5.3 Characterizing midday ΨL regulation
Periodic midday ΨL measurements (10:00–16:00 local time) were compiled from a dataset of over 1600 observations collected throughout the growing seasons of 2011–2017. On each measurement day, one to five samples were collected from one to three trees per species from the upper third of the canopy. Leaves were bagged for ~15 minutes prior to excision to allow Ψ of the leaf cells and stem xylem to reach equilibrium (Leach et al. , 1982; Roman et al. , 2015); this approach was conducted in every site except for IN 35yo where canopies were inaccessible from the ground or by cherry picker. After excision, ΨL was measured using a pressure chamber (PMS Instruments, Corvallis, OR, USA) (Turner, 1988) immediately in the field, or after leaves were transferred to the lab in humidified bags stored in a cooler. All together, we made 704, 178, and 757 ΨL observations of L. tulipifera , A. saccharum , and Q. alba , respectively. The number of ΨL observations and sampling days varied across regions, but ΨL was measured on 4–51 different days at each stand, including sampling at the beginning (June) and end (September) of the growing season to permit observation throughout dynamic seasonal changes of moisture conditions.
While regulation of plant water status is frequently characterized as the sensitivity of ΨL to declining ΨS(McDowell et al ., 2008; Klein, 2014; Martínez-Vilalta et al ., 2014; Matheny et al., 2015; Meinzer et al.,2017), this metric of isohydricity can change temporally as drought evolves (Hochberg et al., 2018; Wu et al.,2021), and is often inconsistent for the same species from one stand to the next (Martínez-Vilalta & Garcia-Forner, 2017). These inconsistencies likely reflect the fact that the degree of isohydricity, when defined as\(\ \partial\Psi_{L}/\partial\Psi_{S}\), is complicated by environmental interactions (Hochberg et al ., 2018), including variability in D which can also affect ΨL (Domec & Johnson, 2012; Novick et al ., 2019), or when the magnitude of soil water deficit during the sampling period is insufficient to capture stress responses (Martínez-Vilatla & Garcia-Forner, 2017). Another proposed metric – the “hydroscape” concept (Meinzer et al., 2016; Li et al., 2019) based on the integrated area between the observed \(\Psi_{L}-\Psi_{S}\) curve – can overcome some of the conceptual difficulties associated with\(\partial\Psi_{L}/\partial\Psi_{S}\). However, the hydroscape is still fundmentally informed by the relationship between \(\Psi_{L}\ \)and\(\Psi_{S}\). Thus, the hydroscape does not directly account for variability in ΨL driven by D and can be hard to quantify in mesic sites where \(\Psi_{S}\) may be relatively stationary even while temperature-driven variation in D may be large.
Negative excursions in \(\Psi_{L}\ \)driven by D may be especially important in eastern US forests, where limitations to stomatal conductance from D have been shown to dominate over soil water limitations, at both the stand (Novick et al ., 2016) and tree-scale (Yi et al., 2019; Denham et al ., 2021). While substantial soil water deficits occurred in some of our sites (e.g., MO, NC_E, IN), the more mesic NC_W stands rarely experience soil water limitations, and soil water deficits were not observed during the study period (Fig. S1); however, ΨL reductions during periods of elevated D occurred routinely (Fig. S2). For these reasons, we quantified isohydricity as the variability in seasonal midday ΨL to capture ΨL sensitivity to both declining soil moisture and increasing D (See SI.1 for further discussion). To minimize error associated with uncharacteristic behavior during spring leaf out and fall senescence, ΨL data used for this analysis were constrained to a period of relatively stationary leaf area index (days of year 150–270).