1 Introduction
Global climate change has led to and is continuously resulting in increases not only in drought intensity but also in the frequency and duration of drought events globally (IPCC, 2013). Two main hypotheses have been proposed and are currently debated to explain the mechanisms for the widespread forest dieback caused by increased drought events (Hartmann et al. , 2018; Spinoni et al. , 2018): trees would die (1) due to hydraulic failure or (2) as a result of carbon starvation (McDowell & Allen, 2015; Gessler et al. , 2018). The hydraulic failure hypothesis proposes that the tree mortality results from the embolism of the xylem vessels under high evaporative demand and restricted soil water availability (Nardini et al. , 2013) whereas the carbon starvation hypothesis suggests that tree mortality would be caused by a carbon supply limitation due to stomatal closure and thus reduced photosynthesis that cannot cover the carbon and energy demand for maintenance processes (McDowell et al. , 2008; Rowland et al. , 2015). Concerning carbon starvation, several studies have shown that the contents of starch, a compound that serves as carbon storage and is build up when assimilation exceeds plant’s demand for carbon, were strongly reduced by severe soil water deficit (McDowell, 2011), implying carbon limitation (Reinhardt et al. , 2015). Previous studies also suggested that hydraulic perturbation could prevent phloem transport (Sevanto, 2014) and consequently constrains carbon accessibility (Sala et al. , 2012; Hartmann & Trumbore, 2016), even though the carbon availability in the tree crown area might not be restricted. Thus, hydraulic failure and carbon starvation are often seen to be associated with each other in the cascade of drought events leading to tree mortality (Adams et al. , 2017).
Droughts, however, are very tougher to be defined, and thus a universally accepted drought definition considering both water deficit intensity and duration is still lack (Buitink et al ., 2021). Previous drought-related forest and agricultural studies, especially manipulation experiments, mainly focused on the effects of drought intensity (e.g. various levels of watering such as very limited, limited and optimum watering) on trees or plants (Allen et al. , 2010; Yuet al. , 2021), while less is known about the effects of drought duration on trees’ physiology, growth and mortality. A long-lasting drought event, may cause irreversible changes in plant physiology which are different from those found in relatively short-term severe droughts. For example, recurrent or short-term lasting drought events may allow trees to recover and even acclimate to water restriction (Gessleret al. , 2020), and thus, permit them to survive in the long-term. However, longer term lasting drought events may strongly affect the recovery ability of trees. Currently, unexpected whole-season drought occurs more frequently in many regions around the world. For instance, since the beginning of the 21st century Europe already experienced severe drought summer of 2003, 2010, 2013, 2015, and 2018 (Hanelet al. , 2018; Brunner et al. , 2019). Therefore, mechanistic understanding of tree and forest responses to various drought duration is particularly critical for sustainably forest management under future climate change.
Theoretically, trees’ resilience and resistance to drought stress and the recovery and thus survival ability should be associated with the resource storage and availability. Non-structural carbohydrates (NSC = soluble sugars + starch) among other resources (e.g. nutrients, see Gessler et.al 2017) are known to contribute to tree resilience after stress (Li et al. , 2002; Li et al. , 2008b). For many years, reserve storage was considered as a passive process resulting from an accumulation of resources when uptake and assimilation exceeded growth demand (Sala et al. , 2012; Wiley & Helliker, 2012). An alternative hypothesis was proposed in which reserves storage would be rather an active process during the growing season and would act as a sink competing with other sinks (e.g. growth and reproduction) for available resources (Wiley & Helliker, 2012). Several studies investigated the NSC levels of trees during and at the end of the growing season following a growing-season-long drought (Liet al. , 2013; Schönbeck et al. , 2018; Schönbeck et al. , 2020a; Schönbeck et al. , 2020b). They found a drought-induced growth reduction but did not observe a drought-induced NSC decrease of trees, suggesting an active NSC storage under drought at the expense of growth (Wiley & Helliker, 2012; Li, M-H et al. , 2018). Recent evidence indicates that stress actively induces NSC transfer from aboveground tissues to roots stored (Kannenberg et al. , 2018; Li, W et al. , 2018). In contrast, Li W. et al.(2018) analyzed 27 case studies and found that drought decreased NSC concentration by 17.3% in roots, while it did not change NSC in aboveground tissues in the current season. To our knowledge, even less is known about the changes of NSC over winter (post- vs. pre-winter) in trees previously stressed by drought.
Recently, it was proposed that nutrient addition (i.e. fertilization) will affect the fitness of trees under dry conditions, showing intensifying or mitigating effects on trees’ tolerance to drought (Kreuzwieser & Gessler, 2010; Gessler et al. , 2017; Schönbecket al. , 2020b). Nitrogen (N) deficiency can increase the sensitivity of stomata to low leaf water potential (Radin & Ackerson, 1981; Ghashghaie & Saugier, 1989), which further increases the risk of drought-induced carbon starvation (McDowell, 2011). N itself is a main growth limiting nutrient in temperate terrestrial ecosystems and is also a major component of Rubisco and other photosynthetic enzymes and structures which regulate the photosynthetic activity and thus carbon gain and the NSC level of trees in response to environmental factors such as drought (Bondet al. , 1999; Meng et al. , 2016). Schönbecket al. (2020a) found that negative effects of moderate drought intensity (but not of severe drought) could be compensated by increased nutrient availability in Scot pine saplings. In contrast, Jacobs et al. (2004) reported that fertilization with blended fertilizer impaired the root system development and drought avoidance ability of drought-stressed Douglas-fir seedlings. Similarly, Dziedek et al. (2016) found that nitrogen addition increased the drought sensitivity of saplings of several deciduous tree species (Dziedek et al. , 2016). Despite these studies, there is a strong knowledge gap about drought and nutrient interaction and especially about whether and to what extent nutrient addition affects winter NSC consumption and thus growth recovery in the season following drought.
The species, Quercus petraea (Matt.) Liebl. (oak ) andFagus sylvatica L. (beech), are two coexisting species in European forests. According to (Ellenberg, 2009), oak will become more competitive than beech at sites where as July-temperatures increase to >18°C and precipitation decreases to <600 mm/year as a result of climate change. We carried out a greenhouse experiment to explore the effects of different drought duration in combination with fertilization on the physiology, mortality and growth of these two species. We mainly focused on the following research questions: 1) How does drought duration influence trees’ performance; 2) How does various drought duration in previous growing season affect tree’s winter NSC consumption and thus the over-winter NSC change? 3) Does post-winter NSC level (at the early beginning of next growing season) rather than pre-winter NSC level determine growth recovery of previously drought-stressed trees? 4) Can fertilization mitigate the negative drought effects on trees as proposed by Gessler et al. (2017)?