How to Choose a Hydrological Recovery Mode for Degraded Semiarid Wetland in China? A Case Study on Restoration of Phragmites australis Saline-Alkaline Wetland

Hydrological recovery is the basis for restoring the structure and function of wetlands in semiarid and arid areas of China. Selecting an appropriate hydrological recovery mode may be helpful for improving the effectiveness of wetland restoration. We conducted pot experiments to study the effects of the flooding frequency, duration, depth, and occurrence time on the height, biomass, ion contents, and photosynthetic physiology of Phragmites australis in degraded saline–alkaline marsh in the West Songnen Plain, China. At the end of the growing season, we found that the biomass, photosynthetic parameters, and water use efficiency (WUE) of the leaves increased, whereas the Na+ concentration decreased, and the K+ content remained unchanged under an increased flooding frequency treatment. As the flooding depth increased, the plant height increased, but there were no differences in the photosynthetic parameters, biomass, and WUE under flooding at 5 cm and 10 cm. Under different flooding duration treatments, the plant height and biomass were greater, but the photosynthetic parameters and Na+ and K+ contents were lower under a flooding duration of three months. The flooding occurrence time had little effect on the growth of P. australis. Our results indicate that the flooding frequency and duration had greater effects than the flooding depth and occurrence time in the hydrological recovery model for P. australis restoration. The biomass accumulated by P. australis was related to lower Na+ contents and the maintenance of a high K+/Na+ contents, and WUE increased by adjusting photosynthesis under a moderate flooding frequency and duration. These results have important implications for the restoration of degraded semiarid wetlands with man-made channel systems in conditions with limited freshwater resources.


Introduction
Wetlands are considered the "kidneys" of the Earth and they have many important ecological functions, e.g., carbon sequestration, nutrient cycling, water purification, erosion control, and wildlife habitats [1]. However, wetland degradation has occurred, including area reductions, water quality  [20], and (B) the sampling location in Niuxintaobao wetland in this study.
We conducted pot experiments in 2013. Soil clods (20 × 20 × 20 cm 3 ) containing P. australis rhizomes, buds, and young seedlings were collected and sprayed with water to keep them moist before transplanting into pots. Soil with a depth of 20-30 cm, but without P. australis was prepared in May. In an open space, saline soil was mixed evenly and then used to fill pots (diameter of 30 cm and height of 35 cm) to a depth 5 cm, before adding one plant. Finally, the spaces in each pot were filled with mixed soil. The distance from the soil surface to the bottom was 25 cm after watering. The remaining depth of 10 cm was required for the flooding depth treatment. All pots were weighed using digital scales. Freshwater irrigation was applied for 2 weeks to restore the growth of P. australis before the experiments started. The same numbers of plants were prepared for the different water treatments.

Flooding Treatments
We set different flooding frequencies, depths, durations, and occurrence times in separate pot experiments, as follows (   [20], and (B) the sampling location in Niuxintaobao wetland in this study.
We conducted pot experiments in 2013. Soil clods (20 × 20 × 20 cm 3 ) containing P. australis rhizomes, buds, and young seedlings were collected and sprayed with water to keep them moist before transplanting into pots. Soil with a depth of 20-30 cm, but without P. australis was prepared in May. In an open space, saline soil was mixed evenly and then used to fill pots (diameter of 30 cm and height of 35 cm) to a depth 5 cm, before adding one plant. Finally, the spaces in each pot were filled with mixed soil. The distance from the soil surface to the bottom was 25 cm after watering. The remaining depth of 10 cm was required for the flooding depth treatment. All pots were weighed using digital scales. Freshwater irrigation was applied for 2 weeks to restore the growth of P. australis before the experiments started. The same numbers of plants were prepared for the different water treatments.

Flooding Treatments
We set different flooding frequencies, depths, durations, and occurrence times in separate pot experiments, as follows ( Flooding for one month starting from June, July, August, and September, (C1, C2, C3, C4), respectively. The experimental period lasted from May until 1 October. Except for the flooding treatments with different depths, drought was avoided by keeping the water content at 35-40% of the field capacity using a syringe (weight-based method). All pots were kept in a greenhouse to avoid disturbance by rainwater. The greenhouse was ventilated to keep a similar temperature inside to that outside. Three replicates were tested for each treatment, with a total of 36 pots.
Sustainability 2020, 12, x FOR PEER REVIEW  4 of 13 respectively. The experimental period lasted from May until 1 October. Except for the flooding treatments with different depths, drought was avoided by keeping the water content at 35-40% of the field capacity using a syringe (weight-based method). All pots were kept in a greenhouse to avoid disturbance by rainwater. The greenhouse was ventilated to keep a similar temperature inside to that outside. Three replicates were tested for each treatment, with a total of 36 pots.

Sampling and Determination
Before harvesting, we sampled five random plants per treatment to determine the net photosynthetic rate (Pn), transpiration rate (E), and stomatal conductance (Gs) in September. We measured the fully expanded youngest leaves from each plant at 9:00-11:00 am on each sampling date using a portable open flow gas exchange system (LI-6400; LI-COR Inc., Lincoln, NE, USA), which simultaneously analyzed the field light intensity (1400 μmol m -2 s -1 ). The leaf measurements were repeated five times and the averages were recorded. The water use efficiency (WUE) was calculated for the leaves as follows: WUE (g kg −1 ) = Pn μmol (CO2 m −2 s −1 )/E mmol (H2O m −2 s −1 ).
At the end of the growing season, we determined the heights of five P. australis plants for each treatment. The absolute height was determined using a measuring tape. The above-and belowground organs were separated from the plants in each pot, washed with tap water and distilled water, and oven dried at 105 °C for 15 min and then 70 °C for 48 h, before determining the dry weight. In addition, 100 mg of each dried crushed sample was treated with 10 mL deionized water at 100 °C for 1 h, and the extract was used to determine the Na + and K + contents with an atomic absorption spectrophotometer (TAS-990, Purkinje General, Beijing, China) [28].

Sampling and Determination
Before harvesting, we sampled five random plants per treatment to determine the net photosynthetic rate (P n ), transpiration rate (E), and stomatal conductance (G s ) in September. We measured the fully expanded youngest leaves from each plant at 9:00-11:00 am on each sampling date using a portable open flow gas exchange system (LI-6400; LI-COR Inc., Lincoln, NE, USA), which simultaneously analyzed the field light intensity (1400 µmol m -2 s -1 ). The leaf measurements were repeated five times and the averages were recorded. The water use efficiency (WUE) was calculated for the leaves as follows: WUE (g kg −1 ) = P n µmol (CO 2 m −2 s −1 )/E mmol (H 2 O m −2 s −1 ).
At the end of the growing season, we determined the heights of five P. australis plants for each treatment. The absolute height was determined using a measuring tape. The above-and belowground organs were separated from the plants in each pot, washed with tap water and distilled water, and oven dried at 105 • C for 15 min and then 70 • C for 48 h, before determining the dry weight. In addition, 100 mg of each dried crushed sample was treated with 10 mL deionized water at 100 • C for 1 h, and the extract was used to determine the Na + and K + contents with an atomic absorption spectrophotometer (TAS-990, Purkinje General, Beijing, China) [28].

Statistical Analyses
We performed one-way analysis of variance using SPSS 22.0 (SPSS Inc, Chicago, IL, USA). A transformation was necessary to satisfy the statistical requirement if the data were not normally distributed. The means and standard errors were calculated. Different treatments were analyzed using the least-significant difference method for multiple comparisons (p < 0.05).

Plant Height and Biomass
The height of P. australis was significantly affected by the flooding frequency, depth, and duration (p < 0.05, Table 1), but not by the flooding occurrence time. However, there was no significant difference in the height between M1 and M2, and no significant differences among M1, M3, and M4 ( Figure 3C). The plant height was greater under D10 than D5 ( Figure 3B). Increasing the flooding duration affected the plant height, but there were no significance differences between that under M1 and those with the other treatments. The average heights were higher under M3 and M4 than M1 and M2 ( Figure 3C). P. australis grew higher with a greater flooding higher frequency and flooding depth, with an average height of 80 cm in the pots. Table 1. ANOVA analysis of hydrological condition effects on P. australis.
The biomass of P. australis was influenced significantly by the flooding frequency, depth, duration, and flooding occurrence time (p < 0.05, Table 1). As the flooding frequency increased, the aboveground and belowground biomasses both increased markedly, where the highest increases under F2 were 9.3 and 3.9 times greater than those under F0, respectively ( Figure 4A). There was no significant difference in the biomass under the D5 and D10 flooding depth treatments ( Figure 4B). As the flooding duration increased, the biomass varied and it was higher under M3. The aboveground and belowground biomasses were highest under M3, i.e., 6.6-and 2.8-times higher than those under D0, respectively ( Figure 4C). As the flooding occurrence time advanced, the plant biomass did not change significantly from C1 to C3, but it decreased significantly under C4 ( Figure 4D). The accumulated biomass was higher under C2, where the aboveground and belowground biomasses were 6.8 and 2.8 times higher than those with no flooding (D0), respectively. The biomass of P. australis was greatest under the increased flooding frequency treatment (F2).

Na + and K + Contents
The Na + contents of plants were affected by the flooding frequency and flooding duration (p < 0.05), but not by the flooding depth and flooding occurrence time (Table 1, Figure 5). Under the different flooding frequency treatments, the Na + concentrations in the aboveground and belowground organs were lower under F2 compared with F0, which were 22% and 71% lower, respectively ( Figure 5A). Under the different flooding duration treatments, the Na + content of the aboveground organs was lowest under M3, and the Na + contents of the belowground organs were lowest under M1 and M3 ( Figure 5B). 0.05), but not by the flooding depth and flooding occurrence time (Table 1, Figure 5). Under the different flooding frequency treatments, the Na + concentrations in the aboveground and belowground organs were lower under F2 compared with F0, which were 22% and 71% lower, respectively ( Figure 5A). Under the different flooding duration treatments, the Na + content of the aboveground organs was lowest under M3, and the Na + contents of the belowground organs were lowest under M1 and M3 ( Figure 5B).
The K + contents of the aboveground organs were affected by the flooding duration, flooding depth, and flooding occurrence time, but the K + contents of the belowground organs were only affected by the flooding duration (Table 1, Figure 6). Under the different flooding duration treatments, the K + contents of the aboveground organs were lower under M1 and M3 compared with the control D0, but higher in the belowground organs ( Figure 6B). As the flooding depth increased, the K + contents of the aboveground organs increased significantly up to D10 ( Figure 6C). As the flooding occurrence time advanced, the K + contents of the aboveground organs increased, where the highest value was obtained under C4, and it was 29-238% higher than those under the other treatments ( Figure 6D). The K + contents of the aboveground organs were affected by the flooding duration, flooding depth, and flooding occurrence time, but the K + contents of the belowground organs were only affected by the flooding duration (Table 1, Figure 6). Under the different flooding duration treatments, the K + contents of the aboveground organs were lower under M1 and M3 compared with the control D0, but higher in the belowground organs ( Figure 6B). As the flooding depth increased, the K + contents of the aboveground organs increased significantly up to D10 ( Figure 6C). As the flooding occurrence time advanced, the K + contents of the aboveground organs increased, where the highest value was obtained under C4, and it was 29-238% higher than those under the other treatments ( Figure 6D).

Photosynthetic Parameters and WUE
The photosynthetic parameters and WUE were affected by all hydrological recovery modes (p < 0.05, Table 1). The photosynthetic parameters were higher under F1 than F2 (

Photosynthetic Parameters and WUE
The photosynthetic parameters and WUE were affected by all hydrological recovery modes (p < 0.05, Table 1). The photosynthetic parameters were higher under F1 than F2 (Table 2), whereas the opposite was found for WUE, where WUE was 7.2% higher under F2 compared with F1 and P n was 23.8% higher under F1 compared with F2. There were no significant differences in P n , E, and WUE under D5 and D10. The photosynthetic parameters comprising P n and E varied as the flooding duration increased, with higher values under M0 and M2, but lower values under M1, M3, and M4. WUE exhibited the opposite trend to P n , where the photosynthetic rate was higher when WUE was lower. Similarly, the photosynthetic parameters and WUE varied under different flooding occurrence time, with higher values under C2 and C4 than C1 and C3, where WUE again exhibited the opposite trend to the photosynthetic rate.

Discussion
The preferred habitats of P. australis are described as un-shaded, low-lying areas that are intermittently or permanently flooded with shallow still water [29], and it appears to be well adapted to growing in areas with a widely fluctuating hydroperiod [30]. An important characteristic of most wetland plants in semiarid regions is their capacity to withstand fluctuations between extended dry phases and floods [31]. However, the applied anthropogenic adjustment methods rarely consider these fluctuations when restoring wetlands. In general, the water depth and cover periods are the main factors considered by policy makers or ecological engineers when restoring marshes. Our results indicate that the flooding depth treatment had a greater effect on the plant height compared to the other three treatments, but increasing the flooding depth did not enhance the biomass accumulation (Figures 3 and 4). The effects of the flooding duration and frequency on the establishment of plant species have been shown to be greater than that of the flooding depth [32]. We also found that the effects of the flooding frequency and duration were greater than those of the flooding depth on enhancing the biomass accumulation by P. australis. However, the aboveground biomass did not have a linear correlation with the flooding duration (R 2 = 0.30), but instead the relationship was described by a parabolic binary equation (y = −2.6357x 2 + 18.284x − 12.34, R 2 = 0.77, where x is the flooding duration and y is the aboveground biomass), which showed that the highest biomass would be achieved with a flooding duration of 2.5 months.
Our results indicated that the effects of the flooding frequency were more significant and positive compared with the other treatments (Figures 3 and 4). The F2 treatment involved flooding and draining twice, and this treatment reflected hydrological conditions comprising intermittent flooding or water level fluctuations between dry phases and floods. Under F2, both the aboveground and belowground biomasses accumulated significantly, thereby indicating the positive adaptation of P. australis to growth under these conditions. The pot experiments restricted the expansion of the plant rhizomes, so the experimental heights and biomass were lower than those expected in the field, but it would be possible to lower the water depth and increase flooding frequency when restoring P. australis in wetlands, and this approach may also save water.
If the water levels are low in saline-alkaline marshland, plants will suffer from Na + and high pH stresses. High sodic salt levels induce the expression of an amiloride resistant Na + /H + antiporter, which could account for the remarkable tolerance of NaCl by some plants [33]. Plants then compartmentalize the sodium into one organ to reduce its transport and the harmful effects of ions on other organs [30]. In these organs, plants restrict the excess Na + to vacuoles to facilitate their metabolic functions [30,34]. This salt tolerance mechanism is also active in P. australis when growing in degraded saline-alkaline marshes. High Na + and low H + concentrations outside the root cells make it easier for Na + /H + antiporters to export H + and import excess Na + to support the growth of P. australis. More Na + is then allocated to the belowground organs (rhizomes and roots) compared with the aboveground organs (stems and leaves) to reduce the damaging effects of salinity on the shoots and leaves, as found under the F0 and D0 treatments in the present study. This sodium localization disappears after hydrological recovery. Some water treatments (C1, C3 and C4), may have disrupted the salt tolerance capacity of P. australis ( Figure 5D). The flooding duration and frequency markedly affected the Na + level in the soil to reduce the absorption capacity by P. australis ( Figure 5) because the Na + ions may have been forced deeper into the soil as the flooding duration increased.
K + is essential for all plant life and is the major cationic inorganic nutrient in most terrestrial plants [35]. When plants grow in saline soils, one of the key factors that determine salinity tolerance is the capacity to maintain a high cytosolic K + /Na + [36]. The different hydrological recovery modes caused intermittent saline stress in P. australis, except flooding for the whole growing season ( Figure 6). Compared with D0 (no water), the K + contents decreased under the flooding duration treatments, but increased under the flooding occurrence time treatments, and remained unchanged under the flooding frequency treatments. However, the Na + /K + contents of P. australis were higher under F2, M1, and M3, especially in the belowground organs, which coincided with the trend in the belowground biomass (Table S1). The maintenance of high Na + /K + contents is also a positive response of P. australis to hydrological changes.
Photosynthesis is highly sensitive to the environment, where the immediate sensing of minimal environmental changes triggers a series of adjustments, thereby eventually leading to adaptation through changes in primary production [37]. P. australis exhibits high phenotypic plasticity in terms of its photosynthetic response to climatic change [38]. Our results indicated that P. australis also exhibited growth plasticity in response to water availability changes, where the height and biomass changes were mainly adjusted by P n and WUE (Table 2). Flooding slightly affected the physiological behavior of P. australis in terms of leaf gas exchange and water relationships [39]. Biochemical and physiological measurements indicate that the leaves of P. australis utilize a C3 mechanism for carbon fixation. Zheng et al. [40] demonstrated the existence of a C3-like ecotype in swamp environments. However, the photosynthetic carbon assimilation characteristics of P. australis shift to C4 metabolism in response to increasing drought stress [41]. When P. australis adapted to long-term drought in habitats such as swamp or marshland, together with high temperature, the photosynthetic membrane structure and photosynthetic characteristics changed compared with the swamp ecotype [42].
In general, the P n and G s values in P. australis plants treated with hypoxia are positively correlated and P n also has a strong positive relationship with E [39]. Clevering [43] suggested that the water depth may be a much stronger selective force for wetland plants than the nutrient availability when the water supply is restricted. In wetter habitats, the growth of P. australis is more limited by the nutrient status than the water depth [24], which may explain why we found that P n increased in P. australis under water restriction, whereas P n decreased when sufficient water was available (Table 2), for which a decreasing trend was found as the water depth increased under D0, D5, and D10. Relatively long-term water restriction and sufficient phases were provided under F1, and thus, P n was higher and the biomass accumulation was greater. M4 involved flooding for the whole season and the sufficient water led to decreases or no changes in P n and WUE compared with the control. Under different flooding durations, P n was lower, but WUE was higher under M1 and M3, and thus, the water demand may have been greatest in these development phases for P. australis. Under C1, C2, and C4, P n increased but WUE decreased or remained unchanged in P. australis. The four photosynthetic responses exhibited by P. australis under different hydrological recovery modes, demonstrate its high adaptability to environmental water changes.
Wetlands are increasingly being restored in many areas and these efforts aim to recover the hydrological regime in wetlands mainly by focusing on the flooding depth and time [32,44]. Our results indicated that the flooding frequency and duration are more important than the flooding depth and occurrence time, and these factors may have critical effects on the hydrologic conditions to allow the rapid recovery of P. australis populations. These findings may be integrated into methods for effectively improving saline-alkali soil to facilitate the restoration and conservation of wetlands. Our findings indicated that precipitation may be the main issue on Songnen Plain because it may have led to different flooding frequencies. Decreased precipitation is a key factor that affects wetland systems by changing hydrological regimes [44]. Thus, we consider that policy makers should not focus only on the flooding depth and time, but instead flooding cycles should be artificially implemented. Our findings may help to meet the demand for constructing an ecological civilization in China, and the Central Committee of the Communist Party of China and the State Council of the People's Republic of China clearly stated that the wetland area should not be less than 800 million acres by 2020.
This study had some limitations. In particular, the experimental period was short and the plants were grown in pots, so the treatments did not compare with the actual situations in wetlands. Thus, we could not determine restoration thresholds for different hydrological variables, and our results might not be applicable to actual wetland restoration situations. However, we found that the flooding frequency and duration may be crucial for the successful restoration of channeled man-made systems, and our results also have implications for wetland management in the presence of P. australis.

Conclusions
In this study, we investigated the effective hydrological recovery of P. australis populations in a degraded semiarid wetland in China by considering the Songnen Plain saline-alkaline wetland as a case study. Our results indicated that all hydrological recovery modes had significant effects on the plant biomass, saline ion concentrations, photosynthesis, and WUE. Increasing the flooding duration and flooding depth had positive effects on the height of P. australis, and increasing the flooding frequency enhanced the accumulated biomass. A moderate flooding duration (3 months in the early growing season) and flooding frequency (flooding and draining twice) markedly reduced the absorption of Na + by P. australis, where the former decreased the K + level, whereas the latter left the K + level unchanged. P. australis exhibited positive physiological responses to different hydrological recovery modes by adjusting its photosynthetic rate and WUE. Our results indicate that it may be possible to lower the water depth, reduce the flooding duration, and increase the flooding frequency to enhance P. australis production and save water. Our results suggest that the flooding frequency and duration are more important than the flooding depth and occurrence time, and these findings may facilitate the implementation of more effective wetland restoration strategies.