Increased drought frequency and severity are a pervasive global threat, yet the capacity of mesic temperate forests to maintain resilience in response to drought remains poorly understood. We deployed a throughfall removal experiment to simulate a once in a century drought in New Hampshire, USA, which coupled with the region-wide 2016 drought, intensified moisture stress beyond that experienced in the lifetimes of our study trees. To assess the sensitivity and threshold dynamics of two dominant northeastern tree genera (Quercus and Pinus), we monitored sap flux density (Js), leaf water potential and gas exchange, growth and intrinsic water-use efficiency (iWUE) for one pretreatment year (2015) and two treatment years (2016-17). Results showed that Js in pine (Pinus strobus L.) declined abruptly at a soil moisture threshold of 0.15 m3 m-3, whereas oak's (Quercus rubra L. and Quercus velutina Lam.) threshold was 0.11 m3 m-3-a finding consistent with pine's more isohydric strategy. Nevertheless, once oaks' moisture threshold was surpassed, Js declined abruptly, suggesting that while oaks are well adapted to moderate drought, they are highly susceptible to extreme drought. The radial growth reduction in response to the 2016 drought was more than twice as great for pine as for oaks (50 vs 18%, respectively). Despite relatively high precipitation in 2017, the oaks' growth continued to decline (low recovery), whereas pine showed neutral (treatment) or improved (control) growth. The iWUE increased in 2016 for both treatment and control pines, but only in treatment oaks. Notably, pines exhibited a significant linear relationship between iWUE and precipitation across years, whereas the oaks only showed a response during the driest conditions, further underscoring the different sensitivity thresholds for these species. Our results provide new insights into how interactions between temperate forest tree species' contrasting physiologies and soil moisture thresholds influence their responses and resilience to extreme drought.
