Experimental design.

The experiment was a 2 (aggregate particle size) × 2 (mineral temperature pretreatment) factorial in a randomized complete block design with three replications with 10 plants in each replication. The treatments were pine bark-amended with palygorksite–bentonite mineral aggregate from Ochlocknee, GA (Oil-Dri Corp. of America, Chicago) with one of either two particle sizes (0.85 to 4.75 mm or 0.25 to 0.85 mm) that had been pretreated at one of two temperatures (RVM = 121 °C or LVM = 800 °C or less). All clay was added as an 8% (by vol.) substrate amendment. An additional substrate was included in the experimental design to represent the industry standard (control), which contained 11% sand (by volume) (8 pine bark:1 sand). All substrates were amended with 0.6 kg·m⁻³ ground dolomitic limestone [CaMg(CO₃)₂]. The 112-d experiment was conducted from 16 May 2002 to 5 Sept. 2002 at the Horticulture Field Laboratory (lat. 35°47′37″N; long. 78°41′59″W) located at North Carolina State University, Raleigh. Research plots received 41.2 cm of rain (average of 1.5 cm rainfall per event) for a total of 27 d.

Water and nutrient management.

One hundred fifty rooted stem cuttings of Skogholm cotoneaster (Cotoneaster dammeri C.K. Schneid. ‘Skogholm’) liners with a 212-cm (5.4 cm × 5.4 cm × 7.3 cm) root ball were potted into 14-L containers (C-2000; Nursery Supplies, Chambersburg, PA) and placed on 15 separate plots, 10 containers per plot, which were designed to enable collection of all leachate leaving each plot. Leachate was unable to be collected when rainfall exceeded 0.8 cm. Plots were 8 × 1 m with a 2% slope. Containers were top-dressed with 60 g of 17N–2.2P–9.6K fertilizer (5- to 6-month 17N–5P₂O₅–12K₂O Fast Start CRF with micronutrients; Harrell's Fertilizer Co., Lakeland, FL) that was hand-incorporated into the top 2 cm of the substrate at the time of treatment initiation (day 0). Irrigation water was applied to each container through pressure-compensated spray stakes [Acu-Spray Stick (200 mL·min⁻¹); Wade Mfg. Co., Fresno, CA]. Irrigation was applied in a cyclic manner with the irrigation volume divided equally among three applications applied at 1100, 1400, and 1700 hr eastern standard time (EST). Irrigation volume to maintain a target leaching fraction (LF, Eq.[1]) of 0.2 was applied to each container based on effluent values monitored daily and influent volumes that were measured biweekly.

View Expanded

These data were used to determine water volume applied, water volume retained, and water use as affected by each treatment.

Effluent and substrate handling.

The amount of effluent from each plot was measured daily and an aliquot of the effluent was analyzed colorimetrically using a ultraviolet-visible spectrophotometer (Spectronic 1001 Plus; Milton Roy Co., Rochester, NY) for NH₄-N (Chaney and Marbach, 1962), NO₃-N (Calado et al., 1975), and dissolved reactive P (DRP) (Murphy and Riley, 1962).

Because there are no standard methods for mineral nutrient extraction from organic substrates, two extractants were used: deionized water and 0.1M KCl. Substrate total N (NH₄-N + NO₃-N) and DRP were extracted from the substrate using a 1 substrate:1.5 extract (115 cm³ substrate:175 mL extractant) (Sonneveld et al., 1974). The substrate dilution was shaken for 1 h and filtered through 0.43-μm filter paper (Whatman No. 1; Whatman, Florham Park, NJ). An aliquot of the filtered solution was then analyzed colorimetrically for NH₄-N, NO₃-N, and DRP as described previously.

In addition, macro- and micronutrients, cation exchange capacity (CEC), and exchangeable acidity [Al (aluminum), H (hydrogen)] were determined for each substrate on a volume basis (Tucker, 1984) using crushed (mortared and pestled) substrate samples dried at 60 °C. The North Carolina Department of Agriculture (NCDA), Agronomic Division, Raleigh, performed standard analytical methods using Mehlich III extractant (Mehlich, 1984a), photometric determination (Mehlich, 1984b), and Mehlich buffer method (Mehlich et al., 1976).

Ammonium-N, NO₃-N, and DRP remaining in the fertilizer prills at the end of the study were measured as follows. Nutrients were extracted from the fertilizer prills by blending the prills in 200 mL deionized water. After blending, the liquid was transferred quantitatively to a 1-L volumetric flask that was adjusted to volume before removing an aliquot of the extractant supernatant. Ammonium-N, NO₃-N, and DRP in fertilizer prill extracts were quantified using the spectrophotometer as described previously.

Crop response.

At 112 d after treatment initiation (DAI), two containerized plants were randomly chosen per plot (total of six plants per treatment) and removed. Tops were severed at the substrate surface and roots were placed over a screen and washed with a high-pressure water stream to remove substrate. Tops and roots were dried at 65 °C for 5 d and weighed. Because of their size, tops were ground initially using a Model 4 bench, 1 HP Wiley Mill (Thomas Scientific, Swedesboro, NJ) to pass a 6.0-mm sieve. The material from the tops and roots was further ground separately through a Foss Tecator Cyclotec 1093 sample mill (Analytical Instruments, LLC, Golden Valley, MN) to pass a 0.5-mm sieve. Roots and tops were analyzed for N, P, K, Ca, Mg, S, Fe, Mn, B, and Mo by the Agronomic Division of the (NCDA).

A nutrient budget was developed for each treatment to quantify the fate of recovered nutrients (RN; Eq. [2] from the applied CRF: absorbed by plant, lost in leachate, remaining in substrate (Mehlich III extractable), or remaining in the fertilizer prill.

View Expanded

The nutrient budget was used to calculate nutrient uptake efficiency (NUE, Eq. [3]) of N and P.

View Expanded

To determine if the clay treatments increased the water-buffering capacity of the substrate, irrigation of the remaining plants was discontinued at the end of the study. Leaf gas exchange was measured on one plant from each replication between 1030 to 1130 hr EST and 1530 to 1630 hr EST for 4 d with a portable IR gas exchange system (LI-COR 6200; LI-COR, Lincoln, NE). Data were recorded for 30 s on the terminal 5 cm of an intact stem with ≈5 fully expanded leaves. Measurements were taken under natural light in which photosynthetic active radiation remained greater than 1896 μmol·m⁻²·s⁻¹. Average CO₂ concentration within the chamber was 325 ± 4 se μmol·mol⁻¹ and relative humidity was maintained at 43% ± 1 se. CO₂ exchange rate (CER) and g _S were calculated using the LI-COR 6200 programs. In addition, dry weight and irrigation data were used to calculate water use efficiency of productivity (WUE_P, Eq. [4]).

View Expanded

Substrate physical properties.

Ten cylindrical aluminum cores, five 347.5 cm³ and five 100 cm³, were placed in fallow containers of each substrate. These containers were placed adjacent to the plants in the research study and received equivalent irrigation and rainfall as the corresponding treatment. After 9 weeks, the 347.5-cm³ cores were extracted and total porosity, container capacity (CC), AW, and air-filled porosity were determined using the NCSU Porometer as described by Fonteno and Bilderback (1993). Unavailable water, held in the substrate at 1.5 MPa or greater, was determined with the 100-cm³ cores through a procedure developed by Milks et al. (1989). Bulk density (D_b) was determined using oven-dried (110 °C) substrate in 347.5-cm³ cores. Particle size distribution of ≈500 cm³ oven-dried substrate (110 °C) was determined gravimetrically using 6.30, 2.00, 0.71, 0.50, 0.25, or 0.106-mm soil sieves. Particles 0.106 mm or less were collected in a pan. Sieves and pan were shaken for 5 min with a RX-29/30 Ro-Tap test sieve shaker (278 oscillations/min⁻¹, 150 taps/min⁻¹) (W.S. Tyler, Mentor, OH).

Data analysis.

All data were subjected to analysis of variance procedures (SAS Institute, 2001) with α ≤ 0.10 to reduce the risk of a type II error (Marini, 1999). Treatment means were compared using orthogonal contrast statements. PROC REG and PROC NLIN were used to further investigate linear and nonlinear segmented trends associated with water and nutrient data. Join points for segmented lines denoted as X_n were determined using PROC NLIN.

Substrate physical properties.

CC and AW were unaffected by mineral aggregate amendments, whereas RVM increased air space 3% (by vol.) compared with LVM (Table 1). However, CC and AW pooled over mineral aggregate amended substrates increased CC 5% (by vol.) and AW 4% (by vol.) compared with the control (sand-amended substrate). This increase represents ≈500 mL of AW per container of aggregate-amended substrate more than the control. Thus, mineral aggregate-amended substrates increased container water-holding capacity without decreasing air space. Increased CC and AW were likely a result of the properties of the mineral aggregate other than particle size, because there were no notable differences in particle size distribution among the sand-amended and mineral aggregate-amended substrates (data not presented). Spomer (1998) hypothesized substrates containing expanded clay have intrapores containing water that is available to the plant, thus increasing water-buffering capacity or the capacity to hold more AW. The mineral used in this study, palygorskite, is a fibrous mineral with zeolitic pores that retain water. In addition, Spomer (1998) reported a substrate amended with clay had increased air space and tortuosity, possibly resulting in greater water retention in the container system. Riviere et al. (1990) also reported a 4% and 42% increase in total porosity and easily available water (available at 10 to 50 cm suction), respectively, when the clay rate increased in a brown peat-based substrate from 15% to 20% (by volume). Warren and Bilderback (1992) found amending pine bark with arcillite (0, 27, 54, 67, or 81 kg·m⁻³) changed pore size distribution, which increased AW curvilinearly and decreased air space linearly but had no affect on total porosity.

Table 1.

Physical properties of substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments.^z

View Table

D_b (mean = 0.21 g cm⁻³ ± 0.001 se) was unaffected by mineral aggregate treatments, but D_b pooled over industrial aggregates was less than the control (mean = 0.30 g cm⁻³ ± 0.001 se) (data not presented). Total porosity (mean = 84% ± 0 se) and unavailable water (mean = 32% ± 0 se) were also unaffected by any of the mineral aggregate treatments nor were they significantly different from the control (data not presented).

Substrate water-buffering capacity.

LF was not different between treatments with an average of 0.22 ± 0.006 se for the entire study (data not presented). However, the sand-amended substrate required 1.5 L·d⁻¹ from 0 to 69 DAI to maintain a targeted LF of 0.2, whereas substrates amended with either aggregate particle size range required an average of 1.25 L·d⁻¹ between 0 and 74 DAI (Fig. 1). From 69 to 112 DAI, irrigation volume required to maintain a LF of 0.2 increased 33% to 2.0 L·d⁻¹ (5.6 mL·min⁻¹) for the control reflecting the increase in plant size resulting from crop establishment and growth. From 74 to 112 DAI, irrigation volume for substrates amended with either aggregate particle size range increased to 1.7 L·d⁻¹. This is a 13% to 15% decrease in daily water application volume. Mineral aggregate substrates, regardless of particle size, required 22 to 26 L less water per container over 112 d compared with the sand-amended substrate to produce equivalent Skogholm cotoneaster growth.

Cumulative irrigation water applied to Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes. Segmented linear models were calculated for 0.25 to 0.85 mm: if x ≤ 74, then y = −0.5 + 1.2x, R ² = 0.99; if x ≥ 74, then y = −37.7 + 1.7x, R ² = 0.98; 0.85 to 4.75 mm: if x ≤ 74, then y = −1.3 + 1.3x, R ² = 0.99; if x ≥ 74, then y = −35.1 + 1.7x, R ² = 0.98; control: if x ≤ 69, then y = −4.3 + 1.5x, R ² = 0.98; if x ≥ 69, then y = −41.3 + 2.0x, R ² = 0.98.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

View Full Size

Cumulative irrigation water applied to Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes. Segmented linear models were calculated for 0.25 to 0.85 mm: if x ≤ 74, then y = −0.5 + 1.2x, R ² = 0.99; if x ≥ 74, then y = −37.7 + 1.7x, R ² = 0.98; 0.85 to 4.75 mm: if x ≤ 74, then y = −1.3 + 1.3x, R ² = 0.99; if x ≥ 74, then y = −35.1 + 1.7x, R ² = 0.98; control: if x ≤ 69, then y = −4.3 + 1.5x, R ² = 0.98; if x ≥ 69, then y = −41.3 + 2.0x, R ² = 0.98.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Cumulative irrigation water applied to Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes. Segmented linear models were calculated for 0.25 to 0.85 mm: if x ≤ 74, then y = −0.5 + 1.2x, R ² = 0.99; if x ≥ 74, then y = −37.7 + 1.7x, R ² = 0.98; 0.85 to 4.75 mm: if x ≤ 74, then y = −1.3 + 1.3x, R ² = 0.99; if x ≥ 74, then y = −35.1 + 1.7x, R ² = 0.98; control: if x ≤ 69, then y = −4.3 + 1.5x, R ² = 0.98; if x ≥ 69, then y = −41.3 + 2.0x, R ² = 0.98.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Chemical buffering capacity.

Mineral aggregate amended substrates decreased NH₄-N and DRP effluent load an average of 39% and 34%, respectively, compared with the control (Figs. 2 and 3). Verhagen (personal communication, RHP Foundation, 2001) found the addition of a calcined clay component into peat-based substrates decreased water extractable P ≈400% and increased NH₄ retention 200% to 300%.

Cumulative effluent ammonium for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments (LVM or RVM). Each container had 5.1 g of NH₄-N surface incorporated at the initiation of the experiment. Segmented linear plateau models were calculated for each treatment: LVM: if x ≤ 7, then y = −7.3 + 9.9x, R² = 0.84; if 7 ≤ x ≤ 33, then y = 62.7 + 0.28x, R² = 0.97; if x ≥ 33, then y = 72.4 + 0.02x, R² = 0.80; RVM: if x ≤ 7, then y = 2.4 + 10.3x, R² = 0.86; if 7 ≤ x ≤ 37, then y = 76.7 + 0.52x, R² = 0.95; if x ≥ 37, then y = 94.9 + 0.06x; R² = 0.92; control: if x ≤ 7, then y = −16.8 + 18.5x, R² = 0.86; if 7 ≤ x ≤ 39, then y = 117.2 + 0.57x, R² = 0.94; if x ≥ 39, then y = 139.4 + 0.06x, R² = 0.82. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

View Full Size

Cumulative effluent ammonium for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments (LVM or RVM). Each container had 5.1 g of NH₄-N surface incorporated at the initiation of the experiment. Segmented linear plateau models were calculated for each treatment: LVM: if x ≤ 7, then y = −7.3 + 9.9x, R² = 0.84; if 7 ≤ x ≤ 33, then y = 62.7 + 0.28x, R² = 0.97; if x ≥ 33, then y = 72.4 + 0.02x, R² = 0.80; RVM: if x ≤ 7, then y = 2.4 + 10.3x, R² = 0.86; if 7 ≤ x ≤ 37, then y = 76.7 + 0.52x, R² = 0.95; if x ≥ 37, then y = 94.9 + 0.06x; R² = 0.92; control: if x ≤ 7, then y = −16.8 + 18.5x, R² = 0.86; if 7 ≤ x ≤ 39, then y = 117.2 + 0.57x, R² = 0.94; if x ≥ 39, then y = 139.4 + 0.06x, R² = 0.82. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Cumulative effluent ammonium for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments (LVM or RVM). Each container had 5.1 g of NH₄-N surface incorporated at the initiation of the experiment. Segmented linear plateau models were calculated for each treatment: LVM: if x ≤ 7, then y = −7.3 + 9.9x, R² = 0.84; if 7 ≤ x ≤ 33, then y = 62.7 + 0.28x, R² = 0.97; if x ≥ 33, then y = 72.4 + 0.02x, R² = 0.80; RVM: if x ≤ 7, then y = 2.4 + 10.3x, R² = 0.86; if 7 ≤ x ≤ 37, then y = 76.7 + 0.52x, R² = 0.95; if x ≥ 37, then y = 94.9 + 0.06x; R² = 0.92; control: if x ≤ 7, then y = −16.8 + 18.5x, R² = 0.86; if 7 ≤ x ≤ 39, then y = 117.2 + 0.57x, R² = 0.94; if x ≥ 39, then y = 139.4 + 0.06x, R² = 0.82. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Cumulative effluent dissolved reactive phosphorus (DRP) for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments (LVM or RVM). All treatments had 1.3 g of elemental phosphorus surface incorporated at the initiation of the experiment. Segmented linear models for 0.25 to 0.85 mm: if x ≤ 44, then y = 1.3 + 0.52x, R² = 0.99; if x ≥ 44, then y = 16.5 + 0.17x, R² = 0.96; 0.85 to 4.75 mm: if x ≤ 45, then y = −1.6 + 0.70x, R² = 0.99; if x ≥ 45, then y = 19.0 + 0.24x, R² = 0.97; LVM: if x ≤ 43, then y = 0.2 + 0.50x, R² = 0.99, if x ≥ 43, then y = 15.4 + 0.15x, R² = 0.96; RVM: if x ≤ 46, then y = −0.5 + 0.71x, R² = 0.99; if x ≥ 46, then y = 20.1 + 0.26x, R² = 0.97; control: if x ≤ 49, then y = 5.1 + 0.91x, R² = 0.98; if x ≥ 49, then y = 39.5 + 0.21x, R² = 0.97. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

View Full Size

Cumulative effluent dissolved reactive phosphorus (DRP) for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments (LVM or RVM). All treatments had 1.3 g of elemental phosphorus surface incorporated at the initiation of the experiment. Segmented linear models for 0.25 to 0.85 mm: if x ≤ 44, then y = 1.3 + 0.52x, R² = 0.99; if x ≥ 44, then y = 16.5 + 0.17x, R² = 0.96; 0.85 to 4.75 mm: if x ≤ 45, then y = −1.6 + 0.70x, R² = 0.99; if x ≥ 45, then y = 19.0 + 0.24x, R² = 0.97; LVM: if x ≤ 43, then y = 0.2 + 0.50x, R² = 0.99, if x ≥ 43, then y = 15.4 + 0.15x, R² = 0.96; RVM: if x ≤ 46, then y = −0.5 + 0.71x, R² = 0.99; if x ≥ 46, then y = 20.1 + 0.26x, R² = 0.97; control: if x ≤ 49, then y = 5.1 + 0.91x, R² = 0.98; if x ≥ 49, then y = 39.5 + 0.21x, R² = 0.97. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Cumulative effluent dissolved reactive phosphorus (DRP) for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments (LVM or RVM). All treatments had 1.3 g of elemental phosphorus surface incorporated at the initiation of the experiment. Segmented linear models for 0.25 to 0.85 mm: if x ≤ 44, then y = 1.3 + 0.52x, R² = 0.99; if x ≥ 44, then y = 16.5 + 0.17x, R² = 0.96; 0.85 to 4.75 mm: if x ≤ 45, then y = −1.6 + 0.70x, R² = 0.99; if x ≥ 45, then y = 19.0 + 0.24x, R² = 0.97; LVM: if x ≤ 43, then y = 0.2 + 0.50x, R² = 0.99, if x ≥ 43, then y = 15.4 + 0.15x, R² = 0.96; RVM: if x ≤ 46, then y = −0.5 + 0.71x, R² = 0.99; if x ≥ 46, then y = 20.1 + 0.26x, R² = 0.97; control: if x ≤ 49, then y = 5.1 + 0.91x, R² = 0.98; if x ≥ 49, then y = 39.5 + 0.21x, R² = 0.97. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Across all treatments, 80% to 90% of total NH₄-N captured in the effluent over 112 d was lost in the first 7 d (Fig. 2). This most likely was a result of the water-soluble N (Fast Start) associated with this specific CRF. From 1 to 7 DAI, the sand-amended substrate lost 18.5 mg· d⁻¹ NH₄-N, whereas the LVM and RVM aggregate-amended substrates lost 9.9 and 10.3 mg·d⁻¹ NH₄-N, respectively. At 8 DAI, the rate of NH₄-N loss decreased across all treatments. The rate of NH₄-N from substrates amended with LVM aggregates decreased 97% to 0.28 mg·d⁻¹, whereas the rate from the sand-amended substrate decreased 94% to 0.57 mg·d⁻¹ NH₄-N. At 33, 37, and 39 DAI for the sand, RVM, and LVM aggregate-amended substrates, respectively, NH₄-N loss from the container was negligible, a trend that continued for the rest of the 112-d study (Fig. 2). This was probably because of rapid oxidation of NH₄ to NO₃, root uptake, or leaching. Foster et al. (1983) reported NH₄-N sorption to pine bark required 20 d to stabilize and equilibrate; however, our data may simply be a result of the time for optimal edaphic conditions (i.e., pH, water content, and temperature) to arise in the substrate to sustain Nitrosomonas sp., the nitrifying bacteria. Establishment of microbes to maximize nitrification occurs usually within 21 to 35 d after potting plants into a pine bark substrate (Niemiera and Wright, 1986).

Aggregate temperature pretreatment and particle size reduced effluent DRP content compared with the control (Fig. 3). There was no significant effect of interaction of temperature pretreatment and particle size. LVM and particle size 0.25 to 0.85 mm further reduced effluent DRP compared with the RVM aggregate and particle size of 0.85 to 4.75 mm. Dissolved reactive P was lost linearly at a rate of 0.50 mg·d⁻¹ from 0 to 43 DAI for substrates amended with LVM aggregate, whereas substrates amended with RVM aggregate lost 0.71 mg·d⁻¹ from 0 to 46 DAI. The rate of DRP lost decreased at 43 or 46 DAI for substrates amended with LVM (0.15 mg·d⁻¹) or RVM (0.26 mg·d⁻¹) resulting in 35% (17 mg) less DRP lost to leaching from 0 to 46 DAI for substrates amended with LVM.

From 0 to 44 DAI, the rate of DRP loss decreased 29% when substrates were amended with 0.25 to 0.85 mm (0.52 mg·d⁻¹) compared with substrates amended with 0.85 to 4.75 mm (0.7 mg·d⁻¹) aggregates. After 45 DAI, the rate of effluent DRP decreased to 0.17 mg·d⁻¹ or 0.24 mg·d⁻¹ for 0.25 to 0.85 mm and 0.85 to 4.75 mm aggregate-amended substrates, respectively. This resulted in a 22% (10 mg) decrease in cumulative effluent DRP in pine bark substrates amended with aggregates of 0.25 to 0.85 mm compared with 0.85 to 4.75 mm from 45 to 112 DAI.

From 0 to 49 DAI, the control substrate lost 0.91 mg·d⁻¹ DRP, which was 29% to 79% greater than DRP loss from the mineral aggregate-amended substrates. This reflects a lower capacity of the industry substrate (pine bark: sand) to retain the oxyanion. After 49 DAI, the sand-amended substrate rate of DRP loss (0.21 mg·d⁻¹) was less than substrates amended with 0.85 to 4.75 mm (0.24 mg·d⁻¹) or RVM (0.26 mg·d⁻¹) aggregates (Fig. 3). The aggregates in these substrates could have been desorbing P resulting in greater quantities of P loss. This is supported by the increase in extractable P that occurs in a mineral versus sand amended substrate (Table 2). However, we speculate 0.25 to 0.85 mm or LVM aggregates possibly continued sorbing P because their rate of P release remained lower than other treatments, DRP at 0.17 mg·d⁻¹ or 0.15 mg·d⁻¹, respectively (Fig. 3). Extractable P averaged across mineral aggregate-amended substrates increased 89% and 87% when extracted with water or 0.1 M KCl, respectively, compared with the control substrate (Table 2). Thus, although the absolute quantities of extractable P varied with extractant, the quantities were consistent between mineral aggregate and sand-amended. It is uncertain why water resulted in greater extractable P than 0.1 M KCl. These differences in extractable P may have been a result a different ionic strength and pH-buffering capacity of the extractant.

Table 2.

Dissolved reactive phosphorus extracted from a pine bark substrate amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two temperature pretreatments and two particle sizes in which Skogholm cotoneaster was grown for 112 d.^z

View Table

Mineral aggregate-amended substrates affected the chemical properties of the substrate by increasing K and S substrate content an average of 250% and 131%, respectively, when compared with the control (Table 3). Calcium and Mg substrate content also increased similarly when amended with a mineral aggregate amendment (data not presented). However, the quantities of these nutrients that were available to the plant were not known. Low S has been reported to limit plant growth in soilless substrates (Browder et al., 2005); however, little research has been conducted on the base cations. In addition, CEC of aggregate-amended substrates were increased an average of 75% when compared with the control. The average mineral aggregate-amended substrate raised pH 0.4 units above the control (Table 3). Furthermore, substrates amended with LVM mineral aggregates increased pH 0.5 above the RVM aggregates.

Table 3.

Nutrient content, cation exchange capacity, and pH of substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two temperature pretreatments and two particle sizes.^z

View Table

The increased nutrient-buffering capacity of substrates amended with mineral aggregates resulted in an average 32% increase in the top and root P content of Skogholm cotoneaster compared with the control (Table 4). The increased water-buffering capacity in association with P retention of the mineral aggregate-amended substrate resulted in a 20% to 48% decrease in effluent DRP load and a 125% to 154% increase in substrate water or KCl-extractable P, respectively, compared with the control. These factors combined to increase P uptake efficiency (PUE) from 27% in sand-amended substrates to an average 36% in aggregate-amended substrate. This was also reflected in the average daily effluent P concentration with the aggregate-amended substrate averaging 1.3 mg·L⁻¹ (Fig. 4). Warren et al. (1995) reported resin-coated P in a CRF resulted in the highest P use efficiency (43%) with a low, constant rate of P loss at ≈1 mg·d⁻¹ when growing Sunglow azalea in 3.8-L containers in a pine bark substrate.

Table 4.

Distribution of P for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two particle sizes and two temperature pretreatments.^z

View Table

Daily effluent nutrient concentration of NO₃-N and dissolved reactive phosphorus (DRP) for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two temperature pretreatments (LVM or RVM). All treatments had 5.1 g NO₃-N and 1.3 g of elemental phosphorus surface incorporated at the initiation of the experiment. Municipal water applied contained an average NO₃-N and P concentration of less than 0.1 mg L⁻¹ and less than 0.05 mg L⁻¹, respectively. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

View Full Size

Daily effluent nutrient concentration of NO₃-N and dissolved reactive phosphorus (DRP) for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two temperature pretreatments (LVM or RVM). All treatments had 5.1 g NO₃-N and 1.3 g of elemental phosphorus surface incorporated at the initiation of the experiment. Municipal water applied contained an average NO₃-N and P concentration of less than 0.1 mg L⁻¹ and less than 0.05 mg L⁻¹, respectively. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Daily effluent nutrient concentration of NO₃-N and dissolved reactive phosphorus (DRP) for Skogholm cotoneaster grown 112 d in pine bark substrates amended (by volume) with 11% sand (control) or 8% Georgiana bentonite–palygorksite clay mineral aggregate with two temperature pretreatments (LVM or RVM). All treatments had 5.1 g NO₃-N and 1.3 g of elemental phosphorus surface incorporated at the initiation of the experiment. Municipal water applied contained an average NO₃-N and P concentration of less than 0.1 mg L⁻¹ and less than 0.05 mg L⁻¹, respectively. LVM = low volatile material; RVM = regular volatile material.

Citation: HortScience horts 42, 5; 10.21273/HORTSCI.42.5.1287

• Download Figure

Regardless of particle size, the LVM aggregate reduced the P lost in the effluent compared with the control, whereas RVM did not (Table 4). The CRF prills retained an average 41% (mean = 539 mg ± 25 se) of the applied 1.3 g of elemental P accounting for 65% of the recovered P (mean = 829 mg ± 29 se) (Table 4). Plant PUE was 5% to 12% greater in a mineral-amended substrate (mean = 36%; P = 0.07, F = 4.0) compared with the control (27%) (Table 4).

Average effluent DRP concentration for the 112 d decreased 26% from 1.8 mg·L⁻¹ to 1.3 mg·L⁻¹ for the sand- versus mineral-amended pine bark substrate, respectively (Fig. 4). Groves et al. (1998b) reported current BMP substrate solution P recommendations (5 mg·L⁻¹ to 10 mg·L⁻¹) could not be maintained when irrigating 3.8-L containerized Skogholm cotoneaster with 800 mL·d⁻¹; however, top and root dry weight were maximized at this irrigation volume, although observed substrate solution P concentrations decreased as low as 1.8 mg P L⁻¹ and 0.1 mg P L⁻¹ at 60 DAI and 119 DAI, respectively (Groves et al., 1998a). Lea-Cox and Ristvey (2003) suggested an 80% reduction in current rate of P application, thus making the optimal P substrate solution concentration 2 mg·L⁻¹ or less, increasing PUE to 75% when adequate N is applied.

Effluent N content, total plant N content, and N uptake efficiency were unaffected by substrate amendment (Table 5). However, mean daily effluent N concentration over the 112-d study was reduced 22% from 8.4 mg·L⁻¹ to 6.5 mg·L⁻¹ when the pine bark substrate was amended with an LVM aggregate versus RVM aggregate (Fig. 4). CRF prills contained 26% of the total applied N after 112 DAI, 42% being NH₄-N and 10% NO₃-N accounting for 55% of the recovered N after 112 DAI (Table 5). Three percent (mean = 149 mg ± 13 se) of N was leached and 38% (mean = 1814 mg ± 30 se) of N was taken up by the plant. No substrate NO₃-N or NH₄-N could be detected with either water or 0.1 mol·L⁻¹ KCl extractant. Nitrogen use efficiency in this study is lower than previously reported, probably because it was calculated based on elemental N and not N species (Tyler et al., 1996). The authors hypothesize that the unaccountable portion of N was primarily lost through denitrification and a minor portion of N was lost to immobilization or volatilization.

Table 5.

Distribution of N for Skogholm cotoneaster grown for 112 d in pine bark-based substrates amended (by volume) with sand at 11% or industrial aggregate at 8%.^z

View Table

Plant response.

Top (160 g ± 4 se) and root (mean = 30 g ± 1 se) dry weight of Skogholm cotoneaster were unaffected by any of the treatments (data not presented). However, water required to produce 1 g dry weight (WUE_P) was affected by treatments (data not presented). Skogholm cotoneaster grown in aggregate-amended substrates required an average of 607 ± 4 mL to produce 1 g of plant dry weight, whereas cotoneaster grown in the control substrate required 731 mL to produce 1 g of plant dry weight, resulting in a 17% increase in WUE_P. We speculate this is a reflection of increased CC and AW in aggregate-amended substrates (Table 1).

Phosphorus (Table 4) and K (data not presented) content of Skogholm cotoneaster were 32% and 49% greater, respectively, when grown in a mineral-amended substrate versus sand-amended substrate. Top or root N, Ca, Mg, S, Fe, Mn, B, Mo, and Cl content were unaffected by mineral aggregate amendments nor were there any differences with the control (data not presented).

CER was 20% greater in Skogholm cotoneaster grown in mineral-amended substrates (mean = 8.2 μmol·m⁻²·s⁻² ± 0.3 se) versus Skogholm cotoneaster grown in the control substrate (mean = 6.9 μmol·m⁻²·s⁻² ± 0.9 se) without induced drought stress (data not presented). CER and g_s was 25% and 33% greater, respectively, after 42 h without irrigation when cotoneaster was grown in a mineral-amended substrate compared with cotoneaster grown in the control substrate (Owen et al., 2003). Mineral aggregate temperature pretreatment had greater influence on CER and g_s than mineral aggregate particle size. After 42 h without irrigation, plants grown in substrates amended with LVM mineral aggregates had 27% higher CER (mean = 8.7 μmol·m⁻²·s⁻² ± 0.8 se) and 43% higher g_s (mean = 0.13 mol·m⁻²·s⁻¹ ± 0.02 se) compared with plants grown in substrates amended with RVM aggregates (mean CER = 6.9 μmol·m⁻²·s⁻² ± 0.6 se; mean g_s = 0.09 mol·m⁻²·s⁻¹ ± 0.01 se) (Owen et al., 2003). Plant response was further separated by particle size, with plants grown in substrates amended with 0.25- to 0.85-mm aggregates (mean = 0.03 mol·m⁻²·s⁻¹ ± 0.006 se) having greater g_s compared with plants grown in substrates amended with 0.85- to 4.75-mm aggregates (mean = 0.02 mol·m⁻²·s⁻¹ ± 0.001 se) after 67 h without irrigation. After 71 h without irrigation, all treatments had similar levels of CER and g_s (data not presented).

Pine bark amended with industrial mineral aggregates increased water and nutrient efficiency without affecting root or top growth of Skogholm cotoneaster. This modified substrate offers a way to reduce environmental impact of nursery crop production compared with traditional means of lowering input quantities. If input quantities become growth-limiting, the result is decreased nutrient uptake efficiency attributable to less plant growth and root exploitation, which results in increased mineral nutrient leaching (Tyler et al., 1996). Mineral aggregate-amended substrates increased WUE_P 17% and decreased water applied 26 L per 14-L container over the growing season when compared with pine bark amended with sand. Substrate CEC was greater than 75% in mineral- versus sand-amended substrates increasing substrate base cation content. Mineral aggregate-amended substrate had a 125% to 154% increase in KCl/water-extractable P, which in turn raised top P content of Skogholm cotoneaster 32% compared with the control. Average daily effluent DRP concentration was 0.8 mg·L⁻¹, which is greater than the USEPA MCL goal for waters. Therefore, with current rate of P fertilization, effluent concentrations that meet USEPA criteria may not be attainable for nurseries recycling water alone. Average daily NO₃-N concentration remained below USEPA MCLs from 20 to 112 DAI. A pine bark soilless substrate amended with 0.25 to 0.85 mm LVM industrial mineral aggregate can reduce effluent DRP and NH₄-N and reduce water application volume when compared with a sand-amended bark-based substrate without sacrificing plant growth.