Water availability drives gas exchange and growth of trees in northeastern US, not elevated CO2 and reduced acid deposition

Dynamic global vegetation models (DGVM) exhibit high uncertainty about how climate change, elevated atmospheric CO2 (atm. CO2) concentration, and atmospheric pollutants will impact carbon sequestration in forested ecosystems. Although the individual roles of these environmental factors on tree growth are understood, analyses examining their simultaneous effects are lacking. We used tree-ring isotopic data and structural equation modeling to examine the concurrent and interacting effects of water availability, atm. CO2 concentration, and SO4 and nitrogen deposition on two broadleaf tree species in a temperate mesic forest in the northeastern US. Water availability was the strongest driver of gas exchange and tree growth. Wetter conditions since the 1980s have enhanced stomatal conductance, photosynthetic assimilation rates and, to a lesser extent, tree radial growth. Increased water availability seemingly overrides responses to reduced acid deposition, CO2 fertilization, and nitrogen deposition. Our results indicate that water availability as a driver of ecosystem productivity in mesic temperate forests is not adequately represented in DGVMs, while CO2 fertilization is likely overrepresented. This study emphasizes the importance to simultaneously consider interacting climatic and biogeochemical drivers when assessing forest responses to global environmental changes.


Supplementary Information
Article title: Water availability drives gas exchange and growth of trees in northeastern US, not elevated CO 2 and reduced acid deposition Authors: Mathieu Levesque, Laia Andreu-Hayles, Neil Pederson The following Supplementary Information is available for this article: Table S1. Tree characteristics and expressed population signals. Methods S1. Calculations of Δ 13 C and i WUE. Methods S2. δ 18 O in precipitation data. Figure S1. Air mass trajectories and δ 18 O of precipitation. Figure S2. Temperature and precipitation trends in the eastern US Figure S3. Raw and residual basal area increments. Figure S4. Correlations coefficients between Liriodendron tulipifera tree-ring time series and climate variables. Figure S5. Correlations coefficients between Quercus rubra tree-ring time series and climate variables. Methods S1

Calculations of Δ 13 C and i WUE
Carbon isotope discrimination (Δ 13 C) corresponds to the preferential use of lighter 12 C atoms over the 13 C heavier atoms during photosynthesis and it is calculated as the difference between the δ 13 C of the air (δ 13 C air ) and δ 13 C in tree ring (δ 13 C tree ). We used the linear interpolation technique from Leuenberger 2 to estimate values of δ 13 C air .
Δ 13 C is linearly related to the ratio of intercellular c i to atmospheric (c a ) CO 2 mole fractions 3 via: where a (4.4‰) is the fractionation during CO 2 diffusion through the stomata 4 , and b (27‰) is the fractionation during carboxylation in C 3 plants 5 . We used c a values based on ice core data for the period 1950-1957 and a yearly average of direct observations after 1957 (http://scrippsco2.ucsd.edu) 6 . i WUE was estimated from values of Δ 13 C and c a according to Farquhar & Richards 5 : where 1.6 is the ratio of diffusivities between water vapor and CO 2 in air.
Previous studies have reported that trends in Δ 13 C and iWUE could potentially be related to increased tree height over time as Δ 13 C can decline with increasing height 7 as a result of: adjustments in hydraulic conductivity 8 ; assimilation of δ 13 C-depleted air near the forest floor 9 ; and changes in irradiance and photosynthetic capacity with height in the canopy 10 . To avoid this potential bias, we sampled mature trees (>125 yrs old, Table S1). Those trees were at least 65 year old at the beginning of the investigation period in 1950 and according to site index curves tree height of Q. rubra and L. tulipifera has only increased by 2-6 m. Such increase in tree height had no or only minor effects on tree-ring Δ 13 C values 7 . Additional evidence is provided by studies showing that before the rise in atmospheric CO 2 concentration, trees generally do not show any age-related δ 13 C trends after an initial juvenile phase of circa 50 years 11 .

Methods S2 δ 18 O of precipitation
Variation in tree-ring δ 18 O is primarily due to evaporative enrichment at the leaf level, biochemical fractionation during oxygen incorporation, and isotopic signature of tree source water, which is mainly influenced by the δ 18 O of precipitation, and to a lesser extent from soil evaporative enrichment 12,13 . Therefore, when isolating the leaf level effect on oxygen isotope enrichment in tree rings a good understanding of δ 18 O of tree source water / precipitation is necessary. Unfortunately, such records do not exits at our study site, so we had to rely on δ 18 O of precipitation data from another site in northeastern US. We used the newly developed and longest (1968-2010) continuous record of precipitation isotope from the Hubbard Brook Experimental Forest (43º56′N, 71º45′W) 14 to help interpreting our tree-ring δ 18 O results. The Hubbard Brook Experimental Forest is located ca. 300 km northeast from Black Rock Forest (Fig. S1a). The analysis of the air mass back trajectories  indicates that at both sites the moisture sources show some similarities despite some monthly offsets in δ 18 O of precipitation due to differences in air temperature, latitude, and precipitation type (rain / snow) (Fig. S1b). Based on these exploratory analyses, we considered that the mean yearly δ 18 O of precipitation records from Hubbard Brook Experimental Forest reflect those at Black Rock Forest. For the period 1968-2010, a significant reduction in δ 18 O of precipitation (− 0.089‰ yr −1 ) was recorded in northeastern US (Fig. S1c) 14    an average ontogenetic growth curve for L. tulipifera and Q. rubra (i.e., the regional curve) was calculated by aligning the raw BAI measurements of each tree to the biological age of the rings, and individual BAI series were then divided by this average curve 18,19 .