Directory UMM :Data Elmu:jurnal:S:Soil & Tillage Research:Vol56.Issue1-2.Jul2000:
Soil & Tillage Research 56 (2000) 3±14
Management of clay soils for rainfed lowland rice-based
cropping systems: an overview
H.B. Soa,*, A.J. Ringrose-Voaseb
a
School of Land and Food Sciences, The University of Queensland, St Lucia, Brisbane, Qld 4072, Australia
b
CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia
Abstract
The problem of concern in this project is that in the dry season following a lowland rice (Oryza sativa L.) crop, yields of
post-rice crops are generally low, despite adequate water commonly being available in the soil pro®le to grow a potentially
high yielding dry season (DS) crop without irrigation. Maize (Zea mays L.) yields are as low as 1 Mg haÿ1 or less, soybean
(Glycine max L. Merr.) and cowpea (Vigna unguiculata L.) at 0.3 to 0.8 Mg haÿ1 in Indonesia and mungbean (Vigna radiata
(L.) Wilzek) around 0.5 Mg haÿ1 in the Philippines. These are all very much below the yield potential of these soils. For
example, mungbean yields of 2.2 Mg haÿ1 have been achieved by IRRI in the Philippines on these soils without irrigation or
additional fertilisers. The causes of low yields of DS crops after rice are mainly poor crop establishment and poor root growth
due to soil physical constraints. These result from the breakdown of soil structure during wet cultivation (puddling) for rice.
Yields are also limited by biological and chemical constraints. As a result of these low yields, farmers are reluctant to invest in
post-rice crops. Therefore, land after lowland rice (at least 51 million ha in Asia according to Huke [Huke, R.E., 1982. Rice
Area by Type of Culture: South, Southeast, and East Asia. International Rice Research Institute, Los BanÄos, Philippines, 32
pp.]) represents an underutilised resource that can be used to meet the food requirement of the ever increasing population of
the developing world. To increase the utilisation of these soils, improved management practices are required to enable dry
season crops to use the stored water in the soil pro®le after the rice crop. This paper outlines a project which was established
with the general objective of developing soundly based soil management technologies that can overcome soil physical
limitations to DS crop production after lowland rice. The speci®c objectives of the program were
1. to test a range of soil management and agronomic practices that have the potential to overcome adverse soil physical
conditions for DS crops after rice, across a range of soil and climates;
2. to evaluate these practices by
2.1. measuring the changes in soil physical conditions throughout the complete cropping cycle from rice to DS crops;
2.2. determining the performance of the DS crop (establishment and growth) and its ability to extract soil water.
3. to determine the mechanisms involved in dispersion due to puddling and in ¯occulation and structural development as the
soil dries after draining surface water from rice ®elds.
Relevant outcomes from this project are described in the following papers in this issue. # 2000 Elsevier Science B.V. All
rights reserved.
Keywords: Soil management; Puddling; Rice; Legumes; Rainfed lowland
*
Corresponding author. Tel.: 61-7-3365-2888; fax: 61-7-3365-1177.
E-mail address: [email protected] (H.B. So).
0167-1987/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 7 - 1 9 8 7 ( 0 0 ) 0 0 1 1 9 - 7
4
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
1. Introduction
To keep pace with rapidly expanding populations,
the production of food legumes and other dry season
(DS) crops must be increased within the lowland rice
growing areas in Indonesia and the Philippines as well
as in other countries of southeast Asia. There are
51 million ha of lowland rice in Asia including 8.2
and 3.5 million ha of lowland rice in Indonesia and the
Philippines of which 3 and 2 million ha, respectively,
are rainfed (Huke, 1982). A feature of lowland rice
culture is that the amount of soil water remaining in
the dry season after the rice crop is usually adequate
for a DS crop. However, when DS crops are grown in
these soils, yields are generally low. For example,
yields are as low as 1 Mg haÿ1 or less for maize, 0.3±
0.8 Mg haÿ1 for soybeans and cowpeas in Indonesia
(Hoque, 1984) and 0.5 Mg haÿ1 for mungbeans is not
uncommon in the Philippines. These yields do not
provide adequate returns to the farmer, so that lowland
rice soils, in particular rainfed lowland rice soils,
represent an underutilised resource during the dry
season. The potential yield for mungbean in the
Philippines is approximately 2.2 Mg haÿ1 (So and
Woodhead, 1987). Therefore, increasing yields of
DS crops would increase the utilisation of land and
residual soil water during the period between rice
crops.
Despite the lower yields of DS crops compared to
rice, Table 1 shows that the lower costs and higher
prices of some DS crops, in particular mungbean, can
result in greater net returns from these crops than from
the rice crop if moderate yields can be obtained
(Maranan, 1986, 1987). Considerable bene®ts could
be expected from growing DS crops, including
increased farmer income and nutrition and a reduction
in imports of food legumes. In 1987, imports of maize
and soybean cost Indonesia about $25 million and $63
million, respectively and the Philippines about $7
million and $2.5 million. Peanut (Arachis hypogaea
L.) imports cost $22 million and $7.5 million to
Indonesia and the Philippines, respectively. In addition, the introduction of food legumes into the rice
rotation could result in substantial savings in nitrogenous fertilisers. These bene®ts would also be applicable to lowland rice areas in other southeast Asian
countries.
Both the Indonesian and Philippine governments
place high priority on raising yields of DS crops,
particularly legume crops, as a means of increasing
farmers' income as well as nutrition. The Indonesian
government expressed this through its Five Year Plans
(PELITAs), of which the Sixth Plan is current. The
Philippine Council for Agriculture, Forestry and Natural Resources Research and Development
(PCARRD) has a Mungbean Development Action
Plan to co-ordinate efforts to increase production of
mungbeans.
The causes of low yields of DS crops after rice are
often poor crop establishment and inferior root growth
due to adverse physical conditions of the soil which, in
turn, are caused by the wet cultivation (puddling)
undertaken for paddy rice (Pasaribu and McIntosh,
1985; So and Woodhead, 1987; Adisarwanto et al.,
1989). Yields are also limited by nutritional and
biological constraints.
Project 8938 was funded by the Australian Centre
for International Agricultural Research (ACIAR)
titled `The Management of Clay Soils for Lowland
Rice-based Cropping Systems' aimed to investigate
the factors that affect the success or failure of DS crops
grown after rice and to make a contribution towards
Table 1
National average yield, actual price, relative pro®tability and the ratio of returns/costs for several crops in the Philippines, 1985a
Crops
Yield
(Mg haÿ1)
Price
(Pesos/kg)
Total return
(Pesos/ha)
Cost of production
(Pesos/ha)
Net return
(Pesos/ha)
Net return/cost
(%)
Rice
Maize
Soybean
Mungbean
Peanut
2.40
1.04
0.99
0.69
0.85
3.24
2.80
7.30
15.40
10.10
7776
2912
7227
10626
8585
5370
2078
3697
3780
6959
2406
834
3530
6846
1626
44.8
40.1
95.5
181.1
23.4
a
Adriano and Cabezon, 1987.
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
the development of stable cropping systems that
incorporate DS crops within rainfed, lowland ricebased cropping systems. It was a synthesis of two
proposals, one from the University of Queensland and
another from CSIRO and the Philippine Bureau of Soil
and Water Management(BSWM) and involved the
University of Brawijaya, Indonesia and two Indonesian Institutes for Food Crops. The factors investigated
are tillage during the preparation for the rice and DS
crop phases, soil amendments, time of sowing the DS
crop, surface drainage and fertilisers.
This paper outlines the background to the project, a
review of the relevant literature and a description of
the experimental design and set-up of the project.
2. Background to the project
Rice in southeast Asian countries is mostly grown
under lowland conditions with 1±3 crops a year
depending on the availability of irrigation water and
the use of modern, short season varieties. After longterm submergence for lowland rice, soil water is
suf®cient to grow a DS crop with reasonable yield
potential. However, under current management practices the yields of DS crops (see above) are generally
well below the yield potential (Pasaribu and McIntosh,
1985; Adisarwanto et al., 1989). The area of lowland
rice is approximately 51 million ha in Asia (Huke,
1982). The untapped potential for food production
from DS crops is a large, underutilised resource.
Furthermore, in Africa there are approximately
100 million ha of land that could potentially be
adapted to rainfed lowland rice with appropriate soil
physical management (Woodhead, 1990).
The realisation that multiple cropping programs are
essential to raise production from rice based systems
lead to the formation of the Asian Cropping Systems
Network. This network coordinates efforts by the
International Rice Research Institute (IRRI) and the
various national programs to jointly develop appropriate rice-based cropping systems in major rice growing environments. (Hoque, 1984). As soil factors are
known to limit yield of DS crops, appropriate soil
management is an essential part of improved cropping
systems. For these systems to be developed, the
dynamics of soil-crop interactions in DS paddy soils
must be better understood.
5
2.1. The importance of legumes in rice-based
cropping systems
Indonesia has been self-suf®cient in rice since 1985
through the success of the government coordinated
BIMAS (mass guidance) and INMAS (mass intensi®cation) programs during past Five Year Plans or
PELITAs. From 1984, PELITA IV gave special attention to the ®rst DS crop (®rst secondary crop within a
lowland rice±DS crop±DS crop cropping system),
with particular emphasis on legumes (Nanseki et al.,
1989). These crops were targeted for increased production with the aim of improving farmers' income
and nutritional status (Vademecum BIMAS, 1987).
The target, in irrigated lowland areas, is to replace the
third rice crop with a DS crop and, in rainfed lowland
areas, is to grow a DS crop before or after the rice
crop. Based on the rate of consumption and imports,
the major DS crops in Indonesia are, in decreasing order,
maize, soybeans, peanuts and mungbeans (FAO, 1984).
The success of the BIMAS and INMAS programs is
partly due to the setting of realistic production targets,
which are negotiated for each province, county and
village which elected to join the program. These
production targets, when agreed to by the parties
concerned, become contracts that must be adhered
to (Agricultural Intensi®cation Program, 1988±1989)
and involve a minimum mandatory set of technology
packages (recipes) that must be carried out. If the
recipe is adhered to, a minimum and achievable
improved yield level is guaranteed. However, these
technology packages do not include soil physical
management recommendations due to a lack of
knowledge in this area. Where adequate irrigation
water is available, improved technology for soybean
has recently been launched through the government
extension program `Supra-insus' (special program for
intensi®cation) with the aim of raising soybean yield
from 1 to 1.5 Mg haÿ1 (Sumarno, 1990). As yet,
satisfactory packages have not been developed for
DS crops after rice.
In the Philippines, the major DS crops are, in
decreasing order, maize (Africa and Marquez,
1989), mungbean and soybean. PCARRD has a Mungbean Development Action Plan aimed at introducing
and studying the impact of new technologies for
mungbeans, in particular new varieties which are
shorter, faster maturing and higher yielding (Cabahug,
6
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
1990). Faster growing varieties are particularly suitable for growing after rice, when the legume is largely
dependent on stored soil water. Mungbean is the major
legume crop in the Philippines partly because, with a
protein content of 20±25%, it is a popular and inexpensive source of protein, often being referred to as
`the poor man's meat' (Cabahug, 1990). Its price is
relatively high and stable and farmer consumption
tends to compensate for any over-production, because
unlike soybean, it does not require processing. In
addition, Table 1 shows that it can be more pro®table
than other legumes or rice, if moderate yields can be
obtained.
Modest increases in mungbean yield in a ricelegume rotation can result in net returns from mungbean being greater than that from the rice component
(Lavapiez et al., 1977; Maranan, 1986, 1987). The
pro®tability of mungbean in Indonesia is also cited as
a major incentive towards their use after rice. However, under current management practices yields of DS
crops are poor and result in a reluctance to invest
management and resources in DS crops so that much
land is underutilised after lowland rice (Varade, 1990).
In addition, there is a social preference for rice.
Therefore, the introduction of management systems
that can stabilise yields of DS crops, particularly
legumes, after rice will have considerable socio-economic bene®ts.
2.2. Physical limitations of puddled soil
The physical limitations imposed by puddled soil
have been recognised as the major cause of poor
establishment and yield of post-rice crops in Asia,
including soybeans in east Java (Adisarwanto et al.,
1989) and mungbeans in the Philippines and other
Asian countries (IRRI, 1984; So and Woodhead, 1987;
Mahata et al., 1990; Varade, 1990). Puddling is associated with the breakdown of soil aggregates during
wet cultivation (Sharma and De Datta, 1985; Adachi,
1990) and results in a massive structure after rice.
After drainage of the surface water prior to rice
harvest, the water content of the surface soil decreases
which is accompanied by a rapid increase in redox
potential (IRRI, 1987; Maghari, 1990) and soil
strength (IRRI, 1985, 1986, 1987, 1988). Puddling
also creates a compacted layer below the puddled
layer, which increases in strength during drying (IRRI,
1986). The effect of seasonal conditions and soil type
on the germination, establishment and root growth of
DS crops after rice is determined by the interactions
between the rates of change of redox potential, soil
strength and available water as the soil dries. To devise
ways of overcoming these limitations, it is important
to quantify the nature of these interactions through a
program of detailed monitoring of the soil physical
conditions.
3. Factors affecting the establishment and growth
of DS crops after lowland rice
3.1. Effect of delay between ®eld drainage and
sowing on germination and establishment of post-rice
DS crops
Successful crop establishment is essential for high
yields, for example, yield of DS mungbean was linearly related to plant population density up to 0.55±0.6
million plants/ha (So and Woodhead, 1987; IRRI,
1988). A key factor determining the success of crop
establishment is the rate of germination. Rapid germination, which depends largely on soil water content
and seed-soil contact, is essential to minimise risks
from adverse factors (So and Woodhead, 1987).
The delay between ®eld drainage and sowing has a
major in¯uence on crop establishment because it
affects soil water content and hence germination
and emergence. Although mungbeans under controlled conditions can germinate at soil water potentials as low as ÿ2.2 MPa (below wilting point), poor
seed-soil contact under ®eld conditions reduces germination rates at low potentials and radicle elongation
rate is reduced at potentials below ÿ0.2 MPa (Fy®eld,
1987; IRRI, 1988). Emergence is slower and falls
below 50% when water potential is reduced below
ÿ0.1 MPa (IRRI, 1986).
Reports on the appropriate period of delay vary and
probably re¯ect differences in climatic conditions
during experiments and in soil type. Under conditions
of little rain after drainage of surface water, it appears
that emergence on silty clay loams is highest when
mungbeans are sown 6±10 days after drainage (DAD)
(Fy®eld, 1987; IRRI, 1987, 1988; Cook, 1989; Cook
et al., 1995). Later sowing tends to reduce emergence,
growth and yield of mungbeans, apparently because of
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
low water potentials and increased seedbed and subsoil strength. On the other hand, growth and yield can
also be reduced after sowing at delays of 0±3 DAD
apparently due to low redox potentials and poor
aeration (IRRI, 1987). It is not clear how these periods
would vary with soil types.
Since rice is generally harvested 7±10 DAD, it is
important that sowing of DS crops be carried out as
soon as possible after harvest. However, in regions
where the probability of rainfall after rice harvest is
high, farmers tend to avoid waterlogging by either
postponing sowing or by providing surface drainage.
Relay cropping, where legumes are sown soon after
draining and before rice is harvested, has been tried as
a means of reducing the sowing delay. However, this
method tends to reduce establishment and yield, as
well as increase problems of weeds and ratooning of
rice (IRRI, 1987, 1989).
3.2. The effect of soil amendments
3.2.1. Surface mulch
The use of surface organic mulch reduces the rate of
water loss from the soil. Mulching with rice straw at
8 Mg haÿ1 over the mungbean rows has been shown to
improve emergence by 17% when sown 17 days after
draining (IRRI, 1988). In the drier regions of the
Philippines a mulch rate of 1.6 Mg haÿ1 increased
yield by 26% (IRRI, 1989). Similarly, in east Java a
surface mulch of 5 Mg haÿ1 rice straw increase yield
by 30% (Adisarwanto, 1985).
Incorporation of organic matter may improve soil in
the long term, but 4 years of organic matter incorporation caused only marginal improvement in topsoil
porosity and in®ltration rate and had no signi®cant
effect on the crop (T. Woodhead, personal communication). This, however, might help to offset further
deterioration in soil structure under intensive ricebased cropping systems (Cass et al., 1994).
3.2.2. Chemical amendments
Calcium amendments, such as gypsum and lime,
have been used successfully to overcome soil physical
problems associated with dispersion of Vertisols (So
and McKenzie, 1984; McKenzie and So, 1989a,b) and
could in¯uence physical properties of clay soils after
drainage. Amendments have also been used in rice
bays in New South Wales, Australia to clear cloudy
7
water by suppressing dispersion (Bacon, 1979). It is
possible that calcium amendments may improve structural development and water relations in drying
puddled soils and may assist establishment of DS
legumes. Gypsum applied to a silty clay loam rice
soil 10 days before draining (20 days before harvest)
resulted in higher seed zone water content over the 30
days after harvest and increased wheat seedling emergence when moisture conditions were sub-optimal
(Zhang, 1990).
The uncertainties surrounding the use of organic
mulch and gypsum or lime as part of soil management
practices and their effect on DS crops after rice
warrants further investigation.
3.3. The effects of tillage
3.3.1. The effect of tillage for the DS crop
The structure of the puddled layer becomes massive
as the soil dries. Puddling also results in the formation
of compacted soil layers below the puddled zone, on
which soil strength increases rapidly as the soil dries
and limits the depth of root exploitation (IRRI, 1986).
The depth of exploitable soil determines the yield of
the crop. For example, mungbean yield has been
shown to be correlated with the depth at which penetrometer resistance increases sharply (IRRI, 1985,
1986). The growth of DS peanuts can also be
adversely affected by the compacted layer and can
be signi®cantly improved by breaking that layer (G.
Wright, ACIAR project 8834, personal communication).
Attempts have been made to overcome these physical constraints using tillage. However, to date the
effects of tillage on yield are unclear. In some experiments tillage caused insigni®cant or no increases in
yield (IRRI, 1986, 1987, 1988, 1989). In other experiments, deep tillage produced signi®cant yield
increases (IRRI, 1988). This uncertainty may be
related to the interaction between tillage and the
length of time tillage was carried out after draining
of the surface water. Tillage during 0±7 DAD may not
be bene®cial because the soil is too wet and would
result in cloddy seedbeds with poor seed-soil contact
(Cook, 1989; Zhang, 1990; Cook et al., 1995). However, when the delay between ®eld drainage and
sowing is increased to improve conditions for tillage,
the yield advantage can become a yield penalty
8
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
because water becomes more limiting and can be lost
faster from tilled soil (IRRI, 1987, 1989; Cook, 1989;
Cook et al., 1995).
The disappointing responses to tillage found in
many experiments may also be because tillage is
not adequately loosening the soil. Results show that
manual loosening of the soil to 1 m using a spade
consistently improved mungbean yield more than
tillage (IRRI, 1987, 1988). The residual effects of
DS deep tillage on increased percolation from the
subsequent rice crop have not been widely investigated, but appear insigni®cant (IRRI, 1984) because
deep cracks developed irrespective of whether tillage
was used or not.
Deep strip tillage, a new technology developed at
IRRI which breaks the compacted layer directly below
the crop rows, can signi®cantly improve soil physical
conditions and root growth of mungbeans (IRRI,
1984, 1985, 1986, 1987; So and Woodhead, 1987;
Woodhead, 1990). However, deep tillage has a high
draft requirement which can be met by hand operated
two-wheel tractors. It requires four-wheel drive tractors or cable winch systems which are generally not
available in southeast Asia (IRRI, 1985, 1986). In
addition, four-wheel tractors can result in greater
compaction. Therefore, this solution does not seem
to be a practical option for the near future.
3.3.2. Effects of puddling intensity on subsequent DS
crops
Wet cultivation or puddling is synonymous with rice
culture in Asia and is used to assist in transplanting of
rice seedlings; to reduce water and nutrient losses and
to control weeds (Sharma and De Datta, 1985). Puddling breaks down and disperses soil aggregates into
individual component particles. The degree of dispersion for a given puddling effort is dependent on the
structural stability of the soil and is likely to affect the
regeneration of soil structure after rice, which, in turn,
will affect the DS crop. The effects of degree of
puddling prior to the rice phase on structure regeneration and growth of a DS crop after rice is related to soil
type. For example, increasing intensity of puddling
resulted in increased maize yields on a Vertisol but
decreased yields in hardsetting, lighter textured Regosols (Trenggono and Willatt, 1988). Similarly, intensive puddling increased DS mungbean yield on a clay
loam but decreased it on a sandy loam (IRRI, 1988).
These differences were attributed to clay content and
mineralogy. The concept of partially controlling soil
structure regeneration after rice through the puddling
treatment prior to the rice phase should be investigated
further by determining which soil types are responsive.
3.4. Seeding techniques
The most commonly used technique for DS
legumes after rice is manual dibbling. However, Cook
et al. (1995) found that dibbling gives variable results,
especially when the soil is wet in the few days after
rice harvest. An inexpensive alternative is manual
furrow seeding, which also gave variable results,
but was better in wet soils. Neither method was
reliable at lower water contents. They also found that
an inverted T seeder (Choudhary, 1985) pulled by a
hand tractor gave better performance in tilled soils
except when very wet or dry.
3.5. Crop/cultivar selection for improved root
performance
A factor in¯uencing the penetration of compacted
subsoils is the pressure that the root system can exert.
Different plants vary in their ability to penetrate
compaction layers. For example, bahia-grass (Paspalum notatum Flugge) penetrated compacted subsoils
better than cotton, which has a taproot, with the result
that cotton grown after bahia-grass yielded better and
extracted more water than cotton after cotton (Elkin
et al., 1977). Similarly, maize after pigeonpea (Cajanus cajan L.) grew better and yielded more than maize
after maize partly because of the superior penetration
by pigeonpea roots (Hulugalle and Lal, 1986).
Roots with the ability to penetrate hard subsoils
would better explore the subsoil provided they are able
to extract available water, which is determined by root
distribution and soil hydraulic characteristics. So and
Jayasekara, (1991) found large differences between 12
cultivars of sorghum (Sorghum bicolor L. Moench) in
their ability to extract water from the subsoil even
when adequate roots were present at depth for all
cultivars. To date, there is only limited information on
the ability of potential DS cultivars species and cultivars to grow through hard soils and to extract soil
water. Such information would assist in selecting
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
crops suitable as DS crops after lowland rice and
possibly in breeding suitable cultivars for that purpose.
9
management technologies that can overcome soil
physical limitations to DS crop production after lowland rice.
3.6. Soil chemical and biological limitations
4.2. The speci®c objectives of the project
Crops grown in lowland soils after rice may suffer
from de®ciencies of plant nutrients and from a lack of
suitable micro-organisms such as rhizobia or VA
mycorrhiza, which may not survive prolonged waterlogged conditions. The availability of residual nutrients from the rice phase is dependent on cultural
practices and soil type. In 1984, IRRI achieved mungbean yields of 2.1 Mg haÿ1 after rice without fertiliser,
inoculum or irrigation and with only 35 mm of DS rain
(IRRI, 1985). However, during a visit to east and
central Java, we saw signi®cant responses of mungbean and peanuts after rice to various combinations of
fertilisers and inoculum. The interaction of phosphorus and zinc has been observed to in¯uence plant
growth in student projects with Vertisols in Indonesia
(S. Setijono, personal communication). Zinc, copper
and boron de®ciencies have been reported for IR 64
rice in some areas of east Java and zinc and copper
applications have increased lowland rice yields
(Suyono, 1990). Therefore, it is possible that these
elements could be de®cient for DS crops as well and
should be evaluated.
3.7. Summary
In summary, it is clear that the limitations to DS
crop growth and yield after lowland rice soils are
complex and still not clearly understood. The need
for solutions to the problems of clay soils after lowland rice received strong endorsement from the 1989
Asian Rice Farming Systems Network workshop in
Bogor which recommended that work on this topic
should be initiated simultaneously in a number of
Asian countries.
4. The project `Management of clay soils for
lowland rice-based cropping systems'
4.1. General objectives
The overall objective of this project was to contribute towards the development of soundly based soil
1. To test soil management and agronomic practices
across a range of soils and climates, that have the
potential to overcome adverse soil physical
conditions for DS crops after rice, including
amendments (calcium or organic matter mulch),
tillage technologies and length of delay periods in
sowing of the DS crop after rice harvest.
2. To evaluate these practices by
2.1. measuring the changes in soil physical
conditions throughout the complete cropping
cycle from rice to DS crop.
2.2. determining the performance of the DS crop
(establishment and growth) and its ability to
extract soil water.
3. To determine the mechanisms involved in soil dispersion due to puddling and the factors controlling
¯occulation and structural reformation as the soil
dries after draining of surface water from rice ®elds.
4.3. The contrasting requirements of rice and DS crops
The project dealt with components of a cropping
system that have vastly different soil structural requirements. The rice phase requires a puddled soil with the
structure largely broken down, whereas the DS crop
requires a soil with good structure to express reasonable
productivity. We recognised that as a result of ameliorative treatments of the soil for the DS crop, detrimental
as well as bene®cial effects to the subsequent rice crop
may follow, e.g. paddy ®elds may become more permeable and leaky; residual N from legumes may be
bene®cial for rice. Therefore, it was important that,
where possible, the changes in physical properties were
monitored throughout the complete cropping cycle.
4.4. Selection of ®eld experimental sites
It was intended that the information and technology
derived from this project should be transferred readily
across a range of soils and climates. Therefore, the
project was designed as a series of benchmark sites
with common treatments, a common DS crop species
10
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
Table 2
Description of the selected ®eld sites in Indonesia and the Philippines with the main relevant soil characteristics
Country
Site
Soil texture
Clay (g kgÿ1)
ÿ1
Swelling clay (%)
Linear shrinkage
Indonesia
Ngale, East Java
Jambegede, East Java
Maros, Sulawesi
Heavy clay
Silty clay loam
Silty clay loam
740 g kg
450 g kgÿ1
460 g kgÿ1
73
15
9
0.19
n.a.a
0.05
Philippines
San ildefonso, Bulacan
Manaoag, Pangasinan
Heavy clay
Silty clay
410 g kgÿ1
520 g kgÿ1
26
33
0.07
0.10
a
n.a.: not available.
(mungbean) and common methodologies covering a
range of sites with different soils and climates. Treatments and species which are speci®c to particular
areas or soils are included at relevant sites.
Table 2 shows the textures of the sites in Indonesia
at Ngale (deep Vertisol) and Jambegede (silty clay
loam) in east Java, where soybean and peanuts are
included alongside mungbean; and a site near Maros
in south Sulawesi (silty clay). The sites in the Philippines are the research station at San Ildefonso,
Bulacan (Vertisol) and a farmer's ®eld near Manaoag,
Pangasinan (silty clay). The sites cover a wide range of
clay contents and clay mineralogies (Ringrose-Voase
et al., 1995). The Ngale site has the greatest clay content
with 740 g kgÿ1 total clay and 73% (on whole soil basis)
swelling clay (smectite and vermiculite), giving it the
greatest shrink/swell potential (0.19 linear shrinkage).
The site at San Ildefonso has 410 g kgÿ1 total clay and
26% swelling clay and intermediate shrink/swell potential (0.07 linear shrinkage). It also has a signi®cant sand
content of 330 g kgÿ1. The Maros and Manaoag sites
have similar particle size distributions with 460 g kgÿ1
and 520 g kgÿ1 total clay, respectively. However, they
have different mineralogies with 9% and 33% swelling
clays, respectively, giving them different shrink/swell
potentials (0.05 and 0.10 linear shrinkage, respectively).
Detailed descriptions of these soil and climates are
given by Schafer and Kirchhof (2000).
5. Experimental methodology
5.1. Field experiments in Indonesia and the
Philippines
At each site in Indonesia and the Philippines, three
®eld experiments were conducted to investigate the
effect of potentially useful management practices on
soil conditions and the resulting growth of DS crops.
Since a complete factorial experiment involving all
the factors of interest will result in a very large experiment which will be dif®cult to manage, treatments
were selected that are relevant to the farmers interest
and incorporated into three ®eld experiments that are
of manageble size. Experiments were also designed to
suit the local expertise and available facilities.
The ®rst two experiments were intended to reduce
the need for large numbers of treatment combinations
by separating treatments applied during the rice phase
(E1) and treatments applied during the legume phase
(E2). A third experiment (E3) measured the dynamics
of changes in soil properties in the period immediately
following ®eld drainage for rice harvest at four of the
®ve sites. The information gained from the latter will
help extrapolate the results from the ®rst two experiments to other soils and climates.
5.1.1. Experiment E1
The objective was to investigate the effects of
degree of puddling on soil physical conditions for
the subsequent DS crop. Previous work has indicated
that increased puddling may increase yields of the
following DS crop on heavy clay soils but decrease
yield on lighter textured soils (Trenggono and Willatt,
1988). If this is correct, puddling intensity can be used
as an inexpensive and readily adoptable practice on
suitable soils. The puddling intensity treatments
imposed during soil preparation for the rice crop were
1. dry cultivation prior to submergence,
2. one wet ploughing and harrowing using draught
animal power,
3. two wet ploughings and harrowings using draught
animal power,
4. two wet cultivations with mechanised rototiller or
hydrotiller.
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
The treatments were combined with two tillage and
sowing treatments (zero-till and dibbled, ZTD versus
ploughing, broadcasting and harrowing Ð PBH) for
the DS crops in the Philippines and two drainage
treatments (without and with surface drainage) in
Indonesia. PBH is the common farmer practice in
the Philippines, and ZTD with surface drainage are
common on heavy clay soils in Indonesia. In this
project, ZTD without surface drainage was the common treatment between the two countries. In both
countries, the most common farmer practice of soil
preparation for rice is two wet ploughings and harrowings using draught animals.
Treatments were replicated four times in a split-plot
design, with puddling intensity as the main plots and
tillage or drainage treatments as sub-plots. The experiment was carried out at each site over a period of 3
years, i.e. three rice-DS crop cycles.
5.1.2. Experiments E2
This experiment investigated the effects of DS crop
management practices on soil properties and crop
performance. To allow comparisons with experiment
E1, soil preparation for the rice phase was done using
two wet ploughings and harrowings (equivalent to
treatment 2 in E1).
Treatments included combinations of amendments
(none, A0; gypsum, AG; and organic matter mulch,
AOM); cultivation (zero-till, C0 and rotovator, C1)
and sowing delay after rice harvest (no delay, D0; one
week delay, D1 and two weeks delay, D2) as follows:
T1
C0
A0
D1
T2
T3
T4
T5
T6
T7
T8
T9
C0
C0
C0
C0
C0
C1
C1
PBH
A0
AG
AOM
A0
A0
A0
A0
A0
D1
D1
D1
D0
D2
D1
D2
D1
No fertiliser (farmers'
practice in Indonesia)
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
No fertiliser (farmers'
practice in the Philippines)
The DS crop was sown by dibbling for all treatments except T9. Treatments were replicated four
times in a randomised block design and the experi-
11
ment was conducted at each site over a period of three
consecutive crop cycles.
Please note that
1. treatments T1 to T8 were used on all sites and
treatment T9 only in the Philippines,
2. treatments T1 and T2 provide a measure of the
nutritional limitations of the soil,
3. treatments T2 to T4 provide a measure of the
effects of gypsum and organic mulch,
4. treatments T2 and T5 to T8 provide a measure of
the effects of cultivation and sowing delay and
their possible interactions.
It should be stressed that all treatments selected had
already been shown as potentially useful for DS crops
after rice in past trials, on speci®c soils and under a
narrow range of climatic conditions. This project
compared these potentially useful treatments on a
wider range of soils and climatic conditions and
attempts to quantify the agronomically relevant
changes in the soil conditions resulting from these
treatments, to enable extrapolation to other sets of
environments.
5.1.3. Experiment E3
A third short-term ®eld experiment was conducted
at one or two sites each season. This experiment
monitors changes in soil mechanical and hydrological
properties and soil structure with time as the soil dries
after draining surface water prior to rice harvest. The
information obtained for each soil will provide functional relationships required to interpret the conditions
encountered at planting and during early seedling
growth in the ®rst two experiments which involve
only three sowing delays. These experiments were
expected to last 2±4 weeks each.
This experiment was carried out on soils with
standard puddling treatments (two ploughings and
two harrowings). The following properties were measured at pre-determined intervals: soil water content,
soil strength (shear and penetrometer resistance),
macropore structure development (using crack measurements and resin impregnated samples).
5.1.4. Selection of dry season crop species
In experiments E1 and E2, mungbean was the DS
crop common to all sites. Mungbean is the major DS
legume crop in the Philippines and the third major
12
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
legume in Indonesia. A second crop of soybean was
included in east Java with soybean on the Vertisol
(Ngale) and peanuts on the lighter textured soil (Jambegede). Thus the crop rotations are as follows:
Ngale
Jambegede
Maros, San
Ildefonso, Manaoag
Rice±mungbean/soybean±
mungbean/soybean
Rice±mungbean/peanut±
mungbean/peanut
Rice±mungbean
5.1.5. Measurements
This paper provides an overview of the measurements undertaken in this project. However, detailed
descriptions of the methodologies will be found in the
relevant sections of this special issue.
5.1.5.1. Initial characterisation of the site. The
uniformity of each site was investigated using a 20±
25 m grid system. Morphological descriptions were
made to determine any gradation in soil characteristics
and to avoid any unrepresentative areas. Composite
samples from sections of the field at different depths
were analysed to provide a baseline data-set prior to
imposition of treatments. These were
soil texture, bulk density, cation exchange capacity,
cations, pH, electrical conductivity, soil organic
carbon,
soil water characteristics,
plastic and liquid limits,
soil structural stability (wet sieving and dispersibility).
5.1.5.2. Measurements during the rice phase. The
following were measured for some sites at the start,
middle and end of the rice phase:
sinkage capacity,
infiltration rates,
dispersibilty of the soil,
yield of rice on all sites.
phases of the crop: emergence, vegetative, flowering,
pod formation and maturity. Measurements included
soil physical measurements;
soil bulk density profiles,
soil strength profiles Ð penetrometers,
soil water content profiles and water use by the
crop,
root distribution/root length densities (at flowering only),
plant measurements;
emergence as a measure of establishment,
plant density at harvest as a measure of survival,
plant biomass and its components,
yield and yield components,
climatic measurements(local weather station data);
rainfall,
evaporation (E pan),
temperature,
radiation.
5.1.6. Simulation of soil puddling and drying in the
laboratory
Laboratory experiments were conducted at the University of Queensland to measure the degree of dispersion after the soil is subjected to a range of standard
puddling treatments similar to those in the ®eld
experiments (Kirchhof et al., 2000b). Degree of dispersion was adopted as a measure of the degree of
puddling and the decrease in structural stability. Since
one major objective of puddling is to reduce percolation rate, soil hydraulic conductivities were measured
for each puddling treatment and related to the degree
of dispersion. The effects of repeated wetting and
drying on soil structural development were also investigated. Soil changes occurring during the rice-DS
crop sequence were observed and quanti®ed in large
lysimeters (approximately 1.2 m1 m1 m) in which
rice was grown under puddled conditions followed by
a DS crop. Soils used covered a similar range of
textures to those used in the ®eld experiments. Details
are described by Kirchhof and So (1994, 1996).
6. Summary
5.1.5.3. Measurements during the dry season crop
phase. Soil measurements were made at the same time
as plant measurements at the appropriate phenological
In summary, this paper presented the background
and an overview of this international collaborative
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
project in Indonesia and the Philippines. Details of the
soils and climate of the ®ve experimental sites are
described by Schafer and Kirchhof (2000). The results
and implications of the E1 experiments are described
by Kirchhof et al. (2000a,b) and those for the E2
experiments by Kirchhof et al. (2000a). Results on the
changes in soil physical properties in experiment E3 is
discussed by Ringrose-Voase et al. (2000) and crop
establishment aspects are discussed by Rahmianna
et al. (2000). The paper by Cabangon and Tuong
(2000) was not part of the collaborative project, but
is relevant as it deals with the in¯uence of cracks on
soil preparation for rice and how shallow tillage can
result in early sowing of rainfed lowland rice, an
important factor during low rainfall seasons.
Acknowledgements
The project was funded by the Australian Center for
International Agricultural Research and their contribution and support is gratefully acknowledged.
References
Adachi, K., 1990. Effects of rice-soil puddling on water
percolation. In: Transactions of the 14th International Congress
of Soil Science, Kyoto, Vol. 1, 146±151.
Adisarwanto, T., 1985. The in¯uence of planting methods and
mulching on soybean seed yield. In: Shanmugasundaram, S.,
Sulzberger, E.W. (Eds.), Soybean in Tropical and Subtropical
Cropping Systems. Asian Vegetable Research and Development Centre, Taiwan, pp. 215±217.
Adisarwanto, T., Santoso, B., Manshuri, G., Floyd, C., 1989.
Agronomic Survey of MK1 Soybean on Sawah in East Java
(Jember to Madiun), 21±24 June 1989. Malang Research
Institute for Food Crops, East Java, Indonesia.
Adriano, M.S., Cabezon, V.E., 1987. Economic policies affecting
crop diversi®cation in the Philippines. In: Irrigation Management for Diversi®ed Cropping. International Irrigation Management Institute, Sri Lanka.
Africa, T.P., Marquez, N.R., 1989. Philippine yearbook. National
Statistics Of®ce, Manilla, Republic of the Philippines.
Agricultural Intensi®cation Program, 1988/89. Special province of
Yogyakarta (translated from: Keputusan Gubernur Kepala
Daerah Istimewa Yokyakarta No 165/KPTS/1988 tentang
Program Intensi®kasi Pertanian Tahun 1988/1989, Propinsi D
I Y, BIMAS Propinsi D I Y)
Bacon, P.E., 1979. Ef®ciency of different calcium application
methods in controlling muddy water. In: Proceedings of the
13
Australian Society of Soil Science Inc. Conference, Yanco,
NSW, Australia, pp. 116±121.
Cabahug, L.V., 1990. In praise of the marvellous mungbean. Food
Legumes Ð Coarse Grains Newsletter 12, 13±14.
Cabangon, R.J., Tuong, T.P., 2000. Management of cracked soils
for water saving during land preparation for rice cultivation.
Soil Tillage Res. 56, 105±116.
Cass, A., Gusli, S., MacLeod, D.A., 1994. Sustainability of soil
structure quality in rice paddy-soyabean cropping systems in
South Sulawesi, Indonesia. Soil Tillage Res. 31, 339±352.
Choudhary, M.A., 1985. Seeding techniques and machinary. In:
Soil Physics and Rice. International Rice Research Institute,
Los BanÄos, The Philippines, pp. 411±412.
Cook, S.M.F., 1989. In¯uence of soil physical properties on
mungbean emergence in rice soils. M.Sc. thesis. University of
Minnesota, USA.
Cook, S.M.F., Gupta, S., Woodhead, T., Larson, W.E., 1995. Soil
physical constraints to establishment of mungbeans (Vigna
radiata L. Wilczek) in paddy rice (Oriza sativa L.) soils. Soil
Tillage Res. 33, 47±64.
Elkin, C.B., Haaland, R.L., Hoveland, C.S., 1977. Grass roots as a
tool for penetrating soil hardpans and increasing crop yields. In:
Proceedings of 34th Southern Pasture and Forage Crop
Improvement Conference, Auburn University, AL, USA,
pp. 21±26.
FAO Yearbook, 1984. Vol 38. Food and Agricultural Organisation
Statistics Series No 63.
Fy®eld, T.P., 1987. Effects of temperature, water and cultivation on
the establishment of a mungbean crop following paddy rice.
Ph.D. thesis, University of Reading, UK.
Hoque, M.Z., 1984. Cropping Systems in Asia: On-Farm Research
and Management. International Rice Research Institute, Los
BanÄos, Philippines.
Huke, R.E., 1982. Rice Area by Type of Culture: South, Southeast,
and East Asia. International Rice Research Institute, Los
BanÄos, Philippines, 32 pp.
Hulugalle, N.R., Lal, R., 1986. Root growth of maize in a
compacted gravelly tropical Al®sols as affected by rotation
with a woody perennial. Field Crops Res. 13, 33±44.
IRRI, 1984. Annual Report for 1983. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1985. Annual Report for 1984. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1986. Annual Report for 1985. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1987. Annual Report for 1986. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1988. Annual Report for 1987. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1989. Annual Report for 1988. International Rice Research
Institute, Los BanÄos, Philippines.
Kirchhof, G., So, H.B., 1996. The effect of puddling intensity and
compaction on properties, rice and mungbean growth: a mini
rice-bed study. In: Kirchoff, G., So, H.B. (Eds.), Management
of Clay soils for Lowland Rice Based Cropping Systems,
ACIAR Proceedings No 70, 1996, pp. 51±70.
Kirchhof, G., So, H.B., 1994. Soil puddling for rice production and
14
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
its effect on soil structure and water percolation. In: Proceedings of International Symposium on Sealing, Crusting and
Hardsetting Soils: Productivity and Conservation, February
1994, The University of Queensland, Brisbane, Qld, Australia,
pp. 7±11.
Kirchhof, G., So, H.B., Adisarwanto, T., Utomo, W.H., Priyono, S.,
Prastowo, B., Basir, M., Lando, T.M., Subandi, Dacanay, E.V.,
Tan-Elicano, D., Sanidad, W.D., 2000a. Growth and yield response of grain legumes to different soil management practices
after rainfed lowland rice. Soil and Tillage Res. 56, 51±66.
Kirchhof, G., Priyono, S., Utomo, W.H., Adisarwanto, T., Dacanay,
E.V., So, H.B., 2000b. The effect of soil puddling on the soil
physical properties and the growth of rice and post-rice crops.
Soil and Tillage Res. 56, 37±50.
Lavapiez, L.M., Duran, B., Nicolas, J., Price, E.C., 1977. Rice-mung
cropping patterns in Manaoag, Pangasinan: a costs and returns
study. Internal Report. Agricultural Economics Department,
International Rice Research Institute, Los BanÄos, Philippines.
Maghari, S.de G., 1990. Mungbean Emergence and Growth as
in¯uenced by Soil Submergence in Previous Rice Crop. M.Sc.
Thesis, University of Minnesota, USA.
Mahata, K.R., Sen, H.S., Pradhan, S.K., Mandal, L.N., 1990. Notill and dry ploughing compared with puddling for wet season
rice on alluvial sandy clay loam in eastern India. J. Agric. Sci.
114, 79±86.
Maranan, C.G., 1986. Agricultural economic research at Guimba
cropping systems site, Nueva Ecija, Philippines. Internal report.
Agricultural Economics Department, International Rice Research Institute, Los BanÄos, Philippines.
Maranan, C.G., 1987. Economic performance of cropping patterns
at Bantug, Guimba, Nueva Ecija CSR site, 1986±87. Internal
report. Agricultural Economics Department, International Rice
Research Institute, Los BanÄos, Philippines.
McKenzie, D.C., So, H.B., 1989a. The effect of gypsum on
Vertisols of the Gwydir Valley, NSW. I. Soil properties and
wheat growth. Aust. J. Exp. Agric. 29, 51±62.
McKenzie, D.C., So, H.B., 1989b. The effect of gypsum on
Vertisols of the Gwydir Valley, NSW. II. Ease of tillage. Aust. J.
Exp. Agric. 29, 63±67.
Nanseki, T., Morooka, Y., Zakaria, A.K., 1989. Comparative
advantage analysis of soybean in an upland area of West Java: a
case study of mathematical programming approach. Palawija
News 6, (2). CGPRT Newsletter.
Pasaribu, D., McIntosh, J.L., 1985. Increasing tropical soybean
production with improved cropping systems and management.
In: Shanmugasundaram, S., Sulzberger, E.W. (Eds.), Soybean
in Tropical and Subtropical Cropping Systems. Asian Vegetable
Research and Development Centre, Taiwan, pp. 1±11.
Rahmianna, A.A., Adisarwanto, T., Kirchhof, G., So, H.B., 2000.
Crop establishment of legumes in rainfed lowland rice-based
cropping systems. Soil and Tillage Res. 56, 67±82.
Ringrose-Voase, A.J., Hutka, J., Beatty, J., Raven, M.D., Rath, H.,
1995. Analysis of some Indonesian and Philippine soils for
ACIAR Project 8938 on Management of clay soils for lowland
rice-based cropping systems. Division of Soils Tech
Management of clay soils for rainfed lowland rice-based
cropping systems: an overview
H.B. Soa,*, A.J. Ringrose-Voaseb
a
School of Land and Food Sciences, The University of Queensland, St Lucia, Brisbane, Qld 4072, Australia
b
CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia
Abstract
The problem of concern in this project is that in the dry season following a lowland rice (Oryza sativa L.) crop, yields of
post-rice crops are generally low, despite adequate water commonly being available in the soil pro®le to grow a potentially
high yielding dry season (DS) crop without irrigation. Maize (Zea mays L.) yields are as low as 1 Mg haÿ1 or less, soybean
(Glycine max L. Merr.) and cowpea (Vigna unguiculata L.) at 0.3 to 0.8 Mg haÿ1 in Indonesia and mungbean (Vigna radiata
(L.) Wilzek) around 0.5 Mg haÿ1 in the Philippines. These are all very much below the yield potential of these soils. For
example, mungbean yields of 2.2 Mg haÿ1 have been achieved by IRRI in the Philippines on these soils without irrigation or
additional fertilisers. The causes of low yields of DS crops after rice are mainly poor crop establishment and poor root growth
due to soil physical constraints. These result from the breakdown of soil structure during wet cultivation (puddling) for rice.
Yields are also limited by biological and chemical constraints. As a result of these low yields, farmers are reluctant to invest in
post-rice crops. Therefore, land after lowland rice (at least 51 million ha in Asia according to Huke [Huke, R.E., 1982. Rice
Area by Type of Culture: South, Southeast, and East Asia. International Rice Research Institute, Los BanÄos, Philippines, 32
pp.]) represents an underutilised resource that can be used to meet the food requirement of the ever increasing population of
the developing world. To increase the utilisation of these soils, improved management practices are required to enable dry
season crops to use the stored water in the soil pro®le after the rice crop. This paper outlines a project which was established
with the general objective of developing soundly based soil management technologies that can overcome soil physical
limitations to DS crop production after lowland rice. The speci®c objectives of the program were
1. to test a range of soil management and agronomic practices that have the potential to overcome adverse soil physical
conditions for DS crops after rice, across a range of soil and climates;
2. to evaluate these practices by
2.1. measuring the changes in soil physical conditions throughout the complete cropping cycle from rice to DS crops;
2.2. determining the performance of the DS crop (establishment and growth) and its ability to extract soil water.
3. to determine the mechanisms involved in dispersion due to puddling and in ¯occulation and structural development as the
soil dries after draining surface water from rice ®elds.
Relevant outcomes from this project are described in the following papers in this issue. # 2000 Elsevier Science B.V. All
rights reserved.
Keywords: Soil management; Puddling; Rice; Legumes; Rainfed lowland
*
Corresponding author. Tel.: 61-7-3365-2888; fax: 61-7-3365-1177.
E-mail address: [email protected] (H.B. So).
0167-1987/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 7 - 1 9 8 7 ( 0 0 ) 0 0 1 1 9 - 7
4
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
1. Introduction
To keep pace with rapidly expanding populations,
the production of food legumes and other dry season
(DS) crops must be increased within the lowland rice
growing areas in Indonesia and the Philippines as well
as in other countries of southeast Asia. There are
51 million ha of lowland rice in Asia including 8.2
and 3.5 million ha of lowland rice in Indonesia and the
Philippines of which 3 and 2 million ha, respectively,
are rainfed (Huke, 1982). A feature of lowland rice
culture is that the amount of soil water remaining in
the dry season after the rice crop is usually adequate
for a DS crop. However, when DS crops are grown in
these soils, yields are generally low. For example,
yields are as low as 1 Mg haÿ1 or less for maize, 0.3±
0.8 Mg haÿ1 for soybeans and cowpeas in Indonesia
(Hoque, 1984) and 0.5 Mg haÿ1 for mungbeans is not
uncommon in the Philippines. These yields do not
provide adequate returns to the farmer, so that lowland
rice soils, in particular rainfed lowland rice soils,
represent an underutilised resource during the dry
season. The potential yield for mungbean in the
Philippines is approximately 2.2 Mg haÿ1 (So and
Woodhead, 1987). Therefore, increasing yields of
DS crops would increase the utilisation of land and
residual soil water during the period between rice
crops.
Despite the lower yields of DS crops compared to
rice, Table 1 shows that the lower costs and higher
prices of some DS crops, in particular mungbean, can
result in greater net returns from these crops than from
the rice crop if moderate yields can be obtained
(Maranan, 1986, 1987). Considerable bene®ts could
be expected from growing DS crops, including
increased farmer income and nutrition and a reduction
in imports of food legumes. In 1987, imports of maize
and soybean cost Indonesia about $25 million and $63
million, respectively and the Philippines about $7
million and $2.5 million. Peanut (Arachis hypogaea
L.) imports cost $22 million and $7.5 million to
Indonesia and the Philippines, respectively. In addition, the introduction of food legumes into the rice
rotation could result in substantial savings in nitrogenous fertilisers. These bene®ts would also be applicable to lowland rice areas in other southeast Asian
countries.
Both the Indonesian and Philippine governments
place high priority on raising yields of DS crops,
particularly legume crops, as a means of increasing
farmers' income as well as nutrition. The Indonesian
government expressed this through its Five Year Plans
(PELITAs), of which the Sixth Plan is current. The
Philippine Council for Agriculture, Forestry and Natural Resources Research and Development
(PCARRD) has a Mungbean Development Action
Plan to co-ordinate efforts to increase production of
mungbeans.
The causes of low yields of DS crops after rice are
often poor crop establishment and inferior root growth
due to adverse physical conditions of the soil which, in
turn, are caused by the wet cultivation (puddling)
undertaken for paddy rice (Pasaribu and McIntosh,
1985; So and Woodhead, 1987; Adisarwanto et al.,
1989). Yields are also limited by nutritional and
biological constraints.
Project 8938 was funded by the Australian Centre
for International Agricultural Research (ACIAR)
titled `The Management of Clay Soils for Lowland
Rice-based Cropping Systems' aimed to investigate
the factors that affect the success or failure of DS crops
grown after rice and to make a contribution towards
Table 1
National average yield, actual price, relative pro®tability and the ratio of returns/costs for several crops in the Philippines, 1985a
Crops
Yield
(Mg haÿ1)
Price
(Pesos/kg)
Total return
(Pesos/ha)
Cost of production
(Pesos/ha)
Net return
(Pesos/ha)
Net return/cost
(%)
Rice
Maize
Soybean
Mungbean
Peanut
2.40
1.04
0.99
0.69
0.85
3.24
2.80
7.30
15.40
10.10
7776
2912
7227
10626
8585
5370
2078
3697
3780
6959
2406
834
3530
6846
1626
44.8
40.1
95.5
181.1
23.4
a
Adriano and Cabezon, 1987.
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
the development of stable cropping systems that
incorporate DS crops within rainfed, lowland ricebased cropping systems. It was a synthesis of two
proposals, one from the University of Queensland and
another from CSIRO and the Philippine Bureau of Soil
and Water Management(BSWM) and involved the
University of Brawijaya, Indonesia and two Indonesian Institutes for Food Crops. The factors investigated
are tillage during the preparation for the rice and DS
crop phases, soil amendments, time of sowing the DS
crop, surface drainage and fertilisers.
This paper outlines the background to the project, a
review of the relevant literature and a description of
the experimental design and set-up of the project.
2. Background to the project
Rice in southeast Asian countries is mostly grown
under lowland conditions with 1±3 crops a year
depending on the availability of irrigation water and
the use of modern, short season varieties. After longterm submergence for lowland rice, soil water is
suf®cient to grow a DS crop with reasonable yield
potential. However, under current management practices the yields of DS crops (see above) are generally
well below the yield potential (Pasaribu and McIntosh,
1985; Adisarwanto et al., 1989). The area of lowland
rice is approximately 51 million ha in Asia (Huke,
1982). The untapped potential for food production
from DS crops is a large, underutilised resource.
Furthermore, in Africa there are approximately
100 million ha of land that could potentially be
adapted to rainfed lowland rice with appropriate soil
physical management (Woodhead, 1990).
The realisation that multiple cropping programs are
essential to raise production from rice based systems
lead to the formation of the Asian Cropping Systems
Network. This network coordinates efforts by the
International Rice Research Institute (IRRI) and the
various national programs to jointly develop appropriate rice-based cropping systems in major rice growing environments. (Hoque, 1984). As soil factors are
known to limit yield of DS crops, appropriate soil
management is an essential part of improved cropping
systems. For these systems to be developed, the
dynamics of soil-crop interactions in DS paddy soils
must be better understood.
5
2.1. The importance of legumes in rice-based
cropping systems
Indonesia has been self-suf®cient in rice since 1985
through the success of the government coordinated
BIMAS (mass guidance) and INMAS (mass intensi®cation) programs during past Five Year Plans or
PELITAs. From 1984, PELITA IV gave special attention to the ®rst DS crop (®rst secondary crop within a
lowland rice±DS crop±DS crop cropping system),
with particular emphasis on legumes (Nanseki et al.,
1989). These crops were targeted for increased production with the aim of improving farmers' income
and nutritional status (Vademecum BIMAS, 1987).
The target, in irrigated lowland areas, is to replace the
third rice crop with a DS crop and, in rainfed lowland
areas, is to grow a DS crop before or after the rice
crop. Based on the rate of consumption and imports,
the major DS crops in Indonesia are, in decreasing order,
maize, soybeans, peanuts and mungbeans (FAO, 1984).
The success of the BIMAS and INMAS programs is
partly due to the setting of realistic production targets,
which are negotiated for each province, county and
village which elected to join the program. These
production targets, when agreed to by the parties
concerned, become contracts that must be adhered
to (Agricultural Intensi®cation Program, 1988±1989)
and involve a minimum mandatory set of technology
packages (recipes) that must be carried out. If the
recipe is adhered to, a minimum and achievable
improved yield level is guaranteed. However, these
technology packages do not include soil physical
management recommendations due to a lack of
knowledge in this area. Where adequate irrigation
water is available, improved technology for soybean
has recently been launched through the government
extension program `Supra-insus' (special program for
intensi®cation) with the aim of raising soybean yield
from 1 to 1.5 Mg haÿ1 (Sumarno, 1990). As yet,
satisfactory packages have not been developed for
DS crops after rice.
In the Philippines, the major DS crops are, in
decreasing order, maize (Africa and Marquez,
1989), mungbean and soybean. PCARRD has a Mungbean Development Action Plan aimed at introducing
and studying the impact of new technologies for
mungbeans, in particular new varieties which are
shorter, faster maturing and higher yielding (Cabahug,
6
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
1990). Faster growing varieties are particularly suitable for growing after rice, when the legume is largely
dependent on stored soil water. Mungbean is the major
legume crop in the Philippines partly because, with a
protein content of 20±25%, it is a popular and inexpensive source of protein, often being referred to as
`the poor man's meat' (Cabahug, 1990). Its price is
relatively high and stable and farmer consumption
tends to compensate for any over-production, because
unlike soybean, it does not require processing. In
addition, Table 1 shows that it can be more pro®table
than other legumes or rice, if moderate yields can be
obtained.
Modest increases in mungbean yield in a ricelegume rotation can result in net returns from mungbean being greater than that from the rice component
(Lavapiez et al., 1977; Maranan, 1986, 1987). The
pro®tability of mungbean in Indonesia is also cited as
a major incentive towards their use after rice. However, under current management practices yields of DS
crops are poor and result in a reluctance to invest
management and resources in DS crops so that much
land is underutilised after lowland rice (Varade, 1990).
In addition, there is a social preference for rice.
Therefore, the introduction of management systems
that can stabilise yields of DS crops, particularly
legumes, after rice will have considerable socio-economic bene®ts.
2.2. Physical limitations of puddled soil
The physical limitations imposed by puddled soil
have been recognised as the major cause of poor
establishment and yield of post-rice crops in Asia,
including soybeans in east Java (Adisarwanto et al.,
1989) and mungbeans in the Philippines and other
Asian countries (IRRI, 1984; So and Woodhead, 1987;
Mahata et al., 1990; Varade, 1990). Puddling is associated with the breakdown of soil aggregates during
wet cultivation (Sharma and De Datta, 1985; Adachi,
1990) and results in a massive structure after rice.
After drainage of the surface water prior to rice
harvest, the water content of the surface soil decreases
which is accompanied by a rapid increase in redox
potential (IRRI, 1987; Maghari, 1990) and soil
strength (IRRI, 1985, 1986, 1987, 1988). Puddling
also creates a compacted layer below the puddled
layer, which increases in strength during drying (IRRI,
1986). The effect of seasonal conditions and soil type
on the germination, establishment and root growth of
DS crops after rice is determined by the interactions
between the rates of change of redox potential, soil
strength and available water as the soil dries. To devise
ways of overcoming these limitations, it is important
to quantify the nature of these interactions through a
program of detailed monitoring of the soil physical
conditions.
3. Factors affecting the establishment and growth
of DS crops after lowland rice
3.1. Effect of delay between ®eld drainage and
sowing on germination and establishment of post-rice
DS crops
Successful crop establishment is essential for high
yields, for example, yield of DS mungbean was linearly related to plant population density up to 0.55±0.6
million plants/ha (So and Woodhead, 1987; IRRI,
1988). A key factor determining the success of crop
establishment is the rate of germination. Rapid germination, which depends largely on soil water content
and seed-soil contact, is essential to minimise risks
from adverse factors (So and Woodhead, 1987).
The delay between ®eld drainage and sowing has a
major in¯uence on crop establishment because it
affects soil water content and hence germination
and emergence. Although mungbeans under controlled conditions can germinate at soil water potentials as low as ÿ2.2 MPa (below wilting point), poor
seed-soil contact under ®eld conditions reduces germination rates at low potentials and radicle elongation
rate is reduced at potentials below ÿ0.2 MPa (Fy®eld,
1987; IRRI, 1988). Emergence is slower and falls
below 50% when water potential is reduced below
ÿ0.1 MPa (IRRI, 1986).
Reports on the appropriate period of delay vary and
probably re¯ect differences in climatic conditions
during experiments and in soil type. Under conditions
of little rain after drainage of surface water, it appears
that emergence on silty clay loams is highest when
mungbeans are sown 6±10 days after drainage (DAD)
(Fy®eld, 1987; IRRI, 1987, 1988; Cook, 1989; Cook
et al., 1995). Later sowing tends to reduce emergence,
growth and yield of mungbeans, apparently because of
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
low water potentials and increased seedbed and subsoil strength. On the other hand, growth and yield can
also be reduced after sowing at delays of 0±3 DAD
apparently due to low redox potentials and poor
aeration (IRRI, 1987). It is not clear how these periods
would vary with soil types.
Since rice is generally harvested 7±10 DAD, it is
important that sowing of DS crops be carried out as
soon as possible after harvest. However, in regions
where the probability of rainfall after rice harvest is
high, farmers tend to avoid waterlogging by either
postponing sowing or by providing surface drainage.
Relay cropping, where legumes are sown soon after
draining and before rice is harvested, has been tried as
a means of reducing the sowing delay. However, this
method tends to reduce establishment and yield, as
well as increase problems of weeds and ratooning of
rice (IRRI, 1987, 1989).
3.2. The effect of soil amendments
3.2.1. Surface mulch
The use of surface organic mulch reduces the rate of
water loss from the soil. Mulching with rice straw at
8 Mg haÿ1 over the mungbean rows has been shown to
improve emergence by 17% when sown 17 days after
draining (IRRI, 1988). In the drier regions of the
Philippines a mulch rate of 1.6 Mg haÿ1 increased
yield by 26% (IRRI, 1989). Similarly, in east Java a
surface mulch of 5 Mg haÿ1 rice straw increase yield
by 30% (Adisarwanto, 1985).
Incorporation of organic matter may improve soil in
the long term, but 4 years of organic matter incorporation caused only marginal improvement in topsoil
porosity and in®ltration rate and had no signi®cant
effect on the crop (T. Woodhead, personal communication). This, however, might help to offset further
deterioration in soil structure under intensive ricebased cropping systems (Cass et al., 1994).
3.2.2. Chemical amendments
Calcium amendments, such as gypsum and lime,
have been used successfully to overcome soil physical
problems associated with dispersion of Vertisols (So
and McKenzie, 1984; McKenzie and So, 1989a,b) and
could in¯uence physical properties of clay soils after
drainage. Amendments have also been used in rice
bays in New South Wales, Australia to clear cloudy
7
water by suppressing dispersion (Bacon, 1979). It is
possible that calcium amendments may improve structural development and water relations in drying
puddled soils and may assist establishment of DS
legumes. Gypsum applied to a silty clay loam rice
soil 10 days before draining (20 days before harvest)
resulted in higher seed zone water content over the 30
days after harvest and increased wheat seedling emergence when moisture conditions were sub-optimal
(Zhang, 1990).
The uncertainties surrounding the use of organic
mulch and gypsum or lime as part of soil management
practices and their effect on DS crops after rice
warrants further investigation.
3.3. The effects of tillage
3.3.1. The effect of tillage for the DS crop
The structure of the puddled layer becomes massive
as the soil dries. Puddling also results in the formation
of compacted soil layers below the puddled zone, on
which soil strength increases rapidly as the soil dries
and limits the depth of root exploitation (IRRI, 1986).
The depth of exploitable soil determines the yield of
the crop. For example, mungbean yield has been
shown to be correlated with the depth at which penetrometer resistance increases sharply (IRRI, 1985,
1986). The growth of DS peanuts can also be
adversely affected by the compacted layer and can
be signi®cantly improved by breaking that layer (G.
Wright, ACIAR project 8834, personal communication).
Attempts have been made to overcome these physical constraints using tillage. However, to date the
effects of tillage on yield are unclear. In some experiments tillage caused insigni®cant or no increases in
yield (IRRI, 1986, 1987, 1988, 1989). In other experiments, deep tillage produced signi®cant yield
increases (IRRI, 1988). This uncertainty may be
related to the interaction between tillage and the
length of time tillage was carried out after draining
of the surface water. Tillage during 0±7 DAD may not
be bene®cial because the soil is too wet and would
result in cloddy seedbeds with poor seed-soil contact
(Cook, 1989; Zhang, 1990; Cook et al., 1995). However, when the delay between ®eld drainage and
sowing is increased to improve conditions for tillage,
the yield advantage can become a yield penalty
8
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
because water becomes more limiting and can be lost
faster from tilled soil (IRRI, 1987, 1989; Cook, 1989;
Cook et al., 1995).
The disappointing responses to tillage found in
many experiments may also be because tillage is
not adequately loosening the soil. Results show that
manual loosening of the soil to 1 m using a spade
consistently improved mungbean yield more than
tillage (IRRI, 1987, 1988). The residual effects of
DS deep tillage on increased percolation from the
subsequent rice crop have not been widely investigated, but appear insigni®cant (IRRI, 1984) because
deep cracks developed irrespective of whether tillage
was used or not.
Deep strip tillage, a new technology developed at
IRRI which breaks the compacted layer directly below
the crop rows, can signi®cantly improve soil physical
conditions and root growth of mungbeans (IRRI,
1984, 1985, 1986, 1987; So and Woodhead, 1987;
Woodhead, 1990). However, deep tillage has a high
draft requirement which can be met by hand operated
two-wheel tractors. It requires four-wheel drive tractors or cable winch systems which are generally not
available in southeast Asia (IRRI, 1985, 1986). In
addition, four-wheel tractors can result in greater
compaction. Therefore, this solution does not seem
to be a practical option for the near future.
3.3.2. Effects of puddling intensity on subsequent DS
crops
Wet cultivation or puddling is synonymous with rice
culture in Asia and is used to assist in transplanting of
rice seedlings; to reduce water and nutrient losses and
to control weeds (Sharma and De Datta, 1985). Puddling breaks down and disperses soil aggregates into
individual component particles. The degree of dispersion for a given puddling effort is dependent on the
structural stability of the soil and is likely to affect the
regeneration of soil structure after rice, which, in turn,
will affect the DS crop. The effects of degree of
puddling prior to the rice phase on structure regeneration and growth of a DS crop after rice is related to soil
type. For example, increasing intensity of puddling
resulted in increased maize yields on a Vertisol but
decreased yields in hardsetting, lighter textured Regosols (Trenggono and Willatt, 1988). Similarly, intensive puddling increased DS mungbean yield on a clay
loam but decreased it on a sandy loam (IRRI, 1988).
These differences were attributed to clay content and
mineralogy. The concept of partially controlling soil
structure regeneration after rice through the puddling
treatment prior to the rice phase should be investigated
further by determining which soil types are responsive.
3.4. Seeding techniques
The most commonly used technique for DS
legumes after rice is manual dibbling. However, Cook
et al. (1995) found that dibbling gives variable results,
especially when the soil is wet in the few days after
rice harvest. An inexpensive alternative is manual
furrow seeding, which also gave variable results,
but was better in wet soils. Neither method was
reliable at lower water contents. They also found that
an inverted T seeder (Choudhary, 1985) pulled by a
hand tractor gave better performance in tilled soils
except when very wet or dry.
3.5. Crop/cultivar selection for improved root
performance
A factor in¯uencing the penetration of compacted
subsoils is the pressure that the root system can exert.
Different plants vary in their ability to penetrate
compaction layers. For example, bahia-grass (Paspalum notatum Flugge) penetrated compacted subsoils
better than cotton, which has a taproot, with the result
that cotton grown after bahia-grass yielded better and
extracted more water than cotton after cotton (Elkin
et al., 1977). Similarly, maize after pigeonpea (Cajanus cajan L.) grew better and yielded more than maize
after maize partly because of the superior penetration
by pigeonpea roots (Hulugalle and Lal, 1986).
Roots with the ability to penetrate hard subsoils
would better explore the subsoil provided they are able
to extract available water, which is determined by root
distribution and soil hydraulic characteristics. So and
Jayasekara, (1991) found large differences between 12
cultivars of sorghum (Sorghum bicolor L. Moench) in
their ability to extract water from the subsoil even
when adequate roots were present at depth for all
cultivars. To date, there is only limited information on
the ability of potential DS cultivars species and cultivars to grow through hard soils and to extract soil
water. Such information would assist in selecting
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
crops suitable as DS crops after lowland rice and
possibly in breeding suitable cultivars for that purpose.
9
management technologies that can overcome soil
physical limitations to DS crop production after lowland rice.
3.6. Soil chemical and biological limitations
4.2. The speci®c objectives of the project
Crops grown in lowland soils after rice may suffer
from de®ciencies of plant nutrients and from a lack of
suitable micro-organisms such as rhizobia or VA
mycorrhiza, which may not survive prolonged waterlogged conditions. The availability of residual nutrients from the rice phase is dependent on cultural
practices and soil type. In 1984, IRRI achieved mungbean yields of 2.1 Mg haÿ1 after rice without fertiliser,
inoculum or irrigation and with only 35 mm of DS rain
(IRRI, 1985). However, during a visit to east and
central Java, we saw signi®cant responses of mungbean and peanuts after rice to various combinations of
fertilisers and inoculum. The interaction of phosphorus and zinc has been observed to in¯uence plant
growth in student projects with Vertisols in Indonesia
(S. Setijono, personal communication). Zinc, copper
and boron de®ciencies have been reported for IR 64
rice in some areas of east Java and zinc and copper
applications have increased lowland rice yields
(Suyono, 1990). Therefore, it is possible that these
elements could be de®cient for DS crops as well and
should be evaluated.
3.7. Summary
In summary, it is clear that the limitations to DS
crop growth and yield after lowland rice soils are
complex and still not clearly understood. The need
for solutions to the problems of clay soils after lowland rice received strong endorsement from the 1989
Asian Rice Farming Systems Network workshop in
Bogor which recommended that work on this topic
should be initiated simultaneously in a number of
Asian countries.
4. The project `Management of clay soils for
lowland rice-based cropping systems'
4.1. General objectives
The overall objective of this project was to contribute towards the development of soundly based soil
1. To test soil management and agronomic practices
across a range of soils and climates, that have the
potential to overcome adverse soil physical
conditions for DS crops after rice, including
amendments (calcium or organic matter mulch),
tillage technologies and length of delay periods in
sowing of the DS crop after rice harvest.
2. To evaluate these practices by
2.1. measuring the changes in soil physical
conditions throughout the complete cropping
cycle from rice to DS crop.
2.2. determining the performance of the DS crop
(establishment and growth) and its ability to
extract soil water.
3. To determine the mechanisms involved in soil dispersion due to puddling and the factors controlling
¯occulation and structural reformation as the soil
dries after draining of surface water from rice ®elds.
4.3. The contrasting requirements of rice and DS crops
The project dealt with components of a cropping
system that have vastly different soil structural requirements. The rice phase requires a puddled soil with the
structure largely broken down, whereas the DS crop
requires a soil with good structure to express reasonable
productivity. We recognised that as a result of ameliorative treatments of the soil for the DS crop, detrimental
as well as bene®cial effects to the subsequent rice crop
may follow, e.g. paddy ®elds may become more permeable and leaky; residual N from legumes may be
bene®cial for rice. Therefore, it was important that,
where possible, the changes in physical properties were
monitored throughout the complete cropping cycle.
4.4. Selection of ®eld experimental sites
It was intended that the information and technology
derived from this project should be transferred readily
across a range of soils and climates. Therefore, the
project was designed as a series of benchmark sites
with common treatments, a common DS crop species
10
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
Table 2
Description of the selected ®eld sites in Indonesia and the Philippines with the main relevant soil characteristics
Country
Site
Soil texture
Clay (g kgÿ1)
ÿ1
Swelling clay (%)
Linear shrinkage
Indonesia
Ngale, East Java
Jambegede, East Java
Maros, Sulawesi
Heavy clay
Silty clay loam
Silty clay loam
740 g kg
450 g kgÿ1
460 g kgÿ1
73
15
9
0.19
n.a.a
0.05
Philippines
San ildefonso, Bulacan
Manaoag, Pangasinan
Heavy clay
Silty clay
410 g kgÿ1
520 g kgÿ1
26
33
0.07
0.10
a
n.a.: not available.
(mungbean) and common methodologies covering a
range of sites with different soils and climates. Treatments and species which are speci®c to particular
areas or soils are included at relevant sites.
Table 2 shows the textures of the sites in Indonesia
at Ngale (deep Vertisol) and Jambegede (silty clay
loam) in east Java, where soybean and peanuts are
included alongside mungbean; and a site near Maros
in south Sulawesi (silty clay). The sites in the Philippines are the research station at San Ildefonso,
Bulacan (Vertisol) and a farmer's ®eld near Manaoag,
Pangasinan (silty clay). The sites cover a wide range of
clay contents and clay mineralogies (Ringrose-Voase
et al., 1995). The Ngale site has the greatest clay content
with 740 g kgÿ1 total clay and 73% (on whole soil basis)
swelling clay (smectite and vermiculite), giving it the
greatest shrink/swell potential (0.19 linear shrinkage).
The site at San Ildefonso has 410 g kgÿ1 total clay and
26% swelling clay and intermediate shrink/swell potential (0.07 linear shrinkage). It also has a signi®cant sand
content of 330 g kgÿ1. The Maros and Manaoag sites
have similar particle size distributions with 460 g kgÿ1
and 520 g kgÿ1 total clay, respectively. However, they
have different mineralogies with 9% and 33% swelling
clays, respectively, giving them different shrink/swell
potentials (0.05 and 0.10 linear shrinkage, respectively).
Detailed descriptions of these soil and climates are
given by Schafer and Kirchhof (2000).
5. Experimental methodology
5.1. Field experiments in Indonesia and the
Philippines
At each site in Indonesia and the Philippines, three
®eld experiments were conducted to investigate the
effect of potentially useful management practices on
soil conditions and the resulting growth of DS crops.
Since a complete factorial experiment involving all
the factors of interest will result in a very large experiment which will be dif®cult to manage, treatments
were selected that are relevant to the farmers interest
and incorporated into three ®eld experiments that are
of manageble size. Experiments were also designed to
suit the local expertise and available facilities.
The ®rst two experiments were intended to reduce
the need for large numbers of treatment combinations
by separating treatments applied during the rice phase
(E1) and treatments applied during the legume phase
(E2). A third experiment (E3) measured the dynamics
of changes in soil properties in the period immediately
following ®eld drainage for rice harvest at four of the
®ve sites. The information gained from the latter will
help extrapolate the results from the ®rst two experiments to other soils and climates.
5.1.1. Experiment E1
The objective was to investigate the effects of
degree of puddling on soil physical conditions for
the subsequent DS crop. Previous work has indicated
that increased puddling may increase yields of the
following DS crop on heavy clay soils but decrease
yield on lighter textured soils (Trenggono and Willatt,
1988). If this is correct, puddling intensity can be used
as an inexpensive and readily adoptable practice on
suitable soils. The puddling intensity treatments
imposed during soil preparation for the rice crop were
1. dry cultivation prior to submergence,
2. one wet ploughing and harrowing using draught
animal power,
3. two wet ploughings and harrowings using draught
animal power,
4. two wet cultivations with mechanised rototiller or
hydrotiller.
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
The treatments were combined with two tillage and
sowing treatments (zero-till and dibbled, ZTD versus
ploughing, broadcasting and harrowing Ð PBH) for
the DS crops in the Philippines and two drainage
treatments (without and with surface drainage) in
Indonesia. PBH is the common farmer practice in
the Philippines, and ZTD with surface drainage are
common on heavy clay soils in Indonesia. In this
project, ZTD without surface drainage was the common treatment between the two countries. In both
countries, the most common farmer practice of soil
preparation for rice is two wet ploughings and harrowings using draught animals.
Treatments were replicated four times in a split-plot
design, with puddling intensity as the main plots and
tillage or drainage treatments as sub-plots. The experiment was carried out at each site over a period of 3
years, i.e. three rice-DS crop cycles.
5.1.2. Experiments E2
This experiment investigated the effects of DS crop
management practices on soil properties and crop
performance. To allow comparisons with experiment
E1, soil preparation for the rice phase was done using
two wet ploughings and harrowings (equivalent to
treatment 2 in E1).
Treatments included combinations of amendments
(none, A0; gypsum, AG; and organic matter mulch,
AOM); cultivation (zero-till, C0 and rotovator, C1)
and sowing delay after rice harvest (no delay, D0; one
week delay, D1 and two weeks delay, D2) as follows:
T1
C0
A0
D1
T2
T3
T4
T5
T6
T7
T8
T9
C0
C0
C0
C0
C0
C1
C1
PBH
A0
AG
AOM
A0
A0
A0
A0
A0
D1
D1
D1
D0
D2
D1
D2
D1
No fertiliser (farmers'
practice in Indonesia)
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
Adequate fertilisers
No fertiliser (farmers'
practice in the Philippines)
The DS crop was sown by dibbling for all treatments except T9. Treatments were replicated four
times in a randomised block design and the experi-
11
ment was conducted at each site over a period of three
consecutive crop cycles.
Please note that
1. treatments T1 to T8 were used on all sites and
treatment T9 only in the Philippines,
2. treatments T1 and T2 provide a measure of the
nutritional limitations of the soil,
3. treatments T2 to T4 provide a measure of the
effects of gypsum and organic mulch,
4. treatments T2 and T5 to T8 provide a measure of
the effects of cultivation and sowing delay and
their possible interactions.
It should be stressed that all treatments selected had
already been shown as potentially useful for DS crops
after rice in past trials, on speci®c soils and under a
narrow range of climatic conditions. This project
compared these potentially useful treatments on a
wider range of soils and climatic conditions and
attempts to quantify the agronomically relevant
changes in the soil conditions resulting from these
treatments, to enable extrapolation to other sets of
environments.
5.1.3. Experiment E3
A third short-term ®eld experiment was conducted
at one or two sites each season. This experiment
monitors changes in soil mechanical and hydrological
properties and soil structure with time as the soil dries
after draining surface water prior to rice harvest. The
information obtained for each soil will provide functional relationships required to interpret the conditions
encountered at planting and during early seedling
growth in the ®rst two experiments which involve
only three sowing delays. These experiments were
expected to last 2±4 weeks each.
This experiment was carried out on soils with
standard puddling treatments (two ploughings and
two harrowings). The following properties were measured at pre-determined intervals: soil water content,
soil strength (shear and penetrometer resistance),
macropore structure development (using crack measurements and resin impregnated samples).
5.1.4. Selection of dry season crop species
In experiments E1 and E2, mungbean was the DS
crop common to all sites. Mungbean is the major DS
legume crop in the Philippines and the third major
12
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
legume in Indonesia. A second crop of soybean was
included in east Java with soybean on the Vertisol
(Ngale) and peanuts on the lighter textured soil (Jambegede). Thus the crop rotations are as follows:
Ngale
Jambegede
Maros, San
Ildefonso, Manaoag
Rice±mungbean/soybean±
mungbean/soybean
Rice±mungbean/peanut±
mungbean/peanut
Rice±mungbean
5.1.5. Measurements
This paper provides an overview of the measurements undertaken in this project. However, detailed
descriptions of the methodologies will be found in the
relevant sections of this special issue.
5.1.5.1. Initial characterisation of the site. The
uniformity of each site was investigated using a 20±
25 m grid system. Morphological descriptions were
made to determine any gradation in soil characteristics
and to avoid any unrepresentative areas. Composite
samples from sections of the field at different depths
were analysed to provide a baseline data-set prior to
imposition of treatments. These were
soil texture, bulk density, cation exchange capacity,
cations, pH, electrical conductivity, soil organic
carbon,
soil water characteristics,
plastic and liquid limits,
soil structural stability (wet sieving and dispersibility).
5.1.5.2. Measurements during the rice phase. The
following were measured for some sites at the start,
middle and end of the rice phase:
sinkage capacity,
infiltration rates,
dispersibilty of the soil,
yield of rice on all sites.
phases of the crop: emergence, vegetative, flowering,
pod formation and maturity. Measurements included
soil physical measurements;
soil bulk density profiles,
soil strength profiles Ð penetrometers,
soil water content profiles and water use by the
crop,
root distribution/root length densities (at flowering only),
plant measurements;
emergence as a measure of establishment,
plant density at harvest as a measure of survival,
plant biomass and its components,
yield and yield components,
climatic measurements(local weather station data);
rainfall,
evaporation (E pan),
temperature,
radiation.
5.1.6. Simulation of soil puddling and drying in the
laboratory
Laboratory experiments were conducted at the University of Queensland to measure the degree of dispersion after the soil is subjected to a range of standard
puddling treatments similar to those in the ®eld
experiments (Kirchhof et al., 2000b). Degree of dispersion was adopted as a measure of the degree of
puddling and the decrease in structural stability. Since
one major objective of puddling is to reduce percolation rate, soil hydraulic conductivities were measured
for each puddling treatment and related to the degree
of dispersion. The effects of repeated wetting and
drying on soil structural development were also investigated. Soil changes occurring during the rice-DS
crop sequence were observed and quanti®ed in large
lysimeters (approximately 1.2 m1 m1 m) in which
rice was grown under puddled conditions followed by
a DS crop. Soils used covered a similar range of
textures to those used in the ®eld experiments. Details
are described by Kirchhof and So (1994, 1996).
6. Summary
5.1.5.3. Measurements during the dry season crop
phase. Soil measurements were made at the same time
as plant measurements at the appropriate phenological
In summary, this paper presented the background
and an overview of this international collaborative
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
project in Indonesia and the Philippines. Details of the
soils and climate of the ®ve experimental sites are
described by Schafer and Kirchhof (2000). The results
and implications of the E1 experiments are described
by Kirchhof et al. (2000a,b) and those for the E2
experiments by Kirchhof et al. (2000a). Results on the
changes in soil physical properties in experiment E3 is
discussed by Ringrose-Voase et al. (2000) and crop
establishment aspects are discussed by Rahmianna
et al. (2000). The paper by Cabangon and Tuong
(2000) was not part of the collaborative project, but
is relevant as it deals with the in¯uence of cracks on
soil preparation for rice and how shallow tillage can
result in early sowing of rainfed lowland rice, an
important factor during low rainfall seasons.
Acknowledgements
The project was funded by the Australian Center for
International Agricultural Research and their contribution and support is gratefully acknowledged.
References
Adachi, K., 1990. Effects of rice-soil puddling on water
percolation. In: Transactions of the 14th International Congress
of Soil Science, Kyoto, Vol. 1, 146±151.
Adisarwanto, T., 1985. The in¯uence of planting methods and
mulching on soybean seed yield. In: Shanmugasundaram, S.,
Sulzberger, E.W. (Eds.), Soybean in Tropical and Subtropical
Cropping Systems. Asian Vegetable Research and Development Centre, Taiwan, pp. 215±217.
Adisarwanto, T., Santoso, B., Manshuri, G., Floyd, C., 1989.
Agronomic Survey of MK1 Soybean on Sawah in East Java
(Jember to Madiun), 21±24 June 1989. Malang Research
Institute for Food Crops, East Java, Indonesia.
Adriano, M.S., Cabezon, V.E., 1987. Economic policies affecting
crop diversi®cation in the Philippines. In: Irrigation Management for Diversi®ed Cropping. International Irrigation Management Institute, Sri Lanka.
Africa, T.P., Marquez, N.R., 1989. Philippine yearbook. National
Statistics Of®ce, Manilla, Republic of the Philippines.
Agricultural Intensi®cation Program, 1988/89. Special province of
Yogyakarta (translated from: Keputusan Gubernur Kepala
Daerah Istimewa Yokyakarta No 165/KPTS/1988 tentang
Program Intensi®kasi Pertanian Tahun 1988/1989, Propinsi D
I Y, BIMAS Propinsi D I Y)
Bacon, P.E., 1979. Ef®ciency of different calcium application
methods in controlling muddy water. In: Proceedings of the
13
Australian Society of Soil Science Inc. Conference, Yanco,
NSW, Australia, pp. 116±121.
Cabahug, L.V., 1990. In praise of the marvellous mungbean. Food
Legumes Ð Coarse Grains Newsletter 12, 13±14.
Cabangon, R.J., Tuong, T.P., 2000. Management of cracked soils
for water saving during land preparation for rice cultivation.
Soil Tillage Res. 56, 105±116.
Cass, A., Gusli, S., MacLeod, D.A., 1994. Sustainability of soil
structure quality in rice paddy-soyabean cropping systems in
South Sulawesi, Indonesia. Soil Tillage Res. 31, 339±352.
Choudhary, M.A., 1985. Seeding techniques and machinary. In:
Soil Physics and Rice. International Rice Research Institute,
Los BanÄos, The Philippines, pp. 411±412.
Cook, S.M.F., 1989. In¯uence of soil physical properties on
mungbean emergence in rice soils. M.Sc. thesis. University of
Minnesota, USA.
Cook, S.M.F., Gupta, S., Woodhead, T., Larson, W.E., 1995. Soil
physical constraints to establishment of mungbeans (Vigna
radiata L. Wilczek) in paddy rice (Oriza sativa L.) soils. Soil
Tillage Res. 33, 47±64.
Elkin, C.B., Haaland, R.L., Hoveland, C.S., 1977. Grass roots as a
tool for penetrating soil hardpans and increasing crop yields. In:
Proceedings of 34th Southern Pasture and Forage Crop
Improvement Conference, Auburn University, AL, USA,
pp. 21±26.
FAO Yearbook, 1984. Vol 38. Food and Agricultural Organisation
Statistics Series No 63.
Fy®eld, T.P., 1987. Effects of temperature, water and cultivation on
the establishment of a mungbean crop following paddy rice.
Ph.D. thesis, University of Reading, UK.
Hoque, M.Z., 1984. Cropping Systems in Asia: On-Farm Research
and Management. International Rice Research Institute, Los
BanÄos, Philippines.
Huke, R.E., 1982. Rice Area by Type of Culture: South, Southeast,
and East Asia. International Rice Research Institute, Los
BanÄos, Philippines, 32 pp.
Hulugalle, N.R., Lal, R., 1986. Root growth of maize in a
compacted gravelly tropical Al®sols as affected by rotation
with a woody perennial. Field Crops Res. 13, 33±44.
IRRI, 1984. Annual Report for 1983. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1985. Annual Report for 1984. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1986. Annual Report for 1985. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1987. Annual Report for 1986. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1988. Annual Report for 1987. International Rice Research
Institute, Los BanÄos, Philippines.
IRRI, 1989. Annual Report for 1988. International Rice Research
Institute, Los BanÄos, Philippines.
Kirchhof, G., So, H.B., 1996. The effect of puddling intensity and
compaction on properties, rice and mungbean growth: a mini
rice-bed study. In: Kirchoff, G., So, H.B. (Eds.), Management
of Clay soils for Lowland Rice Based Cropping Systems,
ACIAR Proceedings No 70, 1996, pp. 51±70.
Kirchhof, G., So, H.B., 1994. Soil puddling for rice production and
14
H.B. So, A.J. Ringrose-Voase / Soil & Tillage Research 56 (2000) 3±14
its effect on soil structure and water percolation. In: Proceedings of International Symposium on Sealing, Crusting and
Hardsetting Soils: Productivity and Conservation, February
1994, The University of Queensland, Brisbane, Qld, Australia,
pp. 7±11.
Kirchhof, G., So, H.B., Adisarwanto, T., Utomo, W.H., Priyono, S.,
Prastowo, B., Basir, M., Lando, T.M., Subandi, Dacanay, E.V.,
Tan-Elicano, D., Sanidad, W.D., 2000a. Growth and yield response of grain legumes to different soil management practices
after rainfed lowland rice. Soil and Tillage Res. 56, 51±66.
Kirchhof, G., Priyono, S., Utomo, W.H., Adisarwanto, T., Dacanay,
E.V., So, H.B., 2000b. The effect of soil puddling on the soil
physical properties and the growth of rice and post-rice crops.
Soil and Tillage Res. 56, 37±50.
Lavapiez, L.M., Duran, B., Nicolas, J., Price, E.C., 1977. Rice-mung
cropping patterns in Manaoag, Pangasinan: a costs and returns
study. Internal Report. Agricultural Economics Department,
International Rice Research Institute, Los BanÄos, Philippines.
Maghari, S.de G., 1990. Mungbean Emergence and Growth as
in¯uenced by Soil Submergence in Previous Rice Crop. M.Sc.
Thesis, University of Minnesota, USA.
Mahata, K.R., Sen, H.S., Pradhan, S.K., Mandal, L.N., 1990. Notill and dry ploughing compared with puddling for wet season
rice on alluvial sandy clay loam in eastern India. J. Agric. Sci.
114, 79±86.
Maranan, C.G., 1986. Agricultural economic research at Guimba
cropping systems site, Nueva Ecija, Philippines. Internal report.
Agricultural Economics Department, International Rice Research Institute, Los BanÄos, Philippines.
Maranan, C.G., 1987. Economic performance of cropping patterns
at Bantug, Guimba, Nueva Ecija CSR site, 1986±87. Internal
report. Agricultural Economics Department, International Rice
Research Institute, Los BanÄos, Philippines.
McKenzie, D.C., So, H.B., 1989a. The effect of gypsum on
Vertisols of the Gwydir Valley, NSW. I. Soil properties and
wheat growth. Aust. J. Exp. Agric. 29, 51±62.
McKenzie, D.C., So, H.B., 1989b. The effect of gypsum on
Vertisols of the Gwydir Valley, NSW. II. Ease of tillage. Aust. J.
Exp. Agric. 29, 63±67.
Nanseki, T., Morooka, Y., Zakaria, A.K., 1989. Comparative
advantage analysis of soybean in an upland area of West Java: a
case study of mathematical programming approach. Palawija
News 6, (2). CGPRT Newsletter.
Pasaribu, D., McIntosh, J.L., 1985. Increasing tropical soybean
production with improved cropping systems and management.
In: Shanmugasundaram, S., Sulzberger, E.W. (Eds.), Soybean
in Tropical and Subtropical Cropping Systems. Asian Vegetable
Research and Development Centre, Taiwan, pp. 1±11.
Rahmianna, A.A., Adisarwanto, T., Kirchhof, G., So, H.B., 2000.
Crop establishment of legumes in rainfed lowland rice-based
cropping systems. Soil and Tillage Res. 56, 67±82.
Ringrose-Voase, A.J., Hutka, J., Beatty, J., Raven, M.D., Rath, H.,
1995. Analysis of some Indonesian and Philippine soils for
ACIAR Project 8938 on Management of clay soils for lowland
rice-based cropping systems. Division of Soils Tech