7.2 The perplexing case of Q. alba
Our finding that Q. alba had the most vulnerable xylem was unexpected. Quercus species are often considered more drought tolerant than many co-dominants, attributed to their morphological and physiological adaptations that allow them to withstand soil moisture deficits (Abrams, 2003). Our results complicate this perspective. We found that Q. alba had particularly high P50 (consistent with previous work: Maherali et al ., 2006; Kannenberg et al.,2019) but were also more anisohydric. We used the variation in ΨL to quantify the degree of isohydricity in order to incorporate stomatal responses to both declining soil water and increasing D , noting that the latter is the predominant factor limiting conductance for these sites and species (Novick et al., 2016; Yi et al., 2019; Denham et al., 2021). However, prior work using other approaches for quantifying isohydricity in these study sites and elsewhere also concludes thatQuercus species are more anisohydric than many of their co-dominant counterparts (Abrams, 1990; Cavender-Barres & Bazzaz, 2000; Ewers et al. , 2007; Meinzer et al., 2013; Roman et al., 2015; Kannenberg et al ., 2019). Here, our results further revealed that Q. alba trees had a high degree of estimated native embolism and negative Ψsafety suggesting they are remarkably vulnerable to drought.
While rooting depth is an important component of a plant’s water use strategy, species-specific differences in rooting depth cannot explain our results. Quercus species tend to be more deeply rooted than cohabiting tree species in eastern US forests (Abrams, 1990), an expectation recently confirmed by our study team in IN 85yo (Lanninget al ., 2020). However, periodic observations of pre-dawn ΨL, a commonly used proxy for integrated ΨS across the rooting zone (Richter, 1997), were less conclusive about the extent to which functional rooting depth varied across species (Table S3). In any event, if Q. alba have deeper roots, then access to more stable moisture pools should keep midday ΨL elevated relative to other species; instead, we findQ. alba typically had more negative midday ΨL(Fig. 6) despite the fact they may have access to deeper pools of water.
While the hydraulic metrics quantified here are widely used to characterize drought-susceptibility, drought impacts on whole-plant physiological function are more complex than stomatal regulation of xylem water tension. For example, drought-susceptibility is determined not only by the risk of xylem dysfunction, but also the plant’s ability to cope with and recover from hydraulic damage (Meinzer & McCulloh, 2013). We therefore consider howQ. alba can exhibit a seemingly risky hydraulic strategy while adhering to drought-tolerance. First, we note that our methodology permits an evaluation of the vulnerability of the entire sapwood depth. However, it is not clear that Q. alba rely on the entire depth of sapwood to actively conduct water (Cochard & Tyree, 1990). In a related study from IN 85yo, Yi et al . (2017) found that the inner sapwood of Q. alba conducted a more significant fraction of water during drought, with water transport largely restricted to outer rings during well-watered periods. Additionally, internal water storage can also play an important role in determining the relationship between leaf gas exchange and stem xylem traits. Ring-porous species are known to use smaller amounts of stored water than diffuse-porous species because of their low number of active rings (Köcher et al ., 2013). Q. alba has much higher wood density than either L. tulipifera orA. saccharum, and species with greater wood density tend to have low capacitance (e.g., Meinzer et al., 2008). Unlike L. tulipifera and A. saccharum that bear large sapwood volume and have low wood density, the small water storage capacity of Q. alba cannot provide enough water to limit the rapid drop in water potential due to stomatal water loss, which could also explain its anisohydric behavior (Matheny et al ., 2015).
Recovery from hydraulic impairment may also explain how Q. albatolerates drought while possessing vulnerable xylem. Refilling of embolized conduits is a possible strategy for ring-porous species to maintain hydraulic function (Brodersen et al ., 2010; Ogasaet al ., 2013; Trifilò et al ., 2019; Zeppel et al ., 2019), although whether xylem refilling routinely occurs in long vessel species is debated (Lamarque et al., 2018). Moreover, Q. alba bears only a few hydraulically active sapwood rings (<10), with the newest rings being the most efficient at moving water (Phillips et al ., 1996). Therefore, Q. albacould potentially repair a 50% loss of conductivity in fewer than five years just by the production of new annual rings. While metabolically costly, growth and assimilation for Quercus species is often less sensitive to water stress than their more isohydric co-dominants (Elliott et al. , 2015; Roman et al., 2015; Au et al., 2020). Thus, the hydraulic strategy of Q. alba may be to maximize carbon assimilation at the risk of hydraulic impairment such that hydraulic function can be readily recovered through new growth.Quercus species also have an abundance of embolism-resistant vasicentric tracheids that can account for as much as 15% of hydraulic conductivity in stems (Percolla et al., 2021). These tracheid networks likely play an important role in sustaining water transport and growth when vulnerable vessels have embolized (Fontes & Cavender-Bares, 2020).
Xylem vulnerability assessments must be conducted with care and a clear recognition of potential sources of methodological bias (Cochardet al., 2013, Johnson et al., 2018; Lobo et al.,2018). The air-injection technique used in this study remains the most popular tool for generating vulnerability curves, though it is sensitive to open vessel artifacts which may produce excessive variability in the derived estimates of P50 (Martin-StPaul et al., 2014). As discussed extensively in our methods, we deployed a thorough set of quality control measures to minimize this source of error in our data. These measures included: [1] limiting samples to young, distal branches which have shorter vessels, [2] direct testing for the presence of open vessels on every sample using the air-infiltration technique, and [3] careful post-facto screening of curves to remove those that were conspicuously ‘r-shaped’. If Q. alba samples were characterized by a greater number of open vessels, then we would have expected a high percentage of Q. alba curves to be ‘r-shaped.’ Instead, variability in P50 was similar across species (coefficient of variation = 0.26, 0.19, & 0.19 for Q. alba , A. saccharumand L. tulipifera , respectively) suggesting that focusing on young branches and directly testing for open vessels were effective at limiting open vessel bias.
We recognize that this study focused only on three tree species and that others have found stomatal regulation and embolism vulnerability to be generally coordinated across species in other temperate regions (e.g., Vogt, 2001). Nonetheless, our results are consistent with other studies employing different strategies to generate ‘s-shaped’ vulnerability curves for Quercus species. Using the cavitron technique, Loboet al., (2018) found that species-specific curves of six EuropeanQuercus species were highly consistent and sigmoidal when branches were screened for open vessels. Johnson et al., (2018) found good agreement between the air-injection and centrifuge methods for Q. fusiform branches when checked for open vessels, as they were in our study. Moreover, Kannenberg et al., (2019) used the air-injection technique to generate xylem vulnerability curves for the entire stem of tree saplings, which should be especially insensitive to open vessel artifacts; that study also concluded that the P50 ofQ. alba was higher than L. tulipifera and A. saccharum . Finally, Skelton et al. (2021) used cutting edge techniques to visually monitor embolism formation. While they concluded that western North American Quercus species that dominate desert/chaparral environments have substantially negative P50s, they also reported that Quercus species growing in more temperate western forests have less negative P50 which were similar to those observed for the Q. alba trees growing in our temperate and mesic study sites (e.g., P50 ≥ −3 MPa).
Altogether, it appears that Q. alba sustain high rates of gas exchange at the cost of operating with damaging water potential gradients and low Ψsafety. Moreover, much of the variability in stomatal conductance and water potential for eastern US trees, and especially Quercus species, are determined by the dynamics of D (Yi et al., 2019, Novick et al.,2019; Denham et al., 2021). Thus, these species may be particularly vulnerable to hydraulic dysfunction linked to future droughts that will be characterized by increasingly high D(Ficklin & Novick, 2017). In that regard, strategies to sustainQuercus dominated forest may need to recognize that they may in fact be quite sensitive to drought stress.