SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 6 May 2005
Published in Soil Sci Soc Am J 69:794-806 (2005)
DOI: 10.2136/sssaj2004.0108
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Caron, J.
Right arrow Articles by Boudreau, J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Caron, J.
Right arrow Articles by Boudreau, J.
Agricola
Right arrow Articles by Caron, J.
Right arrow Articles by Boudreau, J.
Related Collections
Right arrow Hydraulic Conductivity

Soil Physics

Defining Critical Capillary Rise Properties for Growing Media in Nurseries

J. Carona,*, D. E. Elrickb, R. Beesonc and J. Boudreaud

a Département des Sols et Génie Agroalimentaire, Centre de Recherche en Horticulture, Université Laval, Sainte-Foy, Québec, Canada, G1K 7P4
b Land Resource Science Department, University of Guelph, Guelph, Ontario, Canada N1G 2W1
c Assoc. Professor, Mid Florida Research and Education Center, IFAS, University of Florida, Apopka, FL 32703
d Hortau inc, 840 Ste-Thérèse, Suite 300, Québec, Canada G1N 1S7

* Corresponding author (jean.caron{at}sga.ulaval.ca)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Mathematical Model of Capillary...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Water availability for landscape nursery irrigation is foreseen as a major impediment for this industry within the next decade. Among various solutions proposed to increase irrigation efficiency, thereby reducing the water volumes required, are closed and semi-closed subirrigation systems designed to grow plants potted in organic growing media. These systems, however, require organic substrates that have good capillary properties. However, standards for such capillary properties are not available. This study compared substrates composed of peat, bark, and sand having contrasting capillary properties, in a nursery experiment to establish guideline values for the proper and efficient operation on capillary mat devices. It also proposes a theoretical model of capillary rise using the hydraulic characteristics of growing media to predict the suitability of various substrates. Substrates with 60% (per volume) sphagnum peat were found to provide the best capillary rise and best growth, based on empirical measurements, relative to substrates with 30% sphagnum or 30% sedge peat. The proposed theoretical model concurred with these observations.

Abbreviations: MWD, mean weight diameter


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Mathematical Model of Capillary...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
EFFICIENT USE OF IRRIGATION water is becoming a major issue in nursery production. Inefficient use results in increased risk of water shortages during drought periods, and limited water supplies under non-drought conditions, both of which can lead to restrictions on irrigation. Over the years, different systems to conserve water have been proposed for nurseries that utilize overhead irrigation systems. Tailwater recovery and recycling, catchment basins, ebb and flow, capillary mats, and microirrigation, among other systems, appear to be sound procedures for conserving water. For some irrigation systems such as ebb and flow and capillary mats, good substrate capillary properties are needed, but standards for such properties are lacking.

Recently, a new capillary mat system designed for nursery production was tested and found to have low outdoor surface evaporation. Its incorporation in nursery production appears to offer an efficient way of conserving water (Caron et al., 2002a). However, with the introduction of this system, as with other subirrigation devices like ebb and flow and sand beds, substrate properties become extremely important. Common substrates used with overhead irrigation systems have very high airfilled porosities and fast drainage, but poor capillary rise properties and low amounts of available water. These high aeration and drainage requirements have been used for many years. Their wide acceptance is based on trial and error type experiments performed in Mediterranean countries or the Southeast USA, where heavy rainfall can create waterlogged conditions detrimental to crop growth (Raviv et al., 2001). To avoid the risk of root asphyxia during heavy rainfall events or due to over-irrigation, and also to reduce substrate-manufacturing costs, the tendency in these regions has been to reduce the proportion of the fine fraction in nursery mixes. As a result, the amount of available water in most substrates has decreased substantially. The unsaturated hydraulic conductivity at a given potential (or water content), a parameter important to the capillary mat system, is also lowered by decreasing the proportion of fines in the porous medium (Heiskanen, 1999). Preliminary work has indicated that currently used substrates seriously limit the efficient operation of capillary mat systems (Boudreau et al., 2002). A first step to improve performances of such substrates irrigated with overhead systems, as well as ebb and flow and capillary mat irrigation systems, is to increase the proportion of available water to reduce the irrigation frequency.

Apart from adding hydrophilic polymers in mixes (Fonteno and Bilderback, 1993), a more fundamental approach to improve the amount of available water is to manipulate the size and shape of the different components of the substrate. An increase in available water can thus be obtained by decreasing the proportion of large components in a mix. Another way to manipulate the amount of available water in substrates is to change the type of peat. Sphagnum peat is derived from mosses having adapted structures, which store and transfer water through a network of hollow vessels and pores; the remnants of these distinctive structures give sphagnum peat its large water holding capacity and excellent capillary rise properties (Puustjärvi, 1978). On the other hand, sedge peat, often found in warm temperate climates, originates from decayed grasses in marshlands and does not have these specific structures. Water is stored in sedge peats between loose fibers, which are less efficient in storing and transporting water than that of sphagnum mosses (Rivière, 1992).

The second step to improve the performance of these substrates is to increase capillary rise properties, so that when water within a substrate is depleted, additional water from external sources like a capillary mat, ebb and flow or the sand bed can be absorbed into the substrate. Capillary rise is linked to the unsaturated hydraulic conductivity of the substrate and is expected to vary with the type of peat, being reduced at high potential in highly decomposed herbaceous peat relative to slightly decomposed sphagnum peat (Brandyk et al., 2003). In coarse media, adding fine material to substrates increases the unsaturated hydraulic conductivity and therefore is expected to increase capillary rise (Heiskanen, 1999).

Capillary rise properties and available water can be improved, but usually it is thought that this occurs at the expense of substrate aeration. This assumption has been based on changes in air-filled porosity. However, recent research in this area has shown, both in nurseries and greenhouses, that gas relative diffusivity (an index related to the speed of gas movement in the soil relative to that in free air), was a clearly superior index to relate to plant growth than air filled porosity (Allaire et al., 1996, Caron et al., 2001).

Standards for substrate properties need to be defined for nursery growers using capillary mat systems. However little information is available on which parameters are important and how they should be interpreted. While important standards have been defined for available water and air content from water desorption curves (Heiskanen, 1993), little is known about the unsaturated hydraulic conductivity of substrates used in containers to maintain adequate capillary rise (Wallach et al., 1992, Otten, 1994, Heinen, 1997, Raviv et al., 2001). Moreover, this property is not unique since unsaturated hydraulic conductivity curves are subject to strong hysteretic behavior (draining and wetting properties differ, Wallach et al., 1992). For capillary rise, the wetting (or sorption) conductivity properties are most important.

Much of the research on capillary rise in growing media has been limited to numerical model studies used to describe capillary rise in bed or ebb and flow systems under transient state conditions (Otten, 1994, Heinen, 1997). As conditions in outdoor nurseries differ from those in greenhouses (substrates are generally coarser, container geometry differs, the evapotranspiration demand may be much higher and cropping practices differ), models describing this process can assist greatly in the development of guidelines, but cannot be used as they exist. Moreover, many of the parameters used in various existing models are based on experiments performed on disturbed media. Because of the important effects that substrate disturbance, transplanting, and filling may have on final properties (Heiskanen et al., 1996), there is a need to provide that information from an in situ characterization procedure.

Therefore, the first objective of this study was to compare substrates having contrasting physical properties to identify proper substrate guideline values for the optimum operation of a capillary mat system designed for nursery use. A second objective was to identify the in situ characteristics of substrates, which can provide an adequate supply of water for proper operation on subirrigation devices. A third objective of this study was to develop a model of capillary rise to predict the behavior observed with different growing media on a capillary mat system.


   Mathematical Model of Capillary Rise under Evapotranspirative Conditions
TOP
 ABSTRACT
 INTRODUCTION
 Mathematical Model of Capillary...
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The model (Fig. 1) is based on the following assumptions:

  1. The substrate in the pot is assumed to be of a known depth, with the origin set at the surface of the substrate and with z (lL), the vertical space coordinate, defined as positive downward. A water table is established at the bottom of the pot and upward flow is assumed to take place under evaporative conditions at the surface and plant uptake of water within the pot.
  2. The substrate material is assumed to be rigid, homogeneous, and isotropic. This is not a particularly good assumption given the sensitivity of these substrates, but is necessary to solve the appropriate differential equation.
  3. The hydraulic conductivity, K({psi}) (L T–1), of the substrate material is assumed to be two-line exponential (Jarvis and Messing, 1995) as a function of the pressure head (potential), {psi} [L], of water in the unsaturated region.
  4. The total flux density of water, Jw (L T–1), due to the combination of evaporative conditions at the surface and plant uptake within the pot, is constant with time; that is, steady-state flow conditions.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 1. Physical representation of capillary rise in a subirrigation system with free water at the bottom of the pot.

 
The well-known solution of Gardner (1958) was used to obtain the steady-state relationship between soil water head, {psi} (negative in unsaturated soil), and depth z for a constant evapotranspirative flux density at the pot surface. Gardner's equation was modified to include a water-entry value, {psi}w, and with the origin at the surface of the substrate and with z positive downward the solution is given by Elrick et al. (1994):

[1]
where

[2]

In Eq. [1] and Fig. 1, {alpha} [L–1] is the exponential parameter of Eq. [2], Jw(d) = q + Pd is the constant flux density of water entering the pot at z = h (and d), q [L T–1] is the evaporation flux density at the surface of the substrate, P [T–1] is the plant uptake of water per unit of depth, d is the depth to the top of the capillary fringe (where {psi} = {psi}w), Ks [L T–1] is the saturated hydraulic conductivity and {psi}w [L] is the water-entry value. We assume that there is no uptake of water in the saturated zone (d ≤ z ≤ h).

The total flux of water at any given depth, Jw(z) is given by:

[3]

Note that both q and P in Eq. [3] are negative in value and the terms are additive. Negative q indicates that flow is upward and negative P indicates that water is removed from the substrate by the plant roots. Note that P represents plant uptake of water per depth unit and (q + Pd) represents the water flux entering the bottom of the pot. It is also the water flux at z = d as we have assumed that there is no plant uptake of water in the saturated zone.

If h is the depth of the pot, then

[4]
where d < h as {psi}w is negative in value.

In Eq. [2] the rate of decrease in K({psi}) with {psi} is determined by {alpha}, the larger the value of {alpha} the greater the rate of decrease in K({psi}). Earlier reported values of {alpha} have been observed to vary between 0.09 and 0.18 cm–1 (Tardif and Caron, 1993) and between 0.07 and 0.09 cm–1 (Caron et al., 1998), when established from drainage. Tardif and Caron (1993) found much higher values (between 0.45 and 0.83 cm–1) for {alpha} on rewetting and identified a pronounced hysteretic behavior, consistent with Wallach et al. (1992).

As observed previously (Caron and Elrick, 2005) and as shown below, a better representation of K({psi}) can be obtained by assuming that the relationship is two-line exponential on rewetting, with {alpha} being considerably larger in the wet region close to saturation:

[5]
and

[6]
where {alpha}1 is larger than {alpha}2, {psi}b is the break-point value of {psi} where {alpha}1 changes to {alpha}2, {psi}i is the initial pressure head in the substrate and Kb is the value of K({psi}) when {psi} = {psi}b. Note that K({psi}) = Ks in the pressure range {psi}w ≤ {psi} ≤ 0.

If we also assume that {psi} = {psi}w at z = d (negligible drop in head during flow through the capillary fringe), then the solution, based on a procedure similar to that used by Elrick et al. (1994) and Caron and Elrick (2005) and subject to the conditions shown in Fig. 1, is given by:

[7]
where zb ≤ z ≤ d and by:

[8]
where 0 ≤ z ≤ zb.

Note that Kb is the value of K({psi}) at {psi} = {psi}b, Jb is the flux of water at z = zb and can be obtained from Eq. [3]. The variable zb can be obtained from Eq. [7] by solving for z when {psi}1 = {psi}b. Note that {psi}b can be obtained from experimental data or it can be chosen arbitrarily if carrying out a theoretical analysis.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 Mathematical Model of Capillary...
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Preparation of Substrates
In mid-April 2001, container substrates were prepared as follows:

  1. Sp30B60: 30% Sphagnum peat (Premier Peat, Ltd; Rivière-du-Loup, Quebec); 60% composted pine bark fines; 10% coarse sand (volume basis) with 3.46 kg m–3 (2 lbs yd–3) micronutrients (Micromax, The Scotts Co. Marysville, OH) and 17.3 kg m–3 (10 lbs yd–3) dolomite limestone.
  2. Sp60B30: 60% sphagnum peat (Premier Peat, Ltd, Rivière-du-Loup, Québec, Canada) 30% composted pine bark fines: 10% coarse sand with 3.46 kg m–3 (2 lbs yd–3) micronutrients (Micromax) and 17.3 kg m–3 (10 lbs yd–3) dolomite limestone.
  3. Se30B60: A commercial blend consisting of 30% Florida sedge peat; 60% composted pine bark fines:10% coarse sand with 2.6 kg m–3 (1.5 lbs yd–3) Micromax and 10.4 kg m–3 (6 lbs yd–3) dolomite limestone (Florida Potting Soil, Apopka, FL).

The pine bark and sand components used to blend the Sp30B60 and Sp60B30 substrates were collected from the same location in the piles at Florida Potting Soil concurrent with the blending of the commercial substrate, Se30B60. The Sp60B30 substrate was chosen because it is close to the recommended composition for subirrigation systems found in Europe (Wever, 1991) while the other two corresponded to medium formulations widely used in North American nurseries.

Substrate Characteristics
Particle-size distribution. Particle-size distribution was measured on a subsample 500-mL sample of air-dried material, sieved manually for 1 min on a nest of sieves (0.25, 1, 2, 4, 8, 16, 25, 40 mm and a pan), and was replicated three times. The average particle size of the different materials was expressed using the mean weight diameter (MWD) calculated from (Kemper and Rosenau, 1986):

[9]
where xi is the weight retained on the sieve divided by the total medium mass, fi is the average particle size, and n is the number of size classes, here equal to 8.

Water and air properties. Preliminary work was conducted at Laval University, in Québec, Canada, to determine the proper amount of sphagnum peat needed to improve available water in substrates while still maintaining proper aeration (Brisson et al., 2002). Chosen substrates were then characterized in three replicates for their water desorption–sorption curves using a tension table (Topp and Zebchuk, 1979) with a gravimetric determination of container weight, later converted to volumetric water content using bulk density (Paquet et al., 1993). Air-filled porosity was calculated as the difference between total porosity and the volumetric water content at container capacity (–6 cm of matric head), for both rewetting and drainage. Easily available water was determined from the difference between the volumetric water content at container capacity and that at –50 cm of matric head (De Boodt and Verdonck, 1972). These two parameters represent the proportion of the total substrate volume, occupied by air or water that is available to the plant. Gas diffusivity was measured in these substrates using the procedure described by Caron and Nkongolo (2004) on potted substrates. This parameter is simply the ratio of gas diffusion in the substrate relative to the gas diffusion in free air, at a given temperature. Briefly, gas relative diffusivity was calculated from the water desorption curve, assuming the point of air entry was –3.5 cm for Sp60B30, and –2.5 cm for the Se30B60 and Sp30B60 substrates (Caron et al., 2002b), also using the saturated hydraulic conductivity (Ks), measured on the same sample, as described below.

Unsaturated hydraulic conductivity. The unsaturated hydraulic conductivity-matric head relationship K({psi})–{psi} was determined on a separate sample using the Laval tension disc technique (Caron and Elrick, 2005) utilizing the same container and potting procedure as used in the field, at the time of potting. Substrates were saturated, slowly wetted from underneath, then allowed to drain for 30 min. They were rewetted twice more with 1.5 L of tap water and again allowed to reach container capacity. This procedure allows substrates to reach a stable volume within the container. Substrates were transferred to measure water desorption curves, then equilibrated to –100 cm of matric head, and returned to –6 cm. A small quantity of glass beads were spread on top of the substrate surface to improve contact with the tension disc which was subsequently placed on top of the glass beads (Caron and Elrick, 2005). Measurements and calculations of K({psi})–{psi} relationships were determined on rewetting using the piecewise exponential approach described by Caron and Elrick (2005).

Statistical analysis. Comparison of all basic chemical and physical properties were performed as a randomized complete block design for all substrates. For unsaturated hydraulic conductivity, the natural log of the data were used and analyzed as a randomized complete block design comparing substrates, using the matric potential as a covariable (Freund et al., 1986), using the SAS software (SAS Institute, Cary, NC).

Field Experiment
System performances. The experiment was conducted at the University of Florida's IFAS Mid-Florida Research and Education Center in Apopka, FL during the 2001 growing season. The experiment was conducted in a production area consisting of individually controlled and irrigated production pads. Within each pad two Aquamat systems (Soleno Textiles inc, Laval, Canada) were installed. The Aquamat was a two-layer capillary mat specially designed for nursery production (Caron et al., 2002a). This system was selected for its ability to maintain a constant potential at the bottom of the pot and therefore better met one of the assumptions made in the model. Overhead irrigation was supplied to the mat through galvanized pipe risers, with impact sprinklers (Model 25BPJFPADJ RainByrd, Glendora, CA.) attached above 10.2 kg cm–2 (30 psi) pressure regulators (Senninger Irrigation, Orlando, FL). A low volume mechanical water meter (model Kent C700, ABB Ocala, FL) was installed upstream from the risers. Sprinklers were adjusted until the Christensen Coefficient of Uniformity (Haman et al., 1997) within a rectangular area 3.7 x 4.9 m (12 x 16 ft) in the center of the pad was >0.85. Irrigation application rates were calibrated for each pad before installing the plants. The percentage of the applied irrigation, which fell within the confines of the area where the CU was >0.85 was calculated.

Two plant species were grown in each pad, with each species wholly contained on an independent Aquamat in one half the rectangular growing area and treated as a separate experiment. Ligustrum japonicum (common ligustrum) was chosen for its widespread distribution in the USA. It is also relatively a fast growing plant, with a coarse aggressive root system. The second species, Viburnum odoratissimum (Sweet viburnum), is also a relatively fast growing plant but is very responsive to water availability during growth. It is commonly grown in Zones 8 to 10.

All rooted liners of viburnum and ligustrum were potted into #1 black polyethylene containers (010, Lerio Corp., Kissimmee, FL.) measuring 15 cm diam x 18.5 cm high during the week of 23 Apr. to 27 Apr. 2001. Plants were irrigated at night with overhead irrigation. A total of 4 mm (0.16 in) of water was applied from potting until 3 wk after. From May 11th, each pad area received 4.8 mm (0.19 in) beginning at 2300 h and about 3 h later another 3.0 mm (0.12 in) each night.

Within a week after potting and again two weeks later, each plant was drenched with about 150 mL (5 oz.) of a 300 ppm N solution of liquid fertilizer (Peters 20-20-20; Scotts Co., Marysville, OH). Also within a week after potting, an application of the pre-emergent granular herbicide Ornamental Herbicide II (The Scotts Co.) was broadcast over each pad area. Plants were given 12 g of Osmocote 18-6-12 (The Scotts Co.) the second week of May and again the second week of September. Plants were pruned as needed to promote commercially acceptable quality.

Crop yield. When at least 16 of the 18 plants measured per substrate obtained the minimum for marketable size at the Florida Fancy grade, canopies of all 18 plants were removed at the substrate surface and dried to a consistent weight at 65°C and their dry mass recorded. Minimum dimensions for a Florida Fancy grade are 28 cm (11 in) tall x 21 cm (9 in) average width (Florida Grades and Standards, 1995).

Statistical analysis of data was performed using SAS (SAS Institute, Cary, NC). Data from each species was analyzed separately. Data was analyzed as a randomized complete block, with three substrate treatments replicated in three blocks of six plant replicates. Where appropriate, mean separation employed F-protected LSD.

Water use. Irrigation volumes applied to each subplot were calculated by multiplying the difference in initial and final water meter readings when 92% marketability was obtained, by the percentage of the irrigation applied, which fell within the plant production rectangle. Irrigation volumes were analyzed as described for final growth measurements.

Observed potential variation. The evapotranspirative demand was evaluated by taking plants from the Sp60B30 treatment and wetting the substrate for about half an hour from underneath and then excessively irrigating the container/plant system from above. Verification was made that the procedure was enough to bring the sample to container capacity. Volumetric water contents were then measured at 0800, 1400, and 2000 h, on four replicates of the treatment, two days in a row, on May 12 and 13 and August 16 and 17. The lowest volumetric water content corresponded to potentials higher than –30 cm, very likely to be within the potential range of full availability to plants (De Boodt and Verdonck, 1972, Caron et al., 1998). The average transpiration rate was calculated from the 800 to 1400 period, as five times more water was lost during that period than that from 1400 to 2000. This high evapotranspirative demand period was selected because it was likely to be the most limiting conditions for proper testing of the capillary mat-substrate system, relative to the full day period. Water content measurements were performed with a time domain reflectometry (Topp et al., 1980) probe of the same length as the substrate height, to estimate the average volumetric water content throughout the container, using a calibration curve specific to the growing media (Paquet et al., 1993). May 12 corresponded to the start of the proper operation of the mat system while August 16 corresponded to the time of maximum size that plants had reached before first harvest of plants of commercial size. Tensiometers made of PVC access tube with a 5 cm long x 2.5 cm outer diameter porous stone were inserted at mid-height (6–7 cm from the bottom of the pot) of all three substrates and their internal vacuum read with a pressure transducer device. Measurements were performed on May 10 through May 13 and August 16 through August 17 from two to four times a day on all treatments. Statistical analyses of tensiometers data were performed by specific time, considering a randomized complete block design, with substrates as the treatments. Potential data were transformed using the natural logarithm to homogenize the variance.

Laboratory Experiment
A laboratory experiment was conducted on the Se30B60 and Sp60B60 treatment to confirm field observations with greater accuracy. Substrates were packed into 15 cm ID by 25 cm long columns, which were then saturated, to completely wet the substrate, and placed onto a capillary mat (Fig. 2). Three replicated columns were used per substrate. The experiment was performed for 115 d and verification was made that steady-state conditions were obtained by using TDR probes inserted at different heights, to verify that volumetric water content values were constant. Water was supplied to the mat by a tension infiltrometer, which was adjusted to a soil water potential of –0.4 kPa (–4 cm) at the surface of the capillary mat. This slight tension was necessary to allow an accurate operation of the infiltrometer in a room subject to slight barometric changes. The mat was covered with a polyethylene liner to minimize surface evaporation. The experiment took place in a room where ambient conditions were controlled. The relative humidity was low (30%) and the temperature was maintained constant (22 ± 2°C). A constant dry air flux was maintained at the top of the columns to increase surface evaporation, which, at steady state conditions reached values of 0.0035 and 0.020 cm h–1 for the Se30B60 and Sp60B30 treatment, respectively. On 28 May 2001, 115 d after the start of the experiment, columns were removed from the capillary mats, cut into slices of about 1.5 cm thick, dried at 105°C for 24 h and weighted to determine bulk density and gravimetric water content. From these, volumetric water content was calculated and converted to potential using the appropriate water desorption–sorption curve determined as above.



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 2. Experimental set-up used to maintain a constant upward capillary movement in the laboratory experiment.

 
Predictive Model
Potentials were predicted as a function of height (referred to as the matric head profile) in the substrate using Eq. [7] and [8], for both the laboratory experiment and the field experiment. For the lab experiment, the measured evaporative demand was used, P taken equal to zero (no plant uptake), and the bottom boundary conditions changed to correspond to the –4 cm applied to the mat with the tension infiltrometer. For the field experiments, the P values were chosen such that the total evapotranspirative demand corresponded to measured values taken on containers from 0800 to 1400 h and that q corresponded to about 7% of the total evapotranspiration. This proportion was established from the amount of water lost between 1400 and 2000 h, that lost being assumed to be largely dominated by the evaporation process while that lost from 0800 to 1400 h was considered to be largely dominated by plant transpiration. The simulation used also the {alpha}1, {alpha}2, {psi}b parameters calculated from the K({psi})–{psi} curves determined on the average of three pots at different matric heads (see below). Only the two most extreme substrate treatments (Sp60B30 and Se30B60) were used for clarity in representing the simulated data, as the Sp30B60 was close to the Se30B60 treatment.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 Mathematical Model of Capillary...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Substrate Characteristics
Particle-size distribution. Dry sieving indicated significant differences in particle-size distribution among the three substrates (Table 1). The MWD also differed significantly (Table 2). As expected, the Sp60B30 tended to have a higher proportion of finer material (all classes <4 mm) than the Sp30B60 medium and the Se30B60 treatment, while the Sp30b60 and Se30B60 treatment had a significantly higher proportion of coarser material (larger than 4 mm). The MWD of the Sp30B60 (5.13 mm) and Se30B60 (4.27 mm) were indeed significantly much larger than that of the Sp60B30 (2.93 mm), with subsequent differences in their hydraulic properties as discussed below.


View this table:
[in this window]
[in a new window]
  		Table 1. Mass fraction (g g–1) of particles remaining on each sieve. Each value is the mean of three replicates.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Initial substrate characteristics.

 
Water and Air Properties. Substrate manufacturing resulted in differences in physical properties, very pronounced for air-filled porosity. Upon both desorption and rewetting, the substrate with the higher amount of peat had the lowest air-filled porosity (Table 2 and Fig. 3). Easily available water did not differ significantly though, as gas diffusivity (Table 2). Average values suggested that these substrates had no aeration problem as air-filled porosity was well above the 0.20 to 0.30 cm3 cm–3 range (DeBoodt and Verdonck, 1972), and gas diffusivity in the 0.01 through 0.02 optimal range identified for Prunus xcystena (Allaire et al., 1996).



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 3. Water desorption curves of the different substrates at the beginning, before potting. Each point is the mean of three replicates and the error bar is the standard deviation of the data.

 
The easily available water of a substrate (DeBoodt and Verdonck, 1972), likely to accommodate a plant during high evapotranspirative demand, was within the 0.20 to 0.30 cm3 cm–3 ideal value identified for all three substrates on desorption (Table 2). However, it was well below these values on sorption. Plant may therefore have suffer from water stress periodically, since the easily available water was very low on sorption and since the maximum volume of water stored within a pot between each irrigation (7.8 mm a day per unit surface area irrigated, neglecting canopy interception) was insufficient to compensate for the daily plant evapotranspiration (25 mm, averaged over the pot surface). Proper plant growth then very likely relied on proper capillary rise from the mat and plant water stress would have been inversely proportional to the capillary rise of stored water from the mat into a container. This movement is linked to a good pot-mat interface but also to an adequate substrate unsaturated hydraulic conductivity.

Unsaturated hydraulic conductivity. Unsaturated hydraulic conductivities in the –5 to –16 cm matric head range were significantly (P < 0.05) lower for substrates composed of 30% peat (Fig. 4) than for the Sp60B30 treatment. The K({psi}) data for both substrates (Fig. 4a) could be fitted reasonably well with the same two-piece exponential curve [Fig. 4a; K({psi}) differences were not significant]. Also, the curve was more variable between replicates for the substrates composed of 30% peat than the Sp60B30 substrate, possibly because of a lower probability of good particle-to-particle contact in coarse media. Because of these significant differences in K({psi})–{psi}, another two-piece exponential curve was fitted to the Sp60B30 substrate and the shape of the curves of the unsaturated hydraulic data were used to investigate whether or not the unsaturated hydraulic conductivity curve could explain the difference in matric head profile observed both in the field and in the lab.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 4. Observed K({psi})–{psi} relationship in all three repitions for each individual substrate. The lines were fitted by segmented regression, using all the values for a given substrates, after a preliminary estimation of {psi}b, later recalculated from the intersection of the fitted curves. They represents the modeled K({psi})–{psi} relationship used later for the (a) Se30B60 and (b) Sp60B30 treatments.

 
Field Experiment
Time to harvest. Differences in plant size and performance were unlikely a result of differences in chemical characteristics, as initial pH and electrical conductivity did not show significant differences (Table 2), and supplemental fertilization was consistent across all substrates.

By the first week of November 2001, all substrate treatments had obtained sufficient marketable size and were harvested. Weeks to obtain 92% marketable plants ranged from 22 to 28 and were the shortest for Sp60B30 followed by the Sp30B60 treatment, for both species. By increasing the amount of peat from 30 to 60%, the production period was shortened by 3 wk for viburnum and 1.5 wk for ligustrum. By changing from sedge to sphagnum peat, the production period was shortened by 0.5 wk for viburnum and 2.5 for ligustrum.

The volume of irrigation water applied was also significantly (P < 0.05) less for the Sp60B30 treatment relative to the other two (Table 3). Total volumes of irrigation required to obtain this level of marketability were dependent on the substrate. Irrigation requirements decreased as the substrate peat type changed from Se30B60 to Sp30B60 and much further with the increasing proportion of peat from 30 to 60% sphagnum peat (Sp60B30), showing significant differences for both species.


View this table:
[in this window]
[in a new window]
  		Table 3. Weeks of production and irrigation volume required to raise 92% of the ligustrum and viburnum plants to marketable size

 
Field-observed water regime. Mean water potential values oscillated between container capacity (–6 cm) and –20 cm at night during the first period of measurement, from May 11 to May 13 (Fig. 5). Differences between treatments were not apparent at that time, but developed as the plants grew. As a result, data collected on August 16 to 17 showed significant (P < 0.05) differences. Because irrigation occurred at night, potentials returned to container capacity (around –6 cm) every morning. As the plants transpired water, substantial potential drops appeared within the substrates, being more pronounced on the 17th August due to 1.5 times the evapotranspirative demand of that of the 16th of August. Differences in pressure potential were significant between the Sp60B30 and the Sp30B60 treatment, the drop in matric potential being more substantial as time elapsed (Fig. 5). Despite this average difference, the data were variable and about 30% of the Sp30B60 and Se30B60 potential readings were close to the Sp60B30 values, suggesting that in some pots of the Sp30B60 and Se30B60 treatments, the capillary rise was as good as in the Sp60B30 treatment, consistent with some model predictions (see below).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 5. Matric head measured at mid container height within the three substrates at the beginning of the experiment and before harvest of the Sp60B30 treatment. Each value is the mean of three replicates. The star indicates significant differences in matric head between the Sp30B60 substrate and the other two substrates and the error bars are the standard deviation of the raw data, for the first and the last measurements, to provide an idea of the minimum and maximum variability observed between measurements.

 
The difference in pressure potential between the Sp60B30 and the Se30B60 treatments were never significant (P > 0.05), despite the fact that the plant sizes were very different between the two, as likely were the evapotranspirative rates. Plant size also differed between the Sp30B60 and Se30B60 treatments. Plants from the Sp30B60 were larger than those of the Se30B60 treatment. Plants in the Sp60B30 treatment were very close to being marketable in mid-August and therefore the evapotranspirative demand would be close to the maximum expected in production. It is normal for smaller plants (Se30B60) to have a higher potential because of a lower evapotranspirative demand. However, the fact the Sp30B60 were smaller than the Sp60B30 treatment but experienced a lower pressure potential is consistent with the view that water transfer was limited by the physical properties of the media, in the Sp30B60 substrates. While potentials were measured with porous cups at the mid-height within the container, the potential profile was suspected to be much more different close to the surface, as the surface of most of the substrate in the Se30B60 and Sp30B60 treatments were observed to be much drier, relative to the Sp60B30 treatment.

Laboratory Experiment
Observations in the lab, with no plant present, indicated a pronounced desiccation of the surface in the Se30B60 treatment, as the observed data strongly depart from hydrostatic equilibrium (Fig. 6). This is in agreement with what was observed visually at the top surface of most of the Se30B60 pots in the field. This treatment showed a pronounced desiccation from mid-height and above and the effect was very similar for all columns. Data from the Sp60B30 treatment, on the contrary, rather showed a profile that was very close to hydrostatic equilibrium.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 6. Observed and predicted depth distribution of water potentials within the cylinder in the laboratory experiment for the Se30B60 and the Sp60B30 substrates. All matric head values above the horizontal asymptote were far below –250 cm for the Se30B60 substrate. The Sp60B30 curve is identical to the hydrostatic profile. The top Se30B60 predicted curve used the average value of the Se30B60 K({psi})–{psi} data (see Fig. 4 above) while the bottom one used the data from the one replicate with the largest {alpha}1 value in the K({psi})–{psi}.

 
Predictive Model
Using the substrate characteristics of Fig. 4 and the theoretical model, simulated values indicated that the drying in the Se30B60 substrate should be more pronounced that in the Sp60B30 substrate for a same evapotranspirative demand (Fig. 7). At mid-height in a container, the model predicted potential values close to hydrostatic equilibrium for the Sp60B30 substrate, independent of the evapotranspirative demand. The model also predicted that at an evapotranspirative equal to 0, both substrates should have potentials value equal to hydrostatic equilibrium (which occurs during the night watering). However, when a slight evapotranspirative demand induced upward water movement, the potential in the Se30B60 substrate dropped sharply and deeper in profile than with the Sp60B30 substrate. At the maximum daily evapotranspirative demand measured, the Sp60B30 substrate maintained relatively wet conditions (as the potentials were close to hydrostatic equilibrium) in the whole pot. In contrast, the Se30B60 substrate typically could only maintain a high potential in the bottom 2/3 of the pot, leaving a higher proportion of the substrates at lower water potentials than the Sp60B30 substrate. Despite the fact that the evapotranspirative demand is transient rather than constant, the model implies that potential at mid-height in the Sp60B30 substrate would maintain itself around hydrostatic equilibrium. The field data also indicated that the potential was maintained close to hydrostatic equilibrium most of the time in that substrate (Fig. 5). However, the potential in that substrate also dropped as low as –50 cm on one particular day. This may be due to the fact that the model took an evapotranspirative demand averaged over 6 h rather than a peak value of describing the flow. Consistent with the model results, the drop in potential should occur faster in the Se30B60 and Sp30B60 treatments than in the Sp60B30 treatment, when evapotranspiration increases. Such a behavior was observed in the field for the Sp30B60 treatment and was particularly pronounced on August 16 and 17 (Fig. 5). This was not observed for the Se30B60 treatment, most likely because the plants in that treatment were of a smaller size than in the Sp30B60 treatment.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 7. Simulated values of the matric potential in the Sp60B30 and the Se30B60 substrates, using the lowest and highest measured evapotranspirative demand during the course of the experiment.

 
The same {alpha}1, {alpha}2, {psi}b, and Ks parameters of the Se30B60 substrate and the theoretical model were used with the lab experiment under steady state conditions, in which a much finer resolution in the matric head profile was obtained (Fig. 6). Both the observed and predicted potential profiles demonstrated the same drying pattern. However, model results predicted drying to occur 10 cm above than what was observed. This difference was possibly due to the use of average K({psi})–{psi} properties, as using the K({psi})–{psi} of the one replicate with the largest {alpha}1 value produced predicted values close to the observed one. Since the K({psi})–{psi} properties were not measured on the laboratory column themselves, it is difficult to assess if it is the only cause for the observed discrepancy.

Deriving Norms to Guide Substrate Manufacturing
To date, no standards that would be useful to guide the substrate manufacturing process have been proposed for substrates used with capillary mat systems in nurseries. However, it may be useful to use the model, taking advantage of the average measured properties and the mean evapotranspirative demand observed over the 0800 to 1400 h period (0.51 mm h–1), measured in the field in May and August, to identify key parameters that describe the K({psi})–{psi} relationship that controls the drying shape of the matric head profile in 15 cm high pots (Fig. 8). The K({psi})–{psi} curve of the Sp60B30 treatment is proposed as a first reference, in that this substrate maintains a high potential in 15 cm high substrates. However, such reference value remains preliminary. Drying patterns will vary with many factors (container height, evapotranspirative demand, root distribution, etc.).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 8. Simulated values of the matric potential in the Sp60B30 substrate, using the fitted K({psi})–{psi} relationship of Fig. 4b and the average measured evapotranspirative demand during the course of the experiment. The (a) {alpha}1, {alpha}2, (b) {psi}b and Ks value were varied from 0.5 to 1.5 of their fitted value.

 
The parameters of the K({psi})–{psi} relationship curve of Sp30B60 are used to see which parameters ({alpha}1, {alpha}2, {psi}b, and Ks) should be changed to achieve a head profile close to the reference Sp60B30 treatment, which gave the best yield and maintained a wet profile close to hydrostatic equilibrium, with the least volume of supplemental irrigation. A change of 50% in {alpha}1, was chosen because it corresponded to changes observed between replicates for the Se30B60 treatment. It is seen that when changing the two {alpha}'s, the modification of {alpha}1 caused by far the most drastic changes in the matric head profile. A decrease of {alpha}1 by 50% leads to a nearly hydrostatic matric head profile, while a 50% increase caused a substantial drying in the top two thirds of the Se30B60 substrate (Fig. 8a). This also illustrates that due to natural variability, important drying may be observed in the pot with such a substrate. This may explain why in the field experiment about 30% of the Sp30B60 substrate potential data showed value similar to the Sp60B30 potential readings (see above). This may also explain the discrepancy for the Se30B60 treatment in the laboratory experiment. Changing {alpha}2 resulted in much smaller changes.

Modifications to {psi}b caused the second-most substantial changes in the matric head profile. A 50% decrease in {psi}b (from –3.32 to –1.66 cm) lead to a nearly hydrostatic equilibrium, while its increase by 50% (from –3.32 to –4.98) predicted a substantial drying in about 60% of the substrates volume (Fig. 8b). Relatively, changes in Ks brought very little changes in the matric head profile.

It therefore appears that until additional research is performed, the characteristics of the Sp60B30 substrate could be proposed as a baseline for a ‘good substrate’. This substrate had the best plant growth and potential for water conservation among the substrates tested. The composition of this substrate is close in term of properties to the recommended composition for subirrigation systems in greenhouses like flooded bench (75% sphagnum, 25% coarse fragment like perlite) in which peat dominates the composition of the media (Wever, 1991).

The substrate's K({psi})–{psi} was characterized by {alpha}1 = 96 m–1, {alpha}2 = 3.7 m–1, {psi}b = –5.08 cm. The value of {alpha}1 should be at most 96 m–1 and that of {psi}b no less than –5.08 cm, as any increases in {alpha}1 and decrease in {psi}b had drastic consequences in the matric head profile predicted in a container at steady state conditions.

This study also indicates that obtaining estimates of these hydrological parameters ({alpha}1, {alpha}2, {psi}b and Ks) is critical in assessing substrate performance. Such parameters remain difficult to estimate properly because of the particularly sensitive structure of growing media and characterization efforts should be made on containerized substrates, using the same filling procedures to those used with plants. Alternatively, container-grown plants could be used for such characterization, as long as the containers have an open bottom and the plant shoot is removed before conducting the characterization. It is critical to recognize that the relationship is strongly hysteretic by nature, thus the characterization of these physical properties should be performed from dry conditions to saturation (and not in the other direction) for meaningful results. Such a characterization requires though the development of a specific methodology (Caron and Elrick, 2005). This study also points out the need to perform such characterization on many replicates to assess its variability.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 Mathematical Model of Capillary...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
It is shown in this study that increasing the proportion of fines in bark-dominated media increased subirrigation efficiency, likely increasing growth rates, thereby reducing the growth period. The study also derived a first set of norms for ideal substrates on capillary mat systems used in outdoor nurseries, based on in situ characterization of the unsaturated hydraulic conductivity for a substrate on rewetting. Finally, the study proposes a steady state model for describing the expected efficiency of some substrates operating on a capillary mat, which confirmed trends observed in the field and in the laboratory.


   ACKNOWLEDGMENTS
 
The authors thank the University of Florida, Laval University, the Natural Sciences and Engineering Research Council of Canada, Soleno Textiles and Premier Horticulture for their financial support. Thanks are extended to A. Deschênes, P. Bussières, and E. Bloodnick from Premier Horticulture Inc., and R. Théorêt from Soleno for technical support, K. Keller, E.J. Sterling, S. Eshom, from the University of Florida, and J. Grandmaison, D. Brisson, and J.T. Denault from Laval University for their help.

Received for publication March 16, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Mathematical Model of Capillary...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
J. S. Price, P. N. Whittington, D. E. Elrick, M. Strack, N. Brunet, and E. Faux
A Method to Determine Unsaturated Hydraulic Conductivity in Living and Undecomposed Sphagnum Moss
Soil Sci. Soc. Am. J., February 15, 2008; 72(2): 487 - 491.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
M. S. Lamhamedi, L. Labbe, H. A. Margolis, D. C. Stowe, L. Blais, and M. Renaud
Spatial Variability of Substrate Water Content and Growth of White Spruce Seedlings
Soil Sci. Soc. Am. J., December 2, 2005; 70(1): 108 - 120.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
J. Caron and D. Elrick
Measuring the Unsaturated Hydraulic Conductivity of Growing Media with a Tension Disc
Soil Sci. Soc. Am. J., May 6, 2005; 69(3): 783 - 793.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow