Ilex glabra ‘Shamrock’ (‘Shamrock’ inkberry holly), Itea virginica ‘Henry's Garnet’ (‘Henry's Garnet’ sweetspire), and Viburnum nudum ‘Winterthur’ (‘Winterthur’ possumhaw) were flooded for 0 (non-flooded), 3, or 6 d, first in a greenhouse and then outdoors. Flooding treatments were in a factorial combination of greenhouse flooding treatment × outdoor flooding treatment. Following each flooding period, plants were allowed to drain for 6 d and received no irrigation. Plants in the 3 and 6 d flood treatments experienced a total of 7 and 5 flood cycles, respectively both in the greenhouse and outdoors. All taxa maintained 100% survival during greenhouse and outdoor flooding. Following greenhouse flooding, final size index (SI) of I. virginica ‘Henry's Garnet’ and I. glabra ‘Shamrock’ decreased with increasing flood length; V. nudum ‘Winterthur’ had no differences in SI among treatments. Following outdoor flooding, there were no differences in flooding treatment for relative size index (RSI) of I. glabra ‘Shamrock’, no clear effect of flooding treatment on RSI for V. nudum ‘Winterthur’, and RSI of I. virginica ‘Henry's Garnet’ was lowest in plants that were not flooded in the greenhouse. RDW and SDW of I. virginica ‘Henry's Garnet’ and I. glabra ‘Shamrock’ tended to be lowest in plants that were initially flooded in the greenhouse for 6 d, while there was no clear effect of either greenhouse or outdoor flooding on RDW and SDW of V. nudum ‘Winterthur’. When exposed to flooding in the greenhouse I. virginica ‘Henry's Garnet’ seemed to better tolerate outdoor flooding, while the other two taxa did not appear to gain any benefit from previous flood exposure. All three taxa sustained growth throughout all phases of the experiment and maintained good visual quality indicating that all three would be tolerant of repeated, short-term flooding.
Significance to the Nursery Industry
Plants for use in rain gardens should be able to withstand alternating periods of wet and dry conditions, along with periods of anaerobic conditions since rain gardens may remain flooded for up to 2 days. Ilex glabra ‘Shamrock’ (‘Shamrock’ inkberry holly), Itea virginica ‘Henry's Garnet’ (‘Henry's Garnet’ sweetspire), and Viburnum nudum ‘Winterthur’ (‘Winterthur’ possumhaw) were flooded for 0 (non-flooded), 3, or 6 d, first in a greenhouse and then outdoors. If previously exposed to flooding in the greenhouse, then Itea virginica ‘Henry's Garnet’ seemed to better tolerate outdoor flooding, while the other two taxa did not appear to gain any benefit from previous flood exposure. All three taxa sustained growth throughout all phases of the experiment and maintained good visual quality indicating that all three would be tolerant of repeated, short-term flooding.
Rain gardens are attractive additions to landscapes that facilitate stormwater management and infiltration. A rain garden is a shallow depression that collects stormwater runoff from a roof, parking lot, or other impervious surface (9). Rain gardens depend on precipitation for irrigation, therefore, during precipitation events, rain gardens may flood and remain saturated above the substrate level until water permeates the ground. Following draining, rain gardens may remain dry for a period of weeks until the next rain event. Dussaillant et al. (9) found that soil in a rain garden would likely remain saturated for 1–2 d and plant roots would need to be able to withstand at least 2 d of standing water in the root zone for optimal recharge to occur.
Cyclic flooding consists of a cycle of flooding and draining that is repeated over time, similar to what a rain garden might experience in the landscape. Plants for use in rain gardens should be able to withstand alternating periods of wet and dry conditions (12), along with periods of anaerobic conditions since rain gardens may remain flooded for up to 2 d. Native plants are adapted to local environmental conditions (14), and are usually able to persist during periods of low rainfall or drought making them desirable for rain gardens (8). Facultative wetland plants are specified for rain gardens since they are generally tolerant of wet or dry conditions (12) and may occur in wetlands or non-wetlands in nature (18).
The effect of previous flooding on subsequent flood tolerance seems to be inconsistent. Pezeshki (28) exposed Taxo-dium distichum (L.) Rich. (bald cypress), Quercus nuttallii Palmer (nuttall oak), and Quercus michauxii Nutt. (swamp chesnut oak) to short-term flood events in an attempt to precondition seedlings to subsequent flood tolerance. Results indicated that after repeated exposure to flooding, seedling response did not improve during subsequent flood events, but that additional research using more mature plants was needed. Previous work by Dylewski et al. (10) suggested that larger plants may be more tolerant of repeated flooding. If previous flood exposure were to provide greater flood tolerance during subsequent flooding episodes, this technique could be utilized for better establishment of plants in rain gardens in the designed landscape. This research aims to evaluate plants for use in both rain garden and bioretention applications and therefore uses both a mineral and organic substrate. The objective of this study was to determine the effect of previous flood exposure on flood tolerance and growth of three native landscape shrubs exposed to subsequent repeated short-term flooding.
Materials and Methods
Greenhouse flooding. On October 16, 2008, 11.2 cm (4.4 in) rooted stem cuttings [0.25 liter (0.07 gal) liners] (Spring Meadow Nursery, Grand Haven, MI) of Ilex glabra ‘Shamrock’ (‘Shamrock’ inkberry holly), Itea virginica ‘Henry's Garnet’ (‘Henry's Garnet’ sweetspire), and Viburnum nudum ‘Winterthur’ (‘Winterthur’ possumhaw) were planted into 2.5 liter (0.75 gal) containers with drainage holes. All taxa are native facultative wetland plants (36). Liners were approximately 6 months old at planting. Container substrates used were pine bark:peat:perlite (5:3:1, by vol), hereafter referred to as PB, or soil [(Marvyn sandy loam) collected from research field plots on Auburn University campus, Auburn, AL]. Prior to use, soil was sterilized in an autoclave at 129C (263F) for 3 h. PB was amended with dolomitic limestone [1.2 kg m−3 (2 lb yd−3)], controlled release fertilizer [8 kg m−3 (13.5 lb yd−3)] (8 month 18N-2.64P-9.96K; Polyon®, Pursell Industries, Sylacauga, AL), and micronutrient fertilizer [0.9 kg m−3 (1.5 lb yd−3)] (Micromax®, Scott's Company, Marysville, OH). Plants grown in soil were top-dressed with the same controlled release fertilizer [14 g (0.5 oz) per container] but received no lime or micronutrients. Plants were overwintered in a shade house on campus in Auburn, AL, from October 15, 2008, to February 23, 2009. During the overwintering period, plants received daily overhead irrigation. If temperatures decreased to approximately −4C (25F) or lower, plants were covered with white polyethylene plastic with ventilation holes.
Plants were arranged on raised benches on February 23, 2009, [131 days after planting (DAP)] in an 8 mm polycarbon-ate greenhouse on campus in Auburn, AL. The greenhouse was ventilated and maintained at a minimum night temperature of 18C (64F) and a maximum day temperature of 30C (85F) with natural photoperiods. There were 21 single container replications per treatment per substrate per taxon (V. nudum ‘Winterthur’ grown in soil contained 8 single container replications per treatment).
Plants were flooded for 0 (non-flooded), 3, or 6 d. Flooded conditions were created by nesting a plant's container with drainage holes inside a 2.5 liter (0.75 gal) container without drainage holes. Plants were flooded to the level of the substrate by adding approximately 1 liter (3.8 gal) tap water to the substrate on the first day of flooding and adding approximately 50–100 mL (17–34 oz) tap water on each additional flood day to maintain water table at substrate level. At the end of a flood cycle, the exterior container without holes was removed, and plants were allowed to drain for 6 d. After 6 d of draining, the same flooding treatment was initiated again. No water was added to the substrate during the 6 d draining period. Substrate percent moisture was measured daily for non-flooded plants using a handheld Theta moisture probe (Delta- T Devices Ltd., Cambridge, England) adjusted to either the organic or mineral setting as appropriate. Non-flooded plants were irrigated with 500 mL tap water when substrate percent moisture reached 25%. Substrate percent moisture was measured using a Theta moisture probe inserted from the substrate surface in 3 and 6 d flood treatments every other day during the draining period.
Flooding treatments were initiated when all plants within a taxon had produced approximately four new nodes (following placement in the greenhouse). Due to differences in growth rates among taxa, treatment initiation dates varied among taxa. For I. virginica ‘Henry's Garnet’ treatments were initiated March 18, 2009, (155 DAP) and ended May 19, 2009 (217 DAP). For I. glabra ‘Shamrock’ treatments were initiated April 2, 2009, (170 DAP) and ended June 4, 2009 (233 DAP). For I. virginica ‘Henry's Garnet’ and I. glabra ‘Shamrock’, treatments were initiated on the same day for both substrates. Due to growth rate differences between the two substrates in V. nudum ‘Winterthur’, treatments were initiated in PB March 24, 2009, (161 DAP) and ended May 26, 2009 (224 DAP); for plants in soil, treatments were initiated April 8, 2009, (176 DAP) and ended June 10, 2009 (239 DAP).
Plants (all taxa) in the 3 and 6 d flood treatments experienced a total of 7 and 5 flood cycles, respectively. Three plants per treatment per substrate were harvested when greenhouse flooding treatments were discontinued, and shoot dry weight (SDW) [shoots removed from the root ball and dried at 68.3C (155F) for 48 h] and root dry weight (RDW) (root ball rinsed to remove substrate and dried at 68.3C for 48 h) were determined. Due to lack of plant material, RDW and SDW were not recorded for V. nudum ‘Winterthur’ plants in soil. Plant size index [(SI) (height + widest width + width perpendicular to widest width) / 3] was recorded at planting and at treatment termination. Each taxon was treated as a separate experiment, and the experimental design was a split plot design with substrate as the main plot and flooding treatments randomized within each substrate in a randomized complete block design. Data were analyzed using contrast statements in the Glimmix procedure with least square means (P < 0.05) (33).
Outdoor flooding. Plants used in the greenhouse flooding experiment were held in the same greenhouse for 40 d after flooding treatments ended. During this period, plants were irrigated with 300 mL (10 oz) to 500 mL (17 oz) tap water when substrate percent moisture reached 25%. Plants were pruned within each taxon × substrate × treatment combination 40 d after greenhouse flooding treatments ended. The purpose of pruning was to create a uniform size within each taxon × substrate × treatment combination while preserving growth differences that occurred during greenhouse flood-ing. Final size index was recorded for all plants after pruning before outdoor flooding treatments were initiated.
Plants from the greenhouse flooding experiment were planted outdoors into [171 liter (45 gal)] [93 × 53 × 50 cm (36 × 21 × 20 in)] plastic tubs (Sterilite Corporation, Townsend, MA) in a pine bark:peat:perlite (5:3:1, by vol) (PB) or fine textured calcined clay (CC) (Profile Products, Buffalo Grove, IL). This outdoor experiment required too large of a volume of soil to be collected and sterilized, thus, CC was used instead of soil. CC is a mineral substrate that is similar in texture to the sandy loam soil used in the greenhouse flooding experiment (personal observation). Plants grown in PB in the greenhouse were planted in PB outdoors, and plants grown in soil in the greenhouse were planted in CC outdoors. Tubs were modified using basic PVC plumbing parts to include a drain on one end of the tub. A 3.8 cm (1.5 in) screw plug on the outside of the tub was tightened and loosened using a pair of channel lock pliers to enable tub to be flooded or drained.
Tubs were filled with 0.12 m3 (0.16 yd3) substrate to a height of approximately 30 cm (12 in). Each tub contained three plants (same taxon), one each that had been non-flooded, flooded for 3 d, and flooded for 6 d in the greenhouse, arranged randomly within each tub. Plants were planted in each tub in a row approximately 10.2 cm (4 in) from the long edge of the tub and 17.8 cm (7 in) on center from each other and the short end of the tub. Outdoor flooding treatments were randomly assigned to each tub. Tubs were flooded for 0 (non-flooded), 3, or 6 d. Thus, overall experimental flooding treatments were in a factorial combination of greenhouse flooding length × outdoor flooding length for a total of nine flooding treatments. Within each taxon × substrate combination there were 6 tubs per treatment (except V. nudum ‘Winterthur’ in CC which only had 5 tubs). Treatments were initiated for I. glabra ‘Shamrock’ on July 23, 2009, (282 DAP) and ended on 23 September 23, 2009 (344 DAP). Treatments were initiated for I. virginica ‘Henry's Garnet’ on July 7, 2009, (266 DAP) and ended on September 7, 2009 (328 DAP). Treatments were initiated for V. nudum ‘Winterthur’ in PB on July 14, 2009, (273 DAP) and ended on September 14, 2009 (335 DAP). Treatments were initiated for V. nudum ‘Winterthur’ in CC on July 29, 2009, (288 DAP) and ended on September 28, 2009 (349 DAP). Differences in outdoor treatment dates were to allow the same number of days (40) in the greenhouse between the greenhouse flooding treatments and the outdoor flooding treatments.
Tubs were placed under a 7.9 × 4.6 m (26 × 15 ft) shade house. The top of the structure was pitched [3.4 m (11 ft) tall sloping to 1.8 m (6 ft) tall along the short side] and was covered in a double layer of 6 mil white polyethylene plastic to exclude rainfall and 60% woven shade cloth (Cassco Associates, Montgomery, AL). The long sides of the shade house were covered in 60% knitted shade cloth (Cassco Associates, Montgomery, AL). The shade house was oriented north to south with the north and south ends left open to ensure adequate air movement and ventilation. Photosyn-thetic photon flux at the plant canopy was approximately 275 μmol m−2 sec−1 during full sun [(measured using a quantum meter (Model LQM50-3, Spectrum Technologies, Inc., East-Plainfield, IL)]. Within each taxon × substrate combination, tubs were randomly arranged in two rows of nine tubs placed side by side.
Substrate percent moisture was measured every 6 h using two ECH2O soil moisture sensors (model EC-5) (Decagon Devices, Inc., Pullman, WA) per tub in two tubs per treatment within each taxon × substrate combination. Sensors were installed in the substrate approximately 12 cm (4.7 in) deep on opposite sides of a tub and centered between the end and middle plant approximately 10 cm (4 in) from the side of the tub. Irrigation (tap water) was applied by hand using a hose and watering wand with a flow rate of 26.5 liters·min−1 (7 gal·min−1). Each irrigation application was timed, and the volume of water applied was calculated. Non-flooded tubs were irrigated with 20 liters (5.3 gal) tap water when substrate percent moisture reached 20%. At 25% percent moisture, substrates still appeared to be wet in lower depths of the tub, which was likely due to the large volume of substrate within each tub. Tubs were flooded to the substrate level by adding 26.5 liter (7 gal) initially and by adding 1.5 liter (0.4 gal) on each additional flood day to maintain the water table at substrate level. Net photosynthesis (Pn) and stomatal conductance (SC) were measured using a LI-COR 6400 (Model 1000, LI-COR Biosciences, Inc., Lincoln, NE) on the last day of flooding and draining during an intermediate and final flood cycle of both flooding treatments. LI-COR 6400 was set to use ambient temperature and humidity, a reference carbon dioxide of 400 ppm, a stomatal ratio of 0, and a flow rate of 500 μmol·s−1. Net photosynthesis and SC were measured on all plants within a tub in three tubs of each taxon × substrate × treatment combination. Net photosynthesis and SC of non-flooded plants were measured concurrently with flooded plants. Leaves of I. glabra ‘Shamrock’ had to be removed from the plant and immediately placed in the cuvette for measuring due to short length of petiole. Following each Pn and SC measurement for I. glabra ‘Shamrock’, the leaf was sealed in a ziploc bag and immediately transported to a lab where leaf area was measured using a LI-COR 3100 leaf area meter (LI-COR Biosciences, Inc., Lincoln, NE). Net photosynthesis and SC values for I. glabra ‘Shamrock’ were reconfigured based on exact leaf areas using LI-COR Simulator software (19) in file exchange mode.
Plants in the 3 and 6 d flood treatments experienced a total of 7 and 5 flood cycles, respectively. Shoot dry weight and RDW were recorded at experiment termination. Final SI was recorded at experiment termination. Relative size index [(RSI) (Final SI – Initial SI) / Initial SI)] was calculated for each plant to determine the change in plant size that occurred outdoors in tubs with respect to plant size at the beginning of the outdoor phase. Each taxon was analyzed as a separate experiment. The experimental design was a split-split plot design with substrate as the main plot and outdoor flooding treatments as the subplot with greenhouse flooding treatments randomized within each tub. Data were analyzed using the Glimmix procedure with least square means (P < 0.05) (33). If significant interactions were present, then data are presented accordingly. When treatment interactions are not significant, main effects are presented. Data are not presented if not statistically different.
Results and Discussion
Greenhouse flooding. In PB, Ilex glabra ‘Shamrock’ final SI was higher in non-flooded plants than in flooded plants (Table 1); RDW and SDW followed the same pattern in PB (data not shown). In soil, I. glabra ‘Shamrock’ final SI was higher in non-flooded plants than in plants flooded for 6 d, with plants flooded for 3 d being intermediate (Table 1). In PB and soil, Itea virginica ‘Henry's Garnet’ final SI was higher in non-flooded plants than in flooded plants (Table 1); RDW and SDW followed the same pattern in both substrates (data not shown). Lower final SI (Table 1) and RDW and SDW (data not shown) in flooded plants than in non-flooded plants of I. virginica ‘Henry's Garnet’ and I. glabra ‘Shamrock’is consistent with others who observed reduced RDW and SDW in flooded plants (17, 27, 34, 35). In PB and soil, there was no effect of flooding on Viburnum nudum ‘Winterthur’ final SI (40 cm and 21 cm, respectively) or RDW and SDW (data not shown). Viburnum nudum ‘Winterthur’ plants grown in soil were considerably smaller than plants grown in PB, which may have been due to soil compaction in the container. For flooded I. glabra ‘Shamrock’ plants, percent moisture of PB averaged 38% at the end of draining, while percent moisture of soil averaged 19% at the end of draining. For flooded I. virginica ‘Henry's Garnet’ plants, percent moisture of PB averaged 31% at the end of draining, while percent moisture of soil averaged 15% at the end of draining. In flooded V. nudum ‘Winterthur’ plants, percent moisture of PB averaged 30% at the end of draining, while percent moisture of soil averaged 16% at the end of draining. During greenhouse flooding, plants in soil experienced drier conditions during draining compared to plants in PB since soil dried down more quickly than PB. It can be assumed that soil in a container does not behave in the same manner as it might in the field with respect to water holding capacity, draining, and structure. Using soil in containers is not recommended for future rain garden studies since once dry, soil shrank from sides of containers and was difficult to re-wet.
Outdoor flooding. In PB and CC, I. glabra ‘Shamrock’ relative size index (RSI) (0.4 and 0.2 cm, respectively, averaged over flooding treatments) was not affected by flooding treatments in the greenhouse or outdoors; SDW and RDW decreased with increasing greenhouse flooding length (Table 2). In PB, I. virginica ‘Henry's Garnet’ RSI was higher in plants that were flooded for 3 or 6 d in the greenhouse than plants that were non-flooded in the greenhouse (Table 3). In CC, RSI was basically similar in all flooding treatments (Table 3). In PB, I. virginica ‘Henry's Garnet’ SDW and RDW were lowest in plants that were flooded for 6 d in the greenhouse (Table 3). In CC, there was no clear effect of flooding treatment on I. virginica ‘Henry's Garnet’ SDW and RDW (Table 3). In both substrates, I. virginica ‘Henry's Garnet’ Pn and SC were generally higher at the end of the experiment than midway through outdoor flooding (Table 4). In both substrates, there was no clear effect of greenhouse or outdoor flooding on V. nudum ‘Winterthur’ RSI, SDW, or RDW (Table 5). For flooded I. glabra ‘Shamrock’ plants, percent moisture of PB averaged 37% at the end of draining, while percent moisture of CC averaged 22% at the end of draining. For flooded I. virginica ‘Henry's Garnet’ plants, percent moisture of PB averaged 37% at the end of draining, while percent moisture of CC averaged 22% at the end of draining. In PB, V. nudum ‘Winterthur’ Pn was higher at the end of the experiment than midway through outdoor flooding (Table 6). For flooded V. nudum ‘Winterthur’ plants percent moisture of PB averaged 32% at the end of draining, while percent moisture of CC averaged 23% at the end of draining.
Itea virginica ‘Henry's Garnet’ appeared to be more affected by flooding in the greenhouse (Table 1) than by flooding outdoors (Table 3). During outdoor flooding, I. virginica ‘Henry's Garnet’ seemed to recover from greenhouse flooding, with few differences in final plant size, possibly due to a fast growth rate. Itea virginica is noted by Dirr (6) to have a relatively fast growth rate when supplied adequate moisture. This is consistent with Bailey (4) who found that the fast growth rate of I. virginica ‘Henry's Garnet’ made differences among irrigation treatments difficult to observe. Flooding treatment did not appear to have any consistent effect on growth of V. nudum ‘Winterthur’ (Table 5). Relative size index of I. glabra ‘Shamrock’ was not affected by flooding treatments applied in the greenhouse or outdoors (data not shown). Ilex glabra is considered slow growing, and the cultivar ‘Shamrock’ has a slower rate of growth than other cultivars of this species (6) perhaps also making treatment differences more difficult to distinguish. Flushes of new shoot growth in I. glabra ‘Shamrock’ were observed in all treatments by the end of flooding in the greenhouse and outdoor flooding, but root systems of flooded I. glabra ‘Shamrock’ were considerably smaller in comparison to non-flooded plants (data not shown). It is not known what the long-term impacts of repeated flooding at this length and frequency might be, though the conditions in this experiment represent more severe flooding than might normally be expected in a typical rain garden. Root systems of I. virginica ‘Henry's Garnet’ and V. nudum ‘Winterthur’ appeared more healthy and robust than those of I. glabra ‘Shamrock’, and this was reflected in results for RDW (Tables 2–3, 5). Ilex glabra ‘Shamrock’ and I. virginica ‘Henry's Garnet’ that were flooded in the greenhouse had less root growth going into outdoor flooding compared to non-flooded plants (data not shown).
Tolerance to flooded conditions can be associated with maintaining adequate rates of photosynthesis and stomatal conductance (3, 11). Plants intolerant to flooding often show a reduction in shoot growth (17, 32, 34), severe leaf defoliation (30, 31), and stomatal closure following 1 to 2 d of soil flooding (16). Luo et al. (21) concluded that it is imperative for plants under variable hydrologic regimes to acclimate their photosynthetic patterns. Flood tolerant species may close stomata at the onset of flooding, but stomata may re-open, which has been associated with the formation of adventitious roots (16). Net photosynthesis and SC of I. virginica ‘Henry's Garnet’ and V. nudum ‘Winterthur’ were higher in the final flooding and draining cycles (Table 4, 6) which is consistent with others who found higher Pn and SC rates during later flood cycles of intermittently flooded plants that were considered to be flood tolerant (2, 29).
Adventitious root growth was observed on lower stem portions at the soil line of I. virginica ‘Henry's Garnet’ during outdoor flooding. The formation of adventitious roots has been noted to contribute to re-opening of stomata (15, 23), which was speculated to result in higher rates of Pn and SC later in the flood research done by Pezeshki and Anderson (29). Adventitious roots may also develop at the shoot base in aerated soil layers (5) as a result of flooding and signal damage to a plants original root system (24, 32).
Plants having shallow root systems are less likely to experience severe anoxic conditions associated with lower soil layers due to root growth occurring in the upper rhizo-sphere (5, 25). In this research, during outdoor flooding, new root growth occurred in flooded plants of all taxa and was consistently found in the upper portion of the substrate (visual observation) which was likely to be partially aerated. Surface root proliferation is common in flood tolerant plants and allows roots to function in a hypoxic state where they can undergo partial aerobic respiration (25).
Evaluating flood tolerance in a greenhouse allows the control of flooding depth but usually includes some limitations such as reduced volumes of soil or substrate (22) associated with container sizes and space constraints. Using larger tubs outdoors seemed to more closely resemble conditions in a rain garden or bioretention cell. Mesocosms (large tubs) have been used extensively to mimic field conditions for bioretention cells (20) and other wetland and aquatic studies (1). The current study as conducted was better for evaluating tolerance to flooding rather than drought. Substrate percent moisture following draining averaged 22% for all taxa, which did not impose drought conditions since control tubs were watered at 20%. Research conducted on I. virginica ‘Henry's Garnet’ found that plants were still visually acceptable when surrounding soil percent moisture was allowed to reach 15% before irrigating (4).
The taxa evaluated in this study naturally occur in wet-lands, and thus, evaluating their performance under drought conditions would be an important final step to determining their suitability for use in rain gardens and bioretention areas. Calcined clay is a moderately well-drained mineral substrate with properties similar to that of a sandy substrate. In previous repeated flooding experiments conducted in 2.5 liter (#1) containers in a greenhouse, CC dried down more so than in the current experiment (10). The large volume of substrate in the tubs outdoors did not dry down enough to expose plants to drought between flood events. Future research may impose more severe drought stress during draining periods on facultative wetland plants to further evaluate plants for standard rain gardens and bioretention areas. Additionally, the use of a substrate with decreased water holding capacity may aid in evaluating effects of drought when screening plants for these practices (10).
Bioretention substrate recommended is a mixture of mineral and organic substrates that consists of 85–88% sand, 8–12% fines, and 3–5% organic matter (13), but rain gardens often use native soil, organic matter, or some combination of the two. Both a mineral and organic substrate were used in this study to evaluate plants for rain gardens and bioretention cells. Using native field soil in containers makes experiments difficult due to shrink-swell behavior and hydrophobicity during drying. Thus, these authors recommend using CC to approximate mineral substrate conditions in containers (large and small).
For taxa evaluated in this study, only I. virginica ‘Henry's Garnet’ appeared to benefit from prior flooding when exposed to subsequent repeated short-term flooding. Although plant size of I. glabra ‘Shamrock’ and I. virginica ‘Henry's Garnet’ decreased with increasing flood length in the greenhouse, subsequent relative growth rates outdoors did not follow the same trend. All taxa evaluated in this study appear to be tolerant of repeated short-term flooding as would be expected to occur in a rain garden or a bioretention cell. It is suggested that plants for rain gardens be able to withstand 2 d of standing water (9), however, this is dependent on soil characteristics, amount of precipitation received, and the size and depth of the rain garden (12). Bioretention cells are designed to have standing water for a maximum of 12 h, but should drain to 61 cm (24 in) below the media surface in 48 h (26). Therefore, because rain gardens rely on the use of native soils, ponding depth and time is more varied compared to the consistency of bioretention cell media. Thus, tolerance of flooding may be more pertinent to slower draining rain gardens or wet rain gardens, while tolerance of cyclic flooding and drought conditions may be more conducive to screening plants for bioretention areas. Results of this study indicate that although V. nudum ‘Winterthur’ may be more sensitive to flooding than the other two taxa, all taxa in all treatments and substrates maintained good plant visual quality and new growth over the course of the experiment, which suggests tolerance to repeated short-term flooding and their suitability for use in rain gardens and bioretention areas.
The authors thank J. Raymond Kessler for statistical assistance.
2Former Graduate Research Assistant. firstname.lastname@example.org. Current address: Department of Agronomy and Soils, Auburn University, AL 36849.
3Associate Professor. email@example.com.
5Associate Professor, Department of Landscape Architecture, Auburn University, AL 36849