Water usage of old-growth oak at elevated CO ₂ in the FACE (Free-Air CO ₂ Enrichment) of climate change

Abstract. Predicting how increased atmospheric CO2 levels will affect water usage by whole, mature trees remains a challenge. The present study investigates diurnal (i.e. daylight) water usage of oaks within an old-growth forest during an experimental treatment season (April–October, inclusive). Over the years 2017–2021, inclusive (years 1–5 of the experiment), we collected individual tree data from 18 oaks (Quercus robur L.) within a large-scale manipulative experiment at the Birmingham Institute of Forest Research (BIFoR) Free-Air CO2 Enrichment (FACE) temperate forest in central England, UK. Diurnal tree water usage per day (TWU, L d−1) across the leaf-on seasons was derived from these data. Equal tree numbers were monitored in each treatment: FACE infrastructure arrays (+150 µ mol mol−1) of elevated CO2 (eCO2), FACE infrastructure control ambient CO2 (aCO2) arrays, and control “ghost” (no-treatment, no-infrastructure) arrays. TWU was linearly proportional to tree stem radius, Rb (∼ 3.1 L d−1 mm−1; 274 mm ≤ Rb ≤ 465 mm). Rb was also a very good proxy for projected canopy area, Ac (m2), which was linearly proportional to Rb (∼ 617 m2 m−1). Applying the stem-to-canopy relation implied a mean July water usage of ∼ 5 L d−1 m−2 of projected oak canopy in the BIFoR FACE forest. We normalised TWU by individual tree Rb to derive TWUn (L d−1 mm−1). We report whole-season treatment effects, differing year on year, alongside July-only results. In the 2019 and 2021 seasons, after correction for repeated measures, there was a 13 %–16 %, reduction in eCO2 TWUn compared to aCO2 TWUn, with a marginal 4 % reduction in 2020, but these model results were not statistically significant. Control trees exhibited a significant 27 % increase in aCO2 TWUn compared to ghost TWUn in the whole season in 2019, with lesser, nonsignificant fixed effects in 2020 and 2021. Several factors may have contributed: the installation or operation of FACE infrastructure; array-specific differences in soil moisture, slope, or soil respiration; or the mix of subdominant tree species present. Our results showing normalised per-tree water savings under eCO2 align with sap flow results from other FACE experiments and greatly extend the duration of observations for oak, elucidating seasonal patterns and interannual differences. Our tree-centred viewpoint complements leaf-level and ground-based measurements to extend our understanding of plant water usage in an old-growth oak forest.