Substrate preparation.

Aged pine bark (Pinus taeda; Phillips Bark Processing Co., Brookhaven, MS, USA) was limed with 1.77 kg⋅m⁻³ of dolomitic lime (Lime-Rite Pelletized Dolomitic Lime, Roswell, GA, USA) and 0.89 kg⋅m⁻³ of micronutrients (Micromax G90505; ICL Specialty Fertilizer, Dublin, OH, USA). A commercial floriculture substrate (50% Canadian Sphagnum Peat Moss, Aged Pine Bark, Perlite, and Vermiculate; Jolly Gardner Pro-Line C/P, Atlanta, GA, USA) was also procured. Porometer analysis was conducted to measure substrate static physical properties [air space (AS), container capacity (CC), total porosity (TP), and bulk density (D_b)] as described by Fonteno and Harden (2010). Particle size distribution was measured for both the commercial mix and pine bark by separating 100 g of dried bark particles through a nest of sieves. The sieve apertures included 6.3-, 2.0-, 0.7-, 0.5-, 0.3-, and 0.1-mm sifters with a base pan and was agitated for 5 min with a Ro-Tap shaker (Rx-29; W.S. Tyler, Mentor, OH, USA).

Experimental design.

Ninety-six 2.37-L blow-molded containers (C300S; Nursery Supplies, Kissimmee, FL, USA) were filled with either a nonstratified control, which consisted of filling a container uniformly with commercial floriculture substrate, or a stratified substrate. Stratified substrate containers were filled to 50% of the container height (8.25 cm) with aged pine bark and agitated to settle particles; thereafter, the floriculture substrate was layered atop and filled to the inner container lip. Custom cut elbow tensiometers (Soil Measurement Systems, Huntington Beach, CA, USA) were installed in both the upper and lower strata (25% and 75% of the height of the container, 10.17 cm from the container wall) of 12 individual replicates (six nonstratified and six stratified replicates) so that the ceramic tips were in the center of the container profile. Water potential values depicting these tensiometers in the upper and lower strata will be further represented by ψ_upper and ψ_lower, respectively.

A Petunia hybrid ‘Supertunia Honey’ plug was transplanted into each of the containers on 23 Feb 2022. Containers were randomly placed on greenhouse benches into one of 12 irrigation zones (six replicate irrigation zones for nonstratified and six zones for stratified), so that one container from each zone was fitted with two tensiometers. A separate container (one per irrigation zone) was placed on a lysimeter system constructed out of a cantilever load cell (ESP-10; Transducer Techniques, Temecula, CA, USA) and an acrylic cut platform. The tensiometers and lysimeters were connected to a Campbell Scientific Data Logger (CR1000X; Campbell Scientific, Logan, UT, USA) and multiplexer (AM16/32B, Campbell Scientific), which measured and recorded hourly averages.

Thus, there were 12 total irrigation zones (six nonstratified and six stratified), each with eight containers, in which one was affixed with tensiometers, and another was placed atop a lysimeter. Each container was affixed with a 12.11-L⋅h⁻¹ spray stake (Plum; Netafim, Fresno, CA, USA). Crops were irrigated 600 mL every day for 8 d. Thereafter, crops were irrigated 400 mL every 3 d for 19 d to allow roots to establish growth into the substrate. Afterward, irrigation scheduling treatments were initiated. Wherein, three irrigation zones of each substrate were shifted to either regular control irrigation schedule that irrigated the crops 400 mL every 3 d, or a deficit irrigation schedule that irrigated 400 mL every 5 d. This schedule continued for 12 d after irrigation initiation, at which time, unexpectedly warm temperatures resulted in a need to shift in irrigation schedules to every 2 or 3 d for the control and deficit, respectively. All plants were hand fertigated every other week (600 mL) with 200 ppm N liquid fertilizer (20N–20P₂O₅–20K₂O; Peters Professional Fertilizer, Summerville, SC, USA).

Monitoring substrate moisture status.

The lysimetry data were used to calculate the substrate system’s volumetric water content (Θ). This was conducted on culmination of the study, where the substrates placed on lysimetry systems were dried for 48 h at 105 °C. The substrates were then weighed to attain a dry weight, and the substrate dry weight + container weight was subtracted from the lysimetry data. Shoot fresh weight was measured five separate times (one container replicate within each zone) throughout the study. Within each treatment, shoot fresh weight was plotted against days after irrigation initiation of the study and a regression line was fitted using JMP Pro (16.1.0; SAS Institute, Inc., Cary, NC, USA). The equation attained for nonstratified regular irrigation, nonstratified deficit irrigation, stratified regular irrigation, and stratified deficit irrigation (Eqs. [1] to [4], respectively) was used to predict shoot fresh weight each day of the study and the predicted fresh shoot weight was subtracted from the lysimetry data to attain more precise Θ calculations.$\text{Fresh shoot biomass}\hspace{0.17em}=\hspace{0.17em}-\text{3.034}\hspace{0.17em}+\hspace{0.17em}\text{1.709}\hspace{0.17em}\times \hspace{0.17em}(\text{days after transplant})$[1] $\text{Fresh shoot biomass}\hspace{0.17em}=\hspace{0.17em}-\text{10.499}\hspace{0.17em}+\hspace{0.17em}\text{1.708}\hspace{0.17em}\times \hspace{0.17em}(\text{days after transplant})$[2] $\text{Fresh shoot biomass}\hspace{0.17em}=\hspace{0.17em}\text{1.619}\hspace{0.17em}+\hspace{0.17em}\text{1.025}\hspace{0.17em}\times \hspace{0.17em}(\text{days after transplant})$[3] $\text{Fresh shoot biomass}\hspace{0.17em}=\hspace{0.17em}-\text{3.586}\hspace{0.17em}+\hspace{0.17em}\text{1.061}\hspace{0.17em}\times \hspace{0.17em}(\text{days after transplant})$[4]

Harvests.

A total of four harvests occurred throughout the study starting 8 d after irrigation initiation. Each additional harvest occurred every subsequent 15 ± 1 d. Culmination of the study was 50 d after irrigation initiation (10 May 2022). Every harvest consisted of randomly selecting a single plant per zone. For each harvest, representative crop photos were captured, along with plant growth index [(plant height + width + perpendicular width) ÷ 3], flower count, chlorophyll content (SPAD 502 Plus; Spectrum Technologies, Inc., Aurora, IL, USA), shoot fresh weight, and root morphological traits were measured. Shoot and root dry weight were measured by drying the plant samples at 70 °C for 7 d. For the final harvest, all substrate root systems were partitioned in half (at the stratified interface; ∼8.25 cm) and dry root biomass of the top and lower half of the container was measured.

RhizoVision.

Root system morphology was assessed at each harvest event. The total root length, the quantity of root tips and branching points, average root diameter, and total root surface area were estimated via RhizoVision software (Seethepalli et al. 2021). In preparation for root imaging analysis, roots were cleaned completely from substrate debris and segmented to fit in a square container (8244 cm³) on standard white printer paper (21.6 cm × 27.9 cm). The container was filled with ∼400 mL of deionized water to spread apart individual roots. Once the roots were laid carefully and separated clearly on the paper, a photo was taken with an iPhone XR (Apple, Cupertino, CA, USA). To ensure each photo was taken at identical heights in darkness, a light inhibiting root imaging chamber apparatus was constructed with a hole at the top for camera placement.

Root images were scanned with TurboScan (Piksoft Inc, Piedmont, CA, USA) and converted to a PDF format. This was required because a flash from the photo created a glare in the image and RhizoVision was not able to correctly analyze the photo. Thereafter, images were cropped to the paper dimensions. The new files were uploaded to the RhizoVision software, and the dots per inch were calibrated to 326 in RhizoVision (iPhone XR camera specifications). Root systems were larger than the square container used; thus, roots were divided into segments and analyzed. Therefore, measured output from RhizoVision required summing the values within each sample besides average diameter.

Data analysis.

The data presented in tables and figures with associated statistics were analyzed in JMP Pro (16.1.0; SAS Institute, Inc., Raleigh, NC, USA) using analysis of variance (ANOVA) to identify any significant statistical differences across the means of the responses previously described. If significant, post hoc Tukey’s honestly significant difference test (α = 0.05) was used to separate means across two substrates and crop growth performance among irrigation schedules, such as growth index, flower count, chlorophyll content, dry weights, and root growth (total length, number of root tips, number of branching points, average diameter, and total surface area).

All tensiometer and lysimetry data were analyzed in JMP Pro (16.1.0). Weekly mean substrate matric potential values were calculated in the upper and lower strata and Θ. In addition, an ANOVA was used to identify any significant statistical differences in means of matric and Θ values within irrigation schedules across substrate treatments. Pearson’s pairwise correlation was used to determine the strength of these linear relationships.

Substrate physical properties.

There are unique relationships between substrate particle diameter classifications and their respective physical and hydraulic properties (Nkongolo and Caron 1999). For peat-based substrates, 60% of the particle diameters were < 2.0 mm, whereas 70% of the bark substrate particle diameters were > 2.0 mm (Table 1; Fig. 1). Decreasing substrate particle diameters generally exhibits negative and positive relationships to substrate water and air storage capacities, respectively (Kiran et al. 2021). Due to capillary pore proportions (Puustjarvi and Robertson 1975), large particle substrates such as bark have greater pore sizes (Drzal et al. 1999), which increases substrate drainage (Hoskins et al. 2014) in comparison with other soilless substrates. The bark media retained significantly less water (P = 0.0002) than the peat-based media; however, peat substrates are inherently fibrous and have increased microporosity, which resulted in lower air-filled porosity values than bark (P = 0.0002; Table 1). Altland et al. (2011) and Fields et al. (2021b) described that manipulating substrate particle diameters will often change AS/CC balances, while having little to no effect on substrate TP values as observed herein (Table 1).

Table 1. Static physical properties and particle size distribution of two substrates. Components include an unscreened conventional pine bark and a peat-based (50% Canadian Sphagnum Peat Moss, Aged Pine Bark, Perlite, and Vermiculate) substrate.

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Evaluating crop production.

On study initiation, all crops were equivalent in size in both the control (P = 0.7046) and deficit (P = 0.2842) irrigation schedules (Fig. 2). Plant shoot growth increased similarly across treatments throughout the study, where midway through production, it appears that crops grown in nonstratified systems were slightly larger under both irrigation treatments (Fig. 2). However, there were no detectable differences in the final crop growth index on study culmination under either irrigation regimen or substrate system (Fig. 2). Plants grown with more frequent irrigation events had a greater final growth index than plants grown with less frequent irrigation (P = 0.0291), regardless of substrate system (P = 0.7607; Fig. 2). Crop growth in a nonstratified system irrigated regularly appeared to stop growing in the final two growth index measurements, whereas crops grown in a stratified system irrigated regularly continued to grow. Criscione et al. (2022a) received similar results, although in a deficit irrigation, traditionally grown crops near the end of the study began to decrease in growth while all stratified treatments continued to outperform crops grown in nonstratified systems. There was a significant main effect on shoot biomass for both substrate (P = 0.0025) and irrigation (P = 0.0051), where crops grown in stratified substrate systems under deficit irrigations weighed less than all other treatment combinations in the final harvest (Table 2). Nevertheless, despite statistical differences in shoot biomass, there were few differences in crop marketability and growth, providing insight that growers can possibly shift their substrate systems and still produce salable crops (Fig. 3).

View Figure

• Download as Powerpoint
• Download Figure

Table 2. Plant tissue measurements of crops grown in two different substrate treatments (nonstratified and stratified) irrigated under two different schedules (regular and deficit). Tissue measurements included shoot and root biomass, total root length, quantity of root tips and branching points, average root diameter, and total root surface area.

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Crop chlorophyll count was greater in both deficit treatments on 58 d after production (Fig. 4). Thereafter, chlorophyll count remained relatively similar across treatments throughout the remainder of the study (Fig. 4). This was not hypothesized considering drought stress typically reduces chlorophyll count (Yang et al. 2021). Plants grown in nonstratified substrate systems consistently exhibited more blooms and had greater total flower count than crops grown in stratified media (Fig. 4). This was detected in both the control (P = 0.0300) and deficit (P = 0.0600) irrigation schedules (Fig. 4). Drought or water-stressed conditions have been shown to induce reproductive and physiological traits, such as flowering in some fruit crops (Wu et al. 2017) and in petunias (Hatamifar and Samani 2017). In addition, fine roots, which are responsible for water and mineral nutrient absorption (McCormack et al. 2015), have been observed in some container crops to contain greater concentrations in the shallowest substrate depths (Bauerle et al. 2013).

View Figure

• Download as Powerpoint
• Download Figure

Greater flower count in nonstratified-grown crops may be attributed to several factors. Stratified root production visually appears to be greater in the upper strata and takes longer to explore the lower strata than that of crops grown in nonstratified systems (Fig. 5). Moreover, nonstratified substrate systems, regardless of the type of media used, contain a moisture content depth profile, resulting in drier upper layers and wetter lower layers (Owen and Altland 2008). Although not measured directly, it is possible that in the upper layers stratified roots contained greater quantities of fine roots, which enabled better water absorption even under stressful conditions (Cuneo et al. 2021), and resulted in less water-stress-induced flowering (Figs. 4 and 5). Thus, crops grown in stratified substrate systems possibly reduced water-stress symptoms (i.e., increased flowering) as observed in nonstratified-grown plants.

View Figure

• Download as Powerpoint
• Download Figure

Root growth and development.

Root morphological characteristics were estimated through RhizoVision software (Seethepalli et al. 2021) and are displayed in Table 2. There were no treatment effects for any morphological trait until the third crop harvest, where irrigation had strong effects on total root length (P = 0.0334), branch point count (P = 0.0175), and total root surface area (P = 0.0258; Table 2). In the final harvest, there were several statistical differences across treatments. Crops growing in a stratified substrate under the control irrigation contained greater root morphology characteristic values than the nonstratified system in total root length (P = 0.0056) and root tip count (P = 0.0398; Table 2). Balliu et al. (2021) described that increasing irrigation frequency generally promotes root development and health. Crops grown under the control irrigation, regardless of substrate system, typically contained more morphologically defined root traits than those grown under irrigated deficit (Table 2). The lack of differences in average root diameter may indicate that any increase in root mass was driven by increased root lengths (Table 2).

Illustrated in Fig. 6A are rhizosphere profiles partitioned in half, demonstrating that crops growing in stratified systems had increased lower root biomass under both regular (P = 0.0061) and deficit (P < 0.0001) irrigation schedules than nonstratified-grown crops. Savvas and Gruda (2018) stated that substrates that contain greater aeration in the root zone typically result in improved root growth. Moreover, the substrate moisture content depth distribution influences root spatial distribution (Wallach 2008), where root growth typically mimics the substrate moisture profile (Ismail and Ozawa 2009). Considering the stratified system was developed to retain more moisture in the upper strata, it is likely that in the nonstratified system, moisture infiltrated into the lower proportions more readily due to gravitational effect (Owen and Altland 2008). This would subsequently increase aeration and gas diffusivity at the container base of the stratified substrates (Cannavo and Michel 2013); therefore, root growth likely was encouraged in the substrate with increased aeration, resulting in greater biomass (Fig. 6A) and total root length (Table 2; Fig. 6B). Root growth appeared to have occurred at different spatial rates in substrate systems, where stratified-grown crops seemed to have explored the upper portion of the profile longer than nonstratified-grown roots (Fig. 5). However, once stratified roots explored into the lower half, root growth flourished (Figs. 5 and 6).

View Figure

• Download as Powerpoint
• Download Figure

Particle and pore distribution play fundamental roles in water redistribution, hydraulic conductivity, and water retention and availability for plants (Hille 2004). Improved soilless substrate unsaturated hydraulic conductivity has been shown to improve plant vigor and water acquisition, especially in deficit conditions (Fields et al. 2017). Fields et al. (2018) stated that finer materials such as sphagnum peat have greater unsaturated hydraulic conductivity values than larger substrates like bark. Moreover, Bartley et al. (2022) reviewed particle densities of horticultural substrates and concluded that peat-based substrates have greater particle densities than bark. Although nonstratified systems contained more peat-based substrate (by volume), the container base may have experienced oxygen deficit conditions (Cannavo and Michel 2013). Thus, the greater particle density and improved unsaturated hydraulic conductivity in the upper strata helped promote more root growth under deficit conditions in stratified systems, whereas the decrease particle density and less water storage in the lower strata (e.g., bark) promoted more root growth (Fig. 6A).

Monitoring substrate hydraulic properties.

Substrate profile water potential (ψ) and Θ measurements were averaged weekly after irrigation scheduling began and are plotted in Figs. 7 and 8. There were no statistical differences detected across substrate treatments within either irrigation schedule for ψ_upper and ψ_lower (Fig. 7); however, there were differences detected among Θ across each substrate system and irrigation schedule (Fig. 8). Substrate as a main effect had no influence, but irrigation scheduling had an effect on Θ values for most of the study, where substrates irrigated more frequently had greater Θ (Fig. 8).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Nonetheless, there were moderate correlations between substrate hydraulic measurements and plant tissue traits. To expand, root biomass was more correlated with Θ than shoot biomass (r = −0.6370 and −0.4856, respectively; Table 3), which can be attributed to root water uptake in the substrate. In addition, root biomass was more strongly correlated with ψ_lower than ψ_upper (r = −0.3199 and −0.1815, respectively; Table 3). These results align with the data depicted in Fig. 6A, where all crop roots had significantly greater root biomass in the lower container portion than upper portions except for nonstratified-grown crops under regular irrigation. Although not measured directly, roots growing in peat-based substrates may have contained greater fine root proportions, whereas roots growing in bark substrates may have contained larger roots due to the larger substrate pore diameters. Thus, the greater root biomass in the lower stratum may be attributed to larger roots. These results contrast the results observed by Fields et al. (2022), where there were greater root biomass values measured in the upper strata than lower. However, the system used fine bark particles layered over coarse bark; thus, there were different pore diameters and distributions (peat-based substrate vs. screened fine bark particles; unscreened bark vs. screened coarse bark).

Table 3.Pairwise Pearson correlations between plant tissue measurements (shoot and root biomass, total root length, quantity of root tips and branching points, average root diameter, and total root surface area) and substrate hydraulic measurements (upper and lower matric potential values and volumetric water content), regardless of treatment. Values were measured from rooted containers. Crops were grown in one of two experimental substrate treatments [nonstratified (100% floriculture mix) and stratified (50% floriculture mix layered atop conventional bark)] and irrigated under two different irrigation schedules [regular (400 mL/2 d) and deficit (400 mL/3 d)].

View Fullscreen

Volumetric water content measurements were more strongly correlated with ψ_lower than ψ_upper, further validating moisture content distributions due to gravity in a substrate system (Owen and Altland 2008). Among the root morphology traits, root diameter had the strongest correlation with Θ (r = 0.5936). This may indicate that more fine roots (smaller diameters) present in a system can have greater effect on substrate moisture status (i.e., more fine roots = more water uptake = decreased Θ; McCormack et al. 2015). It is possible that crops responded more to Θ than ψ because of the accuracy of the lysimetry systems vs. the elbow tensiometers, although the literature discusses that plants tend to respond more to ψ than Θ because of the amount of “work” or “energy” required to extract moisture from the substrate matrix (Cassel and Nielsen 1986; Hillel 2004).