ArticlePDF Available
Chapter 2
Substrates and Their Analysis
Michael Raviv1, Rony Wallach2, Avner Silber3 and Asher Bar-Tal3
1Agricultural Research Organization, Dept. of Ornamental Horticulture, Newe Ya’ar
Research Center, P.O. Box 1021 Ramat Yishay 30095, ISRAEL
2 The Seagram Center for Soil and Water Sciences, Faculty of Agriculture, The Hebrew
University of Jerusalem, P.O. Box 12, Rehovot 76100, ISRAEL
3Agricultural Research Organization, Dept. of Plant Nutrition and Soil Chemistry, Institute
of Soils and Water, Volcani Center, P.O. Box 6, Bet Dagan 50250, ISRAEL
Abstract
2.1. Introduction
2.2. Physical Properties
2.2.1. Bulk density
2.2.2. Particle size distribution
2.2.3. Total porosity and its components
2.2.4. Air-filled porosity and root respiration
2.2.5. Pore distribution
2.2.6. Water release curve
2.2.7. Hydraulic conductivity
2.2.8. Unsaturated hydraulic conductivity
2.2.9. Effects of substrates’ physical characteristics on irrigation management.
2.3. Chemical Properties
2.3.1. Charge characteristics
2.3.2. Phosphorus retention
2.3.3. Effects of substrates’ chemical characteristics on nutrition management.
2.4. Substrate Analysis
2.5. Description of substrates
2.5.1. Inorganic substrates
2.5.1.1. Sand
2.5.1.2. Rockwool and glasswool
2.5.1.3. Expanded minerals
2.5.1.3.1. Perlite
2.5.1.3.2. Expanded clay
2.5.1.3.3. Vermiculite
2.5.1.3.4. Zeolites
2.5.1.4. Pumice
2.5.1.5. Tuff and other volcanic materials
2.5.1.6. Synthetic organic substrates
2.5.2. Organic substrates
2.5.2.1. Peat
2.5.2.2. Composts
2.5.2.3. Tree waste products
2.5.2.3.1. Coir dust
2.5.2.3.2. Tree bark
2.5.2.3.3. Sawdust, wood fibre and wood chips
2.6. Literature cited
Abstract
Overall profitability of intensive (especially greenhouse) crops, grown in soilless
substrates, is higher than those grown in soil. This is due to their superior physical and
chemical properties, to their initial low infestation rate with pathogenic pests and due to their
ease of disinfestation among growing cycles. The result is a worldwide rapid expansion of the
use of substrates during the last decades. This chapter deals with the physical properties of
substrates and their effect on irrigation regime and with the chemical properties of substrates
and their effects on plant mineral nutrition. It also describes the main substrates that are
currently in use and the way they are analysed.
High content of available water and an adequate air supply have been considered as the
most important physical characteristics required for container media in order to achieve
optimal growth. Water availability to plant roots is strongly related to the hydraulic
conductivity characteristics of the medium, which, in porous materials, drop dramatically with
reduction in water content. Very high water content (near container capacity) is, thus, a
prerequisite for optimal water availability. Means to maintain desired water content are
discussed.
The main factor that distinguishes fertilisation management of substrate- and soil-grown
plants is the limited volume of substrates, which means a lower buffer capacity for pH and
solution composition and a limited supply of nutrients. The limited root zone volume also
results in decreased root size and increased root density, causing higher competition among
roots and a bigger effect of root activity on the rhizosphere.
Most growth substrates possess negative permanent and/or variable charged surfaces.
Surface charge properties of substrates have a major effect on the chemical reactions taking
place in the rhizosphere, on the availability of applied cations and on their uptake efficiency.
Strong interaction between orthophosphate ions and solid phase constituents reduce
solution P concentrations following fertilizer application. The main two mechanisms involved
are: (i) fast - electrostatic reactions of adsorption onto the solid phase; and (ii) slow -
formation of new solid metal-P compounds (precipitation). Thus, availability of phosphorus in
the rhizosphere often determines the growth and productivity of crops.
Frequent irrigation and continuous fertilisation should satisfy nutritional plant demands
under most practical situations.
2.1. Introduction
Various substrates, such as rockwool, polyurethane, perlite, scoria etc, are virtually pest-
free and can easily be disinfested between growth cycles in case of disease contamination.
They also enable relatively easy control over the pH and the nutritional status of the root zone.
The main advantage of soilless- over soil-cultivation is, however, its physical characteristics
and specifically, its ability to provide simultaneously sufficient levels of both oxygen and water
to the roots. The physical properties of porous substrates are more suitable than those of soils
for the production of most horticultural crops. Much weaker matric forces in substrates,
compared to soil, hold the water. Consequently, plants grown in porous media at or near
container capacity require less energy to extract water and experience a lower risk of oxygen
deficiency, than those grown in a soil near field capacity. The above factors lead to improved
yields in terms of quality and quantity.
Lack of suitable soils, disease contamination after repeated use and the desire to apply
optimal conditions for plant growth are leading to the worldwide trend of growing plants in
media, instead of soil. Most media-grown plants are grown in greenhouses under supposedly
near-optimal production conditions. An inherent drawback of soilless vs. soil cultivation is the
fact that in the latter the root volume is unrestricted while in containerised culture the root
volume is restricted. This restricted root volume has several important effects, especially a
limited supply of nutrients (Dubik et al., 1990; Bar-Tal, 1999). The limited root volume also
increases root-to-root competition since there are more roots per unit volume of medium.
One of the purposes of this chapter is to focus on the main issues of the physical and
chemical environment of the rhizosphere and to identify the areas where more research is
needed in order to obtain a clearer picture. The other purpose is to describe the main
substrates, used in horticulture and to identify their main advantages and drawbacks.
2.2. Physical Properties
All media are composed of 3 phases: solid, aqueous and gaseous. In the following sections,
the physical characteristics of these 3 phases will be discussed separately and in combination.
2.2.1. Bulk density
Bulk density (BD) of a medium is defined as its dry mass per unit of volume (in a moist
state) and is measured in g cm-3. Numerous methods for the measurement of BD (as well as
other physical parameters) can be found in the literature (e.g. de Boodt and Verdonck, 1972;
Gruda and Schnitzler, 1999a; Morel et al., 2000; Raviv and Medina, 1997; Wever, 1995).
Some methods are used primarily for research purposes (e.g. the standard ISHS method, as
described by Verdonck and Gabriels 1992). Others are used as industrial standards in certain
countries or regions of the world (e.g. BS EN 12580:2001 in the UK, both the LUFA and
DIN methods in Germany and the CEN method in the EU). All of them, however, are based
on one principle: Wet material is allowed to settle within or compressed using known pressure
into a cylinder of a known volume. It is then dried down completely and weighed.
As many media are composed of more than one ingredient, the characteristics of each
ingredient contribute to the total BD of the medium. These are individual and combined
particles' arrangements, BD of the ingredients and compaction qualities. In particular, media
components that differ significantly in particle size have higher BDs as a mix (Pokorny et al.,
1986). Similarly, they have lower total porosity (TP), water holding capacity and air-filled
porosity (AFP) than media composed of similar particle sizes.
BD affects the choice of media in various ways. For example, outdoor production of tree
saplings requires high BD media to prevent container instability in windy conditions. This can
be achieved by the inclusion of heavy mineral constituents, such as sand, soil, clay or tuff in
the mix. On the other hand, high-intensity greenhouse crops, which are frequently irrigated
and may be exposed to oxygen deficiency if hydraulic conductivity and AFP are not high,
require media of low BD. Another consideration is that the mixing and transportation of low
BD-media are easier than those of high BD-media.
2.2.2. Particle size distribution
Having defined the medium as a three-phase system, the solid phase constitutes the stable
component of the medium and the one that gives substance to the whole. The material of
which the medium solid phase is composed includes discrete mineral particles of various
shapes and sizes, as well as amorphous compounds such as organic matter. The array of
particles can be divided into groups according to size, and the medium solid phase as a whole
can be characterised in terms of the relative proportions of its particle size groups. The particle
size and mineral composition largely determine the nature and behaviour of soil: its internal
geometry and porosity, its interactions with fluids and solutes, as well as its compressibility
and strength. In this section we concentrate on the properties of the medium solid phase that
are intrinsic to the material itself and are unaffected by external variables (fluid fluxes, energy
fluxes, and mechanical stress causing deformation). These include texture and particle size
distribution.
The term medium texture refers to the size range of particles in the medium, that is whether
the particles composing a particular medium are of a wide or relatively narrow range of sizes,
and whether they are mainly large, small or of some intermediate size. As such, the term
carries both qualitative and quantitative connotations. Qualitatively it indicates whether the
material is coarse and gritty, or fine and smooth. Quantitatively, it denotes the precisely
measured distribution of particle sizes and the proportions of the various size ranges of
particles within a given medium. As such, medium texture is an intrinsic attribute of the
medium and the one most often used to characterize its physical makeup.
The particle size distribution attempts to divide what generally in nature is a continuous
array of particle sizes into discrete fractions. The cases of tuff and perlite will be presented
here as examples. Typical particle size distribution curves for two types of tuff (scoria) (data
from Wallach et al., 1992a) and perlite (data from Orozco and Marfa, 1995) are shown in
Figure 2.2.2(1). The particle size distribution is recorded by noting the fraction of the total
mass that is made up of particles smaller than a given size. Note that the graph gives an
integral, or cumulative representation. As is discussed later in detail, the shape of particle size
distribution is very useful in estimating the hydraulic properties of the media, such as the water
retention characteristics and the hydraulic conductivity.
The information obtainable from this representation of particle size distribution includes the
diameter of the largest grains in the assemblage, and the grading pattern; that is whether the
substrate is composed of distinct groups of particles each of uniform size, or of a more or less
continuous array of particle sizes. Poorly graded media have a dominance of particles of
several distinct sizes and are represented by a step-like distribution curve. A medium with a
flattened and smooth distribution curve is termed well graded. The particle size distribution of
a porous medium may also be expressed by way of summary statistics of the distribution.
These statistics are less complete in their information than the full distribution, but are often
sufficiently revealing to be useful and more convenient. The most commonly used statistic is
d50, defined as the diameter of the particle for which half of the mass of the soil is made up of
smaller particles (and half larger). d50 is 1.5 and 4.6 mm for the RTB and RTM tuff,
respectively (Figure 2.2.2(1)a), and 0.82 and 2.95 mm for the B12 and A13 perlite,
respectively (Figure 2.2.2(1)b). The same notation is used to define other parameters of the
distribution, such as d10 and d60, for describing the diameter of particles for which 10 and 60%
of the mass of particles are smaller. Having provided a number that defines the size of the
“average” particles, it is also useful to provide an indication of how widely the particle size
distribution is spread. To fulfil this need, the uniformity coefficient, U=d60/d10 is calculated. U
is 1 for equally sized particles but is bigger than 1 for most soils and substrates. The uniformity
coefficient, U, is 2.7 and 16.8 for A13 perlite and B12 perlite, respectively (Figure 2.2.2(1)b).
According to the calculated U values, A13 perlite is poorly graded compared to B12 perlite.
Based on the shape of tuff particle size distribution (Figure 2.2.2(1)a), one can conclude that
RTB tuff, whose distribution curve is more step-like, is poorly graded relative to RTM tuff. In
fact, RTM is a red tuff consisting of particles ranging in size from 0 to 8 mm, obtained from a
tuff strip mine by sieving the natural product, whereas RTB is a coarser red tuff consisting of
particles in the same size range, obtained by crushing larger particles (diameter of 30 - 50
mm), followed by sieving the crushed material through an 8 mm sieve.
Admixtures of constituents having a variable particle size distribution may show
unexpected relative gas diffusion values. Nkongolo and Caron (1999) mixed 40% wood bark,
50% peat, and 10% coarse gravel, on a volume basis. Wood bark particle size was varied over
a wide range (0-2, 2-4, 4-8 and 8-25 mm in one experiment and 1-2, 2-4, 4-8 and 8-16 mm in
a second experiment). Although AFP was similar in all mixtures, relative diffusivity decreased
considerably when coarse particles (8-16 or 8-25 mm) were used. This is probably a result of
the increased tortuosity that characterises poorly graded mixtures.
2.2.3. Total porosity and its components
Total porosity (TP) and its components are expressed as a percentage of the total volume
of the medium. The combined volume of the aqueous and the gaseous phases of the medium
are defined as its total pore space or total porosity. TP is related to the shape, size and
arrangement of media particles. It can be divided into AFP and water-holding capacity. The
various methods of measuring BD are also relevant to the determination of porosity and its
components and will not be discussed further here.
In many cases BD is inversely related to TP (Bugbee and Frink, 1983, Bunt, 1988).
However, the medium BD cannot accurately determine TP if components that have closed
pores, such as perlite or pumice, are used (Bunt, 1988; Bures et al., 1997b). Not only mineral
substrates contain inaccessible pores. Evans et al. (1996) sampled different coir batches.
Contrary to their expectations, the samples exhibited lower a BD with lower TP. This, of
course, can only happen if a significant fraction of the pores was not saturated with water
during the saturation stage of the measurement, suggesting that those pores were probably of
a strong hydrophobic nature.
The volumetric amount of water, θ, which saturates a given volume of a substrate, is
defined as its effective pore space. The difference between total pore space and effective pore
space constitutes the volume of closed pores that are not accessible to water. Container
capacity is defined as the amount of water remaining in the container after water stops
draining following saturation. The amount of easily available water in a medium is defined as
the difference in water content between container capacity (usually defined at a water suction
of 1 kPa) and water content at 5 kPa. Water buffering capacity is defined as the water content
between 5 kPa and 10 kPa. Water held by media at tensions higher than 10 kPa are usually
considered as unavailable (de Boodt and Verdonck, 1972).
The capacity of a medium to store water and air, as well its ability to provide them to the
plant (via its hydraulic conductivity and rate of gas exchange) are determined by its TP and
porosity characteristics, namely pore size distribution, tortuosity and pore continuity. In
general, TP is described in this section while its characteristics are discussed in section 2.2.4.
Water is mainly held by the micropore space of a growth medium, while rapid drainage and
air entry is facilitated by macropores (Drzal et al., 1999). Therefore, an adequate distribution
of large and small pores is essential for a good medium. Since pore size and distribution
determine the rate of water drainage and gas exchange, these factors are critical in defining a
growing medium with optimum physical characteristics. Most soilless growth media contain
60% to 90% total pore space. The components of TP are discussed in the following and AFP
in section 2.2.4.
Roots of container-grown plants have access to a restricted volume, which is small in
comparison to the amount of water lost via evapotranspiration (ET). For example, under
Mediterranean conditions rockwool-grown roses transpire up to 10 litres per day per m2 while
the amount of easily available water (EAW) is about 3 litres per m2 (Raviv, unpublished
results). For this reason the content of EAW in the medium is of importance.
The static approach, still dominating the scientific horticultural literature, suggests that
volumetric EAW should be in the range of 20-30% (de Boodt and Verdonck, 1972). EAW
may be restricted due to:
the existence of too many macropores, enabling water to leave the medium by
gravity (e.g. in bark and pumice).
the existence of too many micropores, increasing the fraction of unavailable water
(e.g. in media rich in clay or soil).
a combination of the above.
The term EAW is based on the hidden assumption that the difference in water suction
in the range of 1 5 kPa is of no physiological significance, as it is far higher than typical
suction pressures applied by plant roots (normally in the range of 0.1 0.4 MPa). Thus, this
term does not discriminate between water availabilities along the retention curve in the above
range. However, the inherent restriction on water flux in many soilless media is not matric
potential. In fact, water flux is limited by unsaturated hydraulic conductivities [K(h)] within a
large part of this range. As K(h) greatly declines in most porous media with suction, the static
approach of EAW has to be replaced by a far more accurate and meaningful expression of
water availability. This subject is thoroughly discussed in section 2.2.7.
2.2.4. Air-filled porosity and root respiration
Under intensive cultivation with a high canopy/root ratio, maintaining low suction in the
medium has a beneficial effect on plant productivity. This is probably due to the effect of high
tsoil on photosynthate distribution between the root system and the canopy (Dasberg and
Feigin, 1978). However, under these conditions, a lack of oxygen may pose a severe problem
in many soil types, while this risk is rare in most porous media.
Horticulturally speaking, AFP is defined as the volumetric percentage of the medium filled
with air at the end of free (gravitational) drainage. This value varies greatly with the height and
shape of the container, so convention determines that AFP is the volumetric percentage
occupied by air at a pressure head of 10 cm (or water suction of 1 kPa) (de Boodt and
Verdonck, 1972).
Plant roots require a constant supply of oxygen for respiration and at the same time they
release carbon dioxide. Adequate aeration of the root zone is vital for normal plant function.
Active roots are always surrounded by a thin water layer and take up oxygen only in a
dissolved state. Roots are therefore dependent upon constant and fast gas exchange between
the gas and the liquid phases of the medium. Stagnant water cannot supply sufficient oxygen
for most horticultural crops. The process of gas exchange in substrates is firstly driven by
diffusion through the pores of the growing medium along the partial pressure gradient in the
gas phase. Large pores permit air entry into the medium shortly after irrigation. The second
stage of the process is the dissolution of oxygen into the films surrounding the roots. Oxygen
diffusivity in media is about 2 orders of magnitude lower than that in air.
The relations between relative oxygen diffusivity, AFP and biomass production of Prunus x
cistena were measured by Allaire et al. (1996), who found that biomass production correlated
well with relative oxygen diffusivity, but not with AFP. The combination of the two transfer
processes (mass flow and dissolution) can be measured using a combined calomel-platinum
electrode and expressed as the oxygen diffusion rate (ODR). ODR, being a dynamic value,
should be a better parameter than AFP for assessing the degree of aeration of a medium. In
fact, in a study comparing the ODR and AFP of 14 different media, Bunt (1991) showed that a
good correlation exists between these two parameters. Owing to its ease of measurement, the
static parameter of AFP is usually used as a criterion for medium aeration.
Although oxygen diffusion in air is four orders of magnitude higher than in water, oxygen
may be supplied to the rhizosphere in a dissolved state by very frequent mass flow of oxygen-
saturated water. This is one of the main physical bases for hydroponic and aeroponic growing
systems and also for substrate-based hydroponic systems such as the nutrient film technique
(NFT).
Most media and mixes have an AFP of 10% to 30%. Optimal AFPs may vary greatly
according to the size of the container and the irrigation frequency. For the rooting of cuttings
under intermittent mist, AFPs of >20% are essential. A somewhat lower AFP is required for
bedding plants grown in shallow trays or plugs. On the other hand, an AFP as low as 10% may
suffice for deep containers with slow growing, infrequently irrigated tree saplings. For all
types of containers and media, it is important to consider the tendency of most root systems to
grow gravitropically and to form a dense layer of roots at the bottom of the container. This
layer is also the site where a perched water table rests after the end of free drainage. This
might be a reason for less than optimal performance of plants in media that are otherwise
considered adequate in terms of AFP (Raviv et al., 2001).
2.2.5. Pore distribution
Evaluation of the size, configuration, and distribution of media pores is essentially
impossible due to their extremely complicated nature. Large pores generally favour high
infiltration rates and adequate aeration for plant growth. An important problem associated
with the characterisation of soil/soilless media pores is the lack of standard terminology related
to their classification into distinct size ranges. Several researchers have identified the need for
a standard classification scheme, and suggestions for such a classification have been made.
Typical examples are the proposed index for soil pore size distribution by Cary and Hayden
(1973) and the suggested classification of micro-, meso-, and macroporosity by Luxmoore
(1981). Although the pore size distribution of many porous materials could be measured by
different techniques, for example water desorption and mercury intrusion methods (Danielson
and Sutherland, 1986), none of them have been applied to soilless container media.
It is recognized by soil physicists that the soil moisture retention curve (section 2.2.6) is
essentially the pore-size distribution curve. The expression relating pore size to the equivalent
suction of water in porous media is the familiar equation of capillarity:
rg
cos2
(1)
where is the medium water suction, is the surface tension of water, is the contact angle,
is the density of water, g is the acceleration due to gravity, and r is the pore radius. Surface
tension and water density are temperature dependent, but assumed to be practically constant.
The contact angle for water going into mineral materials is often taken to be 00 but it may vary
depending on the organic content of the medium.
2.2.6. Water retention curves
The relationship of soil water content to water suction is of fundamental importance to the
understanding of soil water status and water availability to the plants. Owing to the
characteristic texture of container media and soil substrates, unsaturated condition occurs
soon after or even during irrigation. Therefore, the water holding capacity of the media at
different suctions is a vital tool to analyse water flow, water availability, and irrigation
management. The current section will focus on this feature and its uniqueness for container
media compared with regular soils.
Water availability to roots is largely determined by how tightly water is held by the solid
phase of the medium. The closer a water molecule approaches a solid, the more tightly it is
held through the forces of adhesion and cohesion. The water suction, , combines adhesion
forces between solid soil surfaces and water and cohesion forces among water molecules.
can be measured using a tensiometer. Although a fine medium, having a large specific surface
area may hold more water than a coarse medium, less water may be available to the roots, due
to high m.
When suction is applied incrementally to a saturated media, the first pores to be emptied
are the relatively large ones that cannot retain water against the suction applied. From the
capillary equation (eq. 1) ( = 2γ/r, for =0o), it can be readily predicted that a gradual
increase in suction will result in the emptying of progressively smaller pores, until, at high
suction values, only the very narrow pores retain water. Similarly, an increase in media-water
suction is associated with decreasing thickness of the hydration envelopes adsorbed to the soil
particle surfaces. Increasing suction is thus associated with decreasing media wetness (and a
decrease in the osmotic potential of the aqueous phase). The amount of water remaining in the
media at equilibrium is a function of the sizes and volumes of the water-filled pores and the
amount of water adsorbed to the particles; hence it is a function of water suction. This
function is usually measured experimentally, and is represented graphically by a curve called
the moisture retention curve (RC), also known as the moisture release curve or the media-
moisture characteristic (Klute, 1986).
The traditional method of determining the water RC involves establishing a series of
equilibria between water in the substrate sample and a body of water at known suctions. The
substrate-water system is in hydraulic contact with the body of water via a water-wetted
porous plate or membrane. At each equilibrium, the volumetric water content, , of the
substrate is determined and paired with a value of the matrix pressure head, , determined
from the pressure in the body of water and the gas phase pressure in the substrate. The data
pair (, ) forms one point on a RC. Owing to its unique texture, the RC of a container
medium is usually determined for very low suctions (up to 300 cm) using a suction funnel
(Goh and Mass, 1980), suction table (Ball and Hunter, 1988), or a pressure plate system
(Fonteno et al., 1981). A record of measured RCs for different soilless growing media is given
in Table 2.2.6.(1).
Table 2.2.6.(1): Sources for retention curves (()) and unsaturated hydraulic conductivities
(K()) of representative media.
Material Reference Comments
Sand Riviere (1992) ()
Rockwool da Silva et al. (1995) () and K()
Rockwool Riviere (1992) ()
Perlite Orozco and Marfa (1995) () and K() for different types
available in Spain.
Vermiculite Fonteno and Nelson (1990) ()
Vermiculite Riviere (1992) ()
Zeolite Riviere (1992) ()
Pumice Raviv et al. (1997) () and K()
Tuff - Scoria (RTM, RTB) Wallach et al. (1992a) () and K()
Sphagnum peat moss da Silva et al. (1993a) () and K(), including its
mixtures with tuff
Sphagnum peat moss Heiskanen (1995a, b, 1999) (), including its mixture with
coarse perlite
Sphagnum peat moss Heiskanen (1999) () and K(), including its
mixtures with perlite and sand
Canadian sphagnum peat Fonteno and Nelson (1990) ()
Composted agricultural wastes Wallach et al. (1992b) () and K(), including its
mixtures with tuff
Composted cow manure Raviv et al., 1997, 1998c ()
Coir Raviv et al. (2001) () and K()
Pine bark Fonteno and Nelson (1990) ()
Pine bark mixed with hardwood
bark
Bilderback (1985) ()
Sawdust Goh and Haynes (1977) ()
UC mix Raviv et al. (2001) () and K()
Peat and vermiculite mix (1:1) Fonteno (1989) ()
Pine bark peat and sand mix
(3:1:1)
Fonteno (1989) ()
RCs of container growing media have an intrinsically different shape than field soils, mainly
due to the low suction range that usually exists in container-growing media. As an illustration,
RC for RTM and RTB tuff measured over a 0 to 120 cm suction range (Wallach et al.,
1992a), are presented in Figure 2.2.6(1). A notable characteristic is the high value of saturated
water content, s, as discussed in section 2.2.3. Usually, as suction increases from zero, the
substrate remains saturated up to a critical suction at which the largest pores begin to empty
and their water content is displaced by air. This critical suction is called air-entry suction. The
typical height of containers used for soilless culture ranges from 5 to 20 cm and substrates
having air-entry suctions in this range will keep most of the container volume almost saturated
with low air content after irrigation. Temporary shortage of oxygen can reduce root and shoot
growth, and anaerobic conditions for only a few days will result in the death of some roots.
However, the air-entry value of soilless container media is much lower than that of soils. It is
seen in Figure 2.2.6(1) and in other measured RC for container media (list appears in Table
2.2.6.(1)) that the air-entry suction is very small and can be practically regarded as zero. Since
air content in the porous media depends on the content of water, small values of air-entry
suctions ensure that during free drainage of a saturated medium, many pores become empty
and air invades the pore volume occupied by the draining water.
Beyond air-entry suction, any suction increase is associated with a decrease in water
content. The rate of moisture content decrease per unit suction increase differs significantly
between container media and soils, being steeper in the former. The sharp decrease in Figure
2.2.6(1) is initiated at very low suction since the air-entry value is close to zero. The moisture
content decrease occurs over a very narrow suction range (up to 25 cm) and subsequently
reaches a constant value. The sharp decrease in moisture content toward a final value shapes
the characteristic hydraulic conductivity curve for container media, as will be discussed in the
following.
Dissimilarities in RCs of container media and regular soils are not solely due to differences
in their particle size. This is illustrated in Figure 2.2.6(2), which shows the RCs measured for a
1-2 mm fraction of tuff and similarly textured sand (Wallach et al., 1992a). In spite of the
similar particle size, the tuff porosity (0.58 cm3 cm-3) is much higher than the sand porosity
(0.30 cm3 cm-3). The porosity disparity is attributed to the inner microporosity of the tuff
particles (Chen et al., 1980), which is not present in the sand. Sand has a typical S shape RC,
which characterises soils of different types, whilst the tuff RC lacks the first vertical part of the
S shape and looks more like a decreasing hyperbolic function, similar to the RCs shown in
Figure 2.2.6(1). The value of the air-entry suction of the tuff could not be determined from the
measured RC. It is probably a few millimetres compared to 10 cm for the sand. The difference
in air entry value can be attributed to the particle shapes, which affect their spatial distribution
in the container. Sand particles are regular and smooth compared to tuff particles, which are
rough and irregular. The water content approaches 0.125 and 0.02 for tuff and sand,
respectively, when suction increases beyond 30 cm. The difference in the residual water
content values can be attributed to the inner microporosity of the tuff particles.
Measured values of water content and suction ( - ) are often fragmentary, and are
usually based on relatively few measurements over the wetness range of interest. For
modelling and analysis purposes and for the characterisation and comparison of different
substrates and scenarios, it is essential to represent the RC in continuous and parametric form.
A parametric expression of a RC model should contain as few parameters as possible to
simplify its estimation and describe the behaviour of the RC at the limits (wet and dry ends)
while closely fitting the non-linear shape of the - data.
In recent years, efforts have been made to apply mathematical models to describe the RC of
container media. Milks et al. (1989 a, b, c) used a cubic polynomial function to describe
retention data collected for five porous materials commonly used as container media. The
cubic polynomial function failed to describe data measured on media exhibiting very high
values of s and a sharp decrease in moisture content with increasing suction.
van Genuchten (1980) proposed the following equation to represent the θ() function,
]
)
( + [1 =
Sm
n
e
 (2)
where Se is the volumetric water content in a relative form, which is sometimes called effective
saturation, defined as
=
S
rs
r
e
(3)
In eq. (2), , n and m are parameters that determine the shape of the retention curve. These
parameters are determined by curve-fitting techniques. The subscripts s and r in eq. (3) refer
to the saturated and residual values of θ, respectively. The latter is defined as the water
content at which the gradient dθ/d becomes zero (excluding the region near θs, which may
also present a zero gradient). Both θs and θr can be either measured or estimated along with ,
n and m. As opposed to θs that has a clear physical significance, the meaning of θr and its
estimation has not yet been resolved. Stephens and Rehfeldt (1985) reported improved model
accuracy using a measured value of θr. Fonteno and Nelson (1990) determined θr as the water
content at = 300 cm. However, Ward et al. (1983) and van Genuchten (1978, 1980)
suggested that θr should be viewed as a fitting-parameter rather than a soil property. van
Genuchten and Nielsen (1985) considered not only θr but also θs to be empirical parameters
that should be fitted to the - data. The van Genuchten model can be used in conjunction
with predictive models for unsaturated hydraulic conductivity, as will be discussed in section
2.2.8.
For small m/n ratios, parameter in eq. (2) approximately equals the inverse of the air-
entry value. For large values of m/n, this parameter roughly equals the inverse of the suction at
the inflection point of the retention curve (van Genuchten and Nielsen, 1985). Parameter n is
related to the pore size distribution of the medium and the product mn determines the slope of
the () curve at large suction values. Therefore, n may be viewed as being mostly affected by
the structure of the medium (van Genuchten and Nielsen, 1985).
The equation proposed by van Genuchten (eq. 2) was applied to container media by Milks
et al. (1989a, b, c) who found that it fits the measured - θr data better than the cubic
polynomial. It has also been used by Fonteno (1989), Wallach et al. (1992 a, b), da Silva et al.
(1993a, 1995), Orozco and Marfa (1995) and Raviv et al., (2001). Fitted RCs, using eqs. (2)
and (3) are shown in Figures 2.2.6(1) and 2.2.6(2) for the RTB and RTM tuffs and the 1-2
mm fraction of tuff and sand, respectively.
There are two initial stages at which retention curves are measured: A drainage curve is
obtained by establishing a series of equilibria by drainage from zero pressure head. A wetting
curve is obtained by equilibrating samples wetted from a low water content or high suction.
The retention curve is hysteretic, i.e. the water content at a given suction for a wetting
substrate is less than that for a draining substrate (Wallach et al., 1992a). The drainage curve
that starts on complete saturation of the substrate is called the initial drainage curve. The main
wetting curve is obtained by wetting the substrate from low water content. The features of the
hysteresis of the retention curve are shown in Figure 2.2.6(3) for rockwool (da Silva et al.,
1995). The measured initial drainage and main wetting curves were successfully fitted by eq.
(2), (da Silva et al., 1995), as shown in Figure 2.2.6(3). As the substrate is wetted along the
main wetting curve and the suction approaches zero, the water content approaches a value
that is less than the TP, s, due to the presence of entrapped air (Figure 2.2.6(3). This value is
about 0.8s to 0.9s and is called the natural saturation of the satiated water content (Klute,
1986). The drainage curve starting at natural saturation is called the main drainage curve. The
main drainage curve merges asymptotically with the initial drainage curve as the suction
increases.
The main hysteresis mechanism is referred to as the inkbottle effect, because of its similarity
with the behaviour of bottles of ink with very narrow spouts. Another source of hysteretic
behaviour is hysteresis in the contact angle. The contact angle often exhibits a different value
in advancing and receding cases. A central question is when and why hysteresis must be
included when describing container media processes. The answer to this question is not yet
clear given that few studies have been performed on hysteresis in soilless container media. The
adjustment of the theory used in soil physics and its application to container media is not
straightforward. This is due to the special conditions that exist in the limited container volume
where: 1) wetting and drying frequently occur, 2) special conditions that exist at the container
boundaries (the bottom may be open to the atmosphere or in some cases continuously
saturated), and 3) the unique textural and structural properties of soilless substrates compared
to field soils. Thus, additional research is needed to increase knowledge of the effects of
hysteresis on static and dynamic distributions of water in containers filled with soilless media,
and on the availability of water to the plant.
2.2.7. Hydraulic conductivity
The hydraulic conductivity of a saturated porous medium is better explained by using first
the equation for flow in such porous media, known as Darcy’s law :
L
H
AKQ
(4)
where Q is the volume of water that passes through the medium (column) [cm3 per unit
time], K is the hydraulic conductivity [cm per unit time], A is the cross-sectional area of the
column [cm2], and H is the difference between the hydraulic head at the inlet boundary and
the hydraulic head at the outlet boundary and L is the length of the column or media thickness.
From eq. (4) one obtains:
H
Lq
K
(5)
q is the flux density – the rate of water movement through the medium.
From eq. (4) and (5) it is seen that the hydraulic conductivity, K=kg/, is a function of the
properties of both the medium and the fluid. K can be separated into two factors: fluidity
(defined as /g), and intrinsic permeability. The two fluid properties that directly affect the
hydraulic conductivity are viscosity () and density (). The intrinsic permeability (k) of a
medium is a function of pore structure and geometry. Particles of smaller-sized individual
grains have a larger specific surface area, increasing the drag on water molecules that flow
through the medium. Thus, the result is a reduced intrinsic permeability and K.
For porous media that do not interact with fluids and change fluid properties or vice versa,
the same value for k will be obtained for different fluids. However, if fluid-media interactions
alter the medium structure (e.g. swelling that reduces the pore space available for flow), the
intrinsic permeability can be altered greatly. Another cause of permeability reduction is
entrapped air. As a previously unsaturated medium is saturated with water, some air may be
trapped within the system. This trapped air can form bubbles of varying sizes within the
medium, depending on the amount of air present and pore size; such bubbles can obstruct the
flow of water. Microbial activity may also cause a reduction of hydraulic conductivity. The
bacteria may clog the medium, or may induce chemical reactions that produce clogging slimes.
Measurements of hydraulic conductivity of a saturated medium are based on the direct
application of the Darcy law (eq. 4) to a saturated medium column of uniform cross-sectional
area. A hydraulic head difference is imposed on the column and the resulting flux of water is
measured. A constant or variable head difference is maintained along the experiment (Klute
and Dirksen, 1986). The Darcy equation is not valid for all flows in porous media as can be
seen in Figure 2.2.7(1) where measured flux density as function of the hydraulic gradient
(Wallach et al., 1992a) is shown for RTB and RTM tuffs. The departure of the measured
curve from the linear line obtained by the Darcy law, eq. 4, indicates that the latter does not
apply at large flow velocities. As flow velocity increases, especially in systems with large
pores, the occurrence of turbulent eddies or non-linear laminar flow results in dissipation of
effective energy by the internal mixing of the liquid. As a result, the hydraulic potential
gradient becomes less effective in inducing flow. Darcy’s law is generally applied when the
Reynolds number, defined as the ratio between the inertial forces (as given by the product of
fluid density, flux density, and median pore size), to the viscous forces (as given by the fluid
viscosity) is less than about 1 (Hillel 1998). At a higher Reynolds number, the inertial forces
become significant relative to viscous forces. Owing to the coarse texture of soilless container
media high velocities are obtained even at relatively low hydraulic gradients and Darcy’s law
cannot be used to calculate the saturated hydraulic conductivity.
2.2.8. Unsaturated hydraulic conductivity
The conductive properties of unsaturated media are markedly dependent on their texture
and structure. At saturation, all of the pores are water-filled and conductive. The most
conductive media are those in which large and continuous pores constitute most of the overall
pore volume, whereas the least conductive are media in which the pore volume consists of
numerous micropores. When the growing medium desaturates, some of the pores become air-
filled so that the conductive portion of the cross-sectional area diminishes. Furthermore, as
suction develops, the largest pores are the first to empty and become nonconductive. Thus, a
steep decrease in the initially high hydraulic conductivity is expected in soilless container
media that have a high percentage of large pores.
Although the saturated hydraulic conductivity, Ks, is usually determined with little
difficulty, many technical problems are involved in the measurement of the hydraulic
conductivity under unsaturated conditions. It would be desirable to determine K() or K() by
direct measurement. However, this is often not possible because the values of the hydraulic
conductivity may vary by several orders of magnitude within the water content range of
interest. In addition, the measurements are tedious and expensive, and most measurement
systems cannot efficiently cover such a wide range of variation. This is one of the reasons why
the hydraulic characterisation of container media has been based on measured retention
curves, (), and saturated hydraulic conductivity, Ks.
Given a curve-fitting equation for (), the unsaturated hydraulic conductivity can be
directly calculated by means of a predictive equation based on the fitted () curve, and a
single measurement of the saturated hydraulic conductivity, Ks (section 2.2.7). The
combination of a curve-fitting equation for () with a predictive model for K produces the
structure of a combined model for the determination of the substrate’s hydraulic properties.
To obtain an accurate predictive equation for the unsaturated hydraulic conductivity, an
analytical expression that accurately describes the medium water retention curve, over the
whole relevant suction range, is required. This prerequisite is essential, and any attempt to
predict the unsaturated hydraulic conductivity from retention data will fail if the assumed
function cannot describe the data over the whole range of interest (van Genuchten and
Nielsen, 1985). The accuracy of the predictive equation for unsaturated hydraulic conductivity
also depends on the theory and assumptions on which that equation is ultimately based.
Many methods have been developed to estimate unsaturated hydraulic conductivity from
the more easily measured water retention curve. In the following we shall limit our discussion
to the K() function developed by Mualem (1976). Other expressions could be found in soil
physics textbooks (e.g. Hillel, 1998). According to Mualem (1976), the K() function is:
dx
(x)
1
dx
(x)
1
S
= )
S
(
K
2
1
0
S
0
½
eer
e
(6)
where Kr = K/Ks is the relative hydraulic conductivity, Se is the effective saturation (eq. 3), and
x is an integration variable.
Of the various possibilities for restricted m-n cases, the relation m = 1-1/n was shown to
perform better than other cases (van Genuchten and Nielsen, 1985). Solving eq. (2) for =
(Se), substituting the resulting expression into eq. [6] and assuming the above unique
relationship between the parameters n and m leads to:
])
S
- (1 - [1
S
= )
S
(
K2
m
1/m
e
½
eer [7]
where 0 < m < 1. In terms of , eq. (7) becomes:
]
)
( [1
]]
[1
)
( [1
= )(K m/2n
2m-
n
1-n
r

(8)
Equations (7) and (8) describe the unsaturated relative hydraulic conductivity model of
Mualem (1976) as combined with the retention curve model of van Genuchten (1980).
Multiplying the value of Ks by Kr() will provide the K() function.
Predicted hydraulic conductivity curves, K(), for RTB and RTM tuff are shown in Figure
2.2.8(1). The measured retention curves of these substrates shown in Figure 2.2.6(1) were
used for this prediction. Predicted hydraulic conductivity curves for other soilless growing
media can be found: in Wallach et al. (1992b) for composted agricultural wastes and their
mixtures with tuff; in da Silva et al. (1993a) for sphagnum peat moss and its mixture with tuff;
in da Silva et al. (1995) for rockwool (whose retention curve is given in Figure 2.2.6(3)); in
Orozco and Marfa (1995) for perlite, in Raviv et al. (1999) for two types of pumice and in
Raviv et al. (2001) for coir and UC mix. For both types of tuff, K() decreased by more than
three orders of magnitude for an increase in suction from 0 to 25 cm (Figure 2.2.8(1)). The
predicted K() curves for the growing media mentioned above exhibit a similar trend, with a
decrease of several orders of magnitude in K over a narrow range of suctions. This hydraulic
characteristic of soilless growing media has a tremendous effect on water dynamics and
availability to container-grown plants. This subject will be discussed in the following.
Any suggested model for the hydraulic characteristic of porous media can be used for
different container growth media only after its validation for media of similar texture and
structure. As mentioned above, measurement of - data is relatively simple and validating a
mathematical model to predict retention curves by its comparison with the measured data is
straightforward. However, validation of models for K() prediction has been rarely carried
out due to the difficulty of obtaining K - data. Wallach et al. (1992a) used the steady-state
flux control method (Klute and Dirksen, 1986) to determine K - data pairs. According to
this method, a constant flux of water, q, is established at the upper end of a vertical uniformly
packed substrate column, while its lower end is maintained at atmospheric pressure. The flux,
q, which should be lower than the saturated hydraulic conductivity, Ks, is applied to a
previously saturated column, which drains to a condition of steady-state downward flow.
Upon reaching this condition, the suction distribution along the column is expected to be
relatively constant throughout its upper region. A unit hydraulic gradient should, therefore, be
established in that part of the column, and under these conditions K is numerically equal to q.
The experimental set-up for this method included a 50 cm long column (10 cm i. d.),
four tensiometers, a pressure transducer, a manual scanning valve system, and a peristaltic
pump (Wallach et al., 1992a). Each tensiometer consisted of a high-flow ceramic porous cup
(6.3 cm long, 1.95 cm o.d., air entry value > 1 bar), which was mounted in a horizontal
position, extending about 6 cm across the column. The vertical distance between tensiometers
was 10.5 cm. Each tensiometer was filled with deionised, deaerated water and the top was
sealed. A pressure transducer operating in a range of 0 to 100 cm water was mounted on a
fixed position and connected to the scanning system. The transducer (constant voltage supply
of 12 V) was adjusted and calibrated precisely to obtain the relation between voltage and
water suction. Starting at saturation, a controlled flow was maintained until the tensiometer
readings stabilized and the volumetric outflow rate was constant. Due to the unit-gradient, the
constant flux yielded K, while the suction was measured to obtain the related suction. This
process was then repeated at a series of decreasing flow rates, and each time q and h were
recorded. The experiment was performed in duplicate for RTB and RTM. For each flux,
measurements were made until a steady-state condition was attained. This usually took from
several hours to several days.
For an indirect validation of the calculated K(h) relationship, the same experimental
set-up was used. The measured values of h along the column, for each value of q, were
compared with simulated profiles, (z), obtained by solving the steady-state one-dimensional
flow equation:
1) +
dz
d
)(K( = q
[9]
Rearranging eq. [9] and integrating between the bottom of the column (z = 0) and its top (z
= Z), gives:
K(y)
q
+1
dy
= z =dx
(Z)
o
Z
0
[10]
where x and y are integration variables. Introducing the calculated K(h) values (drying
curve) into eq. [10] and using a standard numerical integration procedure to solve the
equation, the suction was calculated as a function of the column's height.
For both RTB and RTM tuff, measured K() values (Wallach et al, 1992a) verified the
calculated K() relationship over the 0 to 25 cm suction range (Figure 2.2.8(2)). For higher
values of q, the agreement between measured and predicted values seems to be very high. As
q was lowered, however, fluctuations in tensiometer readings were observed. These
fluctuations could have resulted from the difficulty in establishing and controlling constant low
fluxes with the equipment used. Very low fluxes (< 10-4 cm min-1) were required to achieve
suctions of about 20 to 25 cm, depending on the medium. However, these fluxes are much
lower than those of drippers or mini-sprinklers used for the irrigation of container media. We
therefore conclude that the steady-state flux control method is applicable to container media
such as tuff, at least in the tested range of water suctions.
2.2.9. Effects of substrates’ physical characteristics on irrigation management
Water flows along the potential gradient from the medium, where high (slightly negative)
water potential () exists, through the plant vascular system to the leaves and to the
atmosphere, where the water potential is lower. The flow rate is proportional to the potential
gradient and inversely proportional to the sum of resistances imposed by the medium,
medium/plant interface, plant, plant/atmosphere interface and atmosphere.
Although water constitutes most of the biomass, only about 1% of the applied water is
actually retained by the plant as water per se, or as a source of hydrogen for the reduction of
CO2. The remainder is lost to the atmosphere through the process of transpiration. This
apparently wasteful process is, in fact, indispensable since dry matter production is
proportional to the ratio of actual to potential ET (de Wit, 1958). This relation results from
the following factors:
The transpiration process is a key factor in plant life in general and in photosynthesis
in particular.
Water, being a solvent, enables the translocation of ions (including essential nutrients)
and organic molecules into, within and out of the plant, driven by the transpirational
stream.
Transpiration helps to maintain plant temperature, which permits normal plant life.
It is clear from the above that maximising transpiration through increased water availability
(unlike increased transpiration due to high vapour pressure deficit [VPD]) is in the best
interests of the grower; thus high water availability is a must for profitable crop production. In
horticultural terms, this means maximizing transpiration through optimal irrigation control.
Irrigation control involves the determination of both the timing and quantity of water to be
applied. It can be approached from the point of view of potential ET according to an existing
climatic model, by using one of several sensors to report the state of substrate water, or by
using various plant response parameters, or any combination of the above (Norrie et al.,
1994). The first approach can generally be described as:
ETm = Kc • ETa [11]
Where: ETm = maximum ET under the specific climatic conditions, Kc = crop coefficient,
ETa = actual ET
The Penman - Montieth equation is the most widely used method for ETm prediction, based
on the relevant climatic data such as net radiation absorbed by leaves, temperature, VPD and
wind speed. A simplified but still accurate version of the Penman - Montieth equation was
developed for Ficus benjamina (Bailey et al., 1993). A simpler equation for use in commercial
units that are not equipped with sophisticated, computerized control, and which is based on
pan evaporation and plant-canopy height and width as input variables, has been generated
(Stanley and Harbaugh, 1989) with partial success. However, the applicability of the Penman -
Montieth equation, and presumably other climatic models for irrigation control in
greenhouses, is still questionable (Munoz-Carpena et al., 1996). Possible reasons for poor
irrigation demands by climatic models include sharp fluctuations in canopy area (e.g. in cut
flowers), the effects of pests and diseases on transpiration etc. Inputs to intensive greenhouse
cultivation, such as supplementary lighting, heating pipes, air humidity and CO2 concentration,
are not well predicted by classical, field-orientated models (Blom-Zandstra et al., 1995; Baille
et al., 1996).
The plant-related approach (the “Speaking Plant” theory) is based on parameters that can
be used to determine irrigation timing, such as: leaf water potential (with pressure chamber);
stomatal resistance for water (with diffusion porometer); canopy temperature (with
thermocouples or infrared thermometers); flow of water in the stem (with the heat pulse
method); changes in stem diameter (with dendrometer); xylem potential and soil-root interface
potential (Caron et al., 1998).
The main advantage of the plant-related approach is its ability to report water shortage
before any visual sign is apparent. The main problem with it is that it identifies events of water
shortage post factum, in terms of horticultural damage. Analysing such events can, however,
help the grower in the future. The use of these parameters is, therefore, more suitable for
research purposes and their applicability to intensive commercial growing systems is doubtful.
Their practical use is restricted to slow-growing plants due to their relatively slow response
time.
The use of medium-related parameters seems to offer the optimal solution for irrigation
scheduling. Tensiometers, which measure suction in the medium near their location, are
reliable, handy, and relatively cheap. They have been used to monitor moisture status in
containers and for decisions on irrigation scheduling. Frequent suction readings at different
locations throughout the container by electro-tensiometers provide continuous on-line
information on suction distribution throughout the container. Suction data can be translated
into moisture content by using retention curves. Means for the direct measurement of water
content have recently been introduced (Kritz and Khaled, 1995; de Groot, 1995; da Silva et
al., 1998). However, the proper interpretation of the measured suction and water content so
as to increase water availability and the efficiency of irrigation management has not been fully
resolved.
A high content of available water and an adequate air supply are considered the most
important physical characteristics of container media in order to achieve optimal growth (De
Boodt and Verdonck, 1972; Spomer, 1974; Beardsell et al., 1979a, b; Chen et al., 1980;
Fontento et al., 1981; Hanan et al., 1981; Fontento and Bilderback, 1983; Bilderback, 1985;
Tilt et al., 1987; Riviere et al., 1990). Knowledge of transpirational demand and the selection
of a container medium that contains sufficient air at low suction values, have been considered
a good basis for irrigation management. Quantitatively, De Boodt and Verdonck (1972)
introduced the concept of ‘easily available water’, EAW, which they defined as the difference
between the water content at 1 kPa and 5 kPa. These authors also defined ‘water buffering
capacity’ (WBC) as the difference between the water content of the medium at 5 and 10 kPa.
Ever since Bunt (1961) reported water retention curves for container media, much effort has
been devoted to determining the usefulness of these curves in relating water availability to
plant growth, and in quantifying such data both for descriptive and predictive purposes.
However, it is now becoming clear (see previous discussion in section 2.2.8.) that the concept
resulting from the work of Bunt (1961) and de Boodt and Verdonck (1972) namely that the
content of EAW can serve as a guiding tool for irrigation management is largely invalid.
An additional term for the relation of water availability to container grown plants is
‘container capacity’, an adjustment of the ‘field capacity’ notion, which is widely used by soil
physicists and irrigation scientists for field-irrigated crops. White and Mastalerz (1966)
defined container capacity as the amount of water retained in a containerised medium after
drainage from saturation has ceased, but before evaporation has started. A combination of
mathematical functions for the water characteristic curves and container geometry was later
shown to provide a more consistent description of container capacity (Bilderback and
Fontento, 1987; Fonteno, 1989). According to this approach, container capacity is the total
volume of water in the container, as given by its water retention curve, divided by the
container volume. This term describes the maximum water retention capacity of the medium.
Bilderback and Fonteno (1987) defined ‘available water’ as the difference between container
capacity and the water held at permanent wilting point. However, although such a value may
represent an end-point for plant survival, the proposed range does not represent the optimum
for plant growth in containers, and water is not equally available over the range from
container capacity to permanent wilting point.
Moisture extraction by roots depends on the momentary atmospheric conditions and
medium water availability, which, in turn, depends on the resistance to water flow from the
medium bulk to the medium-root interface, and water flow across the root tissues into the
root xylem. Moisture depletion in the rhizosphere induces water flow from the surrounding
medium volume to replenish the depleted water. This flux depends on the momentary value
of the hydraulic conductivity within the medium volume that encircles the roots and on the
hydraulic head gradient between this medium volume and the surface of the root:
ds
d
)(Kq
[12]
where q [L/T] is water flux through the root-medium interface, K() is the unsaturated
hydraulic conductivity, which depends nonlinearly on suction, , or water content, , (section
2.2.8), and ‘s’ is an axis pointing in any direction at the proximity of the media-root interface.
Owing to the non-linear relationship between the hydraulic conductivity and moisture content,
K(), a deviation from a certain value of induces a moderate decrease in K, while a similar
deviation from a lower value of induces a sharp decrease in K. Equation (12) is written
under the assumption that gravity has a negligible effect on water flux from the medium to the
root at the medium-root interface proximity.
The momentary balance between the actual water flux from the medium volume
surrounding the root and the potential water extraction rate (which depends mainly on the
momentary atmospheric conditions) determines whether medium moisture is fully available
to the plant. By “fully available” we mean that the water extracted from the vicinity of the
root can be fully replenished at a rate that coincides with the potential ET. When the root
rhizosphere cannot be immediately replenished, water uptake by the root will be limited by
the transport ability of medium water toward the roots, i.e. by the medium’s conductivity
near the rhizoplane.
To demonstrate the actual implication of the water-availability related parameters
mentioned above, we consider two media, peat moss and tuff, which are uniformly packed in
containers of 25 cm height. The maximum potential water uptake rate is at ‘container
capacity’, i.e. following irrigation, drainage, and water redistribution. Under such a
condition, the water content at the container bottom is close to saturation, and gradually
decreases towards the top. Once equilibrium has been reached on the completion of
drainage, the suction at any location in the container will be equal to the gravitational force
exerted by the height of that location from the bottom of the container. For such conditions,
the moisture content distribution along the container height is the retention curve of the
particular medium. These distributions for peat and tuff containers are shown in Figure
2.2.9(1). For such suction and water content distributions, one can add the characteristic
hydraulic conductivity curves of the relevant medium to the water content distribution in
order to evaluate the hydraulic conductivity at each height along the container (Figure
2.2.9(1)). The peat and tuff differ significantly in their water retention properties, peat
exhibiting a higher water-holding capacity. The two media show a high value of saturated
hydraulic conductivity, Ks (Figure 2.2.9(1)). However, as suction (expressed in the figure as
height from the container bottom) increases from 0 to 25 cm and water content decreases
accordingly, the hydraulic conductivity decreases by approximately three orders of
magnitude for peat and by approximately four orders of magnitude for tuff. As such, the
hydraulic conductivity variation in the pot has a major influence on water availability and
should be considered when criteria for water availability are postulated. The enormous and
K variation along the container height demonstrates that water is not evenly distributed
throughout the container volume; potential water availability is much higher at the container
bottom than at the top. This should also be considered prior to the installation of measuring
devices (tensiometers, TDR etc.) in the container and during the interpretation of the
measured data for operative decisions, such as irrigation scheduling, leaching, etc.
It is interesting to compare the values of and K at h=10 cm, which was considered to be
the lower suction limit of EAW definition (de Boodt and Verdonck, 1972). The moisture
content, , for peat is 0.6 (compared to 0.9 at saturation) and for tuff 0.19 (compared to
0.51 at saturation). The hydraulic conductivity at this location is 1.5·10-2 cm min-1 for peat
(compared to 4.91 cm min-1 at saturation) and 5.5·10-3 cm min-1 for tuff (compared to 7.3 cm
min-1 at saturation). These values were estimated by the arrows shown in Figure 2.2.9(1).
Note that both and K of the peat at 10 cm height are three times larger than those of the
tuff. The lower hydraulic conductivity values in the upper part of the container indicate that
water is barely available there, in spite of the fact that the water at this location is within the
range of EAW.
A method to estimate water availability by directly measuring or K() near the roots does
not yet exist. Thus water availability needs to be related to other measurable soil-bulk
properties. It was hypothesized (Wallach et al., 1992a, b; da Silva et al., 1993a, b) and
verified (Orozco and Marfa, 1995; Raviv et al., 1999, 2001) that K() of the medium bulk
indicates the availability (amounts and rates) of medium water to plant roots and significantly
affects the plant’s performance.
A precise balance between ET and irrigation is rarely practical, mainly due to increased
salinity in the root zone. Plants do not take up some non-essential ions, thus causing a gradual
build-up of salinity (Baas and van den Berg, 1999). To prevent this, excessive amounts of
water must be supplied in order to leach accumulated salts. If the root zone is leached
frequently, the roots are never exposed to an excessively low osmotic potential. This practice
assures that o of the medium solution is maintained close to that of the irrigation water.
Frequent irrigation is beneficial also from a consideration of nutrient availability. A test case
for this issue is presented in section 2.3.3.
2.3. Chemical Properties
2.3.1 Charge characteristics
Surface charge properties of growing media have a major effect on the chemical reactions
taking place in the rhizosphere. Even inert substrates accumulate organic compounds excreted
from plant roots or decomposed during the growth period, which build up surface charges
(Tate and Theng, 1980; Huang and Violante, 1986; Tan, 1986; Silber and Raviv, 1996).
Hence, quantification of the charge properties of the solid phase is essential in order to
characterise substrates.
Growth substrates commonly used in horticulture, such as volcanic and organic materials,
possess negative permanent and/or variable charged surfaces. The permanent negative charge
results mainly from isomorphous substitutions in layer-silicates (Gast, 1977; Rhoades, 1982;
Mc Bride, 1989). The extent of cation adsorption by the surfaces is referred to as the cation
exchange capacity (CEC; mmolc kg-1) and is used to characterise the exchange properties of
the medium (Gast, 1977). The variable charge (both negative and positive, depending on the
solution pH) is generated mainly from the adsorption of H+ and OH- on solid surfaces such as
metal oxides, hydroxides, microcrystalline silicates (allophane and imogolite), or on functional
organic groups.
The effects of solution pH on the extent of the positive and negative charge may be
determined experimentally by measuring the CEC and the anion exchange capacity (AEC) at
several pH values. The pH at which the AEC is equal to the CEC is referred to as the point of
zero net charge (PZNC) (Parker et al., 1979; Sposito, 1981) and is frequently used to
characterise the charge properties of the medium. An alternative experimental method is to
conduct a pH titration curve using an indifferent electrolyte (Gast, 1977; Mc Bride, 1989). A
typical potentiometric titration for woody peat is presented in Fig. 2.3.1(1) (Bloom, 1979).
The pH at which a series of titration curves obtained at different electrolyte concentrations
intersect is referred to as the point of zero salt effect (PZSE; Parker et al., 1979; Sposito,
1981). However, common growth substrates rarely consist of a single well-defined component
and typically contain a mixture having both permanent and variable charged surfaces. The
interpretation of the experimental CEC, AEC or the pH titration data obtained from these
mixtures is, therefore, complicated. Moreover, the existence of additional sources or sinks for
H+/OH- derived from the dissolution/precipitation of solid phase minerals during the CEC,
AEC or the pH titration analyses makes direct interpretation of analytical results even more
ambiguous.
The difficulties involved in the determination of the actual surface charge of a
heterogeneous material having pH dependent and permanent charge surfaces together with
primary and secondary minerals are demonstrated for the case of tuff (Figs. 2.3.1(2) and
2.3.1(3)). Tuff was chosen because most of the work on this subject has been done with this
material. The CEC of three types of tuffs, widely used in greenhouses and having different
weathering stages (yellow>red>black), was found to be pH dependent Fig. 2.3.1(2)).
However, it was not possible to determine the AEC of these tuffs because P, released from
indigenous Ca-P minerals during analysis, was re-adsorbed on the tuff surfaces and interfered
with the AEC determination (Silber et al., 1994). The CEC of tuff at pH 3.5, especially that of
the weathered yellow tuff, was relatively high (350 mmolc kg-1) and may indicate the existence
of components with a permanent charge. The pH titration curves of the tuff types significantly
differed from those obtained from homogenous variable charge materials. In the cases of the
un-weathered tuffs (red and black), the pH titration lines obtained from three different NaCl
concentrations overlapped (not presented). Distinction of an ionic strength effect was possible
only for the yellow tuff at a pH above 7 (Fig. 2.3.1(3a)). Overlapping of potentiometric
titration curves at low pHs in volcanic material from Chile was explained by the exchange of
added H+ ions with cations associated with a permanent negative charge (Espinoza et al.,
1975; Gast, 1977). Wann and Huehara (1978) reported that specific adsorption of P caused
overlapping of potentiometric titration curves of soil, and therefore, re-adsorption of P may be
partially accountable for the overlapping of the pH titration curves in the case of tuff.
Nevertheless, assuming H+/OH- consumption in cation/anion exchange, the net charge at each
pH value can be evaluated by subtracting the net quantity of cation/anion accumulated in the
solution phase (charge) from the quantity of acid/base added at the pertinent pH. Hence, charge
(mmolc kg-1) is defined as:
charge = (cat-anion)pH - (cat-anion)ZPT [13]
where ZPT is the suspension pH prior to the addition of protons or hydroxyls (7.1, 7.2 and
7.6 in 0.1, 0.02 and 0.006 M, respectively) and cat and anion include all the cationic and
anionic species in the solution, based on speciation calculations. Comparison of tuff’s pH
titration curves to the net surface charge (Figs. 2.3.1(3a) and 2.3.1(3b), respectively)
demonstrates that the major quantities of acid/base added to tuff solutions were consumed for
exchange with cations/anions adsorbed and for dissolution of indigenous Ca-P mineral and
very fine amorphous particles (Silber et al., 1999).
The CEC of common organic growth substrates, such as peat, compost and reused waste,
is generally high (800-1600 mmolc kg-1) and is pH dependent (Puustjarvi, 1977). Charge
characteristics of composts were found to be dependent on the composting (Inbar et al., 1989;
Iglesias-Jimenez and Perez-Garcia, 1992; Jokova et al., 1997; Saharinen, 1996), mainly
because of changes in the actual CEC sites during composting process (Saharinen, 1998). The
C/N ratio as well as the lignin concentration correlated with the composting time (Raviv et al.,
1987; Tarre et al., 1987; Inbar, 1989; Inbar et al., 1989), and can be used for CEC estimation
(Fig. 2.3.1(4)) and maturity indices of compost quality.
2.3.2. Phosphorus retention
Availability of phosphorus in the rhizosphere often determines the growth and productivity
of crops. Strong interaction between orthophosphate ions and solid phase constituents
continuously reduce solution P concentrations following fertilizer application. The main two
mechanisms involved are: (i) fast (seconds, minutes) electrostatic reactions of adsorption onto
the solid phase; and (ii) slow (hours, days) formation of new solid metal-P compounds
(precipitation). Thus, low P availability may restrict crop productivity even shortly after P
application. The effects of solution pH and metal concentrations on P partitioning between the
solid and aquatic phases are presented below, using tuffs (widespread growth substrates for
horticultural crops) as an example.
Solution P concentration (CP) in three tuff solutions was governed by different surface
charge properties and metal activities (Al3+ at pH below 5.5 and Ca2+ at pH above 7.0). CP in
yellow tuff above pH 4.2 increased exponentially as the pH increased (Fig 2.3.2(1)), because
of reducing P adsorption on the pH-dependent surface charges (Silber, 1991). The higher CP
values in yellow tuff at pH 4.2 resulted from the relatively higher quantities of P contributed
by the pH-dependent dissolution of indigenous hydroxyapatite (Silber et al., 1999). By
contrast, the slope of CP vs. pH in black and red tuff, decreased and the curves became almost
flat above pH 7 and 5, respectively (Fig. 2.3.2(1)). Decreased CP as the pH rises is not a
typical adsorption mechanism and indicates a different phase controlling P in solution, e.g. the
precipitation of metal-P compounds. Metal concentrations in the tuff solutions differed
because of dissimilar chemical and mineralogical composition. The activity products of Al3+
and PO43- in black and red tuff solutions below pH 5.5 exceeded the solubility product (Ksp)
of variscite (AlPO4.2H2O), although it was below the Ksp of berlinite (AlPO4) (Silber, 1991).
In yellow tuff solutions, low Al concentrations at identical pH values kept ion activity
products below the Ksp of variscite (Silber, 1991). The formation of an X-ray amorphous
analogue of variscite or AlOHNaPO4 (Veith and Sposito, 1977) is also possible in the solid
phase, which may control CP in tuff solutions at low pH. It is possible to assume, therefore,
that below pH 5.5 adsorption reactions governed CP in tuff solutions for a short time
(minutes, hours), but as time increased, CP declined until the activity product of Al3+ and PO43-
attained the Ksp of a thermodynamically stable compound (Silber et al., 1999). The activity
products of Al3+ and PO43- above pH 5.5 were below the Ksp of any Al or Fe-P minerals.
At pH 7.1 and with a low quantity of added P (250 mg kg-1), the activity of Ca2+ and
PO43- in tuff solutions exceeded the Ksp of hydroxyapatite (HA: Ca5(PO4)3OH) (Fig.
2.3.2(2)). At high P levels (1000 mg kg-1; encircled signs in Fig. 2.3.2(2)) these activity
products shifted towards the solubility line of TCP (Ca3(PO4)2), while at pH 8.2 towards
OCP (Ca4H(PO4)32.5(H2O) (Fig. 2.3.2(2)). Thus, even if HA was the most stable Ca-P
mineral, more soluble compounds, like TCP or even OCP, temporarily governed P in tuff
solutions (Sanyal and DeDatta, 1991, Imas et al., 1996). If the conclusions of Sanyal and
DeDatta (1991) that TCP governs the dissolution of P in the soil solution are accepted, then
at pH 6-7 (at the pertinent Ca concentration) and as long as CP does not exceed 1.6, 1.5 and
0.9 mmol l-1 in black, red and yellow tuff solutions, respectively, for a short time, adsorption
reactions may govern P solubility in the three tuff solutions (Silber, 1991). As under acid
conditions, kinetic considerations had to be taken into account when evaluating the relative
importance of adsorption vs. precipitation reactions in tuff solutions above pH 7.
2.3.3. Effects of substrates’ chemical characteristics on nutrition management
The main factor that distinguishes fertilisation management of substrate- and soil-grown
plants is the limited volume of substrates, which means a lower buffer capacity for pH and
solution composition and a limited supply of nutrients. The limited root zone volume also
results in decreased root size and increased root density, causing higher competition among
roots and a bigger effect of root activity on the rhizosphere. In a review of this subject, Bar-
Tal (1999) showed that frequent irrigation and continuous fertilisation could satisfy plant
demands under the lowest practical container volumes, and discussed the effects of medium
volume and root size on plant nutrition. Fertilisation management is discussed, here in light of
these principles.
Solution pH and N source - High pH values (>7.5) in the irrigation water are undesirable,
because precipitation of Ca and Mg carbonates and orthophosphate may occur in the tubes
and drippers. High substrate pH may reduce Zn, Fe and P availability to plants. Consequently,
the use of ammonia in fertigation is not recommended, since it raises the pH when injected
into irrigation water. Compounds that may reduce the irrigation water pH are nitric (HNO3)
and orthophosphoric (H3PO4) acids. Depending on their price and the raw water composition,
these acids may be used to reduce the irrigation water pH to 5. The acid dose that is required
to attain a target pH depends primarily on the HCO3 concentration in the raw water. However,
at HCO3 concentrations above 2-3 mM, it is not possible to use only H3PO4 to adjust pH to 5-
5.5 without adding excessive amounts of phosphate to the solution. For instance, if the raw
water contains 3 mM HCO3, about 2-2.5 mmol H3PO4 should be added per litre of nutrient
solution to neutralize as much HCO3 as required attaining the target pH. This would result in a
H2PO4 concentration of 2-2.5 mM in the nutrient solution. However, the H2PO4 concentration
in the nutrient solution should normally not exceed values above 1.0 to 1.5 mM. The use of
nitric acid is not limited by such considerations, since, normally, the target NO3 concentration
in the nutrient solutions is higher than 10 mM. pH values lower than 5.0 are detrimental to
root membranes and may increase the Al and Mn concentrations in the substrate solution to
toxic levels.
Different sources of N fertilizers have different effects on irrigation water and the growth
medium pH. Urea is a highly soluble, chargeless molecule, which easily moves with the irrigation
water and is distributed in the medium similarly to NO3-. Plants cannot utilise urea directly, and
therefore nitrogen deficiency is possible even if total N (urea) is adequate. At 25oC, urea is
hydrolysed to NH4 by soil microbe enzymes within a few days, while in organic media the process
is faster and takes 24-48 h (Wright, 1987). This hydrolysis results in an increase in pH, which in a
medium of pH > 7.5 may reduce P, Ca and micronutrient availability to plants. The ammonium ion
in the solution is in equilibrium with NH3, which is a toxic compound to plants. Ammonia toxicity
increases with pH and the NH4/NO3 ratio in the irrigation water has a strong impact on the medium
pH. When NH4 uptake is predominant, H+ is excreted from roots; when NO3 is the major ion
absorbed, OH- is released into the solution. Fluctuations in medium pH around the roots within the
order of + 1.5 pH units, due to NH4- or NO3-N supply have been reported (Barber, 1984;
Marschner, 1995). In tomato and rose, a stable pH in the nutrient solution has been maintained
when the NH4/NO3 molar ratio was between 1:4 and 1:3. A high N- NH4: N - NO3 ratio in the
fertiliser solution damaged sugar beets (Beta vulgaris L.) through acidification of the
rhizosphere (Breteler, 1973). However, buffering the solution pH improved plant tolerance to
a high N-NH4: N-NO3 ratio (Arnon et al., 1942). The effect of the nutrient solution
composition on the solution medium pH can be reduced by the addition of buffering
components, like zeolites or carbonate lime, to the substrate (Frick and Mitchell, 1993; Elliot,
1996; Argo and Biernbaum, 1997). Increasing the N-NH4: N-NO3 ratio reduced the uptake of
other mineral cations, but increased the uptake of mineral anions in tomato (Kirkby and
Mengel, 1967; Ganmore-Neumann and Kafkafi, 1980b) and pepper (Marti and Mills, 1991a
and b; Bar-Tal et al., 2001b), while Sarro et al. (1995) reported that ammonium reduced Ca
and Mg uptake by pepper, but had no effect on K uptake. A N-NO3: N-NH4 ratio of 1:1 was
found to be optimal for the growth of young tomato plants under a wide range of root
temperatures (Ganmore-Neumann and Kafkafi, 1980a). Feigin et al. (1980) reported that the
solution NH4 fraction in the range of 0 to 30% had no significant effect on the fruit yield of
tomato, but yield decreased as the NH4 fraction increased from 30 to 50%. High N-NH4
reduced the dry weight and fruit yield of bell pepper relative to N-NO3 (Marti and Mills,
1991a; Bar-Tal et al., 2001a). In the Netherlands, the recommended N-NO3: N-NH4 ratio for
greenhouse vegetables is about 6:1 (Roorda van Eysinga and van der Meijs, 1981), whereas in
more recent recommendations the ratio has increased to 12:1 (L. Marcelis, personal
communication). In semi-arid regions, NH4 is used to reduce the pH of the medium to prevent
salt precipitation. Takacs and Tecsi (1992) found that a high N- NH4: N- NO3 ratio in the
nutrient solution caused the destruction of leaf chloroplasts. The N- NH4: N- NO3 ratio affects
blossom-end rot (BER) in tomato. (Wilcox et al., 1973; Wojciechowski et al., 1969) and bell
pepper fruits (Bar-Tal et al., 2001a; Marti and Mills, 1991b; Morley et al., 1993), probably as
a result of its effect on Ca uptake and concentration.
Ammonium was shown to be an undesirable source of nitrogen for tomato and strawberry
plants at root-zone temperatures above 30oC, because it is detrimental to root growth and
development (Ganmore-Newmann and Kafkafi, 1980 a; Ganmore-Newmann and Kafkafi, 1983).
Nitrogen spatial distribution in the growth medium is strongly affected by the source of N in the
irrigation water. In substrates with a high CEC, ammonium is adsorbed by active surfaces and thus
has a restricted mobility relative to the unadsorbed NO3. Ammonium is also nitrified to NO3 by
microbially mediated reactions, at a rate that depends on the medium temperature and moisture
content. The half-life of this process at 25oC and desirable moisture content is about 2 weeks.
When N application rates are temporarily in excess of plant consumption, it is advisable to apply the
excess amount as NH4-N and thus avoid rapid leaching of the unexploited N out of the root zone.
However, this can be done only with a medium having a high buffer capacity, which is a function of
the substrate CEC and the volume of the growing medium.
Organic substrates contain organic N, which serves as a source for mineral N, like slow release
fertilisers (Genevini et al., 1997), but they may immobilise mineral N and to reduce its availability to
plants when the C/N ratio is high (Handreck 1992a, 1992b, 1993b; Gruda and Schnitzler, 1999b;
Sharman and Whitehouse, 1993). This aspect is discussed in sections 2.5.2.
Phosphorus retention by substrates has an important effect on P fertilization management.
For example, a typical greenhouse for vegetables or flowers may contain 50-60 kg of tuff per
m2. The daily ET rate during the growth period of roses in a semi-arid climate is usually 5-10 l
per m2; thus, the daily irrigation supply is 0.1-0.15 l kg-1 tuff. The recommended P
concentration in the irrigation water for cut rose culture is usually 30-45 mg l-1 (Sonneveld and
de Kriej, 1987; Jones, 1997), hence 3-6 mg P kg-1 are supplied per day. Ignoring leaching, the
possible CP at pH 6 according to the P retention curves (Silber, 1991) will be less than 0.05
mol l-1. Based on the data presented in Fig. 5, the CP in tuff solutions at pH 6 were below
βTCP solubility, therefore, adsorption reactions apparently controlled P in tuff solutions. Since
adsorption process are very rapid, the decline of CP in the substrate will be fast, and even
immediately after irrigation solution P concentrations will probably be insufficient for rose
production. The yield of cut roses fell to only 40% following low P application (0.16 mmol l-
1), compared to an adequate P management (1.3 mmol l-1, Johanson, 1978). Thus, low P
concentrations in tuff substrates may restrict plant production. Assuming that 0.5 mmol l-1 is a
proper CP for cut roses in tuff, the amounts of adsorbed P in black, red and yellow tuff at pH
6 should be: 53, 148, and 840 mg P kg-1, respectively. If the daily P quantity added through
fertilization is 4.5 mg kg-1 (0.15 l kg-1 at 30 mg P l-1), then 12, 33 and 187 days are required to
achieve the desired CP in these tuffs, respectively. During this period, cut rose development
will certainly be impaired.
Increasing CP in the irrigation water is useless since the Ca concentration in irrigation
water is usually 1.5-3 mmol l-1, resulting in Ca-P precipitation (Lyndsay, 1979). Lowering the
pH of the tuff solution to below pH 6 will yield lower CP due to an enhancement of tuff’s
positive charge and precipitation of Al-P minerals. Raising the pH to above 7, and assuming
the Ca concentration in the solution to be at the level recommended for cut rose production
(1.5-2 mmol l-1), will lower the CP due to Ca-P precipitation.
Increasing the fertigation frequency while decreasing CP seems to be the preferable
solution for increasing P availability to plants. In addition, exudation of soluble organic acids
by plants during the growth period, may promote a higher CP in tuff (Silber and Raviv, 1996).
Alternatively, a substrate precharged with ions may be used as a source of nutrient.
Williams and Nelson (1997) showed that precharged zeolite could serve as a slow release
source for potassium and phosphate in a pot mixture medium.
Organic substrates usually contain high concentrations of organic P that may serve as slow
release fertilizer, but they may also immobilise mineral P (Handreck, 1996), therefore the C to
P ratio of the organic substrate has to be considered during fertilization management.
Microelements - Composts may supply different micronutrient to plants at a slow rate. For
example, bark may supply manganese to plants at a slow rate, without exerting any negative
effect (Handreck, 1993a). Even an inert substrate like rockwool contains microelements and
can serve as a slow release source of Fe, Mn, Zn and Cu (Rupp and Dudley, 1989). On the
other hand, specific adsorbtion of micronutrients by the substrate may reduce their availability
to plants. Chelating agents like EDTA (ethylene-diamine tetracetic acid), DTPA
(diethylenetriamine pentaacetic acid) and EDDHA (ethylenediamine di (o-hydroxyphenylacetic
acid)) are used to maintain the availability of micronutrient for plants.
Salinity of the fertigation water - A 10-meq l-1 solution has an electrical conductivity (EC) of
about 1 dS m-1 and an osmotic potential of -0.036 MPa (at 25oC). According to the U.S. Salinity
Laboratory, irrigation water with an EC exceeding 1.44 and 2.88 dS m-1 constitutes a moderate and
a high salinization hazard, respectively. According to uptake data (not presented), and assuming a
daily irrigation of 5 mm (50 m3/ha), nitrogen and potassium concentrations in the irrigation water at
the time of maximum consumption rate may reach values of 15 to 20 meq l-1, which correspond to
an EC of 1.5 to 2.0 dS m-1. Under such conditions, especially when the tap water EC is above 1.0,
which is common in arid zones, care should be taken to minimise the amount of accompanying ions
added with the N or K. For example, KCl, which is a cheap source of K, should be replaced with
KNO3 and K2HPO4, whereas NH4NO3 and urea should be preferred to (NH4)2SO4. Sodium-based
fertilizers (e.g., NaNO3 or NaH2PO4) are unlikely sources of N and P, due to the adverse effect of
Na on medium hydraulic conductivity and plant function.
Most of the common substrates are supplied free of salts, but care must be taken when using
new substrates. High salt contents in newly used growth media were reported for expanded clay
mineral (de Boodt et al., 1981), composts, coir and sawdust. The salts can easily be removed by
excessive leaching prior to planting in the medium.
2.4. Substrate Analysis
Various analytical methods have been developed in order to determine different chemical
properties of growth substrates: element analysis, CEC, AEC, surface charge, organic matter
content, carbonate content, pH, EC, nutrient composition and nutrient availability. Some
methods of analysis, element analysis, organic matter content and carbonate content are
adopted from soil analysis and detailed descriptions can be found in Sparks et al. (1996).
Methods to determine CEC, AEC and surface charge are described above, in section 2.3.1.
Specific methods for measuring pH, EC, nutrient composition and nutrient availability in
growing media are given below.
Water extraction methods - Water extracts are used to gain knowledge of the pH, EC and
solution composition in the vicinity of the roots. The most direct method of determining the
composition of the solution in the substrate is to extract the liquid from the substrate under
normal pot moisture. Indeed, in a substrate like rockwool, in which sufficient solution can
easily be obtained by means of simple suction, this is a routine procedure. A variant of this
method can be used in plug substrates, where press extraction is a viable option (Scoggins et
al., 2001). However, in most substrates water is retained much more strongly, and only a
small volume can be obtained in this way. Therefore, a higher volume of water has to be added
to the substrate to obtain an extract. Then the concentration of salts and nutrients in substrate
solution are estimated from those determined in the extracts. As the dilution increases, the
accuracy of the estimates decreases due to dissolution of sparingly soluble salts and ion
exchange reactions (Sonneveld et al. 1990).
A ratio of 1:1.5 volume extract is used to estimate the medium solution for peat and peat
based substrates (Sonneveld et al. 1974). In order to express the nutritional status of different
media on a comparable basis, it was suggested that extraction methods of media should be
based on their container capacity or water holding capacity (WHC) (Raviv and Mordechai,
1987; De Kreij et al., 1995; de Kreij et al., 2001). The main reason for this change is that all
available nutrients (and other soluble ions) can be found in this water volume. Sonneveld and
van der Ende (1994) adopted a similar approach and suggested using a standard suction of 1.0
kPa, which is the desired pressure head throughout crop growth. About 30-50% of the water
is obtained by pressing out the medium at this stage. Sonneveld and van der Ende (1994)
found very high correlation coefficients (0.992) for the relationships between the
concentrations obtained by the two methods. In this method the dilution varied from 3.8 to 4.0
(Sonneveld et al. 1974; Sonneveld and van der Ende, 1994). Some laboratories use higher
ratios of water to substrate like 5:1 and 10:1, as a convenient way to obtain a sufficient
volume of solution for the analyses (International Substrate Manual). However, in this method
the concentrations may be below the detectable limits, especially for microelements.
Throughout the growth of a crop, information on the solution composition can be obtained
from an analysis of the in situ drainage, also known as the pour-through method (Yeager et
al., 1983; Wright et al., 1990). This method is rapid and no special pre-treatment of samples is
required, but the interpretation of the results poses a difficulty since the flow out of drainage
from the substrate is mainly through the larger pores and may provide misleading information
regarding nutrient concentrations in the rhizosphere (Handreck, 1994). An alternative method
is in situ extraction with a ceramic cup inserted into the substrate volume. Argo et al. (1997)
obtained a near 1:1 correlation in pH, EC, N-NO3 and K concentration measured by this
method and by saturated substrate extract. Vincent and Sylvain (1995) found that N-NO3
measurements with a ceramic cup were comparable with free drainage and water extract, but
bigger variations were obtained with the former. These variations indicate the big variability
along the medium profile and in a lateral direction, as affected by plants and the distance from
the water emitters. A disadvantage of the ceramic cup is the adsorption of ions like K+ and
H2PO4- by the ceramic cup leading to selectivity in the transfer of ions.
Extraction with strong extractants - The purpose of this type of extraction is to obtain
information on long-term availability of nutrients in the substrate, including those that are not
readily soluble. Extraction with 0.5 M ammonium acetate for the adsorbed cations is adopted
from soil analysis and a detailed description can be found in Rhoades (1982). De Kreij et al.
(1995) found that the concentration of microelements (Fe, Mn, Zn, Cu, and Si) in ammonium
acetate extract was 7 to 20 times larger than in water, in which the concentrations were close
to or below the detectable value. Only small differences in the concentrations of macro
elements were noted between ammonium acetate and water extracts indicating a high degree
of correlation between the two methods (de Kreij et al., 1995). Thus, in substrates like
expanded clay, ammonium acetate extract is useful for microelements, whereas macro
elements can be determined in a water extract. Ammonium acetate extraction has no
advantage over water extraction for indifferent substrates like perlite and rockwool.
Alternative extractants to ammonium acetate are: 0.1M BaCl2 (Verhagen, 1999), 0.5 M NaCl,
0.05N HCl + 0.025N H2SO4 (Markus and Steckel, 1980) and different variations of common
extractants in soil analysis. Extraction with chelating agents is a common method for the
determination of microelement availability indices. Lindsay and Norvell (1978) concluded that
DTPA was the most useful chelate to use as an extractant for the simultaneous measurement
of available Zn, Fe, Mn, and Cu. In artificial media this method was modified and adopted for
the measurement of microelements as well as macro elements, especially those with low
solubility and mobility like phosphate (Alt and Peters, 1993).
2.5. Description of substrates
There is a big variability in the origin, physical and chemical characteristics of the substrates
used by the horticultural industry. Moreover, new sources of natural and artificial by-products
are being introduced as growing medium every year. Therefore, it is impossible to describe all
the substrates that are used as growth media. Instead, we shall present the major types of
substrates and their properties that affect their use and management. In this chapter, substrates
are primarily divided into organic and inorganic materials. The organic materials comprise
synthetic substrates (like Phenolic resin and Polyurethane) and natural organic matter (peat,
coconut coir and composted organic wastes). Inorganic substrates can be classified as natural
unmodified sources (sand, tuff, pumice), processed materials (expanded clay, perlite and
vermiculite) and mineral wool (rockwool, glasswool). Important properties of the growth
media include their chemical activity and surface charge. Therefore, substrates are
characterised as active (e.g. peat, tuff) or inert (e.g. rockwool and sand) materials.
The description of each substrate includes information on its production and origin, plus
general information on its applications as a growth medium, or for other purposes. The
physical characteristics, bulk density, water retention and hydraulic conductivity are given, as
these properties are essential for proper irrigation management and layout of the substrate.
The chemical characteristics, composition, stability as affected by pH, surface charge
properties, pH and salinity are given as these basic data are required for the proper
management of fertilization and irrigation. Information on substrate sterilisation is given, as
disease control is a major factor for the successive use of growing media. Information on
waste treatment is also presented since this is becoming a central issue in view of
environmental contamination generated by intensive agriculture.
2.5.1. Inorganic substrates
2.5.1.1. Sand
Production, Origin and General Information - Sand is the coarse fraction of the soil
minerals. It is defined by the International Society of Soil Science as particles above 0.02 mm
in diameter, and it is further separated into: (i) coarse sand, 0.2 to 2.0 mm; (ii) fine sand, 0.02
to 0.2 mm. Coarse sand is the preferred as a substrate. Pure sand is widely used in deserts and
coastal plains because it is a cheap, local, natural source. It is often used as a growing bed
situated on the ground above a polyethylene film that separates it from the soil. As a natural
deposit, the particle size and distribution is often not constant. The required depth of the sand
layer depends on the range of particle diameter. The finer the sand, the deeper the required
layer of sand to avoid water logging and poor aeration. Sand is also used as a component of
various growth media mixtures, usually forming the heaviest constituent.
Physical Characteristics - The bulk density of sand is high relative to other growth
substances, 1.48 and 1.80 g cm-3 of fine and coarse sand, respectively. The TP is relatively
low, 0.45 to 0.30 of fine and coarse sand, respectively, and the water content at saturation is
somewhat lower, 0.39 to 0.265, respectively. Sand has a narrow pore size distribution, so the
small pore fraction retains almost constant water volume over increasing suction from 0 to 10
cm water (coarse sand) or 0 to 20 cm water (fine sand). A further increase in water suction
results in a steep decline in water content. The hysteresis phenomenon is negligible (Fig.
2.2.6(2), da Silva, 1991; Wever et al., 1997). This retention curve indicates that aeration
problems are expected when using fine sand in the common pots or beds used in agriculture.
Bunt (1991) showed that the mean oxygen diffusion rate in the profile of a fine sand bed was
10 to 100 times lower than that of peat, perlite, redwood bark and different mixtures. The
saturated hydraulic conductivity of coarse and medium sand is relatively high, 5.1 and 7.1 cm
min-1, respectively (da Silva, 1991). However, the unsaturated hydraulic conductivity of coarse
and medium sand was reduced sharply as the water suction increased above 10 and 20 cm,
respectively (da Silva, 1991).
Chemical characteristics - Quartz (SiO2) is the most common component of the sand
fraction in soils, because, after feldspars, it is the second most common mineral in the earth’s
crust, and it is highly resistant to weathering (Drees et al., 1989). Quartz density is high, 2.6-
2.65 g cm-3, with a relatively low specific surface, 2 m2g-1 (Drees et al., 1989). Quartz is a
stable mineral with a low solubility of 3-7 mg Si l-1, independent of pH in the range of 2.5 to
9.0 (Drees et al., 1989). It is one of the purest minerals known with a very low substitution of
Si by Al, Fe and other trace elements. Thus, the charge deficiency that plays a major role in the
physical-chemical activity of other soil minerals is very low in Quartz (Drees et al., 1989).
Therefore, sand is an indifferent substrate that can serve as diluent to more reactive
components in plant growth media.
Sterilization and waste disposal - Sand can be steam-sterilized well. With thin layers, the
pores may rapidly fill with water, which disturbs the steaming process. Sand is very durable
because it is neither chemically nor biologically affected. Sand waste can be used in
infrastructure and construction, thus it doesn't raise environmental pollution problems. On the
other hand the protection of natural sand dunes limits the use of sand as a growth medium.
2.5.1.2. Rockwool and Glasswool
Production, Origin and General Information - Mineral wool is a light, artificial material,
originally produced for thermal and acoustic insulation in the construction industry. The
application of modified forms of rockwool as a substrate for horticulture began in Denmark in
1969 (Smith, 1987). It is mainly used as slabs or blocks of bonded fibres, but is also available
in granulated form as a component of potting mixtures. Rockwool is manufactured, by heating
a mixture of three natural raw materials: 60% diabase (a form of basalt rock, dolerite), 20%
limestone and 20% coke. These materials are melted together at a high temperature, with the
coke acting as a fuel in the form of a blast furnace through which air is forced to raise the
temperature to 1600OC. The molten mixture is then spun at a high speed into thin fibres of
about 0.005 mm diameter, which are cooled by an air steam. The fibres are heated with certain
additives (a phenolic resin and wetting agents) to bind the fibres together and lower the natural
hydrophobicity of the material. It is then pressed into blocks or slabs of various sizes (Smith,
1987).
Glasswool is made by melting quartz sand in an electric oven at 1200OC. The process of
production of the fibres and slabs is similar to that of rockwool. Mineral wool is inert, sterile,
easily managed and consistent in performance. Therefore, within a decade, rockwool became
one of the major growth media in greenhouses in The Netherlands, Belgium, Germany,
Denmark and other Western European countries, and from where its use has spread to other
regions (Smith, 1987). It is an effective growth medium for horticultural crops, in which the
grower can very easily manipulate the ratio between water and air, and between each of the
nutrients in the root zone. On the other hand it is "unforgiving" to management errors because
it lacks a buffering capacity for nutrients, pH and water, due to the low volume of the
common slabs.
Physical Characteristics - Rockwool is a lightweight substrate with a low bulk density of
about 0.07-0.1 g cm-3 and a pore volume of 92-97%, depending on the producer (Smith,
1987). It shows high water retention at low water tension and the water content declines
sharply as the water tension increases so that virtually no water is retained at suctions higher
than 5 kPa (Fig. 2.2.6(3), da Silva, 1991; da Silva et al, 1995). As a result the water buffer
capacity is low and a steep gradient in water content, occurs from the top to the bottom of the
slab following irrigation and free drainage. The air volume is low at the bottom, just 4% at a
height of 1 cm above the base, while the upper layers are dry. The recommended height of a
slab for optimal water to air ratio is 7.5 to 10 cm (Smith, 1987). A strong hysteresis
phenomenon was observed in the first cycle of wetting and drying, but it was reduced in the
following cycles (Fig. 2.2.6(3), da Silva et al, 1995). The hydraulic conductivity (K) at
saturation is very high (4.6 cm min-1) but is sharply reduced with increased suction (da Silva et
al, 1995). Such reductions in K may lead to poor uptake of water and nutrients and to rapid
development of water stress. Rockwool is a soft and elastic substance that is compressed
under pressure and retains its original height following relaxation, but after plant growth it
becomes softer and less elastic (Wever and van Leeuwen, 1995).
Glasswool is very light and can contain a lot of water and air. In contrast to rockwool, the
fibre diameter of glass wool can vary, which affects the water holding capacity. By having
finer fibres in the upper part of the slab than the lower part, it is possible to obtain better water
content distribution over the height of the slab.
Chemical characteristics - The chemical composition of rockwool, expressed as oxides
(%), is given (Verwer, 1976) in Table 2.5.1.2.(1):
Table 2.5.1.2.(1): Chemical composition of rockwool, expressed as oxides (%).
Material SiO2 Al2O3 CaO MgO Fe2O3 Na2O K2O MnO TiO2
% 47 14 16 10 8 1 1 1 1
The main chemical characteristic of rockwool is that it is totally inert, except for some
minor effects on pH. The initial pH of the commercial material is rather high (7.0-8.0) and
values of up to 9.5 have been recorded (Smith, 1987). Therefore pH adjustment to a more
favourable range, 5.5-6.0, is required, but below pH 5.0 it dissolves. Rockwool itself has no
effect on crop nutrition and all the required macronutrients must be supplied with the water.
However, it was demonstrated that a considerable amount of Fe could be taken up from
rockwool by rose rootstocks (Rupp and Dudley, 1989) and some vegetables (Sonneveld and
Voogt, 1985). Moreover, Rupp and Dudley (1989) reported that considerable concentration
of Fe, Mn, Cu and Zn could be extracted with DTPA from various rockwool products.
Sterilization and waste disposal - Mineral wool is a sterile product. After use it can be
steamed. One of the major problems in the horticultural use of mineral wool is the
environmental issue of the waste, as it is not a natural resource that can be returned back to
nature. In the last decade, different methods of rockwool recycling have been developed. The
waste rockwool and glasswool can be used as a raw material for the production of mineral
wool, or slabs can be chopped into small particles that can be used as a component in mixed
media.
2.5.1.3. Expanded minerals
2.5.1.3.1. Perlite
Production, Origin and General Information - The raw material for perlite is a natural
volcanic mineral. The substrate, named expanded perlite, is produced by heating the ground
and sieved material to 1000oC. Natural perlite contains mineral water, which converts to gas at
the high temperatures in the oven. This causes perlite to expand to 4 to 20 times its original
volume and a lightweight high porosity material is obtained. Perlite is frequently used in
potting soil mixtures and as a growing medium. It is produced in various grades, the most
common being 0 -2 and 1.5-3.0 mm in diameter. The various grades differ in their physical
characteristics.
Physical Characteristics - Expanded perlite is very light with a particle and bulk density of
0.9 and 0.1 g cm-3, respectively. It is very porous, has a strong capillary action and can hold 3-
4 times its weight of water. Bures et al. (1997a) reported that water retained at 10 kPa is
much higher for the coarse fraction (0.5-1.0 mm diameter) than the fine fraction (0.25-0.50
mm diameter) of expanded perlite, but this is not explained by the volume of internal porosity
alone. The slope of the reduction in water content as the water tension increases is moderate
relative to sand and rockwool (Bures et al., 1997a,b). The available and non-available water in
commercial perlite of 0-4 mm diameter was 13.6% and 36.5% of its volume, respectively
(Bures et al., 1997b). The water retention curve of perlite shows moderate hysteresis (Bures
et al., 1997b; Wever et al., 1997). Wever et al. (1997) reported that the saturation of perlite
was very rapid, independent of its initial moisture. The saturated hydraulic conductivity
depends on particle diameter (Bures et al., 1997a). For commercial perlite of 0-4 mm
diameter, having 50% of the particles smaller than 0.5 mm, K was 0.3 cm min-1 (Bures et al.,
1997b). A reduction of 2 orders of magnitude in the hydraulic conductivity was obtained as
the water suction increased from 0 to 30 cm water (Bures et al., 1997a). This change is
moderate in comparison to sand.
Chemical characteristics - Perlite is neutral with a pH of 7.0 to 7.5, but it has no buffering
capacity and contains no mineral nutrient. When the pH is low there is a risk of toxic Al
release into the solution. The chemical composition of the material, as analysed by Olympios
(1992) is given in Table 2.5.1.3.1.(1):
Table 2.5.1.3.1.(1):Chemical composition of rockwool, expressed as oxides (%).
Material
SiO2
Al2O3
CaO
MgO
Fe2O3
Na2O
K2O
% 73.1 15.3 0.8 0.05 1.05 3.65 4.5
Sterilization and waste disposal - Perlite is a sterile product as it produced at a very high
temperature. After use it can be steamed and its stability is not greatly affected by acids or
microorganisms. Chemically, perlite is a stable material, which can last for several years. Marfa
et al. (1993) found that it retains its physical properties for successive crops, although our yet
as unpublished results show a significant shift in physical properties (R.W and M.R.). Perlite is
sensitive to mechanical compression, which may grind particles to powder. Being an inert
material, recycling perlite poses no environmental problems.
2.5.1.3.2. Expanded clay granules
Production, origin and general information - Expanded clay is a granular product with a
cellular structure. It is produced by heating dry, heavy clay to 1100oC, at which temperature
gas is released and expands the clay. The raw material has to have a low content of soluble
salts so that substances, like lime, are not added during the process. Otherwise, salt may be
released during cultivation. The grade size used in horticulture is 3-10 mm diameter.
Expanded clay granules have been used in horticulture since 1936. They are used in hydro-
culture and in buckets for various crops.
Physical characteristics - Expanded clay granules are light with a low bulk density of 0.28
to 0.63 g cm-3 and water content of 11 to 24% at a pressure head of -10 cm (de Kreij et al.,
1995). Wever and van Leeuwen (1995) reported high bulk density, 1.2 g cm-3, and 53%
porosity. They found that the water content was reduced from 48 to 44% as the water suction
increased from 3 to 50 cm, thus the amount of readily available water is very small, just 4%.
By contrast, the material contains a large amount of air. It has been shown that there were no
changes in physical characteristics after 5 years of intensive cropping. The saturation process
of the substrate from a dry condition is relatively slow, but if it starts from a pressure head of
100 cm it is rapid (Wever et al., 1997). The water retention curve shows insignificant
hysteresis. There is no available information on the hydraulic conductivity of expanded clay
granules, but they are relatively resistant to compression (Wever and Leeuwen, 1995).
Chemical characteristics - Expanded clay granules are neutral, with a pH of about 7.0. The
EC is low if low salt clay was used and no salts were added during the baking process;
otherwise, washing is required. It has been classified as an inert material with no cation
exchange or buffering capacity. However, Meinken (1997) and Meinken and Fischer (1994)
demonstrated that nutrients may accumulate by diffusion into the granules and they can be
released back to the solution.
Sterilization and waste disposal - Expanded clay granules are a sterile product as it
produced at a very high temperature. After use it can be washed and sterilized without
deleterious effect. Expanded clay granules are very stable and can last for many years. Waste
material can be used in the construction industry.
2.5.1.3.3. Vermiculite
Production, origin and general information - The raw material for vermiculite is a natural clay
mineral that has a layered structure with water in between the layers. The substrate, named
expanded Vermiculite, is produced in a similar way to perlite by heating the grinded and sieved
material to 1000oC. The water is converted to vapour at the high temperatures in the oven and
push the layers away from each other. As a result, expanded Vermiculite consists of granules
with an accordion shape, a light weight and high porosity. Vermiculite is used as a sowing
medium and as a component of potting soil mixtures. Thanks to its physical properties, it is
considered an excellent rooting medium (Wright, 1989). It is produced in various grades, the
most common being 0 –2, 2-4 and 4-8 mm in diameter. The various grades differ in their
physical characteristics.
Physical characteristics - Expanded Vermiculite is very light with a particle and bulk density
of 0.9 and 0.1 g cm-3, respectively. It is very porous, has a strong capillary action and can hold
3-4 times its weight of water.
Chemical characteristics - Vermiculite is neutral, with a pH of 7.0 to 7.5 and low EC. Like
the raw material, it has a permanent negative charge; consequently, it has a CEC value of 150-
210 meq/g and a buffering capacity for pH and cations. It also adsorbs ions like phosphate due
to its high surface area and some positive charged sites on the edges of the clay. The chemical
composition of the material, expressed as oxides (%), is given in Table 2.5.1.3.3.(1):
Table 2.5.1.3.3.(1): Chemical composition of rockwool, expressed as oxides (%).
Material SiO2 Al2O3 MgO Fe2O3
% 20-25
5-10 35-40 30-35
When the pH is low, there is a risk of toxic Al release into the solution. Mg is an important
component of the mineral structure and is the dominant adsorbed cation.
Sterilization and waste disposal - Vermiculite is a sterile product as it is produced at very high
temperatures. However, it cannot be steam-sterilized as it disintegrates during heating.
Vermiculite expanded structure collapses easily. Therefore, it is not suitable for long period of
use. It is sensitive to mechanical compression, which may grind particles to powder. Disposal
of vermiculite is not hazardous to the environment.
2.5.1.3.4. Zeolite
Production, origin and general information Zeolites are crystalline hydrated
aluminosilicates of alkali and alkaline cations that possess infinite, three-dimensional crystal
structures (Ming and Mumpton, 1989; Mumpton, 1999). Zeolites are usually formed by
metamorphism of volcanic rocks, but may also be formed from non-volcanic materials in
marine deposits or aqueous environments (Ming and Mumpton, 1989). Due to their ion
exchange, adsorption, hydration-dehydration and catalysis properties, zeolites are widely used
in agriculture and in numerous industries for the removal of pollutants from waste and
drinking water, ion exchangers, building stones, lightweight aggregates, and pozzolans in
cement and concrete (Ming and Mumpton, 1989; Mumpton, 1999). Amelioration of the
chemical and physical properties of natural zeolite is achieved by producing synthetic minerals
and since the 50s manufactured zeolites are produced all over the world (Sherman, 1999).
Zeolites (mainly clinoptilolite) are used in soil remediation to adsorb nuclear waste or heavy
metals (Kapetanios and Loizidou, 1992; Rosen, 1996; Chlopecka and Adriano 1997;
Paasikallio, 1998; Krutilina et al., 1999; 2000). They are used in agriculture as soil
amendments for: (i) a source of P, K and NH4 nutrients in infertile soils and substrates
(Hershey et al., 1980; Chen and Gabelman, 1990; Notario del Pino et al., 1994; Allen et al.,
1995; Williams and Nelson, 1997; Dwairi, 1998); (ii) reducing N losses and nitrate
contamination (Weber et al., 1983; Ferguson and Pepper, 1987; Huang and Petrovic, 1994;
Ando et al., 1996; Kithome 1999); and (iii) improving water availability (Huang and Petrovic,
1995; Yasuda et al., 1995). Additional agricultural uses of zeolites were described by Pond
and Mumpton (1984).
Zeolites possess extremely high CEC values (2200-4600 mmolc kg-1) as well as a relatively
high bulk density (1.9-2.3 mg m-3) (Ming and Mumpton, 1989) and therefore the use of zeolite
as a single component growing substrate is not recommended. However, in mixed substrates,
which include organic (peat and compost) or inorganic materials (sand and perlite), zeolites
are widely used for flower and vegetable production all over the world (Bulgaria, China,
Cuba, Italy, Japan, Jersey, Korea, Russia, USA and Yugoslavia). The experimental use of
zeolite as a single growing substrate has been reported for several crops, such as carnation
(Challinor et al., 1995), sweet pepper (Harland et al., 1999), tomato (Rivero-Gonzales and
Rodriguez-Fuentes, 1988), and gerbera (Papadopoulus et al., 1995).
Clinoptilolite is the principal zeolite mineral used in agriculture and therefore only its
physical and chemical properties will be detailed below.
Physical characteristics - Clinoptilolite has a particle density of 2-2.1 g cm-3 (Ming and
Mumpton, 1989) and saturation water content of 34% (Mumpton, 1999). However, grinding
and sieving processes affect the chemical and physical properties of the product. The physical
properties of Turkish zeolites (clinoptilolite) differed by aggregate size, but were almost
identical to that of known materials used for substrates (perlite, basaltic and rhyolitic tuff) at
the same size (Unver et al., 1989) The volumetric pore space, AFP and EAW (defined at a
water suction of: 0, 0-1.0 and 1.0-5.0 kPa, respectively) of these zeolites are detailed in Table
2.5.1.3.4.(1) (from Unver et al., 1989):
Table 2.5.1.3.4.(1): Physical properties of zeolite fractions.
Material
Bulk density
TP AFP
EAW
(mm) g cm
-
3
%
2-6
6-10
10-14
0.69
0.65
0.64
50
40
35
28
20
10
4
2
1
Chemical characteristicsClinoptilolite is very stable, but dissolves at pH 2, or lower in a
short period of time (Ming and Mumpton, 1989). Clinoptilolite has a unit-cell formula of
[(Na3K3)(Al6Si3 0O72)]24H2O and CEC of 2200 mmolc kg-1 (Ming and Mumpton, 1989).
Sterilization and waste disposal – Zeolites are not sensitive to mechanical compression and
keep their physical properties for successive crops. No decrease in performance of sweet
pepper grown on recycled clinoptilolite after steam-sterilization was observed (Harland et al.,
1999).
2.5.1.4. Pumice
Production, origin and general information Pumice is a product of volcanic activity and
usually forms from silicic lavas developed in rhyolitic composition, rich in gases and volatiles
(Challinor, 1996). Rapid releases of pressure during volcanic eruptions lead to gas expansion
and the formation of low-density materials composed of highly vesicular volcanic glass.
Pumice is common in areas rich in volcanic activity, such as the Portuguese Azores, Greek
islands, Iceland, Japan, New Zealand, Russia, Sicily, Turkey, and the U.S.A. The raw material
is mined from quarries, ground and sieved according to customer requirement. Its physical and
chemical properties are affected by its aggregate size. Pumice has been used since Roman
times as lightweight aggregates for building, stonewashing in the clothing industry, polishing
and cleaning metal, wood and glass, and as filler in the paper and plastic industries.
Physical characteristics – Pumice is a lightweight aggregate, having a low bulk density of
0.4-0.8 g cm-3, and a pore space of 70-85 % (Boertje, 1994; Challinor, 1996; Raviv et al.,
1999), depending on its origin and the sieving/grinding processes. Pumice possesses large
pores and consequently its volumetric water content decreases sharply as water tension
increases. (Boertje, 1994; Raviv et al., 1999). The water-holding capacity of pumice is
relatively low compared with rockwool, perlite or organic substrates and may limit water and
nutrient uptake by plants, especially in hot climates (Raviv et al., 1999). The volumetric pore
space, AFP and EAW (defined at a water suction of: 0, 0-1.0 and 1.0-5.0 kPa, respectively) of
three pumices are detailed in Table 2.5.1.4.(1):
Table 2.5.1.4.(1): Physical properties of pumice from several sources.
Material
Bulk density
Pore
AFP
EAW
g cm
-
3
%
Iceland
1
Sicily2
Greece2
0.4
0.7
0.6
85
70
75
40
27
41
5
4
2
1 – from Boertje, 1994; 2 – from Raviv et al., 1999.
The unsaturated hydraulic conductivity of Sicilian and Greek pumice decreased by almost 4
and 6 orders of magnitude, respectively, as water tension increased from 0 to 10 kPa (Raviv et
al., 1999). Therefore very frequent irrigation is required for plants grown in pumice.
Chemical characteristics Pumice is an inert aluminosilicate material composed primarily
of silica and Al-oxide, but may also contain metal oxides, calcite or salts. The chemical
composition of pumice is shown in Table 2.5.1.4.(2):
Table 2.5.1.4.(2): Chemical composition of pumice, expressed as oxides (%).
Material SiO2 Al2O3 CaO MgO Fe2O3 Na2O K2O
% 70-75 12-14 1-3 0.1-0.6 0.8-2.0 3-6 4-5
Pumice has no buffering capacity and possesses a very low surface charge, derived mainly
from impurities of carbonate and metal content (Silber, unpublished data). The material is
stable even at pH 2.5 (Silber, unpublished data). However, caution is recommended when
using new pumice material because high concentrations of Na are leached out at the beginning
of use.
Sterilization and waste disposal – Pumice is biologically inert and contains no pathogens or
weeds (Challinor, 1996). It is stable and can be reused practically indefinitely. Being a natural
product, it can be disposed of without causing environmental pollution.
2.5.1.5. Pyroclastic materials (tuff)
Production, origin and general information – Tuff is a common name for pyroclastic
(Greek pyro “fire”, and klastos “fragment”) volcanic material, characterized by high porosity
and surface area. Volcanic rocks are classified according to their silica content as follows
(silica percentage in the solid phase): rhyolitic (more than 65), andesitic (50-65) and basaltic
(less than 50). Rhyolitic lavas are formed during eruption at relatively low temperatures (800-
10000 C), and therefore contain predominantly light elements such as Si and Al, whereas the
Fe, Mn, Ca and Mg content are low. The viscosity of rhyolitic rocks is high and they have a
light colour. Basaltic lavas are formed at high temperatures (above 10000C), and therefore
have a higher Fe, Mn, Ca and Mg content, which induces low viscosity and dark colour.
Andesitic lavas are formed at intermediate temperatures and have intermediate colour and a
chemical composition between that of rhyolitic and basaltic lavas.
Rapid cooling of magma during eruption prevents the formation of primary minerals and,
therefore, pyroclastic materials contain mainly vesicular, volcanic glass. The physical and
chemical properties of tuff are determined mainly by its mineralogical composition and
weathering stages (Silber et al., 1994; 1999). Three tuffs erupted from the same volcano and
having almost the same chemical composition, but a differing primary mineral composition and
weathering stage, have different surface charge characteristics (Fig. 2.3.1(1)), P adsorption
capacity and dissolution kinetics (Silber et al., 1994; Silber et al., 1999). In addition, grinding
and sieving processes may affect tuff’s physical and chemical properties.
Physical characteristics – Tuff has a bulk density of 0.8-1.5 g cm-3, and a pore space of 60-
80 %, depending on its origin and the sieving/grinding processes. The water retention curves
and hydraulic properties of tuff are presented in detail in Figs. 2.2.6(1) and 2.2.8(1),
respectively.
Chemical characteristics – Tuffs have permanent and variable charge surfaces, resulting
mainly from amorphous materials. The mineralogical composition (g per 100 g-1) and surface
characteristics of three tuffs from Northern Israel differing by their weathering stage are
shown in Table 2.5.1.5.(1) (After Silber et al., 1994):
Table 2.5.1.5.(1): Mineralogical composition of various tuff types.
Mineralogical composition Black Red Yellow
Hydroxyapatite 4.5 4.6 5.0
Magnetite+TiO2+Fe2O3 12.3 11.1 8.5
Olivine 7.8 1.8 10.2
Pyroxene 7.3 14.3 10.3
Volcanic glass 62.2 29.4 3.5
Halloysite 5.0 10.0 15.0
Halloysite-like allophane - 18.8 51.5
Surface characteristics
Specific surface area (m
2
g
-
1
)
1
pH in H2O2
pH in 1 M-KCl2
CEC in pH 7 n(mmolc kg-1)3
7
9.3
7.8
107
28
8.1
6.6
285
174
7.9
6.6
601
1 - N2 adsorption; 2 90 min shaking 0.5 g 25 mL-1; 3 - The two tuffs have variable charge
surfaces as shown in Fig. 2.3.1(1).
Tuffs possess a buffering capacity and may adsorb or release nutrients, especially P, during
the growth period (Silber et al., 1999; Silber and Raviv, 1996). The chemical stability of tuffs
depends on their mineralogical composition. Volcanic glass dissolution is very rapid while that
of secondary minerals, such as kaolinite and halloysite, is slower. Hence, non-weathered
materials containing a high concentration of volcanic glass, like black tuff are unstable and
dissolve easily in solution below pH 6, while red and yellow tuffs are more stable (Silber et al.,
1999). Introducing plants to black, or even red tuff, after equilibration with an acidic electrolyte
(pH below 5) may be risky due to Al and Mn toxicity.
Sterilization and waste disposal Tuff is a stable material, which can last for several years.
Growing plants may even improve the chemical properties of tuff due to the accumulation of
organic matter and low-molecular-weight fulvic acid (Silber and Raviv, 1996). Steaming, solar
heating or chemical treatments can be used for disinfestation. Disinfestation treatments of tuff
after rose growth resulted in yields (cut roses), which were superior to those of unused tuff
(Raviv et al., 1998a). The disinfestation process per se did not affect the solid phase or stability of
the tuff, but significantly affect the solubilization of organic matter. The soluble organic compounds
released during disinfestation are adsorbed later by the tuff surfaces, thereby changing its surface
charge. (Silber and Raviv, 2001). Thus, beyond their beneficial effects on pathogen populations, the
disinfestation treatments enhanced nutritional element availability.
2.5.1.6. Synthetic organic substrates
Production, origin and general information - Mineral oil is the raw material for the
production of various foams, slabs and granules used for the furniture and construction
industries. Mineral oil products, like di-isocyanate, may be mixed with glycol to produce
polyurethane. The product is a polymer that contains excess di-isocyanate groups, which react
with water to release CO2 and induce foaming of the polymer. The resulting foams are used
for the furniture industry and the cutting residues provide the raw material for substrates. They
are ground to granules, which are pressed together with additives to form a slab. Steam at
140OC is blown in during the process. Foam is used as a growth medium for vegetables and
flowers in Belgium and The Netherlands. The granules are often used as a component in
mixtures.
Physical characteristics - Polyurethane is a very light material, with a low particle density
and very low bulk density, 1.19 and 0.078 g cm-3 respectively. Consequently, the pore volume
porosity is very high, 0.95. As most of the pores are relatively large, it holds a lot of air and
very little water under low suction (Kipp et al., 2000). The available water in the range of 10
to 100 cm suction is about 2% of its volume. There are no available data on its hydraulic
conductivity characteristics.
Chemical characteristics - Polyurethane is an indifferent substrate with a low EC and a pH
of 6. It does not contain or release any important nutrient, except Fe, Zn and B. In the past the
foam contained harmful organic substances, but this problem has now be eliminated.
Sterilization and waste disposal - Polyurethane can last for a long time, at least 10 years. It
is resistant to acids and cannot be composted. It is a sterile product that can be steamed after
use. Waste can be reused in the production of polyurethane slabs for construction, or it can be
burnt.
2.5.2. Organic substrates
2.5.2.1. Peat
The value of peat for gardening and plant production was recognized as early as the 18th
century (Perfect, 1759; Wooldridge, 1719). The origin and natural history of peat were
thoroughly described in the beginning of the 19th century (Rennie, 1807; Steele, 1826). Since
that time peat has been the subject of research conducted by botanists, pedologists, chemists
microbiologists and other scientists. Production, properties and use of peat in horticulture
were reviewed by Robinson and Lamb (1975), Puustjarvi (1977) and Bunt (1988).
Production, origin and general information - Peat is formed as a result of the partial
decomposition of plants, including species of sedges, grasses and mosses, under cool
temperatures and anaerobic or semi-anaerobic conditions. This process occurs in poorly
drained areas (peat bogs), where low nutrient levels and low pH prevail. Under these
conditions lignin and sphagnol (the lignin-like substance of mosses) cannot be decomposed,
thus the main structure of peat-forming plants remain unaltered (Given and Dickinson, 1975).
Different types of peat vary in their degree of decomposition. Plant species, climate, and the
quality of water affect the distinct characteristics of peat. Von Post (1937) suggested a
classification of peat types, based on their degree of decomposition, as presented in Table
2.5.2.1.(1).
Table 2.5.2.1.(1): Peat moss classification based on degree of decomposition.
Class Degree of Decomposition
Light peat
H1-H3
Dark peat
H4-H6
Black peat
H7-H10
The texture of peat is affected by the method in which it was harvested and processed. The
technique used to harvest peat depends on the climate and bog characteristics, such as the
presence of tree stumps in the bog. Peat is harvested from bogs by hydraulic mining or block
cutting. Peat compacts more through hydraulic mining. In this harvest method, peat is
shredded and removed from the bog by dredging, so its aeration porosity decreases.
Peat derived by the block-cutting method is cut into slabs as it is excavated from the bog
and shredded to a coarse texture. Block-cut peat has a higher TP and aeration porosity and
holds more available water than mined peat; however, water-holding porosity is lower in
block-cut peat as opposed to mined peat (Wilson, 1985).
In addition to classification on the basis of decomposition, four horticultural classifications
of peat exist, although variations may exist within any of these types. Peats of the same
classification often differ in quality, and even peats from the same bog taken from separate
layers can possess different chemical and physical properties. The type of plant material and
degree of decomposition largely determine peat’s value for use in a growing medium.
Sphagnum peat moss is the dehydrated remains of acid-bog plants of the genus Sphagnum.
Approximately 335 species of Sphagnum exist throughout the world, but mostly in northern,
cool regions. The upper layer of the bog’s profile consists of the least decomposed material. It
is typically light tan to brown in colour. It has a very light BD and is commonly used for plant
shipment, propagation or to line hanging baskets. Somewhat deeper layers, usually termed as
peat moss are somewhat darker in colour, more decomposed and slightly heavier. Sphagnum
moss is perhaps the most desirable form of organic matter for the preparation of growth
media. The Baltic States, Finland, Germany, Canada and Ireland are the principle regions of
Sphagnum moss production.
Hypnum peat moss. This type of peat consists of the partially decomposed remains of
Hypnum, Polytrichum and other mosses of the Hypanaceae family. In its stable form it
contains a high concentration of humic acid in a colloidal form. Drying out of this material is
difficult to reverse due to the hydrophobic nature of the resulting clods. This situation can be
prevented by a process of freezing and thawing, which breaks down the clods and creates a
granular structure. Although hypnum peat moss is usually less expensive than sphagnum peat,
it may contain plant pathogens or weed seeds as a result of the conditions under which it was
produced. To meet the criteria of this classification, over 90% organic matter must comprise
the oven dry weight of hypnum peat; 50% of this must represent plant material from the genus
Hypnum. Container tree seedlings should not be grown in media consisting of large quantities
of hypnum peat, but it is often used as a medium component for acid-intolerant crops.
Reed and sedge peat and Peat humus are more decomposed, and are characterised by a
high BD and low TP. They are not recommended as major components of growing media and
therefore will not be discussed further.
Main applications - Peat is, by far, the most important organic ingredient of growing media
for both nursery and greenhouse mixes and as a stand-alone substrate.
Physical characteristics The physical characteristics of sphagnum peat were thoroughly
reviewed and analysed by Heiskanen (1993). After decomposition of the readily biodegradable
parts, the resistant skeleton of the sphagnum leaves forms a very porous structure, mainly
composed of sphagnol. Sphagnol, containing numerous phenolic and hydroxylic radicals,
applies strong electrostatic forces on water molecules, thus conferring peat a very high water
holding capacity. Therefore peat is usually included in a mix to increase the water-holding
capacity or to decrease the weight of the mix. Peat is inherently hydrophobic and its wetting
may be problematic. Frequently, wetting agents or lime are used to reduce hydrophobicity.
Chemical Characteristics - Sphagnum peat contains at least 95% organic matter on a dry
weight basis (Table 2.5.2.