Content uploaded by Chris Curtis
Author content
All content in this area was uploaded by Chris Curtis on Aug 15, 2021
Content may be subject to copyright.
Wicking bed design
The effects of different reservoir media on plant growth, water use and
soil moisture in wicking beds using capillary watering
A dissertation submitted to Charles Sturt University
for the degree of Bachelor of Science (Honours)
Chris Curtis
Bachelor of Horticulture
October 2020
Wicking bed design
ii
Chris Curtis
CSU student id: 11559223
email: chris_curtis@ozemail.com.au
web: roogulli.com
Wicking bed design
iii
I , Christopher James Curtis, certify that I am the author of the dissertation titled:
which is submitted to the University on this day 26 October 2020.
I certify that this dissertation has not been previously submitted by me or any
other person to Charles Sturt University or another tertiary institution in order
to obtain an award of the University or another tertiary institution.
Should the dissertation be favourably assessed and the award for which it is
submitted approved, I agree that if requested by the Faculty or if I so choose,
I will provide at my own cost a bound copy of the dissertation as specified by
the University to be lodged in the University Library.
r I consent to the dissertation being made available for circulation and
copying for the purposes of study and research in accordance with the
normal conditions established by the University.
OR
r I have applied to the Dean of the Faculty to restrict access to the
dissertation for a period of ______ years (insert number which would not
normally exceed 2 years) on the grounds that the dissertation contains
matters of a commercial or sensitive nature.
Tick as appropriate.
Signed Date 26/10/2020
CERTIFICATE OF AUTHORSHIP OF
THE DISSERTATION AND
AGREEMENT FOR THE RETENTION
AND USE OF THE DISSERTATION
ü
Wicking bed design
iv
I ________________________________ consider that the dissertation titled
submitted by (name of candidate) for the
award of (name of the award)
is ready / not ready (delete as appropriate) for assessment and that the candidate has
completed all requirements of the program leading to the above award.
Additional comments:
Signature of Supervisor Date
CERTIFICATION OF
DISSERTATION READINESS
FOR EXAMINATION
Wicking bed design
v
ACKNOWLEDGMENTS
Firstly, I would like to acknowledge the support and encouragement I have received
from my wife, Jennie, without which none of this would have happened.
My supervisors, Dr Ben Stodart and A/Prof Phil Eberbach have provided excellent
advice in reviewing my work and discussing my results.
Peter Mills provided the inspiration to undertake this research and valuable advice in
designing the project.
Dr Richard Lang answered many questions and provided feedback on my writing and
research results.
Bob Stevenson from WaterUps from DownUnder provided the WaterUps® modules
used in the experiments at no charge and with no conditions attached.
And finally, I would like to thank members of Canberra City Farm and various online
gardening and permaculture communities for their keen interest and probing
questions about my research.
Wicking bed design
vi
ABSTRACT
Wicking beds are planting containers that have a reservoir of water in the lower
portion providing moisture to plants using capillary action. The scientific literature has
findings from a few studies that wicking beds have a higher yield and greater water use
efficiency than top watered containers but no research has been found about the
effects of different materials in the reservoir layer. This study investigated the capillary
rise, water holding capabilities and performance in wicking bed reservoirs of several
materials.
Capillary rise of water in various materials was measured in Perspex tubes.
Crusher dust had the greatest capillary rise, followed by sand, fine perlite and a
cocopeat/compost/sand mix. Gravel and scoria had poor capillary rise.
Wicking beds were constructed with four reservoir treatments – cocopeat mix,
sand, gravel and WaterUps® with medium grade perlite as the wicking medium. A
cocopeat/compost/sand mix was used as the growing medium for each reservoir
treatment. A commercial potting mix was also used with a sand reservoir. Three
replicates of each treatment were performed. Two crops were grown sequentially:
spinach then butterhead lettuce.
For the spinach crop, the cocopeat and sand/cocopeat beds grew the greatest
plant weight followed by WaterUps®, gravel, and sand/potting mix. Soil moisture at
150mm depth was lowest in gravel, followed by WaterUps®, sand/cocopeat, cocopeat
and sand/potting mix.
For growing lettuce, the wicking material in the WaterUps® was changed to sand.
There was no significant difference in the weight of lettuce grown in any of the
treatments. Soil moisture at 150mm depth remained reasonably constant throughout
the growing period for WaterUps® and cocopeat. Gravel and sand/potting mix dried
the most. The potting mix remained wettest of all treatments at 200mm depth but was
driest at 100 and 50mm depths indicating that it had poor capillary rise capabilities.
This study found the reservoir material has an effect on the soil moisture and
plant growth in wicking beds. Although often used, gravel appears a poor choice since
it results in the driest growing medium. The reservoirs with cocopeat mix and sand
delivered better soil moisture and plant growth. WaterUps® with sand as the wicking
material also delivered a high level of soil moisture.
Wicking bed design
vii
TABLE OF CONTENTS
1!INTRODUCTION 1!
2!REVIEW OF LITERATURE 4!
2.1!Wicking beds 4!
2.2!Sub-irrigation 6!
2.3!Capillary rise 8!
2.4!Saturated water holding capacity 13!
2.5!Growing media 14!
2.6!Soil moisture content 17!
2.7!Surface salt accumulation 19!
3!METHODS 20!
3.1!Water holding capacity 20!
3.2!Capillary rise 21!
3.3!Wicking bed trials 22!
3.3.1!Wicking bed trial 1 (WBT1) 22!
3.3.2!Wicking bed trial 2 (WBT2) 29!
3.3.3!Wicking bed trial 3 (WBT3) 31!
3.4!Data analysis 32!
3.5!Electronic tensiometers and data logger 33!
4!RESULTS 34!
4.1!Water holding capacity of potential reservoir media 34!
4.2!Reservoir capacity of wicking beds 35!
4.3!Capillary rise in reservoir materials 36!
4.4!Capillary rise in growing media 39!
4.5!Wicking bed trial 1 (WBT1) - Spinach 40!
4.5.1!Plant weight 40!
4.5.2!Water use 42!
4.5.3!Soil moisture 43!
4.5.4!Reservoir water levels 46!
4.5.5!Plant canopy area 48!
4.6!Wicking bed trial 2 (WBT2) - Lettuce 49!
4.6.1!Plant weight 49!
4.6.2!Water use 50!
4.6.3!Soil moisture 51!
4.6.4!Reservoir water levels 57!
4.6.5!Plant canopy area 59!
4.6.6!Plant root growth 60!
4.6.7!Soil electrical conductivity 60!
4.6.8!Temperature and humidity 61!
4.7!Wicking bed trial 3 (WBT3) - lettuce in small wicking beds 63!
4.7.1!Plant weight 63!
4.7.2!Water use 63!
4.7.3!Soil moisture 63!
4.7.4!Soil electrical conductivity 65!
4.7.5!Soil temperature 66!
4.8!Comparison between WBT1 and WBT2 66!
5!DISCUSSION 69!
5.1!Media wicking ability and effect on choice of reservoir medium 70!
Wicking bed design
viii
5.2!Water holding capacity 73!
5.3!Growing medium 74!
5.4!Plant growth 75!
5.5!Soil moisture 76!
5.5.1!Soil moisture measured by tensiometer 76!
5.5.2!Soil moisture at different depths 79!
5.5.3!How often to refill wicking beds 82!
5.5.4!Plant water use 82!
5.6!Wicking bed size 83!
5.7!Effect of geotextile separating reservoir and growing medium 84!
5.8!Electrical conductivity 85!
6!CONCLUSION 86!
LITERATURE CITED 89!
APPENDIX 1 - SELECTION OF WICKING BED DESIGNS FROM POPULAR LITERATURE 96!
APPENDIX 2 - ARDUINO CODE FOR DATA LOGGER 101!
Wicking bed design Introduction
1
1 INTRODUCTION
Wicking beds are planting containers that have a reservoir of water in the lower
portion and a growing medium above. Water from the reservoir wicks up through the
growing medium to supply the plants. Figure 1 shows a typical wicking bed design.
Various wicking media such as gravel, scoria, sand, woodchips, soil and plastic
frameworks have been proposed for filling the reservoir layer. In some designs a layer
of geotextile is used above the reservoir layer to prevent the growing medium from
mixing with the reservoir. Other common features of wicking beds include an overflow
outlet at the top of the reservoir layer to prevent excess water from flooding the
growing medium, and a fill pipe that allows delivery of water directly into the reservoir
layer.
Figure 1 - Cross section of typical wicking bed design showing the major components
The beds are a popular way of growing vegetables in home gardens and have been
used in some small scale urban farms. Apart from in these urban farms, wicking beds
do not appear to be used to any significant extent in commercial horticulture.
Little has been written in the scientific literature about wicking beds but there are
many articles about them in the popular press. These range from publishers that may
be thought to have some authority, such as ABC’s Gardening Australia, to people with
unknown experience publishing blogs or YouTube videos.
Only three scientific papers describing research into wicking beds have been found. All
these papers concentrated on comparing yield and water use efficiency of wicking
beds with conventional planters and did not investigate alternate wicking bed designs
or media to any great extent. However, papers dealing with subirrigation of plants in
containers in the nursery industry and aspects of hydroponic production may provide
useful guidance for the design of wicking beds. A review of literature about capillary
irrigation noted the almost complete lack of published papers on wicking beds and
commented that:
Wicking bed design Introduction
2
“Due to the lack of research on wicking beds, we contend that there is a specific -
and very important - knowledge gap relating to the verification of performance,
and of design guidance, of wicking beds.” (Semananda, Ward, & Myers, 2018)
One of the major differences in wicking bed designs that appear in the internet and in
popular publications is the question of what to use in the reservoir layer of a wicking
bed. To be effective, the reservoir medium must meet a number of criteria including:
• sufficient structural strength to support the growing medium above the reservoir
• the ability to wick water from the full depth of the reservoir to the growing
medium
• the ability to wick water at a sufficient rate to match evapotranspiration of the
plants growing in the wicking bed
• sufficient pore space to store an adequate volume of water within the reservoir;
the larger the pore space the more water can be held and the longer the interval
between watering events
• not having any detrimental effect on the growth of plants in the wicking bed.
Because of the popularity of wicking beds among sections of the gardening community
and the diversity of unverified designs that are being proposed in the popular
literature and social media, there is a need for rigorous research to compare various
wicking bed designs. This research was undertaken to, at least partially, fill that
knowledge gap.
Research aims
The main aim of this project was to examine and compare plant growth and water
movement by capillary action in wicking beds using different media in the water
reservoir layer in order to determine the best reservoir media for use in wicking beds.
The hypothesis that was tested by this research was that the choice of material used in
the reservoir layer of a wicking bed affects plant growth and moisture distribution
within the growing media.
To achieve this aim, the research objectives were:
1) to determine the rate and extent of upward water movement by capillary
action in the various potential reservoir media.
Wicking bed design Introduction
3
2) to compare saturated water holding capacity of common reservoir media and
assess the impact of this on how often the reservoir needs to be refilled.
3) to identify the effects of different wicking media on soil moisture and water use
within wicking beds.
4) to assess the impact of wicking bed design on plant growth.
Wicking bed design Review of literature
4
2 REVIEW OF LITERATURE
This chapter presents a summary of the published literature on wicking beds. Because
of the limited amount of wicking bed research that has been published, literature on
other topics that are relevant to the design and operation of wicking beds is also
reviewed. These topics include: subirrigation of container plants, capillary rise in soils,
saturated capacity of soils, growing media for container planting, the effects of soil
moisture content on plant growth, and accumulation of salts on the surface of planting
containers.
A summary of selected wicking bed articles in the popular press is attached at
appendix 1.
2.1 Wicking beds
Three papers present research into wicking beds. Semananda, Ward, and Myers (2016)
compared the yield and water use efficiency of tomatoes grown in wicking beds and
surface irrigated containers, Sullivan, Hallaran, Sogorka, and Weinkle (2015) compared
the yield of tomatoes grown in wicking beds with top watered raised planters and the
number of irrigation events for each, and Semananda, Ward, and Myers (2020)
compared yield and water use efficiency of lettuce and radish in wicking beds and
surface watered pots both with and without a mulch layer. Beyond this, there appears
to be no published research specifically involving wicking beds. In a review of
subirrigation literature, Semananda et al. (2018) recommend that further research
should be done into soil properties and depths of the reservoir and growing media for
wicking beds.
Semananda et al. (2016) used wicking beds with gravel in the reservoir covered with
geotextile and a “poorly graded sandy soil” as the growing medium. Reservoir depths
were 150mm and 300mm and soil depths were 300mm and 600mm. One wicking bed
treatment had a wick of soil extending into the reservoir layer. The yield of marketable
tomatoes was higher in all wicking bed treatments than the surface irrigated
containers, but the deeper soil produced no more marketable fruit than the shallower
soil. Water use efficiency (WUE) was greater in the wicking beds with 300mm deep soil
than the 300mm deep surface watered containers, but there was no difference in WUE
between the 600mm deep wicking beds and surface watered containers.
Wicking bed design Review of literature
5
The depth of the reservoir layer (150mm or 300mm) used by Semananda et al. (2016)
did not affect the WUE of the wicking beds. Water was added to the wicking beds
when the average soil moisture dropped below 75% of field capacity and the reservoir
depth did not affect the number of watering events required. Although the wicking
bed design included a clear tube as a manometer to view the water level in the
reservoir, no information was provided about the water levels in the reservoir when
the soil had dried to a point where irrigation was required. This may suggest that the
gravel in the reservoir layer could not wick water higher than 150mm and that the
water in the bottom half of the 300mm deep reservoir was not used. It is reasonable
to assume that the finer textured soil would have greater capacity for capillary rise
than the gravel and that using a column of soil as a wick into the reservoir layer would
allow more of the reservoir water to be used and thus require less frequent water
applications. However, the wicking bed with a 300mm deep reservoir and a soil wick
needed the same number of irrigation events as the wicking beds with just gravel in
the reservoir. Semananda et al. (2016) did not measure the wicking capability of the
soil so it is unknown how effective the soil wick would have been at moving water into
the upper layers of the soil compared with gravel.
The productivity of wicking beds compared with conventional raised beds was also
investigated by Sullivan et al. (2015) who described them as "subirrigated planters"
rather than wicking beds but they met the definition of wicking beds by having an
integrated reservoir layer in the base of the beds. These wicking beds had a void in the
reservoir layer created by a coil of flexible drainage tubing to create a 100mm deep
reservoir. Wicking beds were tested both with and without a layer of landscape fabric
separating the reservoir and growing medium. The growing medium was a 460mm
deep layer of commercial potting mix comprised of a mix of compost, peat and perlite.
It was not explicitly stated by the authors, but it can be inferred from their design
diagrams, that some of the potting mix would have filled spaces around the outside of
the drainage pipe and thus provided a capillary path from the base of the reservoir
into the growing medium.
There was no significant difference in the production of cherry tomatoes between the
wicking beds with and without the landscape fabric barrier, and the wicking bed
without landscape fabric produced a slightly larger crop of tomatoes than the
Wicking bed design Review of literature
6
conventional raised bed planter (Sullivan et al., 2015). This study found that the
wicking beds required one fifth of the watering events of the conventional beds but
did not record the quantity of water applied to the wicking or conventional beds.
Wicking beds with a 150mm deep gravel reservoir and 300mm of soil had greater
water use efficiency and plant yield than conventionally watered containers growing
lettuce and radish (Semananda et al., 2020). The wicking beds were refilled every two
weeks with an amount of water based on the drop in soil moisture, but the variation in
water levels within the reservoir is not reported. Yield was also higher in beds with a
layer of gravel mulch on the surface.
The research to date on wicking beds has been very limited and leaves many areas
unexplored. For example, there has been no comparison of different reservoir
materials or growing media, little reporting of moisture levels in the growing media,
and no measurement of how much of the water contained in the reservoir is used in
the wicking bed before more water needs to be added.
2.2 Sub-irrigation
Subirrigation systems provide water and often fertiliser solutions to the base of
containers. Water and nutrients are then transported to the roots of the plant in the
containers by capillary action through the growing medium (Ferrarezi, Weaver, Van
Iersel, & Testezlaf, 2015). Various forms of capillary or subirrigation are used in the
commercial hydroponics industry and there has been much more research published
about these systems than about wicking beds. Although these subirrigation systems do
not have the integrated water reservoir of wicking beds, they are similar to wicking
beds in that they rely on capillary rise to move water into and through the growing
medium and thus much of the research into subirrigation systems is potentially
applicable to wicking bed systems.
Subsurface irrigation includes capillary mats, troughs, and flooded trays, benches and
floors (also called ebb-and-flow systems) (Raviv & Lieth, 2008). In flooded benches and
other ebb and flow systems, plant containers are placed on raised waterproof benches
with low sides. Periodically, the benches are filled with water pumped from a reservoir
and the base of the containers are flooded. After sufficient water has risen into the
pots through capillary action, the benches are drained with the excess water returned
to the reservoir. Flooded floor systems are similar but on a larger scale where the
Wicking bed design Review of literature
7
whole floor of the growing area is periodically flooded (Ferrarezi, Weaver, et al., 2015).
The flooding level does not have to be deep and the growing medium can be fully wet
by a depth of about one fifth of the container height (Evans, Barrett, Harbaugh, &
Clark, 1992). In trough systems, containers are placed in gently sloped troughs and
water pumped into the higher end of the trough. Any water not absorbed by the
plants is drained from the lower end of the trough into the reservoir (Semananda et
al., 2018). Capillary mat systems consist of an absorbent fabric with an underneath
layer of waterproof plastic and a covering of perforated plastic. The mat is placed on
horizontal benches and flooded with water. Plant pots with holes in their base are
placed on the perforated surface of the mat and water wicks into the growing medium
by capillary action. The perforated covering allows water through into the pots while
preventing evaporation from unused areas of the mat. Providing the media used is fine
enough to allow capillary rise, pots up to 21cm high are suitable for use with capillary
mats (Schuch & Kelly, 2006).
Subirrigation systems have been developed to capture and recycle unused nutrient
solutions. This has become necessary due to increasing environmental and legal
requirements to minimise leaching and drainage of excess nutrients away from plant
production sites (Roeber, 2010; Yeager & Henley, 2004). Systems have been developed
to capture and recycle drainage from top-watered containers but the risk of spreading
pathogens between containers in this type of system is high. There is less risk in
recycling nutrient solutions from subirrigated containers because most of the unused
nutrient solution has not been in contact with the media in the containers (Biernbaum,
1992; Roeber, 2010). George, Biernbaum, and Stephens (1990) grew geranium
seedlings in an ebb and flow system with a shared nutrient solution reservoir. Some
pots were inoculated with varying levels of Pithium ultimum and other pots were not
inoculated. After eight weeks, all plants in inoculated pots showed effects of the
disease but none of the uninoculated plants were affected. However, the reservoirs of
the trays containing pots with the highest level of inoculation did contain P. ultimum
and this may have spread to the unaffected plants if the experiment was continued for
longer. Wicking beds address the issue of disease spread firstly by not having any
excess water that needs to be prevented from leaving the site, and secondly by having
a separate reservoir for each container and thus having no risk of spreading pathogens
through sharing of solutions.
Wicking bed design Review of literature
8
As well as reducing the amount of water used compared with top watering,
subirrigation can result in greater shoot growth and higher shoot/root ratio than
overhead watering (Frangi, Amoroso, Piatti, & Faoro, 2011; Piatti, Frangi, & Amoroso,
2011). Greater shoot and root growth was observed in wheat plants watered from
below compared with top-watered plants (Singh, 1922). The continuous supply of
water and nutrients from the capillary mat limits water stress and reduces the need for
root growth, thus allowing greater shoot growth. Frangi et al. (2011) also found that
leaves of the plants grown on the capillary mat had a higher nitrogen, phosphorus and
potassium levels than the overhead watered plants.
One key difference between wicking beds and most subirrigation used in commercial
hydroponic systems is that the water reservoir is constantly supplying water to the
growing medium in a wicking bed whereas subirrigation systems typically use
intermittent watering that allows some drying of the growing media between
irrigation events. What effect this difference may have on growing media selection or
plant growth in wicking beds does not appear to have been examined.
2.3 Capillary rise
Capillary rise of water through the reservoir and growing media is the method by
which water travels from the reservoir to the plant roots in a wicking bed. Therefore,
the capillary rise capability of the media used should affect the operation and
productivity of a wicking bed. Important characteristics of capillary rise are the height
of capillary rise within the medium, the water storage capacity of the capillary system,
and the amount of water that can be transported to the root zone via capillary rise in
a given time (Liu, Yasufuku, Miao, & Ren, 2014). A review by Salim (2016) indicates
that there were contradictory values for the height of capillary rise in sands, silts and
clays, and that most quoted figures are estimates based on mathematical models,
rather than measured values.
Wicking bed design Review of literature
9
Figure 2 - Capillary rise (h) of liquid in a tube or radius (r) showing contact angle (
Q
)
Factors that affect capillary rise include the diameter of the tube formed by the grains
of the medium, the contact angle between the liquid and the solid surface, the density
and viscosity of the liquid, and the amount of surface tension. Capillary rise is inversely
related to the size of the tube; large pores formed between large particles will have
less capillary rise than small pores formed by small particles. The
hydrophobic/hydrophilic properties of the surface will affect the contact angle. If the
surface is hydrophilic the contact angle will approach 180o and the surface will wet and
capillary rise will occur. If the surface is hydrophobic, the contact angle will approach
0o, the surface will remain dry and no capillary rise will occur. However, soils are not
uniform; tube diameters are not consistent, tubes are not continuous, and the nature
of the particle surfaces vary. This creates difficulties in applying mathematical models
of capillary rise to real soils (Salim, 2016).
Keen (1919) gives a formula (1) for estimating capillary rise in an ideal homogenous
soil with spherical grains.
!" #"$
%
&'
()*
(1)
where
!
is the height of capillary rise,
+
is surface tension,
,
is density of water,
-
is
the force of gravity and
.
is the side of an equilateral triangle approximating the pore
size. After applying some assumptions about values for surface tension, gravity and
density, equation 1 is simplified to:
!" #"/012
3
(2)
where
4
is the radius of the soil particles and the units are centimetres.
Wicking bed design Review of literature
10
Keen notes that values are likely to be less in field conditions due to non-uniformity of
soil particles. Washburn (1921) developed equation (3) to define the distance a liquid
will flow through a capillary tube was developed by Washburn (1921) which has been
widely used as the basis for further work.
5 #
6
73809:;<
=>
(3)
where
5
= distance of flow,
?
= surface tension,
4
= pore radius,
@
= contact angle and
A"
= viscosity.
The Young-Laplace equation (4) is widely used to calculate vertical capillary rise (
!9
). It
is based on work by Thomas Young in 1805 and Pierre-Simon Laplace in 1806 (Masoodi
& Pillai, 2012).
!9#=709:;<
()3
(4)
Lu and Likos (2004) compare the Washburn equation and his experimental results with
calculations based on Terzaghi's 1943 book (Theoretical Soil Mechanics) and develop
their own equations to predict capillary movement of water in soils.
Capillary rise in sand and loam is affected by moisture content and hydrophobicity of
the surfaces. Examination of movie films of the wetting front in soils sandwiched
between two glass plates showed that capillary rise is not continuous at a constant
rate, but rather proceeds in a series of irregular movements (Wladitchensky, 1966).
The process leading to this irregular movement, as described by Wladitchensky (1966),
is:
"According to all these considerations one may picture the mechanism of capillary
movement of water in a single pore as following. The water penetrates into a
pore, wets with a thin film the pore walls, the wetted pore wall surface being
greater when the wetting of walls is more perfect. The film 'tows' the water
column. As the weight of water column increases the curvature of the water
surface on the upper boundary of the water column decreases. This process
proceeds until the surface reaches the curvature of static wetting angle. At this
moment the capillary movement ceases. If the pore wall above the static meniscus
edge can adsorb moisture, the capillary rise does not end. The pore surface
adsorbs water vapour, water condensation occurs, and a film of adsorbed water is
formed on the pore-wall. The wetting of the solid phase increases. A water film is
formed again. This film gives rise to a new jump of capillary water to the boundary
of the next more hydrophobic part of pore wall surface."
(Keen, 1919) reports results from Loughbridge on capillary rise in four Californian soils
that showed a faster rate of rise in coarser textured (sandy) soil, slower rise in fine
Wicking bed design Review of literature
11
textured (clay) soil, and a higher total rise in fine textured soil (50 inches) than in sandy
soil (16.5 inches). The level of water in the finer soils was still rising after 195 days.
Results from field trials with in-situ soil showed that capillary rise was slow and not to
a great height but also showed a greater capillary rise in sandy loam than clay loam
(Bouyoucos, 1947). A further report on these field trials (Bouyoucos, 1953) concludes
that capillary rise is too small and too slow to provide water to plants in field
conditions. The author suggests that rather than plants obtaining water from capillary
rise, roots grow fast to find available water. Bouyoucos (1953) suggests reasons why
capillary rise doesn't happen in practice: 1) soil holds on to water more as moisture
tension rises preventing free movement of water; and 2) water movement is impeded
by friction in capillaries, translocation and swelling of colloids blocking capillaries, air
pockets blocking capillaries, and water films and capillaries becoming discontinuous.
Laboratory experiments have demonstrated that the height of capillary rise is affected
by the cross-sectional area of the container. Wadsworth and Smith (1926) conducted
an experiment with square glass-fronted columns filled with screened soil placed in a
container of water. Total capillary rise increased with an increase in the cross-section
of the soil column from one square inch up to 25 square inches. The height of
maximum water content above the free water table also increased with cross-sectional
area up to 16 square inches. The rate of capillary rise was faster in the larger columns
and the smaller columns reached the maximum height in less time than the larger
columns. The authors were not able to explain why these results occurred.
Research conducted by Albaho (2006) demonstrated that the cross-sectional area of
wicks supplying water to potted plants affected the amount of water uptake.
Tomatoes growing in pots suspended over a water reservoir with wicks extending from
the base of the pots down to the water produced more high quality fruit when the
cross-sectional area of the wicks and the growing medium was larger. During high
water demand at flowering and fruit expansion growth stages, more water was taken
up by wicks with a larger cross-sectional area. These findings may be significant when
evaluating the performance of wicking beds that use a limited wicking area (for
example: WaterUps® - www.WaterUps.com.au) compared to sand or gravel wicking
beds where the entire cross-sectional area of the bed can wick water.
Wicking bed design Review of literature
12
Raes and Deproost (2003) note that calculating the rate of upward movement of water
in soil depends on, as well as the capillary properties of the soil, the depth to the water
table, the water content in the root zone, water up-take characteristics of the plant
roots, and the evapotranspiration rate. They have developed a computer program
UPFLOW (Raes, 2004) available from iupware.be/?page_id=883 that uses input
parameters for these and other factors and predicts the order of magnitude of the
likely capillary rise. Utilities such as these may be useful in determining the efficacy of
different wicking bed designs.
Organic compounds derived from decomposing or living microorganisms or plants can
cause soil to become water repellent or hydrophobic and resist wetting for a period.
This can reduce the amount of capillary rise or result in preferential water pathways
and uneven wetting through the growing medium (Raviv & Lieth, 2008). No capillary
rise occurred in air dried hydrophobic peats but hydrophobicity dropped as water
potential (wetness of media) increased (Michel, Rivière, & BellonFontaine, 2001).
Pike (1979) recommended that to establish a water column through the medium and
start capillary watering it was necessary to first water the pots from overhead. This
may serve to evenly wet the medium in the pots and avoid these problems described
above. However, most researchers experimenting with subirrigation do not mention
this. Exceptions are Hoffman, Buxton, and Weston (1996) and Wilfret and Harbaugh
(1977) who moistened the soil in pots before placing the pots on capillary matting.
Capillary rise is related to the unsaturated hydraulic conductivity of the soil. Substrates
that have poor unsaturated hydraulic conductivity will not distribute water adequately
above the level to which the base of the root zone is flooded (Raviv & Lieth, 2008).
Unsaturated hydraulic conductivity depends on many factors such as the pore-size
distribution of the medium, and the tortuosity, shape, roughness, and degree of
interconnectedness of the pores. The hydraulic conductivity decreases considerably as
soil becomes unsaturated since less pore space is filled with water, the flow paths
become increasingly tortuous, and drag forces between the fluid and the solid phases
increase (Lu & Likos, 2004; van Genuchten & Pachepsky, 2011). Unsaturated hydraulic
conductivity at a given potential can be raised by increasing the proportion of fines in a
porous medium. Adding fine material to coarse substrates can therefore be expected
to increase capillary rise (Caron, Elrick, Beeson, & Boudreau, 2005; Heiskanen, 1999).
Wicking bed design Review of literature
13
The scientific literature on capillary rise debunks a "fact" that frequently appears in the
popular articles about wicking beds: that water will not wick up more than 250 or
300mm. Providing a suitable medium is used, that is clearly not true. The evidence
presented in the literature demonstrates that large particle sizes reduce wicking, and
this casts doubt on the common practice of using gravel in the reservoir layer of
wicking beds. However, the actual performance of such media in wicking beds has not
been measured.
2.4 Saturated water holding capacity
The reservoir layer of a wicking bed has two main functions: to wick the water up to
the growing media so it is available to plants (as described above) and to store water
that can be used to wet the growing media. The amount of water that can be stored in
the reservoir media is related to the saturated water capacity of the media.
Water is stored in the pores between the solid particles in the media. The porosity of a
medium can be calculated from its bulk density and particle density and, in theory,
when the medium is saturated all the pore space is occupied by water (Rose, 1966).
Most horticultural, agricultural and engineering applications are more interested in
how fast water drains from a medium once the water content is above field capacity
rather than the total amount of water a medium can store. However, the water
holding capacity is of interest when designing drainage trenches for applications like
septic tanks or stormwater drainage where it can be advantageous to temporarily
store water when inflows exceed the infiltration rate of the soil surrounding the trench
(Quisenberry, Brown, & Smith, 2006; Sieker, 1998).
Quisenberry et al. (2006) measured the storage capacity of several subsurface waste
storage products including a gravel filled trench. They found that the measured
capacity of the gravel filled trench was the same as the calculated capacity based on
the volume of the trench and the measured porosity of the gravel. However, they also
stated, without explanation, that 10% difference between the measured and
calculated capacities of gravel systems can be expected.
There is a high correlation between the saturated water capacity of a soil and both the
bulk density and organic content of the soil. This was demonstrated by Yi, Li, and Yin
(2013) who developed artificial intelligence systems to predict the saturated and field
Wicking bed design Review of literature
14
capacity of soil based on the amount of clay, silt and sand, the bulk density and the
organic matter content of the soil.
2.5 Growing media
The growing media layer in a wicking bed needs to provide a suitable rhizosphere
environment for the growing plants. It should provide physical support for the roots
and the above ground parts of the plants, have adequate water holding capacity,
contain air-filled pores to provide oxygen and provide a nutrient supply to the plants
(Wilkinson, Landis, Haase, Daley, & Dumroese, 2014). In hydroponic growing, plants
are grown in nutrient enriched water and therefore the nutrient holding capacity of
the media is of little concern. For wicking beds, the growing media must hold and
supply the nutrients needed by the plants.
Wicking beds are a form of container growing and the media used in them must, in
some way, be artificially constructed (even if it is just removing soil from the ground
and placing it in the wicking bed). Media developed for container growing where
overhead irrigation is used generally have good drainage characteristics and high air-
filled porosities. The larger particle sizes that provide these characteristics result in
poor capillary rise properties that make them less than ideal for subirrigation (Caron et
al., 2005). In 2002, Boudreau et al. (cited in (Caron et al., 2005) found that the
efficiency of capillary mat systems was being limited by the substrates then in use. In
very light and open mixes water does not reach the upper levels of the pots due to a
reduced number of capillary pores. A tighter mix with more fine particles provides a
better capillary path and the greater the height of the pot, the tighter the mix needs to
be to transport water to the top (Evans et al., 1992). Peat, perlite, rockwool, bark,
compost and mixes of these materials are all used in subirrigation systems
(Semananda et al., 2018). Therefore, careful consideration is required when
developing the growing media to be used in wicking beds.
Early work in developing artificial mixes for container growing was done at University
of California. (Matkin & Chandler, 1957) described five soil mixes with varying
percentages of fine sand, sphagnum peat moss and fertiliser. The combination of fine
sand and peat provided a reproducible medium that had similar water and nutrient
retention capabilities to loam. They recommend either 50:50 or 25:75 sand:peat moss
mix for use in pots. "Peat-lite" potting mixes were developed at Cornell University in
Wicking bed design Review of literature
15
the 1960's and this work provides the background for many modern potting mixes
(Boodley & Sheldrake, 1972). The peat-lite mixes vary based upon intended use, with
50% sphagnum peat combined with either 50% vermiculite (mix A), 50% perlite (mix B)
or 25% vermiculite and 25% perlite (mix C). However, the capillary rise properties have
not been reported for the UC or Cornell peat-lite mixes.
When considering the properties of peat and its role within the growing medium, the
source plant may be a critical consideration. Sphagnum peat, derived from mosses, is
composed of structures that have a network of hollow vessels and pores that provide a
large capacity for holding water and good capillary rise properties. Sedge peat is
decayed grasses and does not have the same hollow fibrous structure. Water retention
and capillary movement in sedge peat is largely in the spaces between, instead of
within, the fibres and this is less effective than in sphagnum peat (Caron et al., 2005).
Because of these properties, sphagnum peat is more widely used in horticultural
applications than sedge peat, although both are used.
The advantages of peat moss include its ability to hold large amounts of water, air and
nutrients and it is widely used in container growing media, particularly in North
America and Europe but concerns have been raised about the environmental impacts
of continuing to mine ancient peat bogs (Robertson, 1993). Alternatives to peat
include composts, coconut coir, fresh or composted rice hulls, and pumice. Composted
bark is now widely used as the main ingredient in potting mixes (Landis & Morgan,
2009). In a trial using composted municipal solid waste (MSW) to grow woody plants
(Cotoneaster), the plants performed poorly in 100% composted bark or MSW (40g
shoot dry weight) but good growth occurred in mixes of MSW and at least 25% peat
(60-70g shoot dry weight) (Hicklenton, Rodd, & Warman, 2001).
Coir is frequently used as a substitute for peat. It is a by-product of the coconut
industry and therefore may be regarded to have some improved environmental
credentials, although transport impacts from tropical regions where it is produced are
significant (Landis & Morgan, 2009). If experimental results using peat are to be
interpreted with a view to substituting coir for peat, it is important to understand the
comparative performance of the two media. Compared with peat based mixes, coir
has higher aeration, lower water holding capacity, less easily available water and, at
water tension above 2.5kPa or volumetric water content below 40%, greater hydraulic
Wicking bed design Review of literature
16
conductivity (Abad et al., 2005; Londra, 2010; Raviv, Lieth, & Wallach, 2001). No
studies have been found that investigated capillary rise properties of coir. Londra
(2010) surmised that due to the higher hydraulic conductivity, the coir based mix
would be better able to meet plant water demands during high periods of
evapotranspiration. This may indicate that coir would be a good ingredient in a
substrate mix for use in wicking beds.
Much work has been done in comparing top watering with various types of
subirrigation (Colla et al., 2003; Cox, 2001; Elia, Santamaria, Parente, & Serio, 2003;
Frangi et al., 2011; Klock-Moore & Broschat, 2001; Santamaria, Campanile, Parente, &
Elia, 2003). Results from these studies vary and it is not possible to definitively
conclude that one method of irrigation is better in all circumstances. Usually, the same
medium has been used in both the top-watered and subirrigated samples and no
consideration has been given to the different physical requirements that the different
irrigation systems may impose on the media. Comparisons of an ideal medium for
subirrigation with an ideal medium for top watering may produce different results.
Some work has been done to compare the effectiveness of different media with
subirrigation. Water content of two media (peatmoss and perlite 7:3 and 5:5 (v/v)) in
three sub-irrigated systems (nutrient flow wick, nutrient stagnant wick and ebb-and-
flow) was higher in media with greater amount of peatmoss in all irrigation systems
and was higher in smaller pots (6.3cm high) than larger pots (9 and 13.5cm high) (Oh,
Cho, Kim, & Son, 2007). Comparison of three peat/composted bark/sand mixes found
that a mix of 60% sphagnum peat with 30% bark and 10% sand had better capillary
rise, faster plant growth, less water use and more even water potential throughout the
medium than mixes with 30% sphagnum peat/60%bark/10%sand or 60% sedge
peat/30% bark/10%sand (Caron et al., 2005). The best performing medium had the
highest percentage of fine particles with a mean weight diameter of 2.93mm and 50%
of the particles less than 1mm diameter.
Wesonga, Wainaina, Ombwara, Masinde, and Home (2014) conducted capillary rise
tests on four mixes (SSM - 3 parts forest soil, 2 parts sand and 1 part manure; SCMP - 2
parts forest soil, 4 parts cocopeat, 1 part manure and 1 part pumice; SCMS - 2 parts
forest soil, 4 parts cocopeat, 1 part manure and 1 part sand; CMP - 4 parts cocopeat, 1
part manure and 2 parts pumice). They found that SSM>SCMP>SCMS>CMP based on
Wicking bed design Review of literature
17
amount of water absorbed, but did not report the height of capillary rise achieved.
Neither did they report the physical properties of the forest soil they used which limits
the applicability of these results to other soil samples.
Perlite has shown good performance in subirrigated systems. Perlite absorbs more
than twice the amount of water than pumice, peat or mixes of pumice and peat (Elia et
al., 2003). However, differences between substrates can be compensated for by
measuring the water tension in the substrate and adjusting the watering regime
accordingly (Elia et al., 2003).
A number of questions about the best growing media for wicking beds remain.
Because most subirrigated systems use pots that are not as deep as most wicking beds,
the capillary rise capabilities of media to sufficient height for wicking beds has not
been investigated. Most media used in subirrigated systems use significant proportions
of sphagnum peat, perlite or vermiculite which have little nutrient content and
reasonably high environmental impacts. The suitability and performance of media that
are more sustainable and are able to supply nutrients to plants without the constant
irrigation with a nutrient solution is still to be researched.
2.6 Soil moisture content
A good supply of water is essential for plant growth and soil water potential (Y) is a
reflection of the energy that plants need to extract water from the soil. As Y drops
below -0.25MPa, cell expansion slows; below -0.5MPa cell wall and protein synthesis
slows; and photosynthesis and stomatal conductance drop at -1MPa. Around
Y = -1.5MPa is generally regarded as permanent wilting point below which plants will
not recover, although some plants are adapted to survive much lower Y values. Field
capacity is the maximum amount of water a soil will hold after free-draining water has
drained away. In natural soils this is generally Y between -10 and -33kPa. Too much
water is also detrimental to plants. In saturated soils, pores that normally would be
filled with air become filled with water, oxygen is unavailable to plant roots and root
tissues become hypoxic (Taiz, Zeiger, Moller, & Murphy, 2015).
Although wicking beds can use natural soil as a growing medium, they are most
frequently filled with soilless substrates (for example commercial potting mixes) or
mixes of soil and soilless components. Soilless substrates differ from natural soils in
Wicking bed design Review of literature
18
that they have lower bulk densities and fewer micro-pore structures. Common practice
is to irrigate soilless media to maintain much higher water potentials than is usual with
soils. At a tension of -1kPa, soilless media generally has sufficient air-filled pores, and
plant growth slows below optimal when tension is below -10kPa (de Boodt &
Verdonck, 1972).
Oh et al. (2007) reported that, at the time, few studies had investigated an optimum
water content for soilless media. However, numerous researchers have reported the
water contents that were measured during subirrigation experiments. Maximum
volumetric water content in 6cm diameter pots (pot height or volume not stated) filled
with 7:3 mixture of peat moss and perlite subject to different subirrigation methods
varied between 60% (ebb and flow irrigation) and 30% (nutrient flow wick system).
These moisture levels produced good growth in Kalanchoe blossfeldiana with plants
after 10 weeks having an average height of 24.9cm, 468.7cm2 leaf area and 4.24g
shoot dry weight (Son, Oh, Lu, Kim, & Giacomelli, 2006).
Ferrarezi, van Iersel, and Testezlaf (2015) grew rooted hibiscus cuttings in 12cm high
pots, with a peat/perlite mix and ebb and flow subirrigation. Irrigation was provided
for a fixed time when the volumetric water content of pots reached a set threshold
between 10% and 42% volumetric water content. Shoot height and dry weight
increased with increasing thresholds for triggering irrigation events. The maximum
water content after irrigation was 70% at the 10% threshold, and 90% at the 42%
threshold. Capillary rise was slower in the drier pots due to soilless media having a low
hydraulic conductivity at low water contents.
Spinach yield and nutritional quality is affected by water availability. Spinach grown in
sandy soil with a closely controlled soil water matric head of -20 to -30 cm had faster
growth, higher yield and better quality than spinach grown at matric heads of -10 to
-20 cm or -30 to -40 cm (Nishihara, Inoue, Kondo, Takahashi, & Nakata, 2001).
Despite the practice being to grow plants in soilless substrates at high water
potentials, they can survive very low potentials. Mature marigold plants in a
peat/vermiculite/perlite mix exhibited first signs of wilt at -0.6Mpa (12.8% volumetric
moisture content) and severe wilt at -2.2MPa (4.8% moisture content). After re-
watering, all plants appeared to completely recover (Fields, Fonteno, & Jackson, 2014).
Wicking bed design Review of literature
19
Small changes in water tension do not affect yield of tomatoes. Capillary mat irrigation
from a trough with water 7, 10 and 14cm below the growing medium maintained
water tensions of -1.0, -1.6 and -2.0kPa in the medium but did not cause any
difference in yield (Saarinen & Reinikainen, 1995).
There have been no papers published on the soil moisture content achieved in wicking
beds and whether or not wicking beds can supply sufficient moisture for optimum
growth. Semananda et al. (2016) refilled their wicking beds when moisture content
dropped below 75% of field capacity but did not report what that moisture content
was.
2.7 Surface salt accumulation
A potential problem with subirrigated containers is that salts rising with the irrigation
water may accumulate near the surface of the growing medium. In top watered
containers, excess irrigation can be periodically applied to leach salts out of the
medium but there is no mechanism for this in subirrigation systems (Evans et al.,
1992). This is potentially a constraint for plant production systems in wicking beds as
excess salinity can stunt the growth of plants (Bernstein, 1975).
Several researchers have measured high electrical conductivity indicating high salt
levels in subirrigated containers (Argo & Biernbaum, 1996; Colla et al., 2003; Elia et al.,
2003; Kent & Reed, 1996) but high salt levels on the surface generally do not affect
plant growth because in subirrigated containers most root growth occurs lower in the
growing medium (Colla et al., 2003; Cox, 2001; Kent & Reed, 1996). High surface salt
levels may restrict reuse of the medium because small seedlings planted in the surface
layer may be affected (Cox, 2001; Evans et al., 1992).
It would appear that accumulation of salts on the surface of wicking beds could be a
problem but it has not been measured. The UPFLOW program (Raes, 2004) can be
used to estimate salt movement due to capillary rise. Where salt accumulation has
been reported in subirrigated containers, the irrigation has been done with a nutrient
solution. It is not known if wicking beds that are irrigated with plain water and use
nutrients stored in the growing medium for plant growth will suffer from this problem.
Wicking bed design Methods
20
3 METHODS
Two experiments were conducted to measure water holding capacity and capillary rise
of various media, and three wicking bed trials were conducted in a 6m x 4m polytunnel
located in Bywong, NSW (35o10.558'S, 149o 21.425'E, elevation 735m). The methods
used to conduct these experiments are described in the following sections.
3.1 Water holding capacity
To measure the water holding capacity of various materials, rectangular plastic take-
away food containers (111 x 172 x 52mm, nominal volume 650ml, actual volume
670ml) were filled with the media to level with the top of the container then weighed
on certified commercial scales (CAS SW1C-10). Water was added to the media-filled
containers until there was free water level with the top of the container. The water-
filled containers were re-weighed to determine the total water holding capacity. The
containers were then covered with aluminium flyscreen mesh (wire diameter 0.37mm,
holes 1.5 x 1.5mm) and inverted onto a wire rack and left to drain until water ceased
dripping through the rack. The drained containers were weighed to determine the
amount of water remaining the give the field capacity of the media. The media used
are listed in Table 1 and shown in Figure 3. Three replicates for each medium were
performed.
Table 1 - Materials used in water holding capacity experiment
Medium
Description
Cocopeat mix
6:3:1 mix of cocopeat, compost and washed sand
Crusher dust
fine crushed rock graded from 5mm down to dust
Crushed gravel
10mm screened crushed aggregate rock sold as "blue metal"
River gravel
10mm rounded gravel
River sand
Scoria
15mm screened volcanic rock
Washed sand
Figure 3 - Materials used in water holding capacity experiment: A cocopeat mix, B
crushed gravel, C crusher dust, D River gravel, E river sand, F scoria, G washed sand
Wicking bed design Methods
21
3.2 Capillary rise
This experiment measured the rate of capillary rise of water through the selected
media. One end of clear Perspex tubes (46.2mm inside diameter, 50.25mm outside
diameter, 500mm length) was covered with aluminium flyscreen mesh (wire diameter
0.37mm, holes 1.5 x 1.5mm) and the tubes were filled with air dried samples of the
selected media. The media was settled in the tubes by gentle manual vertical shaking
of the tubes as they were being filled but not otherwise compacted. The top ends of
the filled tubes were left uncovered.
The tubes were placed vertically in 750ml plastic take-away containers with the ends
of the tubes supported 5mm above the base of the containers by metal spacers. The
containers were filled to a depth of 35mm with 500mm of water. The containers were
covered with a lid with a hole for the tube to pass through to reduce evaporation from
the containers.
The height of capillary rise (wetting front) was identified by visually identifying a colour
change in the media. In the case of perlite, it was not possible to see the wetting front
using pure water so four drops of green food colouring was added to each container.
Periodic measurements were taken of the height of the wetting front in each tube.
Frequency of measurement varied with the rate of rise. At the start of each
experiment measurements were taken approximately every 0.5 to 2 hours, and then
approximately daily. The wetting front did not always rise evenly around the
circumference of the tube. Where the rise was uneven, the median height between
the highest and lowest point was used.
Table 2 - Materials used in capillary rise experiment
Medium
Description
Cocopeat mix
6:3:1 mix of cocopeat, compost and washed sand
Crusher dust
fine crushed rock graded from 5mm down to dust
Crushed gravel
10mm screened crushed aggregate rock sold as "blue metal"
Perlite(fine)
Fine grade perlite, mean particle diameter 4.67mm
Perlite(med)
Medium grade perlite, mean particle diameter 6.37mm
River gravel
10mm rounded gravel
River sand
Scoria
15mm screened volcanic rock
Washed sand
Woodchips
pine woodchips, 2-10mm
Three replicates were performed for each material except woodchips (2 replicates),
scoria (5 replicates) and crushed gravel (4 replicates). Woodchips are rarely suggested
as a reservoir material, showed only moderate wicking ability and were not to be used
Wicking bed design Methods
22
in the wicking bed experiments so an additional replicate was not deemed necessary.
Scoria and gravel are frequently specified as reservoir material in popular literature
and gravel was used by Semananda et al. (2016) and Semananda et al. (2020). The first
three replicates of these materials showed poor wicking performance so additional
replicates were done to ensure robustness of the results.
Figure 4 - Experimental apparatus used for capillary rise experiment
3.3 Wicking bed trials
Three wicking bed trials were conducted in a polytunnel in Bywong, NSW from
October 2019 to April 2020. The first two experiments used the same wicking beds
while the third used a different design. The three experiments are described in the
following sections.
3.3.1 Wicking bed trial 1 (WBT1)
Five different combinations of reservoir and growing medium were used. Four
reservoir media (cocopeat mix, 10mm crushed gravel, washed sand and WaterUps®
modules) were used in conjunction with a cocopeat mix growing medium. A washed
sand reservoir was also used with a commercial potting mix growing medium. Three
replicates of each treatment were tested in a completely randomised design. The
treatments and their allocation to beds are presented in Table 3.
Wicking bed design Methods
23
The cocopeat mix was a 6:3:1 mix by volume of cocopeat fibre, municipal green waste
compost and washed sand. When used in the growing layer, Scotts Osmocote
Vegetable Tomato and Herb fertiliser (NPK 13.3:1.8:4.4) was added to the cocopeat
mix (110g/100L of cocopeat mix). The commercial potting mix used was Scotts
Osmocote Professional Premium Potting Mix. No additional fertiliser was added to the
potting mix.
The WaterUps® modules are a commercial product designed to be used as the
reservoir layer in wicking beds. They consist of 400 x 400mm plastic trays with four 60
x 60mm, 140mm deep hollow 'legs'. The trays support the growing medium above a
water filled reservoir and the legs, filled with a wicking medium, provide a path for
water from the reservoir to move to the growing medium.
Table 3 - Treatment (reservoir and growing media) and container type used for wicking
beds
Bed
Treatment
Reservoir medium
Growing
medium
Container1
1
cocopeat
cocopeat mix
cocopeat mix
B
3
sand.cp
washed sand
cocopeat mix
T
4
sand.pm
washed sand
potting mix
T
5
gravel
10mm crushed gravel
cocopeat mix
B*
6
cocopeat
cocopeat mix
cocopeat mix
B
7
sand.cp
washed sand
cocopeat mix
T
8
WaterUps®
WaterUps®
cocopeat mix
T
9
sand.pm
washed sand
potting mix
T
10
cocopeat
cocopeat mix
cocopeat mix
T
11
sand.cp
washed sand
cocopeat mix
B
12
WaterUps®
WaterUps®
cocopeat mix
B*
13
sand.pm
washed sand
potting mix
B*
14
gravel
10mm crushed gravel
cocopeat mix
B
15
WaterUps®
WaterUps®
cocopeat mix
T
16
gravel
10mm crushed gravel
cocopeat mix
T
(1)T: top half of IBC, B: bottom half of IBC, B*: bottom half of IBC with outlet
Sixteen wicking beds were constructed from Intermediate Bulk Containers (IBCs) cut in
half and with a dividing panel placed vertically across each half giving beds of 950mm x
580mm and 470mm deep. Due to the shape of the IBCs, there was a small variation in
volume between beds made from the top and bottom half of the IBC, and beds that
incorporated the IBC outlet had approximately 5 litres less volume than other beds.
Table 3 indicates which portion of the IBC was used for each bed. One of the sixteen
beds was not used in the experiments.
The bottom 250mm of each bed was lined with PVC pond liner to ensure that the
dividers in the middle of the IBCs were watertight. A 15mm diameter clear plastic tube
was attached to an outlet in the base of each bed and extended 200mm up the side of
Wicking bed design Methods
24
the bed. This tube served both as an overflow outlet to ensure that the water level in
the reservoir layer remained below the growing medium and as a visual indication of
the level of free water in the reservoir.
Beds with cocopeat, gravel and sand reservoirs had a 50mm diameter PVC tube fill
pipe to direct water to the base of the reservoir. This was connected to a 50mm
diameter slotted PVC pipe that ran along the centre of the base of the reservoir to
distribute water through the media (Figure 5). Where gravel or sand was used in the
reservoir, a layer of geotextile fabric (Grunt Landscape Fabric, needle punched
polyester fibres) was placed across the surface of the reservoir layer to separate it
from the growing medium. The beds with cocopeat in both the growing and reservoir
layers did not have geotextile.
Figure 5 - Design of wicking beds with cocopeat, gravel and sand filled reservoirs
The reservoir layer was filled with the reservoir medium to a depth of 200mm. The
growing layer consisted of 300mm depth of growing medium, which compacted to
approximately 250mm depth after watering.
The beds with WaterUps® had a 60mm deep layer of washed sand placed in the base
to raise the tops of the WaterUps® modules to 200mm above the bottom of the
container. Two complete WaterUps® modules were used in each bed plus partial
modules to fill the remaining area of the bed. Each bed had 12 legs filled with wicking
material (Figure 6, Figure 7). In the first wicking bed experiment (WBT1), medium
grade perlite was used as the growing medium in the legs of the WaterUps® modules.
In the second experiment (WBT2) the perlite was replaced with washed sand.
Wicking bed design Methods
25
Figure 6 - Design of wicking beds using WaterUps modules in the reservoir layer
Figure 7 - (A) WaterUps® module showing hollow legs to be filled with wicking medium
and top to support growing medium above water-filled reservoir. (B) WaterUps® modules
placed in the bottom of a wicking bed prior to filling the legs with wicking medium.
The sixteen wicking beds were placed inside a 4 x 6m poly tunnel. Figure 8 shows the
layout of the beds within the polytunnel and Figure 9 is a photo of the polytunnel.
The reservoirs of the wicking beds were filled with material to 200mm depth and the
surface of the material was levelled. Water was added to the reservoir until free water
was just visible on the surface of the reservoir medium and the water level was at the
top of the indicator/overflow tubes on the sides of the beds.
For sand and gravel reservoirs, a layer of geotextile was placed over the reservoir
material. Growing medium was added to 300mm depth on top of the reservoir layer
(Figure 10).
Wicking bed design Methods
26
Figure 10 - IBC wicking beds for WBT1 and WBT2 during construction showing container,
waterproofing pond liner, PVC fill pipe and reservoir media in place: (A) cocopeat mix, (B)
sand, (C) gravel, (D) WaterUps® with perlite wicking medium. (E) Geotextile covering
reservoir media prior to adding growing media. (F) Indicator tubes to show water level in
reservoir. (G) Beds filled with growing media and tensiometers in place ready for planting.
Figure 9 - Polytunnel in Bywong, NSW
containing wicking beds used in this trial
Figure 8 - Layout of wicking beds within polytunnel.
Rectangular beds 1-16 used in WBT1 and WBT2.
Circular beds A-I used in WBT3.
Wicking bed design Methods
27
The growing medium in all beds was watered from on top until the medium was
saturated and water flowed out of the overflow tubes. The beds were left to
equilibrate for 12 hours before planting.
For WBT1, each bed was planted with twelve spinach (Spinacia oleracea 'Ironman')
seedlings on 17/10/2019. The seedlings were grown tightly packed in punnets and
suffered some root damaged when they were separated for transplanting. After
planting, each bed was watered from on top with 1.5L of dilute Seasol solution (30ml
Seasol in 9L water) then another 2L water. An additional 2L/bed of water was applied
in the same manner on each of the next four days. Five days after planting each bed
was watered with 1.5L of dilute Powerfeed fertiliser (50ml Powerfeed in 9L water) as
the seedlings did not appear to be thriving. Following this, all beds relied on capillary
watering from their reservoir.
Figure 11 - WBT1 wicking beds with spinach seedling after planting
Soil water measurements were made with electronic tensiometers with their tips
buried 150mm from the surface in the centre of each bed. See section 3.5 for details
of the tensiometers. Measurements from the tensiometers were made manually using
a regulated 5V power supply and a handheld multimeter.
Reservoir water levels were measured by the drop in the level of water in the indicator
tubes on the outside of each bed. However, this was only effective for the reservoirs
Wicking bed design Methods
28
containing gravel or WaterUps®. Because the indicator tube showed the level of free
water in the reservoir it was not an effective measure for media with a large field
capacity. Once the free water was removed from the sand and cocopeat reservoirs by
capillary rise, the level in the indicator tubes dropped to the bottom even though the
reservoirs would have still held a considerable amount of available water.
The reservoirs were refilled between one and three times during the course of the
experiment with the volume of water required to fill each reservoir recorded (Table 4).
Table 4 - Schedule of the number of days after transplant that the reservoir of the
wicking beds were refilled during WBT1
Treatment
Bed
Refill
(days after transplant)
cocopeat
1, 6, 10
28, 32, 42
gravel
5,14,16
42
sand.cp
3,7,11
28, 32, 42
sand.pm
sand.pm
4
9,13
28, 42
28, 32, 42
WaterUps®
8,12,15
42
Measurements of the percentage of the area of each bed covered by spinach leaves
were made using the Canopeo app (Patrignani & Ochsner, 2015) on an iPhone SE.
Measurements were taken until the leaf canopy covered most of the surface area of
the bed, after which this method was determined not to provide an accurate
measurement of plant growth.
Figure 12 - Example of photo and plant area analysis by Canopeo app (photo of lettuce
plants from WBT2)
The spinach plants were harvested by cutting the stems level with the top of the soil at
101 days after transplanting. By this time the plants were dead from high
temperatures and were desiccated (Figure 13). The plants were regarded as air dried
for the purpose of weighing. The total weight of all plants from each bed was recorded
using certified commercial scales (CAS SW1C-10).
Wicking bed design Methods
29
After harvest, each reservoir was refilled with water and the volume of water added
was recorded.
Figure 13 - Typical bed of desiccated spinach plants prior to harvest (bed 16)
3.3.2 Wicking bed trial 2 (WBT2)
The second wicking bed experiment used the same beds and treatments as WBT1 but
were planted with lettuce instead of spinach. Prior to planting, some refurbishment of
the beds occurred.
In the WaterUps® beds (beds 8,12,15), perlite was removed from the legs of the
WaterUps® modules and replaced with sand. This was done because, although the
manufacturer recommended using perlite, the performance of the WaterUps® beds
during WBT1 was less than expected and the capillary rise experiments had shown that
sand was a better wicking medium than perlite.
In all beds, the growing media was dug over to return all beds to a common state.
Fertiliser was added to all beds (66g/bed of Scotts Osmocote Vegetable Tomato and
Herb fertiliser) and the growing media was topped up with more cocopeat mix or
potting mix as appropriate.
Electronic tensiometers were reinstalled at 150mm depth in the centre of each bed.
The tensiometers were connected to custom built data loggers set to record readings
once per hour. The data loggers also recorded temperature and humidity near the
surface of the beds in the centre of the poly tunnel. Details of the data loggers is
presented in section 3.5.
Prior to planting, the reservoirs were filled and all beds were top watered until the
growing medium was saturated and water overflowed from the indicator tubes. The
beds were left to equilibrate for 15 hours.
Wicking bed design Methods
30
Twelve butterhead lettuce seedlings (Lactuca satvia var. capitata) were transplanted
into each bed on 20/2/2020. Seedlings were five weeks old had had been grown one
per cell in an attempt to avoid the planting shock observed with the spinach seedlings
in WBT1. Figure 14 is a photograph of one of the beds after transplanting the lettuce
seedlings. All beds were top watered with 1.8L of water after planting, and again on
days 1 and 3 after planting. Four days after planting the lettuces in potting mix
(sand.pm treatment) were wilting and 2.4L of water was applied to all beds. Some of
these plants were still showing signs of water stress later that day so an additional 1.8L
of water was applied to the sand.pm beds only (beds 4,9,13). Following this, all beds
relied on capillary watering from their reservoir.
Figure 14 - Typical planting of lettuce seedlings for WBT2 (bed 6)
The reservoirs on all beds were refilled 36 days after transplant.
Measurements of soil moisture, electrical conductivity and soil temperature were
made with a PulseTM meter (Bluelab, www.bluelab.com/pulse). Before the start of the
experiment, the meter was calibrated as per the manufacturer's instructions to read
100% soil moisture in the cocopeat mix at field capacity. The same calibration was
used for measuring both the cocopeat mix and the potting mix. Data from the meter
was recorded on an iPhone using the PulseTM app from Bluelab. Measurements were
made at three equally spaced locations in each bed. Measurements at 50mm, 100mm
and 200mm depth were made at each location.
Leaf area measurements were made using the Canopeo app as in WBT1.
Wicking bed design Methods
31
Lettuces were harvested 45 days after transplant by cutting the stems level with the
top of the soil. Each lettuce was weighed individually then cut in half and spread on
racks in the poly tunnel for air drying. Lettuces were air dried for 36 days then the total
dry weight for each bed was recorded.
3.3.3 Wicking bed trial 3 (WBT3)
The aim of WBT3 was to investigate the effect of a geotextile layer on the movement
of water within a wicking bed. Three treatments were used:
cp.gtex cocopeat mix in both the reservoir and growing layers, but separated
by a geotextile membrane
cp.none cocopeat mix in both the reservoir and growing layers with no
geotextile (equivalent to the cocopeat treatment in WBT1 and WBT2
sand.none washed sand in the reservoir layer and cocopeat mix in the growing
layer, with no geotextile separating the layers
The cocopeat mix was the same as the earlier experiments (6:3:1 cocopeat fibre,
municipal green waste compost, washed sand plus additional fertiliser). The sand and
geotextile were also the same as used in the earlier experiments.
Three replicates of each treatment were used.
Figure 15 - Design of wicking beds based on 9L plastic buckets used for WBT3
The wicking beds for WBT3 were constructed using 9L plastic buckets with extensions
of 600mm wide plastic damp course material to give a total depth of 500mm. Overflow
drainage for the reservoir layer was provided by 19mm diameter plastic pipe through
the sides of the buckets 200mm up from the base. A 19mm plastic pipe to the bottom
of the bucket was used to fill the reservoir with water (Figure 15 and Figure 16). The
Wicking bed design Methods
32
nine wicking beds for WBT3 were placed randomly in a single row along the western
side of the same polytunnel used for WBT1 and WBT2 (Figure 8).
Figure 16 - Wicking beds for WBT3 during construction. (A) components: bucket, plastic
dampcourse, overflow and fill pipes, geotextile. (B) Assembled wicking bed with fill pipe
covered with geotextile and overflow pipe. (C) Wicking bed with sand in reservoir. For
treatment with geotextile, a circle of geotextile would have been placed on top of the
sand. (D) Filled and planted wicking beds in the polytunnel.
Prior to planting, the reservoirs of the beds were filled and the beds top watered until
the reservoirs overflowed. The beds were allowed to drain for three hours. One five
week old lettuce seedling (Lactuca satvia var. capitata) was planted in each bed on
26/2/2020. Beds were top watered with 150ml/bed of water and another 150ml/bed
was applied two days after planting. Following this, all beds relied on capillary watering
from their reservoir.
Soil moisture measurements were made with the same PulseTM meter used in WBT2.
Water was added to all beds until the reservoirs overflowed 34 days after
transplanting. The lettuce plants were harvested 44 days after transplanting by cutting
the stems level with the surface of the soil. The harvest (wet) weight of each lettuce
was recorded. The reservoirs of each bed were refilled after harvest and the beds were
top-watered until the reservoirs overflowed. The amounts of water added to the
reservoirs and growing media were recorded.
3.4 Data analysis
All data was recorded in Microsoft Excel spreadsheets. Some data was also stored in a
MySQL database which was used to extract subsets of the data for analysis. Analysis of
variance was performed using R version 3.5.2 (https://www.r-project.org). Means
were separated by Tuckey's HSD test with P ≤ 0.05 considered to be statistically
significant.
Wicking bed design Methods
33
3.5 Electronic tensiometers and data logger
Electronic tensiometers (Figure 17) and Arduino based data loggers were constructed
based on a method described by Thalheimer (2013). The ceramic tips used were part
of the Blumat® Classic pot plant watering system obtained from Eurolux Australia.
Figure 17 - Electronic tensiometer used for WBT1 and WBT2
Two data loggers that recorded readings from eight tensiometers plus temperature
and humidity were constructed. The parts used in the data loggers are listed in Table 5
and shown in Figure 18. The Arduino code running on the data loggers is available in
appendix 2.
Table 5 - Components used in data logger for tensiometer, temperature and humidity
data
Function
Component used
Microcontroller board
Arduino Nano V3 with ATmega328
SD card and real time clock module
Geekcreit Nano V3.0 Data Logging Shield
Multiplexor chip
74HC4067 High Speed CMOS 16-Channel
Analog Multiplexer/Demultiplexer
Temperature and humidity sensor
Duinotech Temperature and Humidity Sensor
Module Jaycar cat no XC4520
Figure 18 - Arduino-based data logger used for WBT2
Wicking bed design Results
34
4 RESULTS
This chapter presents the results obtained from each of the five experiments
conducted as part of this study (water holding capacity, capillary rise and three wicking
bed trials). The results of each experiment are presented separately. Comparisons
between experiments and the conclusions that can be drawn from these results are
contained in the discussion in chapter 5.
4.1 Water holding capacity of potential reservoir media
The main function of the reservoir layer in a wicking bed is to store water and supply it
to the growing medium above. A greater volume of stored water (provided the
reservoir can supply it via capillary rise) means longer periods between refill of the
reservoir and therefore less work in maintaining the crop growing in the wicking bed.
Water is held in pore spaces between solid particles of the reservoir medium; thus
water holding capacity is effectively a measure of the porosity of the medium. Table 6
presents the volume of water (mean of three replicates) that could be added to a
670ml test container filled with reservoir media.
Table 6 - Total volume of water to fill 670ml test container filled with reservoir material
and volume of water remaining in container after drainage (as field capacity)
Material
Saturated capacity
(ml)
Field capacity (ml)
mean
s.d.
mean
s.d.
cocopeat mix
476
d1
11.00
431
d
12.02
crushed gravel
328
c
16.52
53
a
14.57
crusher dust
226
a
33.45
207
b
17.90
river gravel
264
ab
11.93
59
a
2.52
river sand
266
ab
4.51
251
c
7.55
scoria
321
c
8.08
67
a
4.00
washed sand
273
b
14.01
198
b
9.45
significance
P= 9.32e-10
P= 2.7e-15
(1)Mean separation within columns by Tukey (P<0.05)
The cocopeat mix contained the most water (476ml or 64% by volume) followed by
crushed gravel (328ml or 49%), significantly more than any other medium except
scoria. The least water was held by crusher dust (226ml), followed by river gravel and
river sand.
The media with smaller particles (cocopeat mix, sands, crusher dust) retained more
water when the free water was drained away than the coarser media (presented as
field capacity in Table 6). The cocopeat mix retained the most water due to the
cocopeat fibres being able to absorb water (Ilahi & Ahmad, 2017) as well as water
Wicking bed design Results
35
being retained in pores between the particles. Of the inorganic materials, river sand
retained significantly more water than the other media. Crusher dust and washed sand
retained more water than any of the coarser media (crushed gravel, river gravel and
scoria).
4.2 Reservoir capacity of wicking beds
After the reservoir medium was placed in the wicking beds used for WBT1 and WBT2
to a depth of 200mm, the reservoirs were filled with water (Table 7). The reservoir
with the WaterUps® product held significantly more water (72.6L) than any other
reservoir. This was 65% of the nominal reservoir volume. Sand filled reservoirs held the
least water (27.6 and 26.6L) or approximately 25% of the nominal reservoir volume.
There was no significant difference between the amount of water held in the cocopeat
reservoir(47.2L) and the gravel reservoir (51.3L).
Table 7 - Volume of water required to fill reservoir layer of wicking beds used in WBT1
and WBT2, predicted volume based on experimental results of media water capacity
and difference between predicted and actual volume
(1)Mean separation within column by Tukey (P<0.05)
(2)Predicted water capacity for cocopeat mix, sand and gravel was based on experimental saturated capacity
results in Table 6. Predicted capacity for WaterUps® was calculated from the physical dimensions of the
product and does not include an allowance for water that could be contained within the material filling the
legs of the WaterUps® or the sand base they were standing on.
Table 7 also shows the predicted water capacity for each reservoir based on the
experimental results in Table 6. The capacity of reservoirs filled with crushed gravel
(46% of nominal volume) was reasonably close to the predicted volume based on the
measured saturated capacity (49%) but the actual water capacity of reservoirs filled
with cocopeat mix and sand was much less than predicted (43 and 24% compared with
71 and 41%. It is likely that more compaction occurred when placing the cocopeat mix
and sand in the wicking beds than when the experimental containers were filled.
Compaction would reduce the pore space available for holding water. The gravel
would compress less than sand or cocopeat mix so the difference between the wicking
beds and the experimental containers would be smaller and this is reflected in the
closeness of the predicted and actual water capacity. The actual capacity of
Treatment
Water capacity (L)
Actual mean
s.d.
Predicted2
Difference
(predicted-actual)
cocopeat
47.2
b1
7.39
78.2
31.0
gravel
51.3
b
1.30
53.9
2.6
sand.cp
26.6
a
3.04
44.9
17.8
sand.pm
27.6
a
1.66
44.9
16.8
WaterUps®
72.6
c
1.57
65.1
-7.5
significance
P= 0.000000183
Wicking bed design Results
36
WaterUps® wicking beds was greater than predicted, but the predictions did not
include an allowance for water in the wicking material (perlite) in the hollow legs of
the WaterUps®, or the sand bed that the WaterUps® were placed on.
4.3 Capillary rise in reservoir materials
Water in the reservoir layer of a wicking bed must be delivered to the growing layer for
it to be used by the plants. Capillary rise through the reservoir medium is the
mechanism that provides this water delivery. If water cannot rise by capillary action
through the full depth of the reservoir, then not all water in the reservoir can be used.
Figure 19 - Capillary rise of water above free water level in 50mm diameter perspex
tubes filled with potential reservoir media. The first measurement was taken half an
hour after placing the base of the tubes in water, subsequent measurements taken at
increasing intervals from two hours to approximately daily.
The capillary rise of water through various media that could be used in the reservoir
layer was assessed using 50mm diameter Perspex tubes. Three replicates were
0 2 4 6 8 10
0 100 200 300 400 500
Capillary rise − reservoir media
Capillary rise (mm)
Days
Cocopeat mix
Crusher dust
Crushed gravel
Perlite(fine)
Perlite(med)
River gravel
River sand
Scoria
Washed sand
Woodchips
Wicking bed design Results
37
performed for each material except woodchips (2 replicates), scoria (5 replicates) and
crushed gravel (4 replicates). The greatest capillary rise was obtained in the crusher
dust (Figure 19) with two of the three replicates rising to the top of the tubes (470mm)
within three days and the third replicate reaching this level within 11 days.
Crushed gravel, river gravel, scoria and woodchips all failed to provide a capillary rise
greater than 200mm (Table 8). The cocopeat mix had an mean maximum capillary rise
of 198mm.
Not all samples in the capillary rise experiments had readings taken at exactly 10 days.
The 10 day rise shown in Table 8 has been estimated by assuming a constant rise
between reading taken before 10 days and the one after 10 days and calculating the
proportion of that rise from reading before 10 days (Equation 5).
R10 = R- + (R+ - R-) * (10 - D-) / (D+ - D-) (5)
where R10 is the calculated rise after 10 days, R- is the rise recorded before 10 days, R+ is the rise
recorded after 10 days, D- is the number of days since the start of the experiment when the reading
before 10 days was taken and D+ is the number of days when the reading after 10 days was taken.
Table 8 - Maximum capillary rise above free water in 50mm diameter Perspex tubes
filled with potential reservoir media after 10 days
Material
Maximum rise
(mm)
mean
s.d.
cocopeat mix
198
de1
3.79
crushed gravel
118
c
35.44
crusher dust
467
g
4.62
perlite(fine)
232
e
0.00
perlite(med)
188
de
15.70
potting mix
165
cd
5.77
river gravel
17
a
2.89
river sand
218
de
11.55
scoria
35
ab
19.76
washed sand
364
f
52.56
woodchips
101
bc
12.73
significance
P <2e-16
(1)Mean separation within column by Tukey (P<0.05)
The measurement of capillary rise in river sand was terminated after two days because
it was apparent that it was not showing a greater rise than the washed sand and that
the washed sand would be a better candidate for use in wicking bed trials. The
maximum rise measured for river sand (218mm) was after 1.95 days. The mean
capillary rise for washed sand after 1.95 days was 292mm which is significantly more
(p=0.029) than the rise in river sand after this time.
Wicking bed design Results
38
To give an indication of the comparative rate of rise between media, the number of
days taken to reach 100 and 200mm are presented in Table 9. In general, the materials
that had a higher total rise also had a faster rate of rise.
Table 9 - Number of days for capillary rise in reservoir media in 50mm Perspex tubes to
reach 100mm and 200mm above free water
Medium
Days to rise to 100mm
Days to rise to 200mm
mean
s.d.
mean
s.d.
Cocopeat
0.93
ab1
0.26
10.98
c
1.88
Crushed gravel
2.20
b
1.96
Crusher dust
0.05
a
0.00
0.10
a
0.06
Perlite (fine)
0.28
ab
0.05
5.66
b
0.63
Perlite (med)
0.71
ab
0.16
River sand
0.01
a
0.00
0.96
a
0.74
Washed sand
0.02
a
0.01
0.13
a
0.10
significance
P=0.0289
P=0.000000204
(1)Mean separation within columns by Tukey (P<0.05)
A capillary rise rate of 5mm.day-1 has been suggested as a minimum for subirrigation
(Schindler, Lischeid, & Müller, 2017). Table 10 shows the maximum height at which
each of the materials tested were able to maintain this rate of rise. There is a direct
correlation between this height and the maximum height over 10 days.
Table 10 - Maximum capillary rise above free water in reservoir media in 50mm Perspex
tubes with rate of rise > 5mm.day-1
Medium
Maximum height
(mm)
mean
s.d.
Cocopeat mix
199
d1
7.55
Crushed gravel
114
bc
31.50
Crusher dust
470
f
0.00
Perlite(fine)
252
d
4.61
Perlite(med)
190
cd
18.23
River gravel
10
a
5.00
River sand
239
d
45.39
Scoria
27
a
14.65
Washed sand
368
e
69.76
Woodchips
96
ab
14.85
significance
9.06e-14
(1)Mean separation within column by Tukey (P<0.05)
The maximum reservoir capacity for each material is a reservoir that is as deep as the
maximum height to which the medium can sustain a capillary rise of 5mm per day. This
criteria has been used to calculate the maximum reservoir capacity for several
materials (Table 11).
Wicking bed design Results
39
Table 11 - Maximum reservoir depth, water capacity and days of supply for reservoir
medium to provide capillary flow of 5mm.d-1 from the full depth of the reservoir
Medium
Reservoir
depth
(mm)
Reservoir
capacity
(Litres.m-2)
Days of
supply
Cocopeat mix
199
141
28
Crushed gravel
114
56
11
Crusher dust
470
159
32
River gravel
10
4
1
River sand
239
95
19
Scoria
27
13
3
Washed sand
368
150
30
4.4 Capillary rise in growing media
Figure 20 - Capillary rise of water above free water level in 50mm diameter Perspex
tubes filled with potential growing media. The first measurement was taken half an
hour after placing the base of the tubes in water, subsequent measurements taken at
increasing intervals from two hours to approximately daily.
Capillary rise was also tested in a selection of potential growing media. This included
three soils from a local landscape supplier, the cocopeat mix that was also included in
0 2 4 6 8 10 12
0 100 200 300 400 500
Capillary rise − growing media
Capillary rise (mm)
Days
Cocopeat mix
Garden mix
Potting mix
Super soil
Vegi mix
Wicking bed design Results
40
the reservoir materials results above, and a commercial potting mix (Figure 20). Three
replicates were done for the cocopeat mix and potting mix. Only one replication was
done with each of the soils because it appeared that large particles of undecomposed
bark or woodchip caused air gaps in the media in the plastic tubes. These air gaps
would stop capillary rise and the results may not be representative of these materials
in actual use in wicking beds.
Despite potential problems with hydrophobicity, the cocopeat mix had a significantly
greater maximum capillary rise than the potting mix. The maximum level reached in
ten days for each media is shown in Table 12 and the maximum height reached with a
rise greater than 5mm per day is in Table 13.
Table 12 - Maximum capillary rise above free water in 50mm diameter Perspex tubes
filled with potential growing media after 10 days
Media
Maximum capillary
rise (mm)
mean
s.d.
cocopeat mix
198
a1
3.79
potting mix
165
b
5.77
significance
P= 0.00116
(1)Mean separation within column by Tukey (P<0.05)
Table 13 - Maximum capillary rise above free water in reservoir media in 50mm Perspex
tubes with rate of rise > 5mm.day-1
Medium
Max height (mm)
mean
s.d.
cocopeat mix
199
a1
7.55
potting mix
162
b
8.14
significance
P=0.00433
(1)Mean separation within column by Tukey (P<0.05)
4.5 Wicking bed trial 1 (WBT1) - Spinach
This section presents results from the first wicking bed trial growing spinach in IBC
wicking beds. Results collated include the dry weight of plants, volume of water used,
soil moisture measurements and reservoir water levels. Three replicates of each
treatment were used.
4.5.1 Plant weight
The spinach plants were harvested 101 days after transplanting into the wicking beds.
During the month starting 45 days after planting, there were 11 days with maximum
temperatures above 35oC (BOM, 2020) and all the plants had died and were largely
desiccated by day 72. Spinach is a cool season crop and is intolerant of temperatures
above 25oC (Swiader & Ware, 2002). Plants were harvested 101 days after transplant.
Wicking bed design Results - WBT1
41
All plants in all beds were in the same condition and the weight of the plants as
harvested was used as the dry weight in Table 14. The greatest weight (236g) was in
both the cocopeat and sand.cp beds and the least plant weight (116g) was in the
sand.pm treatment.
Table 14 - Total dry weight per bed of stems and leaves of spinach stems and leaves (12
plants per bed) in WBT1 harvested in a desiccated state 101 days after transplant
Treatment
Dry plant weight (g)
mean
s.d.
cocopeat
472
b1
117.01
gravel
310
ab
40.77
sand.cp
472
b
19.47
sand.pm
232
a
66.76
WaterUps®
392
ab
29.55
significance
P= 0.00404
(1)Mean separation within column by Tukey (P<0.05)
Visually, there appeared to be differences in plant growth between the halves of the
beds next to the aisle in the poly tunnel and the halves on the outer edges of the poly
tunnel (Figure 21). The reasons for this variation are unclear. It affected both the
eastern and western sides of the polytunnel. It may have been due to proximity to the
polytunnel walls, but the beds were not centred in the polytunnel and the inner half of
the eastern beds was the same distance from the polytunnel wall as the outer half of
the western beds. One possible explanation is that the seedlings on the outer halves of
the bed suffered more trauma during transplanting than the seedlings on the inner
half. All seedlings were planted from the isle in the centre of the polytunnel so planting
the outer half involved reaching further to place the seedlings in the soil.
Figure 21 - Photo of one bed in WBT1 showing apparent difference in growth between
plants next to the centre (left of picture) and the wall of the polytunnel (right)
To investigate any possible effects of these growth differences, after harvest the plants
from each bed were weighed in two portions; those from the inside half and those
from the outside half. The weight of plants from all treatments from the inner halves
Wicking bed design Results - WBT1
42
of the beds was significantly greater than the weight from the outer halves (P=0.0463,
Table 15), but there was no interaction between the effects of position and reservoir
treatment.
Table 15 - Total dry weight of stems and leaves of spinach plants from inner and outer
halves of all beds in WBT1 harvested in a desiccated state 101 days after transplant
Position
Dry plant weight (g)
mean
s.d.
in
215
a1
79.04
out
160
b
62.75
significance
P=0.0463
(1)Mean separation within column by Tukey (P<0.05)
4.5.2 Water use
Water was added to the reservoirs of the wicking beds several times during the
experiment. After harvesting the plants the reservoirs were again filled. The total
amount of water added to each bed is shown in Table 16. There were significant
differences between the total water added to each treatment. The cocopeat reservoir
took the most water (105L) and the gravel just more than half this amount (56.7L). The
amount of water added to the reservoir does not represent the total
evapotranspiration from the beds because the experiment was started with the
growing medium in the beds at field capacity but the growing medium was not re-wet
after harvest and contained less moisture than field capacity. This shortcoming was
addressed in WBT2 when the volume of water needed to re-saturate the growing
medium was also measured.
Table 16 - Total volume of water added to each bed in WBT1 during growing period plus
amount needed to refill reservoir after harvest
Reservoir
treatment
Water added (L)
mean
s.d.
cocopeat
105.0
c1
16.82
gravel
56.7
a
6.81
sand.cp
92.0
bc
5.20
sand.pm
64.3
ab
17.21
WaterUps®
77.7
ac
5.51
significance
P=0.00286
(1)Mean separation within column by Tukey (P<0.05)
There was a reasonable positive correlation (r2=0.7061) between the amount of water
used in each bed and the dry weight of plants from that bed (Figure 22).
Wicking bed design Results - WBT1
43
Figure 22 - Correlation by bed between total dry plant weight and volume of water used
to refill reservoir for all treatments in WBT1
Plant water use (WU) and water use efficiency (WUE) were calculated using equations
6 and 7 respectively, and results are shown in Table 17.
WU = I/Y (6)
WUE = Y/I (7)
where I is the volume of water added to the reservoir during the experiment and Y is
the dry weight of the spinach plants.
Table 17 - Water use and water use efficiency (WUE) of growing spinach in WBT1.
Calculated from total dry weight of plants per bed and total water added to reservoirs
during the experiment
Treatment
Dry weight
Water use (L/kg)
WUE (g/L)
mean
s.d.
mean
s.d.
cocopeat
228.2
ab1
42.27
4.5
ab
0.80
gravel
185.1
a
34.25
5.5
b
0.90
sand.cp
194.8
a
3.61
5.1
b
0.10
sand.pm
277.8
b
5.90
3.6
a
0.10
WaterUps®
198.4
a
12.30
5.0
ab
0.31
significance
P=0.00672
P=0.0148
(1)Mean separation within column by Tukey (P<0.05)
4.5.3 Soil moisture
Soil moisture was measured during this experiment by one tensiometer in the centre
of each bed. Most beds with cocopeat, sand.cp and sand.pm were refilled on days 28
and 32 and all beds were refilled on day 42. Tensiometer readings were taken
approximately every two days until day 44 and a final reading was taken on day 72. As
noted in above, the plants were dead by this stage so they were using no more water
60 80 100 120
200 300 400 500
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
r2 = 0.7061
y = 4.6 x + 14
Dry weight of plants (g)
Water used from reservoir (L)
Wicking bed design Results - WBT1
44
Figure 23 - Soil water tension for each wicking bed in WBT1 for reservoir treatments (A)
cocopeat, (B) gravel, (C) sand.cp, (D) sand.pm, (E) WaterUps® from tensiometers placed
150mm below the surface of the growing medium. Days when water was added to
reservoirs indicated by vertical dashed lines. Tensiometer readings were taken on days 1
and 3, then at an average interval of 3 days. No readings were taken between 44 and 72
days after transplant so the shapes of the graphs during this period may not reflect
actual soil water tension in this period.
Wicking bed design Results - WBT1
45
but when water use stopped was not measured. Over time, the soil water tension in all
treatments increased then decreased when the reservoirs were refilled.
Figure 23 shows the soil water tension measured by the tensiometers in all beds.
There are a few anomalous spikes in the tensiometer readings. High readings for beds
1 (cocopeat) and 11 (sand.cp) around day 40 may be due to air leakage into the
tensiometers and the tensiometers were refilled with water. Other sudden, temporary
changes in soil water tension (bed 5 day 41, bed 4 day 38, bed 8 day 41) cannot be
explained, but may be instrument errors rather than real soil moisture changes
because the soil moisture was unlikely to change so rapidly in isolated beds.
On each day that tensiometer measurements were taken after day 3, there were
significant differences in the mean measurement between treatments (Figure 24).
There was greater consistency between beds in the WaterUps® treatment with the
variance between WaterUps® beds being smaller than other treatments on all days
except two.
Figure 24 - Mean soil water tension by treatment in wicking beds in WBT1 from one
tensiometer in centre of bed 150mm below soil surface. Tensiometer readings were
taken on days 1 and 3, then at an average interval of 3 days. No readings were taken
between 44 and 72 days after transplant so the shapes of the graphs during this period
may not reflect actual soil water tension in this period.
There were significant differences in the maximum soil water tension between
treatments (Table 18). The soil in the wicking beds with gravel reservoirs dried out the
most (-65.9kPa) but this was not significantly drier than WaterUps® (-55.7kPa) or
sand.cp (-40.2kPa). There was no significant difference between treatments in the
minimum soil water tension.
10 20 30 40 50 60 70
−100 −80 −60 −40 −20 0
Average tensiometer readings
Gravel
Sand.cp
Cocopeat
Waterups
Sand.pm
Soil water tension (kPa)
Days since transplant
Wicking bed design Results - WBT1
46
Table 18 - Minimum and maximum soil water tension in wicking beds in WBT1 from one
tensiometer in centre of bed 150mm below soil surface
Treatment
Minimum soil water
tension (kPa)
Maximum soil
water tension (kPa)
mean
s.d.
mean
s.d.
cocopeat
-26.0
b1
13.84
-7.1
a
1.70
gravel
-65.9
a
6.58
-6.5
a
3.38
sand.cp
-40.2
ab
15.39
-6.6
a
2.37
sand.pm
-25.9
b
14.83
-7.5
a
1.44
WaterUps®
-55.7
ab
4.30
-8.3
a
0.23
significance
P=0.00678
(1)Mean separation within column by Tukey (P<0.05)
4.5.4 Reservoir water levels
The drop in the level of free water in the reservoir was measured on the indicator tube
on the side of each wicking bed. Figure 25 shows the change in water level over time
for each bed. The lowest water level reached during the experiment is shown in Table
19. As the reservoirs are 200mm deep, a drop of 200mm indicates that there is no free
water in the reservoir. However, because cocopeat and sand can retain a large volume
of water when free water is drained out a 200mm drop does not mean there is no
water in the reservoir available to wick up into the growing media.
Table 19 - Mean maximum drop in level of free water from full level in 200mm deep
reservoir layer of wicking beds in WBT1
Treatment
Water level drop (mm)
mean
s.d.
cocopeat
200
0
gravel
133
20.21
sand.cp
200
0
sand.pm
200
0
WaterUps®
164
51.87
Wicking bed design Results - WBT1
47
Figure 25 - Level of water over course of the experiment in indicator tube of 200mm deep
reservoir of each bed in WBT1 for reservoir treatments (A) cocopeat, (B) gravel, (C) sand.cp,
(D) sand.pm, (E) WaterUps®, where 0=reservoir full and -200=no free water in reservoir.
Measurements made on day 17 then an average of every 3 days until day 44 then on day 67.
Wicking bed design Results - WBT1
48
4.5.5 Plant canopy area
Figure 26 - Photos showing spinach plant growth in bed 3 (sand.cp treatment) of WBT1
on days 21, 28 and 34 after transplanting
Beds were photographed using the Canopeo app (Figure 26) and the percentage of the
surface area of each bed covered by leaves was recorded on days 21, 28 and 32 after
transplant(Figure 27 and Table 20).
Figure 27 - Mean percentage of bed surface area covered by leaves by treatment in
WBT1. Measured using Canopeo app on days 21, 28 and 32 after transplant
Twenty one days after transplanting the gravel treatment had the greatest leaf area
and sand.pm the smallest. On days 28 and 32, sand.pm still had the smallest leaf area
but there was no significant difference between any of the other treatments (Table
20).
20 25 30 35
0 20 40 60 80 100
Average plant area
Gravel
Sand.cp
Cocopeat
Waterups
Sand.pm
Leaf area (%)
Days since transplant
Wicking bed design Results - WBT1
49
Table 20 - Mean percentage of bed surface area covered by plant leaves by treatment in
WBT1 on 21, 28 and 32 days after transplanting. Measured using Canopeo app
Treatment
Plant area (%)
Day 21
Plant area (%)
Day 28
Plant area (%)
Day 32
mean
s.d.
mean
s.d.
mean
s.d.
cocopeat
37
b(1)
2.98
64
b
5.00
70
b
3.22
gravel
47
c
1.67
72
b
7.81
77
b
12.53
sand.cp
39
bc
5.60
65
b
1.35
74
b
4.22
sand.pm
26
a
3.26
40
a
7.95
42
a
11.76
WaterUps®
37
bc
3.66
65
b
3.74
78
b
4.10
significance
P=0.000735
P=0.000483
P=0.00156
(1)Mean separation within column by Tukey (P<0.05)
4.6 Wicking bed trial 2 (WBT2) - Lettuce
WBT2 used the same wicking beds and treatments as WBT1 but grew lettuce instead
of spinach. The same measurements were made (dry weight of plants, volume of
water used, soil moisture measurements and reservoir water levels) plus additional
measurements of wet plant weight, soil moisture, EC and temperature at three depths,
and air temperature and humidity during the experiment. Three replicates of each
treatment were used.
4.6.1 Plant weight
The total plant weight per bed at harvest (wet weight) and weight after air drying for
36 days (dry weight) are shown in Table 21. There was no significant variation across
treatments in either wet or dry weights. Although the lettuce were air dried and not
oven dried, the dry matter percentage recorded is similar to results from Montesano,
Van Iersel, and Parente (2016) who obtained lettuce dry matter percentages of 4.58-
5.67%.
Table 21 - Total wet and dry weight and dry matter as a percentage of wet weight per
bed of lettuce stems and leaves (12 plants per bed) from WBT2. Wet weight is at harvest
45 days after transplant; dry weight is after air drying for 36 days
Treatment
Wet weight (g)
Dry weight (g)
% dry matter
mean
s.d.
mean
s.d.
cocopeat
5761
a1
489.28
240
a
9.61
4.17%
gravel
5696
a
784.23
232
a
26.76
4.07%
sand.cp
5856
a
49.12
244
a
20.60
4.17%
sand.pm
5780
a
286.11
248
a
4.04
4.29%
WaterUps®
6116
a
311.85
210
a
14.01
3.43%
significance
not significant
not significant
(1)Mean separation within column by Tukey (P<0.05)
Because of a variation in soil moisture within beds had been found, each lettuce was
weighed individually. However, there was no significant difference in the weight of
individual lettuces within beds.
Wicking bed design Results - WBT2
50
4.6.2 Water use
At the time of transplant, the reservoirs of all beds were full and the growing medium
was at field capacity. Table 22 shows the total amount of water added to the reservoir
of each bed during the growing period plus the water needed to refill the reservoir
after harvest and to return the growing media to field capacity. This gives the total
amount of water lost from the beds by evapotranspiration during the growing period.
The total amount of water added to the reservoir (excluding the rehydration of the
growing media) is included for comparison with results from WBT1 where the growing
media were not rehydrated after harvest. There was no difference in the significance
of the comparisons between treatments when the rehydrating water was included.
Table 22 - Total water added per bed including additions to reservoir during growing
period, refilling reservoir after harvest and rehydrating growing media to field capacity;
amount of water added to reservoir during growing period and to refill reservoir after
harvest during WBT2
Treatment
Total water added
(L)
Water added to
reservoir (L)
mean
s.d.
mean
s.d.
cocopeat
61.8
ab1
2.82
58.0
ab
2.90
gravel
50.3
a
6.56
42.6
a
6.99
sand.cp
54.4
ab
8.36
50.8
ab
7.95
sand.pm
50.1
a
3.92
44.9
a
4.93
WaterUps®
70.5
b
6.88
66.2
b
6.82
significance
P= 0.0089
P= 0.00489
(1)Mean separation within columns by Tukey (P<0.05)
Plant water use and water use efficiency were calculated using equations 6 and 7
respectively, and are shown in Table 23 for both harvest (wet) and dry lettuce weight.
For this table, the water volume used was the total volume of water added to the
reservoir during the experiment plus the water added after harvest to rehydrate the
soil to its initial state
Table 23 - Water use (WU) and water use efficiency (WUE) of growing lettuce in WBT2.
Calculated from total wet and dry weight of plants per bed and total water added to
reservoirs during the experiment plus water used to rehydrate soil after harvest
Treatment
Wet weight
Dry weight
WU (L/kg)
WUE (g/L)
WU (L/kg)
WUE (g/L)
mean
s.d.
mean
s.d.
mean
s.d.
mean
s.d.
cocopeat
11.1
ab1
0.88
90.8
ab
6.95
257.7
a
21.71
3.9
ab
0.30
gravel
8.3
a
1.39
123.6
b
23.09
216.9
a
4.74
4.6
b
0.10
sand.cp
9.7
ab
1.08
104.3
ab
11.61
225.4
a
54.16
4.6
b
1.00
sand.pm
8.5
a
0.66
118.6
ab
8.98
202.4
a
17.36
5.0
b
0.46
WaterUps®
12.0
b
1.88
85.1
a
14.64
334.9
b
15.99
3.0
a
0.17
significance
P=0.0181
P=0.0302
P=0.0013
P=0.00634
(1)Mean separation within column by Tukey (P<0.05)
Wicking bed design Results - WBT2
51
4.6.3 Soil moisture
Soil moisture in each bed was measured in two ways. Soil water tension was measured
by one electronic tensiometer buried 150mm deep in the centre of each bed. Moisture
was also measured as a percentage of field capacity by a PulseTM meter with
measurements taken at 50, 100 and 200mm depth in each bed.
Measurements from PulseTM meter
Figure 28 shows the mean of three soil moisture readings for each depth in all beds
during WBT2. Water was added to all beds to fill reservoirs on day 36. The average soil
water tension for each treatment is also shown for comparison, but there has been no
calibration between soil moisture and soil water tension values.
There appeared to be a large variation in soil moisture measurements taken by the
PulseTM meter within the same bed, at the same depth on the same day. The meter
manufacturer says that readings may vary and recommends taking three
measurements from each container, which was done for this experiment.
As a check on whether this variation would effect experimental results, a series of 20
measurements at 100mm depth were taken on days 31 and 44 using an evenly spaced
grid pattern across all beds. The day 31 measurements from each bed were compared
with the three measurements taken on each of days 30 and 32 using the assumption
that the moisture within a bed is unlikely to change much from one day to the next.
There were no significant differences in any beds except bed 9 where the mean of one
day was significantly different to one other day. Comparing the data from day 44 with
days 43 and 45, produced a similar outcome; there were only significant differences
within two beds (beds 12 and 16). From this it was concluded that three
measurements were sufficient to produce a valid result.
The minimum soil moisture for each treatment at each depth is shown in Table 24.
Table 24 - Minimum soil moisture as a percentage of field capacity at any time during
WBT2 measured by a PulseTM meter at 50, 100 and 200mm depths
Treatment
Soil moisture (%)
50 mm depth
Soil moisture (%)
100 mm depth
Soil moisture (%)
200 mm depth
mean
s.d.
mean
s.d.
mean
s.d.
cocopeat
48.7
bc1
14.74
56.3
ab
11.58
67.8
b
8.29
gravel
40.8
b
10.57
50.9
ab
10.89
36.8
a
8.79
sand.cp
42.2
b
9.46
60.4
bc
9.15
54.9
ab
11.88
sand.pm
15.9
a
6.31
44.8
ab
13.58
74.2
b
25.91
WaterUps®
60.8
c
12.33
73
c
12.08
70.6
b
19.74
significance
P= 0.0000000436
P= 0.000109
P= 0.0000946
(1)Mean separation within columns by Tukey (P<0.05)
Wicking bed design Results - WBT2
52
Figure 28 - Soil moisture as a percentage of field capacity in each bed in WBT2 at depths of 50, 100 and
200mm below the surface from a PulseTM meter and average soil water tension from a tensiometer
150mm below the surface for reservoir treatments (A) cocopeat, (B) gravel, (C) sand.cp, (D) sand.pm, (E)
WaterUps®. Results displayed are the mean of three measurements at each depth in each bed. PulseTM
measurements were taken on day 6 then an average of every two days to day 44 except measurements
at 200mm depth were not taken until day 24. Days when water was added to reservoirs indicated by
vertical dashed lines. Soil moisture % and soil water tension scales have not been calibrated together.
Wicking bed design Results - WBT2
53
Soil moisture measurements were taken on day 35 before refilling the reservoirs
(Table 25) and day 44 before harvest (Table 26).
Table 25 - Soil moisture as a percentage of field capacity before refilling reservoirs in
WBT2 measured with PulseTM meter at 50, 100 and 200mm depths
Treatment
Soil moisture (%)
50mm depth
Soil moisture (%)
100mm depth
Soil moisture (%)
200mm depth
mean
s.d.
mean
s.d.
mean
s.d.
cocopeat
58
b1
8.10
61
a
11.49
68
b
8.29
gravel
54
b
11.28
58
a
9.01
38
a
8.34
sand.cp
52
b
10.73
66
ab
9.84
55
ab
11.79
sand.pm
24
a
9.76
53
a
16.13
77
b
26.12
WaterUps®
76
c
3.55
81
b
13.33
71
b
19.49
significance
P=5.5e-13
P=0.00028
P=0.0000965
(1)Mean separation within columns by Tukey (P<0.05)
Table 26 - Soil moisture as a percentage of field capacity measured with PulseTM meter
at 50, 100 and 200mm depths before harvesting lettuces in WBT2
Treatment
Soil moisture (%)
50mm depth
Soil moisture (%)
100mm depth
Soil moisture (%)
200mm depth
mean
s.d.
mean
s.d.
mean
s.d.
cocopeat
59
b1
13.94
68
ab
14.98
87
b
7.87
gravel
53
b
16.02
70
ab
14.69
57
a
15.37
sand.cp
58
b
13.23
74
ab
8.43
84
b
9.53
sand.pm
20
a
7.32
57
a
13.34
93
b
11.30
WaterUps®
79
c
15.84
85
b
15.53
80
b
15.09
significance
P=0.00000000141
P=0.00233
P=0.00000255
!1)Mean separation within columns by Tukey (P<0.05)
Measurements from tensiometers
Measurements of soil water tension were made using electronic tensiometers placed
at 150mm depth in the centre of each bed. Measurements from the tensiometers
were recorded hourly. Graphs of the raw measurements from each bed are in Figure
29.
There was a lot of noise and apparent erroneous data in the raw tensiometer data. The
following steps were done to provide clean data for analysis (Figure 30):
• There was a lot of variation in many beds in the days before day 20. This was
possibly due to lack of good contact between the growing medium and the
tensiometer until the growing medium settled. All data before day 20 was
removed from the data set.
• The tensiometer in bed 16 (gravel) suffered from an air leak and gave consistent
measurements near 0kPa so data for this bed was removed.
Wicking bed design Results - WBT2
54
Figure 29 - Soil water tension for wicking beds in WBT2 for reservoir treatments (A)
cocopeat, (B) gravel, (C) sand.cp, (D) sand.pm, (E) WaterUps® from tensiometers placed
150mm below the surface of the growing medium. Hourly measurements were made
using electronic tensiometers placed at 150mm depth in the centre of each bed. Days
when water was added to reservoirs indicated by vertical dashed lines.
Wicking bed design Results - WBT2
55
Figure 30 - Soil water tension (after removing anomalous data and calculating 24 hour
rolling average) for all wicking beds in WBT2 for reservoir treatments (A) cocopeat, (B)
gravel, (C) sand.cp, (D) sand.pm, (E) WaterUps® from electronic tensiometers placed
150mm below the surface of the growing medium recording measurements hourly. Days
when water was added to reservoirs indicated by vertical dashed lines.
Wicking bed design Results - WBT2
56
• A few readings had sudden spikes in the measurements that are unlikely to be
related to real soil moisture changes. Some of these are the first readings after a
flat battery in the data loggers was replaced; the cause of others are unknown.
When spikes were identified, the data recorded at that time was removed from
the datasets for all beds. The data removed was at day 30.71 and the days
between 39.82 and 40.03.
• There was a distinct cycle of daily variation in all tensiometer readings. A rolling 24
hour average was used to smooth out these variations.
The mean soil water tension for each treatment derived from the cleaned tensiometer
data is shown in Figure 31.
Figure 31 - Mean soil water tension for each treatment in wicking beds in WBT2. Data is
from one tensiometer in centre of bed 150mm below soil surface with anomalous data
points removed and a 24 hour rolling average calculated
The mean minimum soil water tension for each treatment derived from the cleaned
tensiometer data is shown in Table 27.
Table 27 - Minimum and maximum soil water tension in wicking beds in WBT2 from one
tensiometer in centre of bed 150mm below soil surface with anomalous data points
removed and a 24 hour rolling average calculated
Treatment
Minimum soil water
tension (kPa)
Maximum soil
water tension (kPa)
mean
s.d.
mean
s.d.
cocopeat
-15.3
ab1
4.7
-3.7
b
1.69
gravel
-29.0
b
7.07
-7.2
ab
0.64
sand.cp
-22.6
ab
2.29
-9.1
a
2.08
sand.pm
-28.2
b
7.00
-9.5
a
0.76
WaterUps®
-10.3
a
0.72
-6.4
ab
0.42
significance
P=0.00486
P=0.0031
(1)Mean separation within columns by Tukey (P<0.05)
10 20 30 40 50
−100 −80 −60 −40 −20 0
Average tensiometer readings
Gravel
Sand.cp
Cocopeat
Waterups
Sand.pm
Soil water tension (kPa)
Days since transplant
Wicking bed design Results - WBT2
57
The tensiometers were located 150mm below the soil surface and soil moisture
readings were made with the PulseTM meter at 50, 100 and 200mm depths. The mean
of the 100 and 200mm depth readings has been compared with the tensiometer
readings from the same day. There is a moderate correlation (r2=0.4456) between the
two different measuring techniques (Figure 32).
Figure 32 - Correlation between the average of pairs of soil moisture measurements in all
beds in WBT2 made with a PulseTM meter at 100mm and 200mm depths and soil water
tension measurements from an electronic tensiometer in each bed at 150mm depth.
4.6.4 Reservoir water levels
The drop in the level of free water in the reservoir was measured on the indicator tube
on the side of each wicking bed. Figure 33 shows the change in water level over time
for each bed. The lowest water level reached during the experiment is shown in Table
28. As the reservoirs are 200mm deep, a drop of 200mm indicates that there is no free
water in the reservoir. As in WBT1, there still would have been available water in the
sand and cocopeat reservoirs when the indicator tube showed empty.
Table 28 - Mean maximum drop in level of free water from full level in 200mm deep
reservoir layer of wicking beds in WBT2
Treatment
Water level drop (mm)
mean
s.d.
cocopeat
200
0
gravel
92
1.53
sand.cp
200
0
sand.pm
200
0
WaterUps®
166
42.76
−25 −20 −15 −10
50 60 70 80 90
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
r2 = 0.4456
y = 1.1 x + 92
Pulse soil moisture (%)
Tensiometer soil water tension (kPa)
Wicking bed design Results - WBT2
58
Figure 33 - Level of water over course of the experiment in indicator tube of 200mm deep
reservoir of each bed in WBT2 for reservoir treatments (A) cocopeat, (B) gravel, (C) sand.cp,
(D) sand.pm, (E) WaterUps®, where 0=reservoir full and -200=no free water in reservoir.
Measurements made on day 19 then an average of every 3 days until day 36 then on day 45.
Wicking bed design Results - WBT2
59
4.6.5 Plant canopy area
Figure 34 shows the mean plant area as a percentage of total bed area for each
treatment up to 24 days after planting. All treatments showed a steady rate of growth.
There were no significant differences in plant area between treatments.
Figure 34 - Mean percentage of bed surface area covered by leaves by treatment in
WBT2. Measured using Canopeo app on days 4, 8, 11, 15, 19 and 24 after transplant
Figure 35 contains photos of all beds immediately before the lettuces were harvested.
Some tip burn on the lettuces is evident in all treatments.
Figure 35 - Lettuce in WBT2 on the day of harvest, 45 days after transplant. Treatments:
beds 1,6,10 - cocopeat; beds 5,14,16 - gravel; beds 3,7,11 - sand.cp; beds 4,9,13 -
sand.pm, beds 8,12,15 WaterUps®
0 5 10 15 20 25 30
0 20 40 60 80 100
Average plant area
Gravel
Sand.cp
Cocopeat
Waterups
Sand.pm
Leaf area (%)
Days since transplant
Wicking bed design Results - WBT2
60
4.6.6 Plant root growth
After conclusion of WBT2, the sand.pm beds were deconstructed and roots were
found to have grown through the geotextile and into the sand reservoir layer. Figure
36 is a photograph of roots on the underside of the geotextile.
Figure 36 - Photograph of the underside of geotextile removed from sand.cp treatment after
WBT2 showing roots that have grown through the geotextile and into the sand filled reservoir
4.6.7 Soil electrical conductivity
Soil electrical conductivity (EC) was lower at the end of WBT2 than at the start at both
50 and 100mm depths except for the WaterUps® treatment at 50mm depth where EC
increased by 0.28dSm-1 (Figure 37, Table 29). The potting mix in the sand.pm
treatment had significantly lower EC at both 50mm (P= 0.0000376) and 100m (P=
0.0000268) depths than the cocopeat mix treatments.
Figure 37 - Electrical conductivity (dSm-1) of growing medium at (A) 50mm and (B)
100mm below surface in WBT2. Measurements were taken using a PulseTM meter on
day 6 then an average of every two days to day 44
Wicking bed design Results - WBT2
61
Table 29 - Difference between electrical conductivity (EC) within beds at the start and
end of WBT2 at 50mm and 100mm below the soil surface
Treatment
EC difference (end-start) (dSm-1)
50mm depth
100mm depth
mean
s.d.
mean
s.d.
cocopeat
-0.46
a1
0.17
-0.59
a
0.25
gravel
-0.36
ab
0.20
-0.63
a
0.18
sand.cp
-0.16
b
0.20
-0.37
ab
0.18
sand.pm
-0.12
b
0.07
-0.11
b
0.39
WaterUps®
0.28
c
0.28
-0.31
ab
0.37
significance
P=0. 875e10-8
P=0.00191
(1)Mean separation within columns by Tukey (P<0.05)
4.6.8 Temperature and humidity
Air temperature in the polytunnel during WBT2 and WBT3 ranged from a minimum of
7.2oC to a maximum of 44.8oC. The mean daily minimum was 13.3oC and the mean
daily maximum was 31.6oC (Figure 38). Humidity also varied on a daily cycle from a
minimum of 10% to a maximum of 95%.
Figure 38 - Hourly temperature and humidity in the polytunnel during WBT2.
Measurements were made hourly except for interruptions on days 18-19 and 28-30
Soil temperatures at 100mm depth fluctuated between 14.7oC and 26.2oC. Figure 39
shows the variation in soil temperatures at this depth throughout WBT2. Soil
temperatures were not taken at the same time each day. Figure 39 also shows the air
temperatures at the time that the soil temperature readings were taken.
10 20 30 40
0 20 40 60 80 100
Air temperature and humidity in polytunnel
0 20 40 60 80 100
Temperature (degrees C)
Humidity (%)
Days since transplant
Humidity Temperature
Wicking bed design Results - WBT2
62
Figure 39 - Mean soil temperature per bed during WBT2 at 100mm below soil surface from a
PulseTM meter on day 6 then an average of every two days to day 44 and air temperature
within the polytunnel housing the beds at the time soil temperature measurements were made
Across all treatments, mean soil temperature was cooler at 200mm than at 50 or
100mm (Table 30). Soil was warmer at all depths in the potting mix in the sand.pm
treatment than the cocopeat mix in other treatments (Table 31). There was no
interaction between treatment and depth affecting soil temperatures.
Table 30 - Mean soil temperature across all treatments at all depths throughout WBT2.
Measurements made with a PulseTM meter
Depth
Soil temperature (oC)
mean
s.d.
50 mm
19.7
b1
2.86
100 mm
19.6
b
2.00
200 mm
18.7
a
1.46
significance
P=6.04e-12
(1)Mean separation within column by Tukey (P<0.05)
The sand.pm treatment was warmer than all other treatments at all depths. Apart
from this, there were no significant differences in soil temperature between
treatments at any depth (Table 31).
Table 31 - Mean soil temperature for each treatment at 50, 100 and 200mm depth
throughout WBT2. Measurements made with a PulseTM meter
Treatment
Soil temperature (oC)
50mm depth
100mm depth
200mm depth
mean
s.d.
mean
s.d.
mean
s.d.
cocopeat
19.4
a1
2.88
19.3
a
1.88
18.6
a
1.14
gravel
19.6
a
2.76
19.4
a
1.94
18.4
a
1.29
sand.cp
19.4
a
2.91
19.2
a
1.98
18.2
a
1.39
sand.pm
20.5
b
2.87
20.6
b
1.95
19.8
b
1.49
WaterUps®
19.4
a
2.74
19.3
a
1.90
18.4
a
1.36
significance
P=0.00119
P=3.26e-13
P=1.44e-14
(1)Mean separation within columns by Tukey (P<0.05)
10 20 30 40
0 10 20 30 40
Soil (100mm depth) and air temperatures
Gravel
Sand.cp
Cocopeat
Waterups
Sand.pm
Air
Temperature (degrees C)
Days since transplant
Wicking bed design Results - WBT3
63
4.7 Wicking bed trial 3 (WBT3) - lettuce in small wicking beds
WBT3 was conducted using small wicking beds growing one lettuce per bed. It was
designed to test the effect of the presence or absence of a geotextile layer between
the reservoir and growing layers. Three replicates of each treatment were used.
4.7.1 Plant weight
Plants were weighed after harvest to give a wet weight (Table 32). There were no
significant differences in plant weight between treatments.
Table 32 - Mean wet weight of lettuce stems and leaves after harvest from WBT3
Treatment
Plant weight (g)
mean
s.d.
cp.gtex
586
a1
79
cp.none
661
a
108
sand.none
597
a
35
significance
not significant
(1)Mean separation within columns by Tukey (P<0.05)
4.7.2 Water use
Table 33 shows the total amount of water added to beds during WBT3 including
refilling the reservoirs and re-saturating the soil after harvest, and the total water
added to the reservoirs before rehydrating the soil (as per WBT1). The sand.none
treatment required significantly less water than the other treatments. The geotextile
layer made no difference to the amount of water added to the cocopeat treatments.
Table 33 - Total water added per bed including additions to reservoir during growing period,
refilling reservoir after harvest and rehydrating growing media to field capacity; amount of water
added to reservoir during growing period and to refill reservoir after harvest during WBT3
Treatment
Total water added
(L)
Water added to
reservoir (L)
mean
s.d.
mean
s.d.
cp.gtex
6.93
b1
0.97
5.99
b
0.88
cp.none
6.99
b
0.47
5.96
b
0.43
sand.none
4.91
a
0.40
4.04
a
0.35
significance
P=0.0135
P=0.0112
(1)Mean separation within columns by Tukey (P<0.05)
4.7.3 Soil moisture
The soil moisture measured by the PulseTM meter at 50, 100 and 200mm depths is
shown in Figure 40. The reservoirs were refilled on day 34 and resulted in an increased
soil moisture at 200mm depth in all treatments. However, the sand.cp treatment did
not recover to the same level as the other treatments. There was a smaller rise in soil
moisture after the reservoir refilling at shallower depths in all treatments.
Wicking bed design Results - WBT3
64
Figure 40 - Soil moisture as a percentage of field capacity in each bed in WBT3 at depths of 50,
100 and 200mm below the surface for reservoir treatments (A) cp.gtex, (B) cp.none, (C)
sand.none from a PulseTM meter. Results displayed are the mean of three measurements at
each depth in each bed. Measurements were taken on day 2 then an average of every 2.5 days
to day 44 except measurements at 200mm depth were not taken until day 21. Days when
water was added to reservoirs indicated by vertical dashed lines.
Just before the reservoirs were refilled on day 34, there was no significant differences
between treatments in soil moisture at 50 and 100mm depths. The sand.none
treatment was significantly drier than the other treatments at 200mm depth (Figure
40, Table 34).
Wicking bed design Results - WBT3
65
Table 34 - Mean soil moisture as a percentage of field capacity measured by PulseTM
meter in WBT3 at 50, 100 and 200mm depth on day 34 before refilling reservoirs
Treatment
Soil moisture (%)
50mm depth
100mm depth
200mm depth
mean
s.d.
mean
s.d.
mean
s.d.
cp.gtex
59
a1
15.43
72
a
14.05
74
b
15.62
cp.none
53
a
9.46
58
a
13.28
62
b
13.91
sand.none
55
a
10.75
60
a
12.67
37
a
7.20
significance
(1)Mean separation within columns by Tukey (P<0.05)
4.7.4 Soil electrical conductivity
Soil electrical conductivity as measured by a PulseTM meter was lower at the end of the
experiment than at the start in all treatments at both 50mm and 100mm depth. The
cp.gtex treatment had a significantly smaller difference than the other treatments at
50mm depth and the sand.none treatment had a significantly greater difference at
100mm depth (Figure 41, Table 35).
Figure 41 - Soil electrical conductivity EC (dSm-1) in WBT3 at (A) 50mm depth and
(B) 100mm depth
Table 35 - Difference between electrical conductivity (EC) within beds at the start and
end of WBT3 at 50mm and 100mm below the soil surface
Treatment
EC difference (end-start)
50mm depth
100mm depth
mean
s.d.
mean
s.d.
cp.gtex
-0.27
b1
0.16
-0.33
b
0.21
cp.none
-0.48
a
0.17
-0.51
b
0.16
sand.none
-0.50
a
0.11
-0.88
a
0.09
significance
P=0.00413
P=0.000000851
(1)Mean separation within columns by Tukey (P<0.05)
Wicking bed design Results - WBT3
66
4.7.5 Soil temperature
Soil temperatures measured by the PulseTM meter in the WBT3 beds fluctuated
between 13.9oC and 40.6oC. Figure 42 shows the variation in mean soil temperatures
at 100mm depth throughout WBT3. There were no significant differences in mean soil
temperatures between depths or treatments in WBT3 (Table 36).
Figure 42 - Mean soil temperature by treatment during WBT3 measured by a PulseTM
meter 100mm below soil surface. Measurements were taken on day 2 then an average
of every 2.5 days to day 44
Table 36 - Mean soil temperature by treatment at 50, 100 and 200mm below soil surface
throughout WBT3. Measurements by PulseTM meter
Mean soil temperature (oC)
cp.gtex
cp.none
sand.none
mean
s.d.
mean
s.d.
mean
s.d.
50 mm depth
21.9
5.90
21.8
6.03
21.9
6.01
100 mm depth
21.4
5.98
21.6
6.28
21.4
6.10
200 mm depth
21.3
6.59
21.6
6.85
21.7
6.98
4.8 Comparison between WBT1 and WBT2
For some of the measurements made, data presented earlier as separate results for
WBT1 and WBT2 have been combined into a single graph for ease of comparison.
The average rates of water level fall in the reservoirs of the gravel and WaterUps®
treatments in WBT1 and WBT2 have been derived from the data used for Figure 25
and Figure 33. The rates of fall are shown in Figure 43.
10 20 30 40
0 10 20 30 40
Soil (100mm depth) and air temperatures
cp.none
cp.gtex
sand.none
Temperature (degrees C)
Days since transplant
Wicking bed design Results
67
Figure 43 - Daily rate of reservoir water level drop in indicator tubes for gravel and WaterUps®
treatments in WBT1 and WBT2 from start of experiment to time of first reservoir refill
Figure 44 shows the mean soil water tension 150mm below the soil surface during
WBT1 and WBT2. These are duplicates of Figure 24 and Figure 31, presented together
for more convenient comparison.
Figure 44 - Mean soil water tension for each treatment in (A) WBT1 and (B) WBT2 from
tensiometers buried 150mm below soil surface
20 25 30 35 40
0 5 10 15 20
Average rate of water level drop
WBT1
gravel
waterups
WBT2
gravel
waterups
Rate of waterlevel drop (mm/day)
Days since transplant
Wicking bed design Results
68
Table 37 shows number of days before soil water tension dropped below -20kPa when
soil can be considered dry for container growing.
Table 37 - Mean number of days after transplant for each treatment before soil water
tension dropped below -20kPa
Treatment
Days after transplant
WBT1
WBT2
cocopeat
-1
-
gravel
32
31
sand.cp
37
35
sand.pm
44
32
WaterUps®2
35
-
(1)one cocopeat bed did drop below -20kPa on day 44 but has been excluded from these results because it did not rehydrate after
the reservoir was refilled
(2)The WaterUps® treatment used medium grade perlite as the wicking medium in WBT1 and sand in WBT2
Wicking bed design Discussion
69
5 DISCUSSION
The function of the reservoir in a wicking bed is to store water and deliver the water to
the growing medium above. One of the main benefits of growing plants in wicking
beds is the reduced workload due to less frequent watering compared with other
growing systems. Watering frequency is affected by the water holding capacity of the
reservoir which is a function of the pore space in the reservoir medium. For a given
volume, a material with a large pore space will hold more water than one with a small
pore space and the reservoir will require less frequent refills.
Transfer of water from the reservoir to the growing medium is done by capillary rise
through the reservoir medium. The reservoir medium should have sufficient wicking
ability to meet two aims. Firstly, it should be able to deliver water at a sufficient rate
and quantity to match evapotranspiration from the plants in the growing medium.
Secondly, it should be able to wick water from the full depth of the reservoir. As the
water level in reservoir drops, the wicking height increases; if the reservoir is deeper
than the maximum wicking height of the material used in the reservoir, not all of the
water in the reservoir will be available to the plants and the time between refills will
be shorter than would be indicated by the total water capacity of the reservoir.
Most of the recommendations in the popular literature for reservoir media for wicking
beds are for inorganic materials such as sand, gravel or scoria. Gravel was used in
wicking bed experiments by Sullivan et al. (2015) and Semananda et al. (2016). Austin
(2011) suggests using a soil-based growing medium to fill both the growing and
reservoir layers with no division between the two layers. The cocopeat mix treatment
was included in the current study based on this recommendation. Although not
included in the current study, it has been the experience of this author that using a
purely organic substrate in the reservoir layer can result in problems as the organic
material decomposes under anerobic conditions and loses its structure. For this
reason, the cocopeat mix which is 90% organics may not be a suitable long-term
reservoir medium and a mix with a greater proportion of inorganic materials may be
more suitable. Nevertheless, the cocopeat mix was an adequate medium for use in the
short term of this study.
Wicking bed design Discussion
70
While the main aim of this study was to investigate the suitability of various potential
reservoir materials for use in wicking beds, it is important to recognise that the
growing medium will also have a significant effect on the growth of plants in the
system. The growing medium must have sufficient wicking ability to transfer water
supplied from the reservoir to the plant roots as well as providing support and
nutrients for the plants.
5.1 Media wicking ability and effect on choice of reservoir medium
The reservoir in a typical wicking bed is 200mm deep so, for all the water in the
reservoir to be available to plant roots in the growing medium, water needs to move
by capillary action 200mm through the reservoir material. If water cannot move this
distance then not all of the water in the reservoir will be available to the plants. Of the
reservoir materials tested for capillary rise, a rise of 200mm or greater was measured
for crusher dust, washed sand, river sand, and cocopeat mix. Scoria, river gravel,
crushed gravel and woodchips all had a capillary rise of less than 200mm, suggesting
that that they would not be suitable choices for use in a 200mm deep reservoir.
The capillary rise results suggest that crushed gravel would only be suitable for
reservoirs up to 100mm deep. In the wicking bed experiments, the water level in the
gravel reservoirs dropped a maximum of 133mm and 92mm which left 34-54% of the
total water still in the reservoir. However, as the water level in the reservoirs dropped,
the soil moisture levels became drier in the gravel beds than the other treatments with
the cocopeat growing medium. This suggests that transfer of water from the reservoir
to the growing medium had stopped or substantially slowed because of the limited
capillary rise in the gravel. This is supported by Semananda et al. (2016) who found no
difference in the number of watering events required for gravel-filled reservoirs 150
and 300mm deep and no difference in the crop yield or water use efficiencies between
the two depths.
River gravel and scoria both had very small capillary rise capabilities (17 and 35mm)
indicating that they would only be suitable for shallow reservoirs. Scoria is often
recommended for use in wicking bed reservoirs in the popular literature (e.g.
Gardening Australia, October 2019, p50) because its porosity allows it to hold more
water. However scoria does not hold more water than 10mm crushed gravel and has a
Wicking bed design Discussion
71
lower capillary rise. It appears that scoria would be a poor choice for a wicking bed
reservoir.
Although the capillary rise test for river sand was terminated earlier than other
materials, there was a significant difference in the rise in washed and river sand at the
time the test was ended. This suggests that different sands have varying performances
in wicking bed reservoirs, although both sands tested were able to convey water to
over 200mm height.
Perlite is known to have good capillary rise properties (Szmidt, Hitchon, & Hall, 1988;
Wilson, 1980) despite having a coarse texture. However, there was a significant
difference in capillary rise in the two grades of perlite tested. After 10 days, capillary
rise in the fine grade perlite was 232mm but the medium grade perlite only reached
188mm. In WBT1 when the legs of the WaterUps® were filled with medium grade
perlite, the water level in the reservoir dropped less and the soil moisture was less
than in the beds in WBT2 when the perlite was replaced by sand. These results show
that medium grade perlite is less suitable for wicking bed reservoirs than other media.
As well as the height to which water will rise, the volume of water that can be
delivered to the growing media is important. The volume of water rising by capillary
action was not measured, but the rate of rise can give an indication of the water
volume available; the faster the rise, the greater the volume delivered to a specified
height. In all media, the rate of capillary rise was greatest at the start and slowed as
the level of rise grew. Crusher dust and the two sands had the shortest time to reach
both 100 and 200mm.
A capillary rise rate of 5mm per day has been used as an indicator of the suitability of a
substrate to supply water to growing plants (Schindler et al., 2017). Of the materials
tested, only crusher dust, perlite (fine) and the sands were able to provide this rate of
rise at 200mm. Water in the cocopeat mix rose at over 5mm per day up to 199mm but
this was not significantly different to perlite(fine) or river sand (252 and 239mm). This
criteria confirms the suitability of cocopeat mix, crusher dust, perlite(fine) and sand for
use in wicking bed reservoirs up to 200mm deep, reaffirms that gravel is only suitable
to 100mm deep, and shows the unsuitability of river gravel and scoria.
The rate of fall in the reservoir water levels in the gravel and WaterUps® treatments
for both WBT1 and WBT2 from the start of the experiment to when water was added
Wicking bed design Discussion
72
to the reservoirs has been calculated. Since the only mechanism for a fall in the water
level in the reservoir is movement of the water into the growing layer the rate of fall in
the reservoir level can be equated to the rate of movement of water into the growing
layer. This was possible for only the gravel and WaterUps® treatments because they
were the only reservoir treatments that contained sufficient free water for meaningful
measurements from the indicator tubes.
In WBT1, the supply of water to the growing layer in the gravel beds as indicated by
the rate of fall in the reservoir was a maximum seven days before refilling. We have
seen earlier that the soil moisture levels were lowest at the time of refilling so it would
appear that the gravel reservoir was unable to meet the water demands of the plants
even though it still contained over half its water. At day 35 (seven days before the
reservoirs were refilled), the water level in the gravel reservoirs were 95mm below the
top and it has been shown in the capillary rise experiment that the rate of rise in gravel
reduced markedly at about this level. In WBT2 the rate of fall in the gravel reservoir
was still increasing when the reservoir was refilled on day 36. However this was done
when the reservoir level was down 92mm so we cannot say that the pattern is
different to WBT1.
In both WaterUps® trials, the rate of drop continued to increase until the reservoirs
were refilled. The rate was much higher in WBT2 with sand compared to perlite in the
legs of the WaterUps® modules. This indicates that the sand was transferring more
water to the growing layer than the perlite and is further supported by the comparison
of soil water tension in WBT1 and WBT2.
It is possible that some mechanism other than capillary rise allows water to move from
the reservoir to the growing medium, particularly when the distance between the top
of the free water and the bottom of the growing medium is greater than the maximum
capillary rise potential of the reservoir medium Popular literature suggests that water
vapour may transfer sufficient moisture to the growing medium and Wladitchensky
(1966) describes the role that water vapour plays in capillary rise in fine pores, but it
seems unlikely that this mechanism could supply sufficient water and no scientific
publication has been found to support this suggestion. Another possibility is that plant
roots may grow down into the reservoir layer and access water below the air gap.
While a geotextile layer would inhibit this, some roots were found to penetrate the
Wicking bed design Discussion
73
geotextile (Figure 36), confirming results of Semananda et al. (2020). In hydroponic
systems, plants can grow successfully with roots crossing an air gap between the
growing medium and a nutrient solution (Kratky, 1993) as would happen if roots grew
into a gravel reservoir layer. However, in a wicking bed where the reservoir contains
plain water rather than nutrient solution, plants are unlikely to thrive if the soil has
insufficient moisture to facilitate nutrient transfer into the roots, even if their roots
have entered the reservoir.
5.2 Water holding capacity
As noted above, the increased time between watering events is one of the main
benefits of wicking beds compared to other growing systems. Provided that the
reservoir can transfer sufficient water for the plant's needs and assuming that the
reservoir has the same plan area as the growing medium, the factors affecting the time
between watering events are the depth of the reservoir container and the pore space
within the reservoir medium.
Pore space in the materials tested ranged from 34% in crusher dust to 71% in cocopeat
mix. The inorganic medium with the greatest pore space was crushed gravel (49%)
while the sands had a pore space of 39-40%. On this criteria alone it would appear
that, of the inorganic materials, gravel would be the best choice to provide the largest
water supply. However, as has been shown above, the wicking ability of gravel is
limited and not all water from a gravel reservoir deeper than about 100mm would be
accessible to the plants.
The maximum reservoir capacity for each material is a reservoir that is as deep as the
maximum height to which the medium can sustain a capillary rise of 5mm per day
(Table 11). Cocopeat, crusher dust and washed sand provide the largest capacity of
approximately 150 litres.m-2 of bed area. Based on an evapotranspiration rate of 5mm
per day, these reservoirs would only require filling approximately once each month.
They would, however, differ greatly in their depths: 470mm for crusher dust, 368mm
for sand and 199mm for cocopeat mix. The greater depth for sand and crusher dust
would add to the cost of constructing the wicking beds. This would need to be
considered against reduced operating costs of less frequent watering. As a point of
comparison, hydroponic systems produce the greatest yield with as many as 5-7
watering events per day (Pires et al., 2011; Suazo-López et al., 2014).
Wicking bed design Discussion
74
Based on the actual capacity of the wicking bed reservoirs used in WBT1 and WBT2,
the expectation was that WaterUps® would provide greatest period between watering
but the effectiveness of WaterUps® depends on an efficient wicking material placed in
legs. Gravel has the next greatest water capacity but due to it wicking only 114mm,
almost half of the 200mm deep reservoir could not be used which would gave it
approximately the same effective capacity as sand.
The percentage of water actually used from the reservoir in the wicking bed trials was
not able to be measured for all treatments. The wicking beds had a clear plastic
indicator tube attached to the reservoir that showed the water level of free water in
the reservoir. This allowed measurements to be made for the gravel and WaterUps®
treatments. For the sand and cocopeat treatments, the water level in the indicator
tube dropped to zero in 19 - 28 days as the free water in the reservoir was used.
Because field capacity of cocopeat and sand was high, these reservoirs would have still
contained a large amount of available water even though the indicator tubes were
empty. To measure the amount of water remaining in cocopeat and sand reservoirs,
soil moisture sensors would have to be placed in the reservoir layers.
5.3 Growing medium
Once water moves out of the reservoir into the growing medium, it continues to rise
by capillary action through the growing medium. So before examining the effect of
different reservoir materials on the soil moisture in the growing layer, the capillarity of
various growing media was examined.
The capillary rise tests performed on the soil mixes cannot be regarded as significant
because only one replicate was conducted, however the differences recorded for the
soil types were greater than the differences within the cocopeat mix and potting mix
tests which demonstrates that further testing of the capillary rise properties of
commercially available soil mixes is needed. The soil mixes were all various
proportions of sand, silt, compost, manure and pine bark. The most expensive mix
(Super soil) had the greatest capillary rise and the cheapest (Vegie mix) had the lowest
capillary rise.
The cocopeat mix that was used in the wicking bed trials appeared to be a suitable
medium for growing plants in wicking beds. Apart from the variation in growth rates
between the inner and outer halves of the beds that occurred with spinach in WBT1,
Wicking bed design Discussion
75
no growth problems that could be attributed to the growing medium were observed in
either the spinach or lettuce plants. The variation in spinach plants between the bed
halves also occurred in the beds using potting mix so the growing medium was not the
cause of this variation.
Maximum capillary rise in the cocopeat mix was significantly more than in the potting
mix (198 and 165mm). However, the capillary rise in both these media is within the
reported range for media used in hydroponic growing, which have a capillary rise of
40-180mm in 48 hours (Kappel & Slezák, 2004). The cocopeat mix rose 131mm in 48
hours and the potting mix 121mm. The cocopeat mix was hydrophobic when dry and it
was suspected that this reduced its wicking ability (Hallett & Gaskin, 2007; Letey,
Osborn, & Pelishek, 1962). However, if the material was moistened before the
experiment to make it less hydrophobic, it was not possible to observe the wetting
front in the tube as there was no discernible colour change. A different experimental
technique would be required to assess this. The potting mix was only able to sustain a
capillary rise of greater than 5mm per day up to 162mm, significantly less than the
cocopeat mix. This would suggest that the potting mix would be more suitable for
shallower growing layers than the 250mm used in these trials.
The sand.cp and sand.pm treatments both used the same reservoir treatment but with
different growing media (cocopeat mix and potting mix respectively), so comparison
between these two treatments can be used as a comparison of the growing media
performances. There was no difference in the plant weights in WBT2 between the
sand.pm and sand.cp treatments, however in WBT1 the sand.pm treatment produced
significantly less plant growth than sand.cp. The smaller capillary rise capability of the
potting mix compared with the cocopeat mix was reflected in the soil moisture
measurements. The potting mix was wetter than cocopeat mix at 200mm depth and
drier at 100 and 50mm depth.
5.4 Plant growth
In the beds with cocopeat mix as the growing medium in WBT1, cocopeat and sand.cp
grew the heaviest spinach plants and gravel the lightest. The weight of plants in
potting mix with a sand reservoir (sand.pm) was less than all the other treatments.
Cocopeat with a sand reservoir (sand.cp) also used the most water from the reservoir.
There was a positive correlation between plant dry weight and amount of water used
Wicking bed design Discussion
76
from the reservoir (r2=0.7061), but no correlation between plant weight and minimum
soil moisture or between water used and minimum soil moisture.
In WBT2, there was no significant difference in the weights of the lettuce between any
of the treatments. As with the spinach trial, the cocopeat and sand.cp treatments used
more water than gravel or sand.pm, although the WaterUps® treatment in WBT2 with
sand as the wicking medium used the most water.
The potting mix in the sand.pm treatment had a lower EC than the cocopeat mix used
in the other treatments which may have indicated a lower level of nutrients in the
potting mix. However, a nutrient solution with EC of 1.4dSm-1 produced the best
growth in hydroponically grown lettuces (Samarakoon, Weerasinghe, & Weerakkody,
2006), which is similar to the EC of the potting mix, and there was no difference
between the ultimate weight of lettuces in the potting mix and cocopeat mix so the
nutrient level of the potting mix was obviously adequate.
5.5 Soil moisture
The discussion above has described the potential for various reservoir materials to
supply water to the growing layer of a wicking bed. How well this supply meets the
needs of the growing plants is shown in the amount of moisture actually in the soil
where the plants are growing. Soil water tension in the experimental wicking beds was
between -6.5 and -65.9kPa (WBT1) and between -3.7 and -29.0kPa (WBT2). In
hydroponic cropping soil water tension is usually above -8kPa, and -10 to -20kPa is
considered dry while in open fields, soil is usually in the range -10 to -75kPa (Raviv &
Lieth, 2008). The soils in the wicking beds were generally wetter than field soils but
drier than hydroponic systems.
5.5.1 Soil moisture measured by tensiometer
Comparing tensiometer measurements between WBT1 and WBT2, most treatments in
both experiments show a drying of the soil over time until the reservoirs were refilled
which caused rehydration of the soil. However, there are differences between the
experiments as well as some similarities.
Some of the differences in soil water tension between trials are:
• The soil in all treatments in WBT1 became much drier over a similar time period
than in WBT2. This could have been due to differing water demands from a
Wicking bed design Discussion
77
different crop (spinach vs lettuce) but was more likely due to a higher ambient
temperature during WBT1 causing greater evapotranspiration.
• The sand.pm treatment was generally the wettest treatment in WBT1 but was as
dry as the gravel treatment in WBT2. There was a large variation in soil moisture
between the three sand.pm beds; in WBT1 two beds remained wet and one bed
dried while in WBT2, two beds dried and 1 bed remained wet. This was probably
due to poor and uneven wicking in the potting mix. The soil moisture
measurements at different depths show that the potting mix at 200mm depth
remained very wet while it was considerably drier at 100mm depth. With the
tensiometer located at 150mm between these two depths, a small variation in
wicking could lead to large differences in the tensiometer measurements.
• The soil in the WaterUps® treatment was almost as dry as gravel in WBT1 but was
the wettest in WBT2. This would have been caused by improved wicking from
changing the wicking material from perlite in WBT1 to sand in WBT2.
Some similarities in results between WBT1 and WBT2 provide support for drawing
general conclusions about wicking bed performance:
• All treatments remained moist (above approximately -10kPa) at 150mm depth
until about day 20 when soil moisture in some treatments started to decline. Both
experiments were started with the soil at field capacity and until about day 20
capillary rise from the reservoirs was able to maintain this level of moisture. After
this time, the increased evapotranspiration rate from the growing plants was
greater than the capillary rise from some reservoir media.
• Beds that dried over time generally recovered soil moisture after the reservoirs
were refilled. The cocopeat treatment in WBT1 did not recover much; two beds
did but one bed continued to dry after refill. This may have been due to a problem
with the tensiometer rather than a true difference in soil moisture because the
plant weight from this bed was not affected .
• The soil in the gravel treatment started drying earlier and dried more than other
treatments in both WBT1 and WBT2. This is in line with the capillary rise results
that showed water in gravel rose less than cocopeat, sand or perlite.
• In both experiments, sand.cp dried about 2/3 as much as gravel.
Wicking bed design Discussion
78
• Cocopeat remained consistently wetter than either sand.cp or gravel.
Considering the results of both WBT1 and WBT2, the following observations can be
made about each reservoir material tested:
cocopeat
Soil remained reasonably evenly moist (between -2.5 and -21.8kPa)
throughout each experiment (excluding the anomalous drying of one
bed in WBT1). This possibly indicates that the cocopeat mix is able to
wick sufficient water to support the growing plants, although this
treatment did not have geotextile between the reservoir and growing
layers so it is possible that plant roots grew deeper than in other
treatments and accessed water without it having to rise as far
through the medium.
gravel
Soil moisture in gravel started falling earlier and dropped at a greater
rate than other treatments even though there was water remaining in
the reservoir. The capillary rise experiments showed limited capillarity
in gravel and this resulted in the gravel reservoir being unable to
sustain sufficient water supply to the plants for as long as the other
reservoir media.
sand.cp
Soil became drier in sand.cp than cocopeat or WaterUps®, but not as
dry as gravel. Capillary rise results indicated sand had a much greater
capillary rise than cocopeat but the cocopeat reservoir held more
water than the sand reservoir so the drying soil may have been due to
water depletion in the sand reservoir. Another possible explanation is
that root growth in cocopeat extended into the reservoir layer and
there were fewer roots extracting water from the growing layer than
in the sand.cp treatment. More investigation into patterns of root
growth in wicking beds is warranted.
WaterUps®
The performance of WaterUps® is dependent on the wicking medium
used. Medium grade perlite was not effective but sand resulted in the
WaterUps® treatment having the highest average soil moisture this
moisture was maintained reasonably evenly throughout the
experiment.
Wicking bed design Discussion
79
WaterUps® with sand wicks maintained better soil moisture than a
reservoir filled with sand. It would appear that WaterUps® are better
than sand because once water starts to be used from the reservoir,
the amount of free water in the sand reservoir drops quickly, while
the columns of sand in the WaterUps® remain surrounded by free
water. Thus there is a much greater tension gradient over a shorter
distance with the WaterUps®, and Darcy’s law (Hubbert, 1956) shows
this will result in greater flow.
Although WaterUps® with sand as the wicking medium maintained a
higher level of soil moisture compared with that provided by other
reservoir treatments, the soil moisture was lower than is typically
used in hydroponic growing systems. It would be interesting to
determine if more or larger wicking columns in a free water reservoir
could keep the soil moisture tension above -8kPa and improve plant
growth rates.
5.5.2 Soil moisture at different depths
Soil moisture measurements were also made during WBT2 at 50, 100 and 200mm
depths. Although there were considerable variations between the three moisture
measurements made in the same bed at the same depth, comparison of twenty
measurements in each bed with the preceding and following days showed that
significant differences were rare and it was considered that the PulseTM measurements
were reliable enough to use. However, it is not known whether the variations that
were observed were real variations in soil moisture at different locations within a bed,
or variations in the accuracy of the PulseTM meter. This needs more investigation. The
meter manufacturer states that measurements may change over time with growth in
roots and that in a commercial nursery setting the meter should be periodically
recalibrated to provide accurate moisture results with a changed root density. That
was not done during this experiment, but the average magnitude of variation between
measurements did not change during the course of the experiment, so if increasing
root density affected measurements, it affected them all equally.
There was only a moderate correlation between the PulseTM and tensiometer
measurements (r2=0.4456 when comparing 150mm tensiometer values with the
Wicking bed design Discussion
80
average of 100 and 200mm PulseTM measurements) but the patterns of change in soil
moisture are generally similar.
Semananda et al. (2020) found a steady increase in soil moisture with increasing depth
in wicking beds. However, the watering regime reported for that study involved
refilling the reservoir based on the calculation of overall soil moisture and the authors
refilled the reservoirs more frequently than in this current study. This may have led to
a more consistent moisture profile than was observed here.
In the current study the cocopeat and sand.cp treatments exhibited similar patterns of
soil moisture variation over time, but in sand.cp the 200mm layer became drier than in
cocopeat, and the sand.cp soil at 200mm became drier than the soil at 100mm. The
moisture in the surface layer of cocopeat was much more variable than sand.cp.
Although sand was shown to have better wicking ability than cocopeat, the cocopeat
reservoir held 77% more water than the sand.cp reservoir and this greater amount of
available water may be the reason why cocopeat remained wetter than sand.cp at
200mm depth.
Soil moisture in the gravel treatment was similar to cocopeat and sand.cp at 100mm
depth but greatly different at 200mm depth. At day 24, when moisture measurements
at 200mm depth were first made, the 200mm depth of the gravel treatment was drier
than 50 or 100mm. The gravel treatment was significantly drier at 200mm than
cocopeat or sand.cp. The soil moisture at 200mm in gravel remained lower than the
soil above until the reservoirs were refilled. Although the gravel reservoir held more
water than cocopeat, gravel exhibited very poor capillary rise compared to cocopeat or
sand. The low soil moisture in the gravel treatment would be caused by lack of water
movement from the reservoir to the growing medium by capillary rise in the gravel.
The beds with WaterUps® had the greatest soil moisture at 100mm depth. Moisture at
50mm depth was highly variable, but was always significantly drier than 100mm depth.
From day 15 to day 35, the 50mm layer became wetter, in contrast to other
treatments where this depth dried out. It may be possible that evaporation from soil
surface reduced as plant canopy covered surface, but plant canopy spread in
WaterUps® was no different to other treatments. It is possible that a greater supply of
water from the WaterUps® reservoir compared to other treatments meant more water
rose to the surface to be evaporated. This could be tested with a comparison of
Wicking bed design Discussion
81
mulched and unmulched beds. Surface mulching of wicking beds can increase surface
soil moisture by 10-15% and lead to better water use efficiency due to lower
evaporation (Semananda et al., 2020).
In all treatments after refilling the reservoir the soil at all depths became wetter. In all
treatments except WaterUps®, the soil moisture at 200mm started to decline again
after rehydration while the soil above was still becoming wetter. The distribution of
roots within the growing medium was not investigated, however it is possible that with
higher moisture in the lower region of the growing medium after planting the plant
roots quickly grew down to 200mm depth. A greater number of roots at this depth
would remove more water from this region. If capillary rise from the reservoir was
insufficient to replenish this moisture, the 200mm depth would become drier than the
layers above. Lettuces grown in wicking beds by Semananda et al. (2020) grew the
majority of their roots in the top 100mm, but these seedlings were top watered every
three days for the first three weeks which would have encouraged development of
surface roots. The seedlings in the current study were top watered only three times in
the first four days which may have encouraged root growth to greater depths.
After refilling the reservoir, none of the 50 or 100mm layers recovered to the moisture
levels recorded at the start of the experiment when moisture was at field capacity. In
all treatments except gravel, the soil moisture at 200mm depth (which is 50mm above
the top of the reservoir) rose to close to field capacity after refilling. Capillary rise does
not appear to have sufficient flow in wicking beds to maintain soil 150mm or more
above top of reservoir at field capacity in beds with lettuce plants approaching harvest
maturity. A longer term study with plants that require several reservoir refills during
their life may shed more light on the movement of water through the growing layer of
a wicking bed. The current results suggest that overall soil moisture may either drop
over time or become steady at a lower level. This may reduce the long term
productivity of wicking beds.
As observed earlier, the potting mix used in the sand.pm treatment remained wetter
than all other treatments at 200mm depth while at 50mm was significantly drier than
other treatments and did not recover moisture after refill. This indicates that the
capillary rise in the wicking potting mix was poor and that this type of potting mix is
Wicking bed design Discussion
82
likely to be better suited to shallower wicking beds, or that transplanted seedlings will
need to be top watered until their roots grow down to the available soil moisture.
5.5.3 How often to refill wicking beds
One frequently asked question about wicking beds is how often they need to be
refilled. Semananda et al. (2020) added water to the reservoir every two weeks and
the amount added was derived by a calculation based on the change in soil moisture
measurements. However, the common practice (and the one adopted in this study) is
to refill the reservoir when needed until it overflows. Semananda et al. (2016) filled
the reservoirs when soil moisture dropped to 75% of field capacity but did not report
how frequently this occurred. In popular literature, suggestions of watering every one
or two weeks are frequently given. Based on the current study, if a minimum soil water
tension of -20kPa is used for growing leafy greens in wicking beds, then gravel
reservoirs would require refilling every 30 days and sand reservoirs every 35 days.
Cocopeat and WaterUps® reservoirs should be able to sustain a higher soil moisture
for longer, but how much longer is not known from this study. The length of time
between refills will also be affected by the weather conditions and ETc of the crop and
may differ somewhat from the results of this study. However, there is a strong
indication that wicking beds do not need to be refilled as often as is commonly
believed.
5.5.4 Plant water use
The lettuce in WBT2 used less water than has been reported by others. Lettuce
requires a minimum of 400L/kg dry weight to avoid tip burn (Both, 1995). Water use in
WBT2 ranged from 202L/kg (sand.pm) to 335L/kg (WaterUps®). Some plants in all
treatments experienced some tip burn on the outer leaves. Average water use
efficiency in WBT2 based on dry weight (4.22g/L) is similar to the best WUE (4.02g/L)
in hydroponic lettuce (Montesano et al., 2016). However, WUE based on wet weight
(85.1 - 123.6g/L) is much greater than WUE for wicking bed lettuce in Semananda et al.
(2020) (12.8 - 23.3g/L). Lettuces grown in a closed hydroponic system weighed 134-
157g/plant with water use of 12.6-14.4litres/kg (Kratky, 1993). The lettuce in the
current study had an average weight of 486g/plant and with water use of 8.3 -
12.0L/kg.
Wicking bed design Discussion
83
The current study was not focussed on growing the highest quality plants; instead it
was designed to investigate water movement from the reservoir and purposefully
stressed the plants by allowing the soil to dry as water supply from the reservoirs
dropped. If water was added to the reservoirs more frequently the quality of plants
may have been higher and water use results may have been closer to those found in
other studies. However, even the WaterUps® beds, which had the highest overall soil
moisture and their reservoirs never ran out of water, suffered some tip burn.
Tip burn in lettuce is caused by a localised calcium deficiency in rapidly growing leaves
(Hartz, Johnstone, Smith, & Cahn, 2007). While a shortage of water may play a part in
producing tip burn, it is likely to be caused by insufficient transpiration, rather than a
lack of soil moisture. Reducing humidity or increasing airflow over the leaves can
reduce tip burn (Goto & Takakura, 1992; Wien & de Villiers, 2005). Daytime humidity
within the polytunnel during WBT2 was consistently around 90%. Although doors at
both ends of the polytunnel were usually open there was no forced air movement or
other ventilation provided. Thus it cannot be concluded that lack of moisture supply
from any of the wicking bed treatments were responsible for causing tip burn.
Measures to provide additional ventilation and reduce daytime humidity within the
polytunnel would likely have reduced the incidence of tip burn.
5.6 Wicking bed size
The main aim of WBT3 was to investigate the effect of a geotextile layer between the
reservoir and growing media. WBT3 used small wicking beds (0.063m2 area) with a
single lettuce plant growing in each bed; WBT1 and 2 used larger wicking beds (0.55m2
area) growing 12 plants/bed.
The small wicking beds produced significantly larger lettuces than the large wicking
beds. The small wicking beds had 37% more soil volume per plant than large wicking
beds. This may account for the greater plant growth. Although the smaller beds used
more water than the larger beds, there was no significant difference in plant water use
(L/g) between the two.
The common treatment between WBT2 and WBT3 was with cocopeat in both the
reservoir and growing layers with no geotextile between the layers (called cocopeat in
WBT2 and cp.none in WBT2). The comparison of this treatment between the large and
Wicking bed design Discussion
84
small beds was the same as comparing all treatments; the small beds grew larger
lettuces and used more water but had the same plant water use.
The patterns of variation in soil moisture over time were very similar at all depths for
the cocopeat (large beds) and cp.none (small beds) treatments. There were no
significant differences in the minimum soil moisture levels measured between the two
experiments for this common treatment.
Thus it is valid to compare soil moisture and water use effectiveness results across
these two wicking bed sizes. It also shows that the smaller wicking beds will be suitable
for conducting future wicking bed experiments at a much lower cost and smaller space
usage than the larger wicking beds.
5.7 Effect of geotextile separating reservoir and growing medium
Because the cocopeat treatment in WBT1 and WBT2 did not use a geotextile between
the reservoir and growing layers but the sand and gravel treatments did, it was
important to investigate whether the presence or absence of geotextile affected the
operation of wicking beds. WBT3 used cocopeat treatments both with and without
geotextile (cp.gtex and cp.none) and a sand reservoir without geotextile (sand.none).
There was no difference in the amount of water used by cp.none and cp.gtex, nor any
difference in the plant weights. From this it can be concluded that geotextile makes no
difference to the wicking, at least when the reservoir and growing layers contain the
same medium.
Sand.none used significantly less water than the other two treatments but produced
plants of the same weight. In WBT1 and WBT2, sand.cp (with geotextile) also used less
water than cocopeat, but in these experiments the difference was not significant
(P>0.05). This result provides some, but not conclusive, support for the findings of
Sullivan et al. (2015) that presence or absence of geotextile makes no difference in
wicking beds.
A geotextile layer separating the growing and reservoir media is often used in wicking
beds to prevent the growing media moving into the reservoir layer and filling pore
spaces that could otherwise be filled with water. This effect was not tested in the
current study, but these results show that the presence of a geotextile layer is unlikely
to have any significant effect on water movement withing a wicking bed. However, this
Wicking bed design Discussion
85
may depend on the particular media used in the reservoir and growing layer. There are
reports that a geotextile layer can form a capillary barrier especially when under fine
grained soils due to differences in pore sizes (Azevedo & Zornberg, 2013; McInnes &
Thomas, 2012) but this was not observed in the wicking beds.
5.8 Electrical conductivity
In agricultural and horticultural systems it is common to occasionally apply excess
irrigation to the soil surface to leach away salts that accumulate on the surface due to
evaporation. Wicking beds are only watered from below, so it is possible that salts will
build up on the surface over time.
Only the WaterUps® treatment in WBT2 experienced a rise in EC in the upper soil
layer. EC in all other treatments declined, presumably as a result of the plants
consuming nutrients from the soil. The average soil moisture at 50mm depth was
slightly higher in the WaterUps® treatment than other treatments, but not significantly
so. There may have been higher evaporation from the surface of the WaterUps® beds
leading to the increase in EC, but there is not convincing evidence from this study that
salt build up is a problem in wicking beds.
The average maximum EC level recorded across all treatments was 2.55 dSm-1 with an
absolute maximum of 2.85 dSm-1. Since yields of most crops are not restricted until EC
rises to 4-8 dSm-1, the EC levels recorded during this study would not have had a
detrimental effect on plant growth.
Wicking bed design Conclusion
86
6 Conclusion
Prior to this study, the limited research that had been published about wicking beds
had focussed mainly on comparing the efficiency of wicking beds with conventional
top watered containers. Little had been done to investigate the effects of different
media in the reservoir and growing layers of wicking beds. This study has started
addressing that gap, and has developed and tested several tools suitable for the
collection of performance related data, including moisture, temperature and EC, that
can be easily implemented in an urban wicking bed production system.
The initial hypothesis was that the choice of reservoir material for a wicking bed would
affect plant growth and moisture distribution within the growing medium. Different
reservoir materials did result in variable growth in spinach but not lettuce, and soil
moisture differed between several treatments in all experiments. Thus the hypothesis
is largely supported, although effects on plant growth may vary depending on the crop
being grown.
The capability of a material used in the reservoir layer of a wicking bed to support
capillary rise of water is fundamental to how well the wicking bed performs. This study
confirms findings in the literature that capillary rise is greater in finer materials than
coarse materials. Of the materials tested, capillary rise was greatest in crusher dust,
followed by sand. 10mm crushed gravel had limited capillary rise and scoria, a material
commonly specified for reservoirs in wicking beds, had very small capillary rise.
Maximum capillary rise in a mix of cocopeat, compost and sand was about two thirds
of the rise in washed sand and almost twice the rise in crushed gravel.
Materials with larger particles had a greater pore space and could store more water in
the reservoir layer of a wicking bed. Gravel held more water than sand which held
more than crusher dust. Test results indicated that the cocopeat mix had the ability to
store the most water of any material tested due to its ability to absorb water within
the cocopeat and compost particles, but in the greenhouse trials the cocopeat mix
held the same amount of water as gravel, probably due to greater compaction of the
material in the wicking bed.
Coarse materials such as gravel or scoria are a poor choice for reservoir unless the
reservoir is shallow. 10mm gravel was able to wick water up approximately 100mm,
Wicking bed design Conclusion
87
but as the water level in the reservoir dropped towards this level there was insufficient
water flow to maintain adequate soil moisture in the growing medium. Sand or
cocopeat are better choices for a reservoir material and provide greater soil moisture
but there remain questions about whether cocopeat is suitable for long term use in a
saturated state.
WaterUps® provided a greater volume of water in the reservoir than other materials
and could maintain higher and more uniform soil moisture levels over time than other
reservoir treatments, but were reliant on use of an appropriate wicking material.
WaterUps® with medium grade perlite performed poorly and resulted in low soil
moisture levels but high soil moisture levels were maintained when using sand as the
wicking medium.
While there was no difference between treatments in the weight of lettuces, the
effects of reservoir material on soil moisture may also have implications for longer
term crops or successive crops if the growing medium is not rehydrated between
cropping cycles. In all treatments, the soil moisture in the middle and upper layers of
the growing medium did not return to their starting levels after the reservoirs were
refilled. Further research is needed to determine if this drying trend continues over
several reservoir refills or if the soil moisture at upper levels stabilises over time. If soil
moisture is concentrated in the lower levels of the growing medium, this may affect
nutrient availability for a longer term crop.
The choice of growing medium can also affect water movement and plant growth in a
wicking bed. Cocopeat mix maintained more consistent soil moisture levels across
depths than the potting mix and grew significantly heavier spinach plants. In the
potting mix, soil moisture was high at the bottom of the growing medium but upper
levels were dry. The cocopeat mix was a better growing medium for wicking beds than
the commercial potting mix that was used.
There was no strong evidence of a build-up of salts on surface of any of the wicking
beds, but the possibly that it could be a problem in the longer term cannot be ruled
out by this study.
Large variations in moisture readings were observed within beds and this introduced a
level of uncertainty to the results. Only three replicates of each treatment were used
in this study which may not have been enough to provide a high level of confidence in
Wicking bed design Conclusion
88
the results. Further trials with more replicates would increase the level of confidence
in the results.
While the outcomes of this current research provided some insight into wicking bed
design and utility, there remain several areas that should be explored by further
research including:
• can wicking beds raise soil moisture to levels similar to that of hydroponic systems
and will this boost productivity?
• what is the best way to determine when to add more water to the reservoir?
• can the tools developed (electronic tensiometers and Arduino based data logging)
and utilised (PulseTM moisture/EC probe) be relied on consistently for both
research and production. For example, are the variations in PulseTM readings
across one bed due to real variations in moisture across the bed or are they
variations in measurement?
• how long can a cocopeat mix be used in the reservoir layer before it decomposes
and becomes detrimental to the system?
• can wicking beds be used in the long term as an organic system relying on
compost and nutrient cycling for plant nutrition and encouraging growth of
mycorrhizal fungi and other soil biota instead of operating them more like
hydroponic systems and relying on manufactured fertilisers?
Wicking bed design
89
LITERATURE CITED
Abad, M., Fornes, F., Carrión, C., Noguera, V., Noguera, P., Maquieira, Á., & Puchades, R.
(2005). Physical properties of various coconut coir dusts compared to peat.
HortScience, 40(7), 2138-2144. doi:10.21273/HORTSCI.40.7.2138
Albaho, M. S. (2006). Cross-sectional area of the root medium affects water uptake of
tomato in a closed system. Journal of Food, Agriculture and Environment, 4(3&4),
175-180. Retrieved from
https://web.archive.org/web/20170809213928/http://world-
food.net/download/journals/2006-issue_3_4/a18.pdf
Argo, W. R., & Biernbaum, J. A. (1996). The effect of lime, irrigation-water source, and
water-soluble fertilizer on root-zone pH, electrical conductivity, and
macronutrient management of container root media with impatiens. Journal of
the American Society for Horticultural Science, 121(3), 442-452.
doi:10.21273/JASHS.121.3.442
Austin, C. (2011). Stones versus organics in wicking beds. Retrieved from
http://www.waterright.com.au/stones-v-organics.pdf
Azevedo, M., & Zornberg, J. (2013). Capillary barrier dissipation by new wicking
geotextile Panamerican Conference on Unsaturated Soils (pp. 20-22). Retrieved
from
http://www.ce.utexas.edu/prof/zornberg/pdfs/CP/Azevedo_Zornberg_2013.pdf
Bernstein, L. (1975). Effects of salinity and sodicity on plant growth. Annual Review of
Phytopathology, 13(1), 295-312. Retrieved from
https://pdfs.semanticscholar.org/23d9/7338ab106339d29962aecaf36a6e9ccd83
bd.pdf
Biernbaum, J. A. (1992). Root-zone management of greenhouse container-grown crops
to control water and fertilizer. HortTechnology, 127-132.
doi:10.21273/HORTTECH.2.1.127
BOM. (2020). Daily maximum temperature Canberra Airport. Retrieved from
http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=122&p_
display_type=dailyDataFile&p_startYear=&p_c=&p_stn_num=070351
Boodley, J. W., & Sheldrake, R. (1972). Cornell peat-lite mixes for commercial growing.
Ithica, New York: New York State College of Agriculture and Life Sciences, Cornell
University. Retrieved from
https://ecommons.cornell.edu/bitstream/handle/1813/39084/1972%20Info%20
Bulletin%2043.pdf?sequence=2
Both, A.-J. (1995). Dynamic simulation of supplemental lighting for greenhouse
hydroponic lettuce production. Cornell University. Retrieved from
https://www.researchgate.net/profile/Aj_Both/publication/334468975_DYNAMI
C_SIMULATION_OF_SUPPLEMENTAL_LIGHTING_FOR_GREENHOUSE_HYDROPON
IC_LETTUCE_PRODUCTION/links/5d2cc2f1458515c11c335ebb/DYNAMIC-
SIMULATION-OF-SUPPLEMENTAL-LIGHTING-FOR-GREENHOUSE-HYDROPONIC-
LETTUCE-PRODUCTION.pdf
Wicking bed design
90
Bouyoucos, G. J. (1953). Capillary rise of water in soils under field conditions. The
Journal of Physical Chemistry, 57(1), 45-49. doi:10.1021/j150502a010
Bouyoucos, J. G. (1947). Capillary rise of moisture in soil under field conditions as
studied by the electrical resistance of plaster of paris blocks. Soil Science, 64(1),
71-82. doi:10.1097/00010694-194707000-00007
Caron, J., Elrick, D., Beeson, R., & Boudreau, J. (2005). Defining critical capillary rise
properties for growing media in nurseries. Soil Science Society of America
Journal, 69(3), 794-806. doi:10.2136/sssaj2004.0108
Colla, G., Rouphael, Y., Saccardo, F., Rea, E., Pierandrei, F., & Salerno, A. (2003).
Influence of salinity and irrigation method on zucchini plants grown in closed-
soilless system. Acta Horticulturae(609), 429-433.
doi:10.17660/ActaHortic.2003.609.66
Cox, D. A. (2001). Growth, nutrient content, and growth medium electrical conductivity
of poinsettia irrigated by subirrigation or from overhead. Journal of Plant
Nutrition, 24(3), 523-533. doi:10.1081/PLN-100104977
de Boodt, M., & Verdonck, O. (1972). The physical properties of the substrates in
horticulture. Acta Horticulturae(26), 37-44. doi:10.17660/ActaHortic.1972.26.5
Elia, A., Santamaria, P., Parente, A., & Serio, F. (2003). Some aspects of the trough bench
system and its performance in cherry tomato production. Acta
Horticulturae(614), 161-166. doi:10.17660/ActaHortic.2003.614.22
Evans, M. R., Barrett, J. E., Harbaugh, B. K., & Clark, G. A. (1992). No-runoff watering
systems for foliage and flowering potted plant production. Florida Cooperative
Extension Service Circular(1059). Retrieved from
https://palmm.digital.flvc.org/islandora/object/uf%3A96245#page/UNNUMBERE
D/mode/2up
Ferrarezi, R. S., van Iersel, M. W., & Testezlaf, R. (2015). Monitoring and controlling ebb-
and-flow subirrigation with soil moisture sensors. HortScience, 50(3), 447-453.
doi:10.21273/HORTSCI.50.3.447
Ferrarezi, R. S., Weaver, G. M., Van Iersel, M. W., & Testezlaf, R. (2015). Subirrigation:
historical overview, challenges, and future prospects. HortTechnology, 25(3),
262-276. doi:10.21273/HORTTECH.25.3.262
Fields, J. S., Fonteno, W. C., & Jackson, B. E. (2014). Plant available and unavailable
water in greenhouse substrates: assessment and considerations. Acta
Horticulturae(1034), 341-346. doi:10.17660/ActaHortic.2014.1034.42. Retrieved
from https://doi.org/10.17660/ActaHortic.2014.1034.42
Frangi, P., Amoroso, G., Piatti, R., & Faoro, M. (2011). High efficiency irrigation systems
for containerized plant production. Acta Horticulturae(922), 157-161.
doi:10.17660/ActaHortic.2011.922.19
George, R. K., Biernbaum, J. A., & Stephens, C. T. (1990). Potential for transfer of
pythium ultimum in production of seedling geraniums with subirrigation and
recirculated solutions. Acta Horticulturae(272), 203-208.
doi:10.17660/ActaHortic.1990.272.29
Goto, E., & Takakura, T. (1992). Prevention of lettuce tipburn by supplying air to inner
leaves. Transactions of the ASAE, 35(2), 641-645. doi:10.13031/2013.28644
Wicking bed design
91
Hallett, P. D., & Gaskin, R. (2007). An introduction to soil water repellency. Proceedings
of the 8th International Symposium on Adjuvants for Agrochemicals (ISAA2007),
6, 9. Retrieved from
https://www.researchgate.net/profile/P_Hallett/publication/228708562_An_int
roduction_to_soil_water_repellency/links/004635236b6bee7f55000000/An-
introduction-to-soil-water-repellency.pdf
Hartz, T. K., Johnstone, P. R., Smith, R. F., & Cahn, M. D. (2007). Soil calcium status
unrelated to tipburn of romaine lettuce. HortScience, 42(7), 1681-1684.
doi:10.21273/HORTSCI.42.7.1681
Heiskanen, J. (1999). Hydrological properties of container media based on sphagnum
peat and their potential implications for availability of water to seedlings after
outplanting. Scandinavian Journal of Forest Research, 14(1), 78-85.
doi:10.1080/02827589908540810
Hicklenton, P. R., Rodd, V., & Warman, P. R. (2001). The effectiveness and consistency of
source-separated municipal solid waste and bark composts as components of
container growing media. Scientia Horticulturae, 91(3-4), 365-378.
doi:10.1016/S0304-4238(01)00251-5
Hoffman, M. L., Buxton, J. W., & Weston, L. A. (1996). Using subirrigation to maintain
soil moisture content in greenhouse experiments. Weed Science, 44(2), 397-401.
Retrieved from https://www.jstor.org/stable/4045696
Hubbert, M. K. (1956). Darcy's law and the field equations of the flow of underground
fluids. Transactions of the AIME, 207(01), 222-239. doi:10.2118/749-G
Ilahi, W. F. F., & Ahmad, D. (2017). A study on the physical and hydraulic characteristics
of cocopeat perlite mixture as a growing media in containerized plant
production. Sains Malaysiana, 46(6), 975-980. doi:10.17576/jsm-2017-4606-17
Kappel, N., & Slezák, K. (2004). Peat substitutes in growing cucumber transplants.
International Journal of Horticultural Science, 10(1), 115-118. Retrieved from
https://ojs.lib.unideb.hu/IJHS/article/download/449/448
Keen, B. A. (1919). A note on the capillary rise of water in soils. The Journal of
Agricultural Science, 9(4), 396-399. doi:10.1017/S0021859600005232
Kent, M. W., & Reed, D. W. (1996). Nitrogen nutrition of New Guinea impatiens
‘Barbados' and Spathiphyllum ‘Petite' in a subirrigation system. Journal of the
American Society for Horticultural Science, 121(5), 816-819.
doi:10.21273/JASHS.121.5.816
Klock-Moore, K. A., & Broschat, T. K. (2001). Effect of four growing substrates on growth
of ornamental plants in two irrigation systems. HortTechnology, 11(3), 456-460.
doi:10.21273/HORTTECH.11.3.456
Kratky, B. A. (1993). A capillary, noncirculating hydroponic method for leaf and semi-
head lettuce. HortTechnology, 3(2), 206-207. doi:10.21273/HORTTECH.3.2.206
Landis, T. D., & Morgan, N. (2009). Growing media alternatives for forest and native
plant nurseries. In: Dumroese, RK; Riley, LE, tech. coords. National Proceedings:
Forest and Conservation Nursery Associations-2008. Proc. RMRS-P-58. Fort
Collins, CO: US Department of Agriculture, Forest Service, Rocky Mountain
Wicking bed design
92
Research Station. p. 26-31. Retrieved from
https://www.fs.fed.us/rm/pubs/rmrs_p058/rmrs_p058_026_031.pdf
Letey, J., Osborn, J., & Pelishek, R. (1962). The influence of the water-solid contact angle
on water movement in soil. Hydrological Sciences Journal, 7(3), 75-81.
doi:10.1080/02626666209493272
Liu, Q., Yasufuku, N., Miao, J., & Ren, J. (2014). An approach for quick estimation of
maximum height of capillary rise. Soils and Foundations, 54(6), 1241-1245.
doi:10.1016/j.sandf.2014.11.017
Londra, P. A. (2010). Simultaneous determination of water retention curve and
unsaturated hydraulic conductivity of substrates using a steady-state laboratory
method. HortScience, 45(7), 1106-1112. doi:10.21273/HORTSCI.45.7.1106
Lu, N., & Likos, W. J. (2004). Rate of capillary rise in soil. Journal of Geotechnical and
Geoenvironmental Engineering, 130(6), 646-650. doi:10.1061/(asce)1090-
0241(2004)130:6(646)
Masoodi, R., & Pillai, K. M. (2012). Wicking in porous materials: traditional and modern
modeling approaches. Boca Raton, FL: CRC Press.
Matkin, O. A., & Chandler, P. A. (1957). The U.C.-type soil mixes. In K. F. Baker (Ed.), The
U.C. system for producing healthy container-grown plants. Los Angeles:
University of California, Division of Agricultural Sciences, Agricultural Experiment
Station, Extension Service. Retrieved from
https://ia800203.us.archive.org/14/items/ucsystemforprodu23bake/ucsystemfo
rprodu23bake.pdf
McInnes, K. J., & Thomas, J. C. (2012). Passive control of downslope capillary wicking of
water in sand-based root zones. HortScience, 47(2), 275-279.
doi:10.21273/HORTSCI.47.2.275
Michel, J. C., Rivière, L. M., & BellonFontaine, M. N. (2001). Measurement of the
wettability of organic materials in relation to water content by the capillary rise
method. European Journal of Soil Science, 52(3), 459-467. doi:10.1046/j.1365-
2389.2001.00392.x
Montesano, F., Van Iersel, M., & Parente, A. (2016). Timer versus moisture sensor-based
irrigation control of soilless lettuce: Effects on yield, quality and water use
efficiency. Horticultural Science, 43(2), 67-75. doi:10.17221/312/2014-HORTSCI
Nishihara, E., Inoue, M., Kondo, K., Takahashi, K., & Nakata, N. (2001). Spinach yield and
nutritional quality affected by controlled soil water matric head. Agricultural
Water Management, 51(3), 217-229. doi:10.1016/S0378-3774(01)00123-8
Oh, M. M., Cho, Y. Y., Kim, K. S., & Son, J. E. (2007). Comparisons of water content of
growing media and growth of potted kalanchoe among nutrient-flow wick
culture and other irrigation systems. HortTechnology, 17(1), 62-66.
doi:10.21273/HORTTECH.17.1.62
Patrignani, A., & Ochsner, T. E. (2015). Canopeo: A powerful new tool for measuring
fractional green canopy cover. Agronomy Journal, 107(6), 2312-2320.
doi:10.2134/agronj15.0150
Wicking bed design
93
Piatti, R., Frangi, P., & Amoroso, G. (2011). Alternative nursery management systems:
closed-loop and high retention mat. Acta Horticulturae(889), 581.
doi:10.17660/ActaHortic.2011.889.74
Pike, J. (1979). Capillary watering. Paper presented at the Proc. Inter. Plant Prop. Soc.
Pires, R. C. d. M., Furlani, P. R., Ribeiro, R. V., Bodine Junior, D., Sakai, E., Lourenção, A.
L., & Torre Neto, A. (2011). Irrigation frequency and substrate volume effects in
the growth and yield of tomato plants under greenhouse conditions. Scientia
Agricola, 68(4), 400-405. doi:10.1590/S0103-90162011000400002
Quisenberry, V., Brown, P., & Smith, B. (2006). In-situ liquid storage capacity
measurement of subsurface wastewater absorption system products. Journal of
Environmental Health, 69(4), 9.
Raes, D. (2004). UPFLOW—water movement in a soil profile from a shallow water table
to the topsoil, reference manual. Leuven, Belgium: KU Leuven University.
Retrieved from https://iupware.be/wp-
content/uploads/2016/03/upflow_manual.pdf
Raes, D., & Deproost, P. (2003). Model to assess water movement from a shallow water
table to the root zone. Agricultural Water Management, 62(2), 79-91.
doi:10.1016/s0378-3774(03)00094-5
Raviv, M., & Lieth, J. H. (2008). Soilless culture : theory and practice. Amsterdam:
Elsevier Science.
Raviv, M., Lieth, J. H., & Wallach, R. (2001). The effect of root-zone physical properties of
coir and UC mix on performance of cut rose (cv. Kardinal). Acta
Horticulturae(554), 231-238. doi:10.17660/ActaHortic.2001.554.24
Robertson, R. (1993). Peat, horticulture and environment. Biodiversity and Conservation,
2(5), 541-547. doi:10.1007/BF00056747
Roeber, R. U. (2010). Environmentally sound plant production by means of soilless
cultivation. Comunicata Scientiae, 1(1). Retrieved from
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.867.6351&rep=rep1
&type=pdf
Rose, C. W. (1966). Agricultural physics. Oxford: Pergamon Press.
Saarinen, J. A., & Reinikainen, O. (1995). Peat substrate and self-regulating irrigation - an
environmentally sound method. Acta Horticulturae(401), 435-442.
doi:10.17660/ActaHortic.1995.401.53
Salim, R. L. (2016). Extent of Capillary Rise in Sands and Silts. (Master's Theses. 688.),
Western Michigan University. Retrieved from
https://scholarworks.wmich.edu/masters_theses/688
https://scholarworks.wmich.edu/cgi/viewcontent.cgi?referer=https://scholar.go
ogle.com.au/&httpsredir=1&article=1713&context=masters_theses
Samarakoon, U., Weerasinghe, P., & Weerakkody, W. (2006). Effect of electrical
conductivity (EC) of the nutrient solution on nutrient uptake, growth and yield of
leaf lettuce (Lactuca sativa L.) in stationary culture. Retrieved from
http://thesis.pgia.ac.lk:8080/bitstream/1/1941/2/PGIATAR-18-13.pdf
Santamaria, P., Campanile, G., Parente, A., & Elia, A. (2003). Subirrigation vs drip-
irrigation: effects on yield and quality of soilless grown cherry tomato. The
Wicking bed design
94
Journal of Horticultural Science and Biotechnology, 78(3), 290-296.
doi:10.1080/14620316.2003.11511620
Schindler, U., Lischeid, G., & Müller, L. (2017). Hydraulic performance of horticultural
substrates—3. Impact of substrate composition and ingredients. Horticulturae,
3(1), 7. doi:10.3390/horticulturae3010007
Schuch, U. K., & Kelly, J. J. (2006). Capillary mats for irrigating plants in the retail nursery:
and saving water. Southwest Hort, 23(5), 24-25. Retrieved from
http://cals.arizona.edu/extension/ornamentalhort/nurseryprod/capillarymat.pdf
Semananda, N. P., Ward, J. D., & Myers, B. R. (2016). Evaluating the efficiency of wicking
bed irrigation systems for small-scale urban agriculture. Horticulturae, 2(4), 13.
doi:10.3390/horticulturae2040013
Semananda, N. P., Ward, J. D., & Myers, B. R. (2018). A semi-systematic review of
capillary irrigation: the benefits, limitations, and opportunities. Horticulturae,
4(3), 23. doi:10.3390/horticulturae4030023
Semananda, N. P., Ward, J. D., & Myers, B. R. (2020). Experimental investigation of
wicking bed irrigation using shallow-rooted crops grown under glasshouse
conditions. Irrigation Science, 38(3), 117-129. doi:10.1007/s00271-019-00651-5
Sieker, F. (1998). On-site stormwater management as an alternative to conventional
sewer systems: A new concept spreading in Germany. 38(10), 65-71.
doi:10.1016/S0273-1223(98)00734-3
Singh, K. (1922). Development of root system of wheat in different kinds of soils and
with different methods of watering. Annals of Botany, 36(143), 353-360.
Son, J., Oh, M., Lu, Y., Kim, K., & Giacomelli, G. A. (2006). Nutrient-flow wick culture
system for potted plant production: System characteristics and plant growth.
Scientia Horticulturae, 107(4), 392-398. doi:10.1016/j.scienta.2005.11.001
Suazo-López, F., Zepeda-Bautista, R., Sánchez-Del Castillo, F., Martínez-Hernández, J. J.,
Virgen-Vargas, J., & Tijerina-Chávez, L. (2014). Growth and yield of tomato
(Solanum lycopersicum L.) as affected by hydroponics, greenhouse and irrigation
regimes. Annual Research & Review in Biology, 4246-4258.
doi:10.9734/ARRB/2014/11936
Sullivan, C., Hallaran, T., Sogorka, G., & Weinkle, K. (2015). An evaluation of
conventional and subirrigated planters for urban agriculture: supporting
evidence. Renewable Agriculture and Food Systems, 30(1), 55-63.
doi:10.1017/S1742170514000131
Swiader, J. M., & Ware, G. W. (2002). Producing vegetable crops (5 ed.). Danville, Illinois:
Interstate Publishers Inc.
Szmidt, R., Hitchon, G., & Hall, D. (1988). Sterilisation of perlite growing substrates. III
International Symposium on Soil Desinfestation, 255, 197-204.
doi:10.17660/ActaHortic.1989.255.23
Taiz, L., Zeiger, E., Moller, I. M., & Murphy, A. (2015). Plant physiology and development
(6th ed.). Sunderland MA: Sinauer Associates.
Thalheimer, M. (2013). A low-cost electronic tensiometer system for continuous
monitoring of soil water potential. Journal of Agricultural Engineering, 44(3),
e16-e16. doi:10.4081/jae.2013.e16
Wicking bed design
95
van Genuchten, M. T., & Pachepsky, Y. A. (2011). Hydraulic properties of unsaturated
soils. In J. Gliński, J. Horabik, & J. Lipiec (Eds.), Encyclopedia of agrophysics (pp.
368-376). Dordrecht: Springer Netherlands. doi:10.1007/978-90-481-3585-1_69
Wadsworth, A. H., & Smith, A. A. (1926). Some observations upon the effect of the size
of the container upon the capillary rise of water through soil columns. Soil
Science, 22(3), 199-212. doi:10.1097/00010694-192609000-00003
Washburn, E. W. (1921). The dynamics of capillary flow. Physical review, 17(3), 273.
doi:10.1103/PhysRev.17.273
Wesonga, J. M., Wainaina, C., Ombwara, F., Masinde, P., & Home, P. (2014). Wick
material and media for capillary wick based irrigation system in Kenya.
International Journal of Science and Research, 3(4), 613-617. Retrieved from
https://www.ijsr.net/archive/v3i4/MDIwMTMxNTQ4.pdf
Wien, H. C., & de Villiers, D. S. (2005). Inducing Lettuce Tipburn with Relative Humidity
Modification. HortScience, 40(4), 1053C-1053.
doi:10.21273/HORTSCI.40.4.1053C
Wilfret, G. J., & Harbaugh, B. K. (1977). Evaluation of poinsettia cultivars grown on two
irrigation systems. Proceedings of the Florida State Horticultural Society, 90, 309-
311. Retrieved from
https://journals.flvc.org/fshs/article/download/97473/93478
Wilkinson, K. M., Landis, T. D., Haase, D. L., Daley, B. F., & Dumroese, R. K. (2014).
Tropical nursery manual: a guide to starting and operating a nursery for native
and traditional plants. Agriculture Handbook 732. Washington, DC: US
Department of Agriculture, Forest Service. 376 p., 732. Retrieved from
https://www.fs.fed.us/rm/pubs_series/wo/wo_ah732.pdf
Wilson, G. C. S. (1980). Perlite system for tomato production. Acta Horticulturae(99),
159-166. doi:10.17660/ActaHortic.1980.99.19
Wladitchensky, S. (1966). Moisture content and hydrophility as related to the water
capillary rise in soils. Trans. Am. Geophys. Un., 38, 222-232. Retrieved from
http://hydrologie.org/redbooks/a082/iahs_082_0360.pdf
Yeager, T. H., & Henley, R. W. (2004). Irrigation and fertilization for minimal
environmental impact. Acta Horticulturae(638), 225-230.
doi:10.17660/ActaHortic.2004.638.30
Yi, X., Li, G., & Yin, Y. (2013). Comparison of three methods to develop pedotransfer
functions for the saturated water content and field water capacity in permafrost
region. Cold Regions Science and Technology, 88, 10-16.
doi:10.1016/j.coldregions.2012.12.005
Wicking bed design Appendix 1
96
Appendix 1 - Selection of wicking bed designs from popular literature
Source
URL
Design type
Reservoir media
Reservoir
depth
Fabric
Growing media
Growing
depth
Comment
VEG - Very
Edible
Gardens
https://www.wicking
beds.com.au/make-
wicking-bed/
horizontal layers
7mm or 1/4inch bluestone
screenings (quarter
minus)are the ticket, or pea
gravel
150-200+ mm
geotextile or
double shade
cloth
fairly porous loam that is not too
heavy in clay or organic matter, a
sandy loam with moderate
organic matter, ask for a good
organic mix for veggies
250-400mm
The Little
Veggie Patch
Co
https://littleveggiepat
chco.com.au/blogs/n
ews/building-a-
wicking-bed
horizontal layers
with wicks of
geotextile
fine grade scoria (the finer
the grade of scoria, the
better)
250mm
geotextile
soils that have high levels of
organic matter and compost, add
1 part perlite to 10 parts soil
250mm
Gardening
Australia
https://www.abc.net.
au/gardening/factshe
ets/building-a-
wicking-bed/9435452
horizontal layers
scoria, gravel or another
aggregate
200mm
Geotextile or
old shade
cloth
good quality vegie garden soil
that's high in organic matter
300mm
Permaculture
Research
Institute
https://permaculture
news.org/2011/06/20
/from-the-bottom-
up-a-diy-guide-to-
wicking-beds/
horizontal layers
gravel
<= 300mm
landscape
fabric
high carbon soil, a combination
of loam, compost and peat
300-320mm
Sustainable
Gardening
Australia
https://www.sgaonlin
e.org.au/wicking-
beds/
horizontal layers
with soil
extending down
into reservoir
15cm gravel or scoria then
15cm soil blend (saturation
layer)
300mm
geotextile on
top of scoria
layer
good quality soil/compost blend,
1/2 mushroom compost and 1/2
organic soil mix
300mm
wicking beds work best
with a higher than usual
compost portion
Deep Green
Permaculture
https://deepgreenper
maculture.com/diy-
instructions/wicking-
bed-construction/
horizontal layers
coarse scoria
200mm
geotextile or
shade cloth
high grade soil with a good level
of organic matter; a mix of 50%
premium soil, 25% organic
compost and 25% organic cow
manure
400mm
Wicking bed design Appendix 1
97
Source
URL
Design type
Reservoir media
Reservoir
depth
Fabric
Growing media
Growing
depth
Comment
Green Life
Soil Co
https://www.greenlif
esoil.com.au/sustaina
ble-gardening-
tips/wicking-beds
horizontal layers
fine bluemetal or gravel,
coarse woody mulch (eg.
tree prunings) or coarse
river sand (If using stones
or gravel, use small ones)
200mm
geotextile or
shade cloth
good, friable organic vegetable
mix
200-300mm
Sophie's
Patch
https://sophiespatch.
com.au/2018/01/26/
wicking-bed-trouble-
shooting/
not blue metal or crushed
limestone (very alkaline)
none
Wicking only works when the soil
is high in organic matter; good
quality commercial vegie garden
soil and then add about 1/3 to
1/2 more compost to it
update on ABC
Gardening Australia
designs; not mushroom
compost - very alkaline
Urban Food
Garden
https://www.urbanfo
odgarden.org/main/w
icking-beds/wicking-
beds.htm
horizontal layers
25mm scoria; 20mm
bluestone
90mm
non-woven
weed mat
light friable soil
310mm
Gaia's
Organic
Garden
http://www.gaiasorga
nicgardens.com.au/h
ow-to-make-wicking-
bed/
horizontal layers
porous stone, such as pea
gravel
150-200mm
geotextile, old
sheets or
fabric
soil mix with good drainage
250mm
tuck fabric into sides so
that it creates a neat and
secure layer
Urban
Agriculture
Australia
http://www.urbanagr
iculture.org.au/infor
mation/design-
systems/building-a-
wicking-bed/
horizontal layers
washed river sand, 6mm
road base, pea shingle,
scoria, straw or autumn
leaves, or similar
<= 300mm
geotextile or
sugar cane
mulch
lightweight free draining
vegetable mix, sandy loam,
compost (or mixture)
Sustainable
Gardening
Australia
https://www.sgaonlin
e.org.au/sustainable-
wicking-worm-bed/
horizontal layers
aged pine bark chips
300mm
shade cloth
good soil
Milkwood
Permaculture
https://www.milkwoo
d.net/2010/05/11/ho
w_to_make_a_wickin
g_bed/
horizontal layers
gravel
300mm
none
300mm
SBS (Costa)
https://www.sbs.com
.au/shows/costa/listi
ngs/detail/i/1/article/
6172/wicking-garden-
beds
horizontal layers
with plastic voids
in reservoir
washed river sand
geotextile
lightweight planter mix
sand to bottom of
reservoir around and
over plastic tanks
Wicking bed design Appendix 1
98
Source
URL
Design type
Reservoir media
Reservoir
depth
Fabric
Growing media
Growing
depth
Comment
Better Homes
and Gardens
https://www.bhg.co
m.au/how-to-make-a-
wicking-bed-trough
horizontal layers
20mm scoria
200mm
geotextile
potting mix
270mm
ABC Organic
Gardener
https://www.organicg
ardener.com.au/articl
es/creating-wicking-
bed
horizontal layers
gravel
100mm
hessian bag
organic potting mix, compost,
worm castings and minerals
200mm
small wicking bed in
polystyrene box
Medium
https://medium.com/
@rosseyre/how-to-
build-a-wicking-bed-
version-2-0-
f5ddf52f4d57
horizontal layers
stones/gravel
<= 300mm
builders fabric
high quality organic compost,
loamy soils, peat, etc.
<= 400mm
growing media to 10-
20cm below overflow;
uses toilet cistern with
float to keep reservoir
full - use 100mm
reservoir depth
Permaculture
College
Australia
https://permaculture.
com.au/water-saving-
wicking-bed/
horizontal layers
50+mm rocks to ~70mm
then gravel
200mm
old shade
cloth
100mm layer woodchips then
mix of compost, crushed basalt,
sand, soil, crushed biochar
350mm
My Smart
Garden
https://www.mysmar
tgarden.org.au/Resou
rces/Water/Build-a-
wicking-bed
horizontal layers
7mm bluestone screenings
150mm
geotextile
fabric or
doubled
shade cloth
fairly porous loam that is not too
heavy in clay or organic matter
250mm
Simple
Savings
https://www.simples
avings.com.au/p/How
-to-make-an-IBC-
Wicking-Bed
horizontal layers
5-7mm screenings
weed mat
30% compost, 70% loam
Colin Austin
www.waterright.com.
au/wicking_%20bed_
technology.pdf
horizontal layers
Wood chips
200mm
none
Soil
200-300mm
Canberra
Permaculture
Design
http://www.canberra
permaculturedesign.c
om.au/wicking-
beds.html
horizontal layers
gravel or scoria, crushed
bricks or sand, or a
combination <10mm dia
1/4 height of
bed eg
100mm
geotextile
well-rotted down compost and
sand; potting mix
eg 300mm
EcoFilms
http://www.ecofilms.
com.au/create-a-
wicking-bed-garden-
for-easy-vegetable-
horizontal layers
vermiculite; coconut coir
fibre
90mm
none
good potting mix
300mm
Wicking bed design Appendix 1
99
Source
URL
Design type
Reservoir media
Reservoir
depth
Fabric
Growing media
Growing
depth
Comment
growing-powered-by-
fishwater/
My Home
Harvest
http://myhomeharve
st.com.au/newmhh/
wicking-bed-project/
horizontal layers
scoria
150mm
shade cloth
good quality soil, compost and
potting mix
200mm
Goodlife
Permaculture
https://goodlifeperm
aculture.com.au/a-
wildlife-proof-no-dig-
garden-wicking-bed/
horizontal layers
7-20mm blue metal
200mm
geo-fabric
no-dig layers - straw, aged chook
poo
400mm
Verge
Permaculture
https://vergepermacu
lture.ca/2011/05/30/
guide-to-wicking-
beds/
horizontal layers
28mm washed rock
300mm or
less
high grade
landscape
fabric
high carbon soil
300-320mm
Mindarie
Regional
Council
https://www.mrc.wa.
gov.au/School-
community/Fact-
Finding-
Information/Wicking-
Bed
horizontal layers
broken bricks and sand or
flowerpots stuffed with old
t-shirts
old carpet or
shade cloth
soil/compost
Permaculture
West
www.willettongarden
.org.au/wp-
content/uploads/R5-
pw-wicking-beds.pdf
blue metal, gravel, crushed
brick, or even coarse mulch
300mm
shade cloth,
geotextile,
carpet, old
sheets etc
good quality organic soil that's a
bit coarse
300mm
Hume City
Council
https://www.hume.vi
c.gov.au/files/shared
assets/hume_website
/environment/live_gr
een_-
_get_involved/garden
ing_fact_sheets/gard
ening_fact_sheets_a
mp_books/wicking_g
arden_bed_fact_shee
t.pdf
horizontal layers
7mm gravel or sand
200mm
70% shade
cloth or
geotextile
Garden soil, blended with
compost
400mm
Wicking bed design Appendix 1
100
Source
URL
Design type
Reservoir media
Reservoir
depth
Fabric
Growing media
Growing
depth
Comment
Tasman
Ecovillage
https://tasmanecovill
age.org.au/happening
s/news/301-wicking-
bed-guide
horizontal layers
sand & scoria/limestone or
other porous stone/rock
200mm gravel
+ 200mm soil
geotech fabric
improved soil (soil+chicken
manure+compost)
200mm
Living
Fundraisers
https://livingfundrais
ers.com.au/wp-
content/uploads/201
6/05/Wicking-Bed-
Activity-Sheet.pdf
horizontal layers
scoria (stones)
geotech fabric
good quality vegie growing soil
300-400mm
Sustainable
Education
https://sustainableed
ucation.com.au/susta
inable-living/wicking-
beds/
horizontal layers
gravel or scoria
300mm
geotextile
fabric
good quality well-draining loamy
soil with high organic content
300mm
Permaculture
Central Coast
https://permaculture
cc.org.au
horizontal layers
lightweight rocks, or gravel
+ plastic pots for voids
<= 300mm
weed
matting/
geotech
fabric/ shade
cloth
soil high in organic matter
300mm
Wicking bed design Appendix 2
101
Appendix 2 - Arduino code for data logger
//***********************************
// Read voltages of 8 tensiometers thru MUX
// write data to SD card
// and display the data on the monitor
//
// There are two versions of this code defined by LOGGER_ID
// 1 has temperature/humidity sensors, 2 does not
// Chris Curtis 24/10/2019
//
// 20/2/20 - added code to read temperature/humidity sensor
// 20/2/20 - added LOGGER_ID to output string
//*************************************
#include <math.h>
#include <Wire.h>
#include <SparkFunDS1307RTC.h> // for rtc
#include <SPI.h> // for sd card
#include <SD.h> // for sd card
#include "DHT.h"
//#define LOGGER_ID 1 // monitors beds 1-8 and includes
temperature/humidity sensor
#define LOGGER_ID 2 // monitors beds 9-16 (no temperature/humidity sensor)
#define NUM_READS 11 // Number of sensor reads for filtering
#define FIRST_BED 1
#define LAST_BED 8
#define FILE_NAME "tensio2.csv"
#define DHTPIN 15 // Digital pin connected to the DHT sensor (A1)
#define DHTTYPE DHT11 // DHT 11
const String version = "soilMoisture v1.4";
String currentTime;
int i;
int bedId;
int tensVal[9]; // median values from tensiometers - use [1] thru [8], [0] not
used
int rawVal[NUM_READS];
float temperature1; // temperature from LM35 sensor
byte dat[5]; //data from temperature/humidity sensor
boolean readNow = true;
File myFile;
const int chipSelect = 4;
float h,t; // fro humidity and temperature readings
// Initialize DHT temperature/humidity sensor.
DHT dht(DHTPIN, DHTTYPE);
// set interval between readings (3600,000 ms = 1 hour)
// const unsigned long sleepTime = 10000L;
const unsigned long sleepTime = 3600000L;
unsigned long previousMillis = 0;
void setup() {
// initialize serial communications at 9600 bps:
Serial.begin(9600);
Serial.println(version);
Serial.println("Initialising...");
// initialise the real time clock
rtc.begin();
// initialise ten temp/humidity sensor
dht.begin();
// initialise the SD card
Serial.print("Initializing SD card...");
Wicking bed design Appendix 2
102
pinMode(SS, OUTPUT);
if (!SD.begin(chipSelect)) {
Serial.println("initialization of SD card failed!");
return;
}
Serial.println("initialization of SD card done.");
// initialize the mux addressing digital pins as an output.
pinMode(6, OUTPUT); // S0
pinMode(7, OUTPUT); // S1
pinMode(8, OUTPUT); // S2
pinMode(9, OUTPUT); // S3
// initial read of sensors so data is ready when requested
//readSensors();
Serial.println("Ready!");
}
void loop() {
unsigned long currentMillis = millis();
// only run code if sleeptime has elapsed
if ((currentMillis - previousMillis >= sleepTime) or (previousMillis == 0))
{
// save the last time the loop ran
previousMillis = currentMillis;
// read 8 tensiometers
readTensiometers();
temperature1 = readTemp();
if (LOGGER_ID == 1) {
// Reading temperature or humidity takes about 250 milliseconds!
// Sensor readings may also be up to 2 seconds 'old' (its a very slow
sensor)
h = dht.readHumidity();
// Read temperature as Celsius (the default)
t = dht.readTemperature();
}
// write results to SD card
saveReadings();
// display results on monitor
displayReadings();
} // end of if currentmillis...
} // end of loop()
// *********************************
// for each bed, read the sensor multiple times then use the median value
// *********************************
void readTensiometers() {
// get time from RTC
rtc.update();
currentTime = String(rtc.date()) + "/" + String(rtc.month()) + "/" +
String(rtc.year()) + "," + String(rtc.hour()) + ":" + String(rtc.minute()) +
":" + String(rtc.second());
for (bedId=FIRST_BED; bedId<=LAST_BED; bedId++) {
setupMux(bedId);
for (i=0; i<NUM_READS; i++) {
rawVal[i] = analogRead(7);
delay(5);
} // end of multiple read loop
// use median value from multiple reads
sortReadings();
tensVal[bedId] = rawVal[NUM_READS/2];
} //end of bed loop
} // end of readTensiometers()
Wicking bed design Appendix 2
103
// *********************************
// set up multiplexor to read from specified input
// *********************************
void setupMux(int bedId) {
// mux addressing
// i S0 S1 S2 S3
// 14 0 1 1 1 - bed 1
// 13 1 0 1 1 - bed 2
// 12 0 0 1 1 - bed 3
// 11 1 1 0 1 - bed 4
// 1 1 0 0 0 - bed 5
// 2 0 1 0 0 - bed 6
// 3 1 1 0 0 - bed 7
// 4 0 0 1 0 - bed 8
int S0 = 6;
int S1 = 7;
int S2 = 8;
int S3 = 9;
switch (bedId) {
case 1:
digitalWrite(S0, LOW);
digitalWrite(S1, HIGH);
digitalWrite(S2, HIGH);
digitalWrite(S3, HIGH);
break;
case 2:
digitalWrite(S0, HIGH);
digitalWrite(S1, LOW);
digitalWrite(S2, HIGH);
digitalWrite(S3, HIGH);
break;
case 3:
digitalWrite(S0, LOW);
digitalWrite(S1, LOW);
digitalWrite(S2, HIGH);
digitalWrite(S3, HIGH);
break;
case 4:
digitalWrite(S0, HIGH);
digitalWrite(S1, HIGH);
digitalWrite(S2, LOW);
digitalWrite(S3, HIGH);
break;
case 5:
digitalWrite(S0, HIGH);
digitalWrite(S1, LOW);
digitalWrite(S2, LOW);
digitalWrite(S3, LOW);
break;
case 6:
digitalWrite(S0, LOW);
digitalWrite(S1, HIGH);
digitalWrite(S2, LOW);
digitalWrite(S3, LOW);
break;
case 7:
digitalWrite(S0, HIGH);
digitalWrite(S1, HIGH);
digitalWrite(S2, LOW);
digitalWrite(S3, LOW);
break;
case 8:
Wicking bed design Appendix 2
104
digitalWrite(S0, LOW);
digitalWrite(S1, LOW);
digitalWrite(S2, HIGH);
digitalWrite(S3, LOW);
break;
default:
digitalWrite(S0, LOW);
digitalWrite(S1, LOW);
digitalWrite(S2, LOW);
digitalWrite(S3, LOW);
break;
}
delay(10);
} // end of setupmux()
// *********************************
// sort array of values read so that the median value can be found
// *********************************
void sortReadings() {
int j;
int tempVal;
for(i=0; i<NUM_READS-1; i++)
for(j=i+1; j<NUM_READS; j++)
if ( rawVal[i] > rawVal[j] ) {
tempVal = rawVal[i];
rawVal[i] = rawVal[j];
rawVal[j] = tempVal;
}
} //end of sort()
// *********************************
// save readings from tensiometers to sd card
// *********************************
void saveReadings() {
myFile = SD.open(FILE_NAME, FILE_WRITE);
if (myFile) {
myFile.print(String(LOGGER_ID) + "," + String(currentTime) + "," +
String(tensVal[1]) + "," + String(tensVal[2]) + "," + String(tensVal[3]) + ","
+ String(tensVal[4]) + "," + String(tensVal[5]) + "," + String(tensVal[6]) +
"," + String(tensVal[7]) + "," + String(tensVal[8]) + "," +
String(temperature1));
// add humidity (%) and temperature (degrees C)
myFile.print (',');
myFile.print (String(h)); // display the humidity
myFile.print (',');
myFile.println (String(t)); // display the temperature
myFile.close();
} else {
Serial.println(String(currentTime) + " Write to sd card failed");
}
} // end of save
// *********************************
// display readings on monitor
// *********************************
void displayReadings() {
Serial.println(String(currentTime) + "," + String(tensVal[1]) + "," +
String(tensVal[2]) + "," + String(tensVal[3]) + "," + String(tensVal[4]) + ","
+ String(tensVal[5]) + "," + String(tensVal[6]) + "," + String(tensVal[7]) +
"," + String(tensVal[8]) + "," + String(temperature1));
Serial.print ("Current humdity = ");
Serial.print (String(h)); // display the humidity
Serial.println ('%');
Serial.print ("Current temperature = ");
Serial.print (String(t)); // display the temperature
Serial.println ('C');
}
// *********************************
// read TMP35 temperature sensor
Wicking bed design Appendix 2
105
// 10mV/degC with 500mV offset
// actual voltage supplied to sensor is 4.74V
// *********************************
float readTemp() {
int sensorVal;
float degC;
sensorVal = analogRead(2);
degC = ((4.740 / 1024 * sensorVal) - 0.5) * 100;
return degC;
}
