This study was conducted at the Fanggan Research Station at Shandong University in Jinan, Shandong Province, China (36°26′ N, 117°27′ E). Seeds from R. pseudoacacia L. were collected from the trees in our common garden, while seeds from Q. acutissima Carr. were collected from the trees on a mountain nearby, and they were stored at 4°C in a refrigerator. These seeds were germinated in a growth chamber filled with humus soil in early April 2018. When most seedlings reached 10 cm, healthy and uniform germinants were sown in plastic pots (32 × 29 cm, height × diameter) with an 8 kg mixed sandy loam and humus soil, whose holding capacity of soil water at full saturation was ∼2 kg. Seedlings in the pots were allowed to grow for 3 months. The average greenhouse conditions for the entire duration after treatments were as follows: air temperature 30.2°C (26.7–33.5°C) in the daytime (06:00–18:00) and 24.8°C (20.2–28.6°C) at night (18:00–06:00) and relative humidity 79.3% (55.2–100%) in the daytime and 93.6% (67.3–100%) at night.

Two planting methods were used in our research, namely, pure planting, two R. pseudoacacia in one pot (RR) or two Q. acutissima in one pot (QQ), and mixed planting, one R. pseudoacacia (R_Q) and one Q. acutissima (Q_R) in the same pot. In both planting methods, two seedlings were planted at one-third and two-thirds of the pot diameter, respectively. Each planting method had 72 well-irrigated pots of seedlings before treatments. In control group (CK), 28 pots had been well irrigated to the end of the experiment; in drought treatment (D), when treatments began, first, 44 pots were withheld from watering on July 1 for 10 days; second, 20 out of the 44 drought treatment pots underwent the first rewatering treatment (R₁) on July 11 for 20 days; and then, 12 out of the 20 that were given first rewatering treatment pots underwent repeated drought treatment on July 31 for 10 days; finally, 8 out of the 12 pots underwent repeated rewatering treatment (R₂) on August 10 for 20 days. Both R₁ and R₂ belonged to repeated drought-rewatering treatment (R). We drew a table to summarize our experimental design (Table 1). In our experiment, 10 days after drought resulted in soil water content of all pots reaching c. 20% of the field capacity (2 kg). We had seven harvest times, namely, July 1 (Day 0), July 11 (Day 10), July 21 (Day 20), July 31 (Day 30), August 10 (Day 40), August 20 (Day 50), and August 30 (Day 60). On July 1, we only harvested the control group; on July 11, we harvested both control group and drought treatment; on other harvest times, we harvested all the groups. Four pots of each planting method and treatment were randomly selected for measurement.

Height (H, m), basal diameter (BD, cm), and total biomass (TBM, g) were recorded at each harvest time. Four repetitions in each treatment were harvested and separated into roots, stems, and leaves. Then, the samples were oven-dried at 80°C for 48 h and weighed. TBM was calculated as the sum of leaf biomass (LBM, g), stem biomass (SBM, g), and root biomass (RBM, g). The allometric growth was calculated as Šrutek (1997):

Fully expanded and healthy leaves of four repetitions in each treatment of both species were chosen for a net photosynthetic rate (A_P, μmol m^–2 s^–1) measurement on each harvest time, and it was measured in situ with an infrared gas analysis system (Li-6800, Li-Cor, Lincoln, NE, United States). The measurements were conducted at 1,000 μmol m^–2 s^–1 photosynthetic photo flux density, which was supplied by an external light emitting diode light. Gas-exchange characteristics were measured between 9:00 and 11:00 on sunny days. During the measurement, temperature, relative humidity, and CO₂ concentration inside the chamber were controlled at 28°C, 50%, and 400 ppm, respectively (Wang et al., 2020; Liu et al., 2021).

Xylem midday water potential (Ψ_md, MPa) of pure planting was measured using one of the seedlings in each plot; Ψ_md of mixed planting was estimated by measuring leaf water potential in a pressure chamber (1505D-EXP, PMS Instrument Company, Albany, OR, United States); before measurement, leaves were covered with aluminum foil for 1 h to allow leaf water potential to equilibrate with xylem water potential. Four repetitions of each treatment were measured one by one as soon as they were cut from the seedlings at harvest times after A_P measurement.

The stem-specific hydraulic conductivity (K_s, kg m^–1 s^–1 MPa^–1) was measured at the same time as xylem midday water potential measurement. Four repetitions of each treatment were cut and immersed into degassed water. Subsequently, the samples were transported promptly to the laboratory with the crowns covered with black plastic bags. All leaves and bark were removed, and each segment of R. pseudoacacia was 30 cm long, while that of Q. acutissima was 20 cm long, to ensure there would never be open vessels (Liu et al., 2020). The segments were connected to a hydraulic conductivity measurement system that contained degassed, filtered 20.0 mmol L^–1 KCl solution. A 30 cm hydraulic head generated hydrostatic pressure to impel water through the segments. The K_s was calculated as follows:

where L is the length of the segment (m), Q_m is the mass of the water per unit of time through a segment (kg s^–1), S is the average cross-sectional area for both ends of the stem (m²), and p is the intensity of the water pressure across the segment (MPa) (Liu et al., 2021).

Data were checked for normality (Shapiro–Wilk test) and homogeneity (Levene test). Four-way analysis of variance (ANOVA) was applied to detect the main effect and interaction of species, planting methods, treatments, and harvest time on plant functional traits. One-way ANOVA was used to test the differences of plant functional traits among different treatments or harvest time within the same species. All ANOVAs were followed by Duncan’s multiple comparison tests at α = 0.05 when significant differences were found. The relationships between height and harvest time, between basal diameter and harvest time, were analyzed by linear regression and quadratic polynomial regression, respectively. The results with significance and larger adjusted coefficient of determination (R²_adj) were retained. Data of height and basal diameter were log-transformed, and then, linear regression was performed to find out the allometric growth patterns. All statistical analyses were performed using the SPSS 26 software package (SPSS Inc., Chicago, IL, United States), and all figures were drawn using Origin 2019b (Originlab Co., Northampton, MA, United States).

The results of four-way ANOVA show that the main effects of species, planting methods, treatments, and harvest time on plant functional traits are significant (P < 0.05). In addition, there were interaction effects of the four factors with plant functional traits (Table 2). In other words, different planting methods, treatments, and harvest time had significant effects on plant hydraulic traits of the two tree species. Then, we compared the differences of plant functional traits of R. pseudoacacia and Q. acutissima in different planting methods, treatments, and harvest time.

Ψ_md was significantly reduced by drought, 10 days after drought, Ψ_md reached nearly to –10 MPa, and could recover to the control level (c. −0.01 MPa) after rewatering. For R. pseudoacacia, Ψ_md of RR reached the lowest level (–10 MPa) at Day 30, while that of R_Q reached the lowest level at Day 40. For Q. acutissima, Ψ_md of QQ reached the lowest level at Day 40, while that of Q_R reached the lowest level at Day 50. In terms of the two drought-rewatering stages, the water potential decreased by the second drought, which was less than that decreased by the first drought; however, recovery of Ψ_md in repeated rewatering treatment of two species was faster than that in the first rewatering treatment, and it was shown that Ψ_md at Day 40 was higher than that at Day 10 (P < 0.05); it was shorter for repeated rewatering treatment to recover to the control level after rewatering than the first rewatering treatment (Figure 1).

K_s of RR reached 0, where plant may die, at Day 30; however, K_s of R_Q reached 0 at Day 40 in R. pseudoacacia. Meanwhile, it was higher in rewatering treatment than that in the other two treatments. Compared with pure planting, mixed planting reduced K_s (P < 0.05). For Q. acutissima in QQ, the K_s reached 0 at Day 40, while that of Q_R never reached 0. In pure planting, K_s was significantly higher than that of the other two treatments, but there was no significant difference in mixed planting group (Figure 2).

For R. pseudoacacia, drought reduced A_P at first, even less than 0, which recovered rapidly after rewatering, but still significantly less than the control group; A_P of R_Q was higher than that of RR within the same treatment (P < 0.05). For Q. acutissima, A_P was reduced by drought, but it did not fall below 0. After rewatering, A_P recovered rapidly, and A_P of QQ could reach the control level, while that of Q_R could recover even significantly higher than the control group after rewatering (Figure 3).

For R. pseudoacacia, first, in drought treatment, H increased first and then decreased; the highest H of RR appeared on Day 20, and that of R_Q appeared at Day 30. Then, in repeated drought-rewatering treatment, H continued to increase after repeated drought-rewatering, and the growth rate was close to or even higher than the control group especially in repeated rewatering treatment. For Q. acutissima, first, the growth rate of H in Q_R was higher than that of QQ in control group. In drought treatment, H of QQ began to decrease after drought treatment, while that of Q_R first increased after drought treatment, and the highest H appeared at Day 20 and then gradually decreased. Finally, in rewatering treatment, the increase of H in the first rewatering treatment was not obvious, but H in repeated rewatering treatment increased significantly. However, the growth rates were lower than the control group (Figure 4 and Supplementary Table 1).

For R. pseudoacacia, first in CK, BD and its growth rate were significantly increased by mixed planting. Then, in drought treatment, BD increased and then decreased, and the maximum BD of RR appeared at Day 40; however, the maximum BD of R_Q appeared at Day 30. Finally, in rewatering treatment, BD of RR increased rapidly during the first rewatering treatment, and the increase rate was even higher than that of the control group. For Q. acutissima, first in drought treatment, BD decreased gradually, and the decreasing rate of QQ was fast and then slowed down, while that of Q_R was slow and then became faster. Then, in rewatering treatment, BD increased rapidly during the first rewatering treatment, and the increasing rate was close to that of control group, but there was no significant change in BD during repeated rewatering treatment (Figure 5 and Supplementary Table 2).

For R. pseudoacacia, the first rewatering treatment made the allometric growth pattern of RR more inclined to the thickening pattern of basal diameter than control group. However, drought made R_Q more inclined to the increase of H; in other treatments, the growth pattern had no obvious trend, and the overall growth rate of H was larger than that of BD. For Q. acutissima, Q_R was more inclined to the higher H pattern than QQ; in other treatments, the growth pattern had no obvious trend, and the overall growth rate of H was larger than that of BD (Figure 6 and Supplementary Table 3).

For R. pseudoacacia, drought made the biomass significantly lower than that of control group. No matter what kind of planting methods and treatments, compared with drought treatment, the biomass of rewatering treatment had no significant change; under the same treatment, the biomass of R_Q was higher than that of RR (P < 0.05). For Q. acutissima, the biomass of drought treatment was little, but it could return to control group after rewatering; under the same treatment, the biomass of Q_R was higher than that of QQ (P < 0.05) (Figure 7).

Different water availability had significant effects on plant hydraulic traits (Figure 1 and Table 2); this may result from the different relative positions of the two tree species on the “isohydric-anisohydric” continuous spectrum, that is, R. pseudoacacia is closer to anisohydric plant, and its hydraulic traits respond quickly with the change of water availability, while Q. acutissima is closer to the isohydric plant, and its hydraulic traits respond slowly with the change of water availability (Moser et al., 2016; Li et al., 2019, 2020; Liu et al., 2020). The decrease of water potential caused by the second drought was less than that caused by the first drought, and there was no significant difference in Ψ_md of the two species between repeated rewatering treatment and control group at the end of repeated rewatering treatment. It may be due to that the first drought made the two species produce stress imprint, so that they could better cope with the repeated drought and recover quickly after rewatering (Munné-Bosch et al., 2003; Jakab et al., 2005; Douma et al., 2017). For the two species, K_s of drought treatment decreased gradually, while it maintained at a high level after rewatering. K_s of rewatering treatment was almost higher than that of drought treatment, which may be due to the formation of stress imprint after drought stress, so that the two tree species could recover quickly after rewatering and provide enough water for other plant functional traits. Moreover, repeated drought-rewatering did not cause pressure fatigue according to our results. Our results do not find the negative effect of repeated drought-rewatering on plant hydraulic traits of the two tree species. On the contrary, the stress imprint induced by repeated drought-rewatering was conducive to the rapid recovery of the seedlings of the two tree species in the changing environment.

Planting methods had significant effects on plant hydraulic traits (Figures 1, 2 and Table 2). In terms of the relationship between drought and planting methods, the differences in root distribution and structure between R. pseudoacacia and Q. acutissima may overall reduce the competition for water resource and enhance the drought resistance (Wang and Zhou, 2000). In addition, mixed planting may reduce K_s to make the limited water be retained, and it is helpful for the mixed planting R. pseudoacacia and Q. acutissima to survive under drought stress. However, species in pure planting have similar functional characteristics, resulting in functional redundancy and niche overlap, which may lead to intensified water competition and increased drought stress (Lebourgeois et al., 2013; Pretzsch et al., 2013; Forrester and Pretzsch, 2015; Schwendenmann et al., 2015). The results show that species composition may have a certain effect on drought resistance, and mixed planting could weaken the ecological effect of drought.

Treatments had significant effects on leaf photosynthetic traits (Table 2 and Figure 3). Combined with the changes of Ψ_md (Figure 1) and K_s (Figure 2), our results illustrate that both two species formed a stress imprint after the first drought, and when facing the repeated drought, they could better respond to it (Munné-Bosch et al., 2003; Jakab et al., 2005; Douma et al., 2017). In contrast, the two tree species grew rapidly by increasing photosynthesis after rewatering, that is, they adopted the compensatory mechanism. R. pseudoacacia adopted an under compensatory mechanism or equal compensatory mechanism (Figures 3A,B), and Q. acutissima adopted an equal compensatory mechanism or even over compensatory mechanism (Figures 3C,D; Trlica and Rittenhouse, 1993; Pinkard et al., 2007; Quentin et al., 2012; Wang et al., 2020), to minimize the impact of drought.

Planting methods had significant effects on A_P (Table 2). Mixed planting delayed the decrease of A_P in drought treatment and increased A_P in the control group and rewatering treatment. This is basically consistent with the change of Ψ_md. Mixed planting prolonged the time for the two species to reach the lowest Ψ_md in drought treatment (Figure 1), resulting in the fact that the decrease of A_P in drought treatment was delayed. This may come from that repeated drought-rewatering, and different planting methods did not break the water-carbon coupling mechanism of plants (Li et al., 2019, 2020). A_P increased as soon as water availability was high. For R. pseudoacacia, A_P was significantly increased by mixed planting (Figures 3A,B). For Q. acutissima, after rewatering, A_P of QQ showed an equal compensatory mechanism (Figure 3C), while that of RQ showed an over compensatory mechanism (Figure 3D). Therefore, mixed planting of R. pseudoacacia and Q. acutissima was not only conducive to drought resistance, but it was also conducive to the recovery and improvement of photosynthetic capacity after rewatering.

Treatments had significant effects on H and BD (Table 2). Drought significantly inhibited the growth of H (Figure 4) and BD (Figure 5) of the two tree species, which made the allometric growth pattern different (Figure 6). The allometric growth pattern of R. pseudoacacia was not changed by different planting methods, but drought made pure planting and mixed planting present the opposite pattern. In other words, RR presented the BD growth pattern (Figure 6A), while R_Q presented the H growth pattern (Figure 6B). It indicated that under drought conditions, RR adopted the survival strategy in the tradeoff between growth-survival strategy. During the fierce intraspecific competition for water resources, RR could increase BD to store more water and nutrients in response to drought stress, while R_Q had less competition for water resources (Schwendenmann et al., 2015), and less drought stress was better for R. pseudoacacia to get more water by accelerating growth in response to drought. Repeated drought-rewatering increased the growth rate of H of the two tree species (Figure 4) but decreased the growth rate of BD (Figure 5). In terms of plant growth traits, repeated drought did not cause pressure fatigue, while the stress imprint formed by the first drought did not make the two tree species adopt the survival strategy of BD growth mode but take the growth strategy of fast growth.

Planting methods had significant effects on H, BD, and the allometric growth pattern. Our results may be due to the difference in root structure of the two species, resulting in less overall competition for water resource (Schwendenmann et al., 2015). Therefore, under the condition of limited water resources, compared with pure planting, mixed planting reduced the competition for water resources, enhanced drought resistance, and delayed the wilting rate of plant growth (Lebourgeois et al., 2013; Pretzsch et al., 2013). However, pure planting of the two tree species had the same functional characteristics, that is, functional redundancy. The overlap of niches led to the intensification of water competition of R. pseudoacacia or Q. acutissima, increased the degree of drought, and led to the wilting of plant growth under drought stress. Although the response of mixed species to water use and drought stress is very complex (Forrester and Pretzsch, 2015; De Cáceres et al., 2021), our results show that mixed planting could effectively reduce the competition of water resources and enhance drought resistance in terms of plant growth traits.

Drought had a significant inhibitory effect on BM. After rewatering, BM of R. pseudoacacia could not return to the control level (Figures 7A,B), while that of Q. acutissima could return to the control level (Figures 7C,D). This may result from that R. pseudoacacia is close to the anisohydric plant on “isohydric-anisohydric” continuous spectrum (Moser et al., 2016; Li et al., 2019; Liu et al., 2020). Faced with drought, the biomass accumulation stopped, due to the rapid decline of plant hydraulic traits (Figures 1A,B) and the rapid decline of A_P (Figures 3A,B), while BM increased slowly after rewatering due to the compensatory mechanism of photosynthesis, especially in RR. However, Q. acutissima was close to the anisohydric plant on “isohydric-anisohydric” continuous spectrum (Moser et al., 2016; Li et al., 2019; Liu et al., 2020). Faced with drought, plant hydraulic traits decreased slowly (Figures 1C,D), leading to a slow decrease in photosynthesis (Figures 3C,D). After rewatering, BM increased rapidly due to the over compensatory mechanism of photosynthesis, especially in Q_R. Planting methods have significant effects on BM (Table 2). Mixed planting increased the BM of the two species in each treatment (Figures 7B,D), which may result from the more reasonable allocation of resources, reducing the competition between the two species and conserving water and soil (Schwendenmann et al., 2015).

Drought had a significant effect on the plant functional traits of the two tree species under different planting methods. It decreased Ψ_md, K_s, and A_P of the two tree species, and ultimately reduced H, BD, and BM. Mixed planting could reduce the competition of light, water, space, and other resources, making the limited resources more reasonable allocation, so as to weaken the inhibition of drought on the two tree species to a certain extent. After rewatering, plant functional traits were recovered. Repeated drought-rewatering did not make the two tree species produce pressure fatigue. On the contrary, it made the two tree species form stress imprint, that is, the storage of stress-resistant substances through physiological and biochemical responses. The stress imprint could better respond to drought when faced with drought stress again. After rewatering, Ψ_md quickly recovered, and photosynthesis of R. pseudoacacia adopted compensatory strategies. Therefore, the BM was accumulated. Photosynthesis of Q. acutissima was overcompensated to make the BM close to or even reach the level without drought. Therefore, the combination of the two species and repeated drought-rewatering can improve the overall drought resistance ability to a certain extent.

Through the experiment of repeated drought-rewatering of R. pseudoacacia and Q. acutissima under pure and mixed planting, it is found that mixed planting of the two tree species could weaken the inhibition of drought to a certain extent; repeated drought-rewatering did not cause pressure fatigue to the two species but made them form stress imprint, which can quickly respond to repeated drought and recover quickly after rewatering. Our research suggested that the mixed planting of the two tree species could better adapt to repeated drought-rewatering under global climate change.