Content uploaded by Budi Indra Setiawan
Author content
All content in this area was uploaded by Budi Indra Setiawan
Content may be subject to copyright.
SMALL SCALE SUSTAINABLE AGRICULTURES WITH THE
APPLICATION OF PITCHER IRRIGATION1
Budi I. Setiawan2, Edward Saleh3 and Hermantoro4
ABSTRACT
Pitcher, a bottle-like irrigation emitter made of baked clayed soils mixed with sands has
been recognized as the oldest traditional irrigation. It has high water efficiency since the
water seeps directly to and concentrated in the root zones. We conducted numerical and
experimental studies to investigate the water flow in the soil surrounding a pitcher and to
figure out the availability of soil water for crops. The Darcy and Richards’ equation of
water flow in a cylindrical coordinate system was applied and was solved using Finite
Element Method to describe soil moisture profiles. Two soil textures were used, one was
silty clay and the other was sand. The hydraulic conductivity of the pitcher was in order
10-6 cm/s which was smaller 100 times than that of the two soils. The pitcher was buried
in the center axis of a soil box and water was given from Mariotte tube to maintain a
constant the water level inside the pitcher. In another studies, we figured out how the
pitcher can also release nutrients when filled with dissolved fertilizers. For this purpose,
we measured hydraulic and hydro-dynamic properties of the pitcher, and simulated solute
transport using the convective-dispersive equation, and observed the effectiveness of
fertigation in which NPK fertilizers were used on bush pepper plants. The advancements
of wetting front in the soil was very slow and somewhat limited to a radius and depth of
no more than 30 cm and 40 cm, respectively for the tested soils. The soil moisture was in
a range available for plant growth. Different depths of pitcher placement in the soil
produced different reaching distances of the wetting front but showed insignificant
different in the water availability. The pitcher was capable to release dissolved solution.
The three nutrients have different distribution patterns. Nitrogen was well distributed,
Phosphor was accumulated close to the pitcher’s wall, and Potassium increased gradually
with distance. These difference patterns were caused by the difference of hydrodynamic
coefficients in which Nitrogen was the largest value among the others. The diffusion
coefficient ranged at 1.01x10-7 – 4.1x10-3 cm2/day for NaCl, and 6.7x10-6 - 3.5x10-3
cm2/day for NPK fertilizers. Bush papers planted surrounding a pitcher enabled to extract
the nutrients as shown by the progressive growth of the crops: height, branches, leaves
and flowers, which were monitored daily. After unearthed, roots of the bush pepper
developed only in the wetter soils. An accurate placement of pitcher in soil is important
to provide effective soil wetness in the root zone and reduce evaporation rate. The right
placement of pitcher must be determined based on the hydraulic characteristics of the
1 Presented at Postgraduate GP Education Workshop on From Environmental to Sustainable Science:
thinking the shift and the role of Asian agricultural science. Ibaraki University, January 12-13, 2009.
2 Bogor Agricultural University, Bogor. E-mail: budindra@ipb.ac.id.
3 Faculty of Agriculture, Sriwijaya University, Palembang.
4 Agricultural Higher School Institute, Yogyakarta.
pitcher and the soil. From this study we found 5 cm placement depth of the pitcher’s
shoulder is an appropriate reference for the application of pitcher irrigation.
Key words: pitcher irrigation, water flow, solute transport, fertigation, bush peppers.
1. INTRODUCTION
Indonesia has approximately 33.3 million hectares of dry land (BPS, 1997) within the D
and E types as categorized by Oldeman’s agro-climatic zone. In these regions the annual
rainfall is no more than 300 mm which falls within three months in the rainy season
(October to December), such as occurring in the eastern of part of Lombok Island. In this
location there are about 120 ha of dry lands which have become arable thanking to the
groundwater development project in the late 1990s. Since then many irrigation schemes
have been developed to gain higher water use efficiency or reducing water loss due to
high evaporation. One of them is pitcher irrigation to grow chilies and tomatoes
(Setiawan et al., 1998; Setiawan, 2000). This irrigation system could reduce the
evaporation and percolation which has been reported earlier by Mondal (1978).
Pitcher irrigation which uses bottle-like baked clayed soils mixed with sands has been
known as the oldest traditional irrigation. Mondal (1974) classified pitcher irrigation into
subsurface irrigation whereas Stein (1990) classified it into local irrigation since water
seeps slowly with low rate into the root zones resulting in partly wetted soil. Later on
Stein (1994) submerged two types of pitcher into clay-sand soil to observe the seepage
which has different saturated hydraulic conductivity of 3.9*10-7 cm/s and 3.6*10-4 cm/s.
The resulted seepage was about 1.25 l/day (0.014 cm3/s) at the initial stage but then
declined until finally reached the values between 0.5 to 0.6 l/day.
Figure 1 shows a pitcher filled with water is buried in soil and plants are surrounding it.
Water seeps through the pitcher’s wall to the soil when the soil is drier but then the water
rate decreases with time and stops occasionally if there is no extraction by plant roots.
This watering mechanism is known as self-regulating (Saleh, 1993).
Figure 1. Scheme of pitcher irrigation and planting layout
In this study, we conducted numerical and experimental works to investigate
performances of pitcher irrigation under two different soil textures and to find out water
availability for crops in the surrounding soils. We also investigated solute’s concentration
in the soil matrices when the solution of fertilizers was filled into the pitcher, and its
effectiveness to the growth of plants.
2. MATERIALS AND METHODS
Pitcher and Soil Properties
In general, parts of the pitcher designed for this experiment consist of body, shoulder and
neck. The body has diameter of 15 cm and height of 15 cm, and its neck has diameter of
5 cm and height of 10 cm. The thickness of the wall is 1 cm. The pitcher was made from
a mixture of clay and sand. The hydraulic conductivity of the pitcher was tested by means
of a modified constant head permeameter and after repetitive measurements its value
ranged between 4.56∙10-6 to 8.78∙10-6 cm/s. The pitcher was then buried in the center of
soil box 30 cm having length of 100 cm and depth of 50 cm. Pressure transducers were
inserted into the soil and were connected to computer for real time measurements.
Mariotte tube was used to supply water and maintained the water level inside the pitcher.
Accumulative infiltration was manually measured from the Mariotte tube made of acrylic
cylinder.
Water Flow Equations
To describe soil moisture profiles in the unsaturated soils, the Darcy and Richards’
equation of water flow in porous medium was applied. The equation in a cylindrical
coordinate system with the assumption that the soil is homogenous, isotropic and
isothermal conditions can be stated as follows:
0
1
tz
K
zhCK
zrhCK
r
rr ww
....................................................... (1)
Where, θ is volumetric soil water content (cm3 cm-3), h is soil water head (cm), K is
unsaturated hydraulic conductivity (cm s-1), Cw is specific water capacity (cm-1), r is
radius (cm), z is depth (cm) and t is time (s).
Two soil types were used. One was Sandy Soil and the other was silty Clay. Water
retention curves of the soils conformed to Genuchten (1980) model which in such way
modified by Setiawan (1992) to facilitate the presence of positive water head inside the
pitcher. Furthermore, the hydraulic conductivity function of the soils was measured by
meant of the instantaneous profile method and the results was well represented by the
following equation (Setiawan and Nakano, 1993).
Initially, the soil moisture was approximately homogenous and hysteretic effects might
be negligible since the water flow would be very slow. To solve the water flow equation
subjected to the boundary and initial conditions, we applied Galerkin weighted residual
of Finite Element Method (Segerlind, 1984).
Solute Transport Equations
The 2-D solute transport in the soil was used to see distribution of the solute in the soil
matrices. This was done by solving the Equation 3 using Finite Difference Method and
applying Alternate Directing Implicit Scheme (Hermantoro, 1999).
2
2
2
2
z
C
D
z
C
v
x
C
D
x
C
v
t
Cs
z
s
x
............................................................ (2)
Where, C is concentration in soil (g/l), Ds is diffusion coefficient (cm2/s), vx and vz are
average water flow velocity in x- and z- axes, respectively (cm/s), θ is volumetric water
content in soil (cm3/cm3), x and z are distance (cm), and t is time (d).
The coefficient Dw in the range of 1.01∙10-7~4.1∙10-3 cm2/d for NaCl and 6.7∙10-6~3.5∙10-3
cm2/d for NPK in water through the wall of the pitcher, witch were varied with
concentration in the form of Dw=1∙10-10C 9.6832 for NaCl solution and Dw=7∙10-9C 6.2739 for
NPK solution. In both cases, the Saturated Conductivity of the pitcher was 1.01∙10-4
cm/d. The coefficient for NaCl solution soilvaried in the range of 0.3 to 0.825 cm2/d with
volumetric water contents of the soil were 0.60 0.04 cm3/cm3.
Experiments on Pepper Bushes
Four bush peppers were planted in the soil surrounding the pitcher in which NPK solution
was inserted. Water level in the pitcher was maintained with the mariotte tube. Changes
of NPK solution in the soil were sampled and measured analytically, and growths of the
plants, height, leaf, flowers, etc., were observed. Finally, the soil then excavated to
observe the distribution of roots. These experiments were conducted in a green-house.
3. RESULTS AND DISCUSSION
Infiltration Rates
Figure 2 shows rates of water volume (infiltration rate) entering the two types of soils
from the pitcher. As it is common, infiltration rate decreases exponentially with time until
then reaches a steady state. In this case, however, the decrease was rather gradual even
though the soil was initially dry. Pattern of the curves was almost similar for both soils
even though the two soils have contrasting soil properties and different initial conditions.
This indicates the effectiveness of the pitcher permeability which was lower than the
permeability of the soil in controlling the infiltration rate. Earlier, Stein (1994, 1997) was
referring it to auto regulative system to explain this infiltration phenomenon in pitcher
irrigation. Figure 4 also shows calculated infiltration rates which are in good agreement
with the measured data for Kpitcher=6.28∙10-6 cm/s and Ksilt=7.70∙10-4 cm/s, and
Kpitcher=7.0∙10-6 cm/s and Ksand=8.95∙10-4 cm/s, respectively.
Figure 2. Infiltration rates in silty clay and sandy soil. Line is calculated and dotted is
measured values.
The values of cumulative infiltration in silty clay are well represented by a straight line
with the slope approaching 1 and the intercept equal 0. Whilst in sand soil, a slight
deviation between them occurred at the longest times but still gained reasonable results.
Soil Moisture Profiles
Figure 3 shows contour lines of water suction equals 450 and 200 cm of water for silty
clay and sandy soil, respectively taken at several elapsed times. At these times, advances
of wetting front for both soils were very limited and can be said attaining quasy steady
states. Wetting front ceased from further advancement and it was confirmed later after
slashing the soils that there was clear demarcation lines between wetted and remaining
dry regions. The radial and vertical advancements of wetting front was up to 14 cm and
20 cm for the Silty Clay, and 20 cm and 30 cm for the sandy Soil, respectively.
Figure 3. Calculated water suction in the silty clay soil (left) and for sandy soil (right)
at several elapsed times.
In general, the radial wetting front is shorter and the vertical wetting front is longer as the
pitcher placement is deeper. These differences however are not so significant and all
cases still provide available water for plant growth. The deeper pitcher placement
produces drier soil surfaces and consequently it will reduce evaporation rates because of
the effect of soil mulches (lower unsaturated hydraulic conductivity). However, too deep
placement of the pitcher that produces less moisture at the surface soils may give
undesired effects to plant growth at the earlier stages due to their shorter rooting systems.
It is clear that advancement of wetting front in pitcher irrigation was very slow and
somewhat limited to form a balloon like moisture profile within a radius and depth of no
more than 30 cm and 40 cm, respectively for both tested soils. This depth has been
recognized elsewhere as an effective zones for the extraction of soil moisture by pant
roots (Igbadun, et.al., 2007). However, once there is a distortion of moisture profiles for
example due to root extraction then water from the Mariotte tube flows immediately into
the soil. This is another explanation of the self-regulating mechanism that guarantees the
availability of water for plant growth any time such as also was reported earlier by
Setiawan (1998).
The velocity varied with location and direction in the range of 0.18144 to 28.5984 cm/d.
Correlation of numerical results (θm) and data measured (θd) in some points close to the
pitcher obtained θm= 0.913∙θd and R2=0.81. A cluster of data lower than 0.4 cm3/cm3
was overestimated by the calculation, while other data was well represented.
Solute distribution in the soil matrices
Figure 4 show concentration profile of NaCl solution in the soil matrices after reaching at
a stable condition 8.34 days from the beginning of the application. The similar form of
distributions of water content and and NaCl concentration was very clear with which it
figures out the important of water as the effective medium of the NaCl solution transfer
in the soil matrices. Here again, NaCl solution moved farther in z-axes than that in x-
axes. The concentration was larger in the regions closer to the pitcher. At 1 cm apart from
the pitcher, the relative concentration reached 0.89 while at the moving front was about
0.10. Correlation of numerical results (Cm) and data measured (Cd) in some points close
to the pitcher obtained Cm= 0.874∙Cd and R2=0.81. Data was unevenly distributed in two
clusters. One cluster was in the range of 0.2-0.4 while the other concentrated at 0.8. There
was no data capable to measure between 0.4-0.8.
Figure 4. Distribution of NPK Concentration in the soil matrices at the early stage and
equilibrium conditions.
Experiments on bush peppers
When one pitcher was surrounded by four bush peppers, the irrigation rate fluctuated
with time. The lowest and highest rates were 0.56 and 1.30 l/d, respectively with the
averaged was 0.81 l/d, which is equal to 2.33 mm/d. While, evapotranspiration rate which
was measured independently were 1.9 to 4.3 mm/d with the average was 2.8 mm/d. With
these values indicate that the irrigation rates much or less equals to the evapotranspiration
of bush peppers, or could meet the water demand of bush peppers for their growth and
developments. Wet regions in the soil matrices formed like a standing oval-ball with a
longer radius of 25 cm and vertical length of 70 cm. Nitrogen content was distributed
evenly in the soil matrices at ranges of 0.09 to 0.12%. Farther from the pitcher, Phosphor
content decreased significantly, from 27 to 6.3 ppm. Potassium content changed abruptly
but tended to decrease with distance, form 4.99 to 4.03 me/100g. Roots were
concentrating up to depths of 15 cm and its density decreased with depth and there were
no roots anymore at 65 cm. Up to the depth of 10 cm, cumulative wet roots amounted to
50 gram.
Bush peppers grew rapidly to 60-70 cm heights in 30 weeks but then slow downed and
only reached 65-75 cm in 5 days later. The maximum height was around 90-120 cm.
Leaves development also followed these tendencies with the total number of leaves was
60-85 pieces. New branches started from 1 up to 35 weeks amounted to 13-16 branches.
Flowers commonly appear when a new braches is coming out but old branches also
produce flowers. Here, flowers increased linearly with time amounted to 7 flowers in 12
days but then constant up to 27 weeks, and amounted to 9 in 35 weeks. From these
figures, the generative phase of bush peppers was higher than that of conventional
farming using spraying irrigation. In another experiment when the dosage of NPK was
decreased to 50%, there were no significant differences of bush peppers’ growth and
developments. Thus, it is possible to save fertilizer that was commonly used
conventionally.
4. CONCLUSIONS
These studies confirmed that pitcher irrigation can provide soil moisture available for
plant growth directly in the root zones. Infiltrated water accumulated in the root zones
with the maximum radius and depth of wetting front was no more than 30 cm and 40 cm,
respectively. Different depths of pitcher placement in the soil produced different reaching
distances of the wetting front but showed insignificant different in the soil moisture
availability. An accurate placement of pitcher depth in soil is important to provide
effective soil wetness in the root zone and reduce evaporation rate. The right placement
of pitcher must be determined by the characteristics of the pitcher it’s self and the
respective soil. In this study we found 5 cm placement depth of pitcher might be an
appropriate reference for pitcher irrigation practices.
The pitcher which had saturated hydraulic conductivity 1.01∙10-4 cm/d, diffusion
coefficient 6.7∙10-6~3.5∙10-3 cm2/d for NPK solution can fertilizer sufficiently for the
growth of the bush peppers. Water as well as fertilizer was concentrated and formed like
a ball in the soil matrices surrounding the pitcher where most of the roots of bush peppers
resided. Bush peppers could grow well as indicated by the developments of roots, leaves,
branches and flowers. It is possible to reduce the dosage of fertilizer application that
conventionally applied without the risk of decreasing yields.
5. REFERENCES
BPS. 1997. Indonesia in Numbers. Central Statistical Bureau (in Indonesian).
Hermantoro. 1999. Effectiveness of pitcher fertigation system in peppers growth.
Dissertation. Graduate School of Bogor Agricultural Unversity. Bogor, Indonesia
(in Indonesian)
Igbadun, H.E., H.F. Mahoo, A.K.P.R. Tarimo and B. A. Salim. 2007. Simulation of Soil
Moisture Dynamics of the Soil Profile of a Maize Crop under Deficit Irrigation
Scheduling. Agricultural Engineering International: the CIGR Ejournal. Manuscript
LW 06 015. Vol. IX. July, 2007.
Mondal, R.C. 1974. Farming with Pitcher: a technique of water conservation. World
crops Vol. 26(2), 91-97.
Mondal, R.C. 1978. More Water for Arid Lands: promising technologies and research
opportunities. National Academy of Sciences, Washington, D.C. Pages 153.
Saleh, E. Performances of pitcher irrigation system in dryland farmings. Graduate School
of Bogor Agricultural Unversity. Bogor, Indonesia ( in Indonesian).
Setiawan, B.I. 1992. Studies on Infiltration in Soil Having a Macropore. Dissertation in
Laboratory of Soil Physics and Hydrology, Division of Agricultural Engineering,
Faculty of Agriculture, The University of Tokyo. Tokyo. 216 p.
Setiawan, B.I. & M. Nakano. 1993. On the Determination of Unsaturated Hydraulic
Conductivity from Soil Moisture Profiles and from Water Retention Curves. Jur.
Soil Science Vol. 156(6) 389-395.
Setiawan, B.I., E.Saleh & Y.Nurhidayat. 1998. Pitcher Irrigation System for
Horticulture in Dry Lands. Proceedings water and land resources development and
management for sustainable use. Vol II-A. The Tenth Afro-Asian Regional
Confrence. ICID-CIID, INACID, Denpasar-Bali, Indonesia.
Setiawan, B.I. 2000. On the Dissemination of Pitcher Irrigation System for Horticulture
Farming in Dry Lands. Proceedings of China International Conference on Dry land
and Water-Saving Farming. Beijing, November 21~23, 2000.
Segerlind, L.J. 1984. Applied Finite Element Analysis. Second edition. John Wiley and
Sons, Inc, New York. 427 p.
Stein, Th.-M. 1990. Development of Design Criteria for Pitcher Irrigation. Cranfield
Institute of Technology, Silsoe College, M.Sc Tesis, August 1990.
Stein, Th.-M. 1994. Hydraulic Conductivity of Pitcher Material for Pitcher Irrigation
(first result). Zeitschrft fur bewasserungswirtschaft 30(1), 72-93.
Genuchten. M.Th.V. 1980. A Close-form Equation for Predicting the Hidraulic
Conductivity of Unsaturated Soils. Soil Sci. Soc. Am. J., Vol 44. Pp. 892-898.
