusions.
Using experimental and observational data, Zani et al. (1) showed that increases in spring and summer productivity correlate with earlier autumn senescence of temperate trees, and that this process may counteract the expected long-term delays in leaf senescence under autumn warming. In his Comment (2), Norby argues that this finding is contradicted by observations from FACE experiments where leaf senescence was not altered or was delayed in trees exposed to elevated CO2. However, although we agree that FACE experiments are critical to study the responses of plants to a CO2-enriched environment, we believe that results from these experiments do not contradict the findings of our study. Indeed, the results from FACE studies lend additional empirical support for the proposed sink-driven mechanism.
Our model does not predict that increased CO2 will generally lead to advanced senescence. The autumn phenology model developed in our study is based on the empirical evidence that earlier spring leaf-out, elevated spring and summer temperatures, and higher irradiation all independently counteract the expected delays in senescence under warmer autumns. Our predictions therefore do not depend on CO2 levels, as can be seen from the similar performance of the photosynthesis-driven model, which accounts for CO2, and the growing-season index model, which does not account for atmospheric CO2 levels [figure 3 of (1)]. The effect of rising atmospheric CO2 levels over recent decades on Central European autumn senescence dates thus appears to be negligible relative to the effects of rising temperatures. In fact, because of the interactive effects of growing-season productivity and autumn temperatures, our model predicts continuing delays in senescence over the coming three decades with strongly increasing CO2 levels under the RCP8.5 scenario.
The sink limitation paradigm predicts that elevated CO2 will cause advanced leaf senescence only if (i) high CO2 leads to increased spring/summer photosynthesis, (ii) these increases in photosynthesis (source strength) outweigh CO2-driven increases in sink strength, (iii) the respective individual is limited in its carbon sink capacity (e.g., through limited nutrient supply), and (iv) autumn warming does not counteract these trends. The FACE studies cited by Norby (2) do not disagree with these predictions, as we explain below.
Asshoff et al. (3), Godbold et al. (4), and Richardson et al. (5) did not measure photosynthesis, and their data therefore cannot be used to directly test the relationship between spring/summer productivity and leaf senescence. Asshoff et al., however, measured growth responses and found that only Fagus sylvatica showed a small response under elevated CO2, whereas “none of the other dominant species (Quercus petraea, Carpinus betulus) showed a growth response to CO2 in any of the 4 years or for all years together” [(3), p. 848]. This “nonresponsiveness” of growth to elevated CO2 and the observation that leaf senescence was only marginally affected by elevated CO2 (4- and 5-day delays, respectively, in Carpinus and Fagus; 5-day advance in Quercus) agree with the sink limitation hypothesis.
The FACE experiment of Godbold et al. [(4), p. 839] showed that “an increase in fine root growth is a common feature in trees under elevated CO2 and [may] be due to high C allocation to roots, but also as a mechanism to increase nutrient uptake to meet the demand of increased aboveground growth.” Root tip numbers in Betula pendula increased by 31% and 41% under elevated CO2 in 2 years, likely resulting in an increased nutrient supply and sink strength that may explain the species’ delayed autumn phenology under elevated CO2 in these 2 years. In agreement with this, Taylor et al. (6) emphasized that only in the absence of sink limitation can there be a positive effect of CO2 on autumn growth, which aligns well with our sink-driven framework of autumn senescence. They “hypothesized that, with no sink limitation, photosynthesis and canopy greenness would be maintained for longer in elevated CO2 […]” [(6), p. 271]. The absence of a sink limitation can likely be explained by the high nutrient availability at the FACE study site. In addition, Populus is associated with nitrogen-fixing bacteria (7), which may reduce nutrient limitation. The role of soil nutrient supply is also clear from a 3-year CO2-enrichment study on Populus trichocarpa (8), showing that elevated CO2 strongly advanced leaf senescence at low (natural) nutrient availability, but much less under high nutrient availability.
Norby (2) suggests that “with both Acer saccharum and A. rubrum saplings in the TACIT experiment [(9)] and the mature, deciduous Larix laricina trees in the SPRUCE experiment [(5)], senescence or abscission was delayed in warmer temperatures, in contrast to the lack of response to warming reported by Zani et al., and there was no effect of elevated CO2.” Instead, we reported that both autumn temperatures and spring/summer productivity had strong, interacting effects on autumn leaf senescence, which matched the observations in the aforementioned studies. Despite the counteracting effect of spring and summer productivity, autumn warming still delayed senescence dates [figure 1 of (1)]. The finding in the SPRUCE experiment (5) that year-round warming of up to 9°C leads to a delay in senescence also does not contradict our results. Although a 9°C warming probably is well above the photosynthetic optimum of boreal species, such high autumn warming likely outweighed sink limitation effects in this study.
In conclusion, we think that there is no disagreement between the results of FACE experiments and our broad-scale analysis. As Norby (2) argues, studies that found advanced autumn phenology in response to elevated CO2 were characterized by sink limitation, which is in full agreement with our proposed mechanism, and we certainly agree that elevated CO2 should not drive earlier leaf senescence if the balance between carbon source and sink strength is maintained. Given the feedback loops between carbon source processes (photosynthesis) and sink processes (the most important of which is nitrogen availability) (10) and the inherent difficulties in quantifying sink versus source strength (11), the effect of CO2 fertilization on autumn phenology remains difficult to predict. The continued combination of experimental and observational approaches will be necessary to generate robust predictions about future changes in autumn senescence, and FACE experiments provide important data for this. Obtaining a global picture of the sink limitation effect will require tests of the relative effects of source and sink activities on leaf senescence in a variety of species from different biogeographic and phylogenetic backgrounds and in different root and soil systems.