Reconstructing the δ 18 O of atmospheric water vapour via the CAM epiphyte T illandsia usneoides : seasonal controls on δ 18 O in the field and large‐scale reconstruction of δ 18 O a

Using both oxygen isotope ratios of leaf water (δ 18 O L ) and cellulose (δ 18 O C ) of Tillandsia  usneoides in situ , this paper examined how short‐ and long‐term responses to environmental variation and model parameterization affected the reconstruction of the atmospheric water vapour (δ 18 O a ). During sample‐intensive field campaigns, predictions of δ 18 O L matched observations well using a non‐steady‐state model, but the model required data‐rich parameterization. Predictions from the more easily parameterized maximum enrichment model (δ 18 O L–M ) matched observed δ 18 O L and observed δ 18 O a when leaf water turnover was less than 3.5 d. Using the δ 18 O L–M model and weekly samples of δ 18 O L across two growing seasons in F lorida, USA , reconstructed δ 18 O a was −12.6 ± 0.3‰. This is compared with δ 18 O a of −12.4 ± 0.2‰ resolved from the growing‐season‐weighted δ 18 O C . Both of these values were similar to δ 18 O a in equilibrium with precipitation, −12.9‰. δ 18 O a was also reconstructed through a large‐scale transect with δ 18 O L and the growing‐season‐integrated δ 18 O C across the southeastern United States . There was considerable large‐scale variation, but there was regional, weather‐induced coherence in δ 18 O a when using δ 18 O L . The reconstruction of δ 18 O a with δ 18 O C generally supported the assumption of δ 18 O a being in equilibrium with precipitation δ 18 O (δ 18 O ppt ), but the pool of δ 18 O ppt with which δ 18 O a was in equilibrium – growing season versus annual δ 18 O ppt – changed with latitude.