Europ. J. Agronomy 27 (2007) 81–88
Intercropping reduces nitrate leaching from under field
crops without loss of yield: A modelling study
A.P. Whitmore a,∗ , J.J. Schröder b
a
Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK
Plant Research International, Wageningen, The Netherlands
b
Received 11 August 2006; received in revised form 25 January 2007; accepted 5 February 2007
Abstract
A model of soil nitrogen dynamics under competing intercrops is described and used to interpret two sets of experimental field data from the
literature. In one series of experiments, maize received slurry and mineral nitrogen (N) fertiliser or mineral N alone and was grown either alone
or intercropped with undersown grass or with a subsequent rye catch crop during 7 years continuously. In the second system, the model compares
field beans intercropped spatially at different densities with winter wheat.
The model suggests that undersowing grass between the rows of an established maize crop can reduce concentrations of nitrate in water draining
from soils during winter by 15 mg l−1 compared with a conventional catch crop and by more than 20 mg l−1 compared with a fallow soil. The model
further suggests that the yield and profitability of mixed stands of commercial crops is inversely related to the residual nitrate at harvest (potential
leaching). It is concluded that intercropping may be a useful means to reduce nutrient pollution from farming while maintaining yields.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Intercropping; Model; Nitrogen; Leaching; Environment
1. Introduction
Farming profitably within increasingly stringent environmental norms is becoming difficult. Although some crops such as
cereals and sugar beet leave little soil mineral nitrogen (N)
behind at harvest, other crops such as maize, potatoes, vegetable crops and legumes either leave considerably more or leave
residues that release N during the winter. Where excess winter
rainfall is less than 100–200 mm, even water leaving land that
has grown cereals is at risk of having a concentration of nitrate in
it that exceeds limits in current water quality legislation (Anon.,
2000; Whitmore and Addiscott, 1986).
Intercropping, defined here as any system of multiple cropping in space, has a long and successful history in tropical
regions (e.g. Trenbath, 1993; Tsubo et al., 2005). Not only
has the technique been shown to increase yields (De Wit and
Van Den Bergh, 1965) but it is also a useful means of spreading risk: if one crop fails another may still provide sufficient
food until the next harvest (e.g. Trenbath, 1999). In developed
∗
Corresponding author. Tel.: +44 1582 763133; fax: +44 1582 469036.
E-mail address: [email protected] (A.P. Whitmore).
1161-0301/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.eja.2007.02.004
countries and conventional cropping systems, monoculture has
proved the rule, with the exception of some grass–clover mixtures, probably because of the ease of combining or lifting a
single crop with machinery. Despite this, theoretical and experimental work has pointed to the potential benefits of mixtures of
species or varieties. De Wit and Van Den Bergh (1965) showed
that where two annual grasses do not compete for a resource,
yields per m2 may be significantly greater than under monocropping; Bulson et al. (1997) and Hauugaard-Nielsen et al. (2006)
have demonstrated this more widely. However, Ghaffarzadeh et
al. (1997) who investigated water supply, and Ayisi et al. (1997)
and Lesoing and Francis (1999) who both investigated N supply, have shown that these benefits may not extend further than
one row where species intercropped with one another do not
alternate row by row. Intercropping has also been shown to control the spread of pests and disease (Trenbath, 1993; Zhu et al.,
2000) and has been suggested as a means to help control erosion
(Lesoing and Francis, 1999). These results point to clear benefits
in productivity by planting intercrops that do not compete with
each other, because resources are used efficiently. This raises
the question as to whether crops that do compete for a nutrient
might be successfully intercropped with one another in the field
in order to control environmental losses of that nutrient.
82
A.P. Whitmore, J.J. Schröder / Europ. J. Agronomy 27 (2007) 81–88
Schröder et al. (1996) have shown that undersowing maize
with a grass crop can recycle N that would otherwise leach in
winter. Known as a relay crop, the grass is an intercrop while the
maize is in the ground, but is largely out-competed. The relay
crop then remains in the soil during the winter as a sole crop.
Schröder et al. also included a conventional catch crop in their
experiments. The extra N-supply from the winter crops could
be detected in maize crops given less N the following year than
expected for maximum growth, implying that the winter crops
could replace fertiliser N. Hauugaard-Nielsen et al. (2003) found
a small reduction in nitrate leaching (kg ha−1 ) from lysimeters
cropped with a pea–barley mixture compared with sole crops,
although much of this difference may be attributable to differences in the N-content and rate of decomposition of roots and
residues. Where the intercrops have a sequential demand for that
nitrogen, yields (and profit) might be maintained but the losses
of N reduced.
This article describes the adaptation of an existing computer
simulation model (Addiscott and Whitmore, 1987; Whitmore et
al., 1991; Whitmore, 1995, 2007) to intercropping systems and
an analysis of the potential of intercropping to reduce nitrate
leaching from modern-day agriculture. Although there are other
potential benefits of intercropping such as the prevention of the
spread of disease, management of risk, or suppression of weeds
(e.g. den Hollander et al., 2007) these benefits will not be quantified here. Others have published models of intercropping and
nitrogen dynamics (Berntsen et al., 2004; Brisson et al., 2004).
We differentiate our research from this earlier work by focussing
chiefly on the soil processes and analysing both soil mineral N
and leaching losses of N.
2. Methods
2.1. The model
The model used in this study has been described by
Addiscott and Whitmore (1987) with some adaptations by
Whitmore et al. (1991) and Whitmore (1995) and by Whitmore
and Schröder (1996) for maize, but will be described here
briefly for completeness sake. The model system now incorporates organic nitrogen dynamics as described by Whitmore
(2007).
2.1.1. Leaching
Addiscott and Whitmore (1987) describe a dual-porosity
model of the leaching of solutes added to soil. Moisture in soil
is held either between aggregates (mobile water, wm ) and can
be displaced by incoming water, or held within aggregate pore
space (retained water, wr ). For the 50 mm layers employed in
the model the following definitions apply, where θ FC is the volumetric moisture content of soil at −5 kPa (mm3 mm−3 ), θ 200 kPa
the volumetric moisture content at −200 kPa and θ 1.5 MPa the
volumetric moisture content at −1.5 MPa:
θ1.5 MPa
wr = 50 θ200 kPa −
wm = 50(θFC − θ200 kPa ),
2
(1)
Solutes move down the profile with the mobile water, wm , as
described by Addiscott and Whitmore (1987). A fast leaching
routine is incorporated (Addiscott, 1977) in order to simulate
bypass flow following heavy rainfall.
2.1.2. Crop growth, N uptake and development
Dry matter production is estimated from a simple relationship with incoming radiation (Whitmore, 1995). N uptake and
rooting depth are estimated using a simplified logistic function
(Whitmore and Addiscott, 1987):
−n
Y = (A−1/n + e−kx )
(2)
where Y is the N uptake or rooting depth, n distorts the symmetry
of the curve and was set at 1.5 for all crops (Whitmore and
Addiscott, 1987), k a rate constant and x is the thermal time
(the accumulation of the average daily temperature above 0 ◦ C).
The parameter A is the maximum value Y is allowed to take:
200, 250, 250 or 250 kg N ha−1 for N uptake with wheat, grass,
maize or beans, respectively. The maximum potential rooting
depth of winter wheat, maize and the rye catch crop was set at
150 cm, and at 50 cm for beans and for grass. The amount of
root in each layer declines with depth of soil exponentially in
the manner proposed by Gerwitz and Page (1974) with values
of the depth containing 1/e of the total potential proportion of
roots being 43 cm for wheat, maize and rye and 20 cm for grass
and beans. For crop N uptake and rooting, k was set at 0.003 and
0.005 ◦ C−1 day−1 , respectively.
Crop development follows that proposed by Weir et al.
(1984). Growth and development of maize and allocation of
assimilate to roots and below-ground exudation for all crops is
as described by Whitmore and Schröder (1996) using parameters derived from Van Diepen et al. (1989). Development of
beans follows that proposed by Bouniols et al. (1991).
Winter crops are susceptible to cold but may also develop
acclimation (Pomeroy et al., 1975). These processes were modelled as proposed by Fowler et al. (1999) who describe the
calculation of the temperature LT50 that kills 50% of a crop.
Because we are interested here in the death of a proportion of
the crop, these LT50 values were scaled back arbitrarily by the
ratio of the actual temperature to the LT50 value and crop dieback calculated whenever the minimum temperature fell below
−3 ◦ C.
2.1.3. Organic matter turnover and
mineralization–immobilization turnover of N
A part of the soil organic matter turnover in soil is envisaged
in the model (Whitmore, 2007; Whitmore et al., 1997; Vinten et
al., 2002) to be physically protected (Hassink and Whitmore,
1997) from microbial attack (Whitmore, 1996a,b; Whitmore
and Groot, 1997). The rate of protection is proportional to the
fraction of the maximum amount, X, of organic carbon that
the soil may stabilise that is already occupied by soil organic
carbon (SOC). Values of X, which was found to be a function
of clay content, are given in Section 2.2. Whether N mineralizes or is immobilized during the turnover of each compartment
depends on the C:N ratio of each compartment (Whitmore and
A.P. Whitmore, J.J. Schröder / Europ. J. Agronomy 27 (2007) 81–88
Handayanto, 1996; Whitmore and Groot, 1997, 1994) but the
rates of decomposition of residues is modified by the lignin and
fibre contents (Whitmore and Matus, 1996). Recalcitrant SOC
is affected by temperature more than labile residues (Whitmore,
1996c). Whitmore and Schröder (1996) have described the use
of this model to simulate the mineralisation of N from cattle
slurry.
under normal planting conditions:
2.1.4. Nitrogen fixation
Bouniols et al. (1991) describe a model of the fixation of N
by soybean in relation to the development parameters of their
model and these have been adopted here, viz.: fixation capability increases from nothing at emergence to a maximum as leaf
nodes develop, remains at maximum until flowering and then
declines to nothing by the end of grain-filling. Actual fixation
is reduced by low temperature, droughtiness and high mineral
N content in soil as described by Bouniols et al. (1991). Stern
(1993) has argued that N is not transferred directly between intercrops when one of them is a legume, although indirect transfer
via the decomposition of residues in soil is possible. Accordingly direct transfer was not permitted in the model but a relative
increase in uptake of N from soil by the non-legume is possible
if it is the more competitive plant. Giller and Day (1985) suggest
that the consumption of carbon needed to fix N by free-living
soil organisms is in the range of 10–100 mg N fixed g−1 C to a
host plant, which at the upper end equates to 25 mg N fixed g−1
dry matter used in the model (assuming 40% C in dry matter);
Berntsen et al. (2004) suggest a maximum value of 27 mg N g−1
net dry matter production. Fixation has a cost in terms of allocation of carbohydrate. We use a factor 0.4 to reduce dry matter
production for the cost of fixation on a gross basis before reduction of rate modifiers for phenology, temperature, moisture or
nitrate content of soil are applied. This seems to be roughly
equivalent on an annual basis to the net value of 18% of dry
matter production cited by Berntsen et al. (2004) if all of the
legume’s nitrogen is derived from the atmosphere.
Ci
Ci = Di
di =
dmi
Ci
(4)
i, Ci is itself
where dmi is the crop dry matter at time t of crop calculated from the sum of the planting densities Di and Ci
is the value of dmi at which the crop first completely covers the
ground when planted at normal density as a sole crop:
(5)
If either d1 or d2 achieves a value of 1 then that crop completely covers the ground surface and competition is inevitable.
The planting densities, Di , are defined relative to a normal density for crop i, i.e. a normally planted
single crop would have
a value of D of 1. With two crops D is unlikely to exceed 2,
but may be less than 1 (where the sowing density of each crop
is 25% of normal, for example).In many intercropped systems
the Di are likely to be 0.5 and Di 1, reflecting a 1:1 mixture
of crops each planted at half normal density. However, some
of the experiments modelled here have crops planted at greater
and lesser densities and account must be kept of the different
proportions of the field that are occupied. Substituting from (5)
into (4), (3) can be expressed:
Li =
dmi /Ci
(dmi /Ci ) + (dmj /Cj )
2.2. Experimental data
Two sources of experimental data were used to evaluate the
model. Schröder et al. (1996) carried out a series of experiments
on a sandy soil (<5% clay, maximum protective capacity for
POM , X is 21.3 g C kg−1 ) at the Heino experimental research
farm in the Netherlands for seven growing seasons between
1988 and 1994. The experiment investigated ways of reducing
the leaching of N from maize crops that received 45 m3 ha−1
of cattle slurry annually in 1988 and 1989 and 33 m3 ha−1 from
1990 on. The slurry was injected (tine width 50 cm) to a depth of
15 cm after ploughing-in of the winter crop but 7–13 days before
the maize crop was sown. The N content of the slurry was on
average 5 kg total N and 2.3 kg NH4 -N m−3 . The plots received
either 140 (N5), 100 (N4), 60 (N3) or 20 (N2) kg mineral fertiliser N ha−1 as calcium ammonium nitrate. A further treatment
received no slurry but 20 kg mineral N ha−1 only (N1). A fallow winter after maize was compared with (i) a rye catch crop
sown immediately after harvest of the maize, which was between
2.1.5. Competition between plants
Processes in the model are programmed to occur sequentially with a daily time-step. Priority is given alternately to each
crop with time-step. In this way no one crop gains exclusive
access to resources by virtue of always being called first in the
programmed sequence.
A factor Li (i = 1,2) describes the relative capture of photosynthetically active radiation (PAR) in relation to the extent of
ground cover that the leaf canopy of each plant achieves. It is
calculated as follows:
⎧
di
⎨L =
; di < 1, dj > 1; where{j = 2 if i = 1} or {j = 1 if i = 2}
i
d1 + d 2
⎩
Li = 1;
otherwise, i.e. no encroachment or competition
Even if there is strong competition, Li does not fall below 0.05. In
Eq. (3), the di are calculated from the proportion of dry matter
achieved at time, t, of potential Ci , the amount of dry matter
at which each crop first achieves full crop cover relative to its
density (i.e. leaf area index, LAI = 1) when planted as a sole crop
83
(3)
22 and 28 September except in 1993 when sowing had to be
postponed until 10 October and (ii) a grass relay crop (spatial
intercrop) sown in between the maize rows in the middle of June.
In the last year of the experiment, however, no grass was sown.
The grass crop is not a strict intercrop since the maize effectively
84
A.P. Whitmore, J.J. Schröder / Europ. J. Agronomy 27 (2007) 81–88
out-competes it until harvest. Nonetheless it provided a good
test of the ability of the model to simulate the dynamics of N
both under two intercropped species during summer and during
the winter following the harvest of the maize. The winter crops
were destroyed mechanically at the end of March or beginning of
April with a rotavator and the residues were incorporated using
a mould board plough to a depth of 25 cm in the second half of
April. Destruction of the rye crop was postponed until the end
of April in 1989. Six replicate measurements of mineral N were
taken from each plot to a depth of 60 cm five times during the
year: in early spring, before incorporation of the winter crop,
in early summer, at harvest of the maize and in early winter.
Simulations are presented here of the mineral N remaining in
the soil profiles of all treatments and of concentrations of N
leaching from under fallow, grass and rye catch crop soils during
each winter. Four replicate porous ceramic cups were installed
in a single replicate of each N rate treatment at a depth of 100 cm
and the concentrations of N measured periodically throughout
each of the winters. In practice, rooting was restricted below
30 cm by a consolidated layer. The experiment was set up as
a split-plot design with four replicates of the three winter crop
treatments on the main plots and the five N application rates on
the sub plots. Sub plot size was 14 m × 6 m. Full details of the
experiments can be found in Schröder et al. (1996), van Dijk et
al. (1995) and Schröder et al. (1992).
In the second series of experimental data used here, Bulson et
al. (1997) published yields and N uptake by intercropped mixtures of organically grown wheat and field beans sown on a clay
loam soil (typically 45% clay, and X, 38 g C kg−1 ) near Pangbourne in southern England. A feature of this data is the varying
ratios of each crop to the other and the different densities at
which the crops were planted. The varieties of each plant sown
were selected to mature at about the same time and all combinations of beans sown at 0, 25%, 50%, 75% and 100% of the
recommended planting density were intercropped with all combinations (apart from 0, 0) of wheat sown at 0, 25%, 50%, 75%
and 100% of recommended planting density in a bivariate factorial experimental design in three fully randomised blocks. Sole
crop densities were 250 plants m−2 (wheat) and 25 plants m−2
(beans). No soil N data is available in this experiment but an
analysis of the profitability was presented. At low planting densities, weeds invaded much of the experiment. These have not
been treated as a separate crop but the amounts of N in the weeds
have been accounted for in the analyses presented where appropriate. Full details of the experiment and its agronomy are given
by Bulson et al. (1997).
3. Results
3.1. Maize grass mixtures
Residual mineral N in soil at harvest was simulated well under
maize and grass and under maize alone to 60 cm depth (Fig. 1)
although the model underestimates small mineral N contents as
well as the total amount in the top 30 cm (data not shown). Note
that the fallow treatment was fallow every winter in this sequential series of experiments. Both fallow and rye treatments grow
Fig. 1. Simulated vs. measured amounts of mineral N found down to 60 cm
depth at harvest of maize between 1998 and 1994 receiving 140 kg N ha−1 (䊉),
or 60 kg N ha−1 mineral fertiliser plus cattle slurry (), or 20 kg ha−1 mineral
fertiliser only, each year (), in (a) soils growing a rye catch crop each winter,
(b) intercropped with grass from the middle of June, and (c) remaining fallow
each winter. Added line is the line of perfect agreement in all cases. No S.E. is
available for the measured data.
A.P. Whitmore, J.J. Schröder / Europ. J. Agronomy 27 (2007) 81–88
Fig. 2. Mean concentrations of nitrate recovered in porous ceramic cups (points)
placed at 1 m depth or simulations (lines) of the concentration of N leaching at
the same depth in soils receiving 140, 120, 60 or 20 kg N ha−1 mineral N plus
cattle slurry each year (N5-N2) or receiving 20 kg N ha−1 mineral N alone (N1);
(—, 䊉), fallow soil; (- - -, ), rye catch crop; (· · ·, ), grass intercrop. S.E. of
measurement 16.2 mg N l−1 .
maize as a sole crop during the summer, but each winter the rye
treatment has a rye catch crop. Since the same amounts of N are
given to the maize each year irrespective of cropping treatment,
total N supply each spring can be larger on the winter crop treatments. Mineral N in soil at harvest is at risk of leaching, but
actual leaching depends on the conditions during the following
autumn and winter. These conditions include the growth of a
rye catch crop in one treatment and continued growth of grass
relay crop in another. The model simulates the average concentrations of N found to leach from sub-optimally fertilised soils
well (Fig. 2), but seems to overestimate the concentrations after
large applications of mineral N in the grass and rye treatments.
Schröder et al. (1996) attribute the smaller loss of N (kg ha−1 )
from the rye and grass plots compared with the fallow to an
extra sink in the former two treatments. The variability of this
sink may reflect the susceptibility of the winter crops to frost.
In almost all years the spring crops contained far less N than the
same crop in the previous autumn. Although this effect is simulated reasonably well, the model is unable to reproduce the large
early growth then die-back of the grass crop during the winters
of 1990–1991 and 1991–1992 (data not shown). Inevitably dieback on this scale will temporarily tie-up and then release much
N with the potential to be leached.
85
Fig. 3. Simulated vs. measured nitrogen uptake in grain in beans (closed symbols), or nitrogen uptake in grain in wheat at harvest (open symbols). Data within
a series of wheat densities represent the different planting densities of the beans:
() wheat density 0%; (䊉, ) wheat density 25%; (, ) wheat density 50%;
(, ) wheat density 75%; (, ) wheat density 100%. S.E. of measurements
in wheat range from 10.0 to 22.4 kg N ha−1 and in beans from 9.58 to 19.2 kg
N ha−1 . Measured values taken from Bulson et al. (1997) where full details of
the error calculations are given. Line on graph indicates perfect agreement.
only but added to observed N in weeds for Fig. 4. It appears that
intercropping at high density has removed N efficiently from
the system, that the soil residual N and by implication the N
uptake of the cereal has hardly been affected by intercropping
with beans, but that the potential loss of N from the beans has
been reduced by planting with wheat.
It is interesting to make use of the profitability data supplied
by Bulson et al. (1997). Not surprisingly high densities of
intercropped species were relatively profitable. The benefits
in yield have been known for many years (e.g. De Wit and
Van Den Bergh, 1965). Profitability can be good for the
environment, because the densely sown-intercrops seem to
3.2. Wheat–bean mixtures
The uptake of nitrogen by both beans and wheat was simulated well by the model (Fig. 3). Although there appear to be
small effects of the density of the bean crop on the wheat crop at
constant density that the model fails to reproduce, these effects
are all smaller than the standard error in the measurements (Fig. 3
and legend) and display no consistent trend. No measurements of
mineral N were made in these experiments and so the amounts of
mineral N at harvest to 60 cm depth are derived from simulation
Fig. 4. Simulated amounts of mineral N remaining in soil to 60 cm depth
(kg N ha−1 ) at the harvest of mixtures of beans and wheat grown at different
densities.
86
A.P. Whitmore, J.J. Schröder / Europ. J. Agronomy 27 (2007) 81–88
Fig. 5. Profit (£ ha−1 ) vs. simulated amounts of N (kg ha−1 ) remaining in the
soil to 60 cm at the harvest of mixtures of beans and wheat (䊉) sown at different
densities. () Beans alone. S.E. of profits between means of both wheat and
bean densities, £45.4 ha−1 ; interactions between wheat and beans, £102 ha−1 .
Data taken from Bulson et al. (1997).
the total loss of N2 through denitrification following Bradbury
et al. (1993) and this is almost exactly the same under maize
or wheat whether on not a grass or beans intercrop respectively
is present or not. Yang and Cai (2006) found more emission of
N2 O from shaded (intercropped) than unshaded soybean plants
at flowering but the situation is reversed during grain-filling.
Volatilisation of ammonia is not considered by the model either
from the soil or through the stomata of the plant but because mineral nitrogen was not applied to the wheat-beans mixtures, losses
of ammonia from soil are unlikely to differ between inter and
sole cropped beans and wheat. The grass was not sown between
the maize rows until almost 2 months after manure and fertiliser
were applied, so there is little reason to think that losses from
soil would be greater where grass was sown than not. The extra
density of crop in the wheat–beans mixtures might have led to
a small additional loss of ammonia through stomata, but even
with a canopy twice as dense, these losses are unlikely to be
greater than 2 or 3 kg N ha−1 (Schjoerring et al., 1993).
4.2. Do intercropped systems promote the efficient use of N?
make use of all the available soil N and leave little behind at
harvest (Fig. 5). The least desirable combination of minimum
profit and large residual N results from low density planting
of beans alone as would be expected. The region in the top
left of Fig. 5, plots increasing profit with decreasing residual
N. Although not shown, profit also increased with cropping
density. Intercropping at rates up to twice the normal cropping
density per unit area clearly increased yields and profits in these
experiments. It seems likely that it decreased residual N (Fig. 4)
and hence the risk of nitrate leaching also.
4. Discussion
Our thesis in this article is that intercropping different species
will reduce nitrate leaching without a reduction in total yield or
profit and this conclusion appears to be justified by the results.
The presence of the grass intercrop reduced leaching at Heino
(Fig. 2) while the yields of wheat at Pangbourne were little
reduced by intercropping with beans (Fig. 3), but the potential for leaching after beans appeared to be greatly reduced in
the wheat mixtures (Fig. 4). Yields of beans were impacted by
wheat (Fig. 3), but mineral N at harvest was probably unaffected
by the presence of beans provided that the wheat density was at
least 50% of normal practice (Fig. 4). Despite the reduction in
the yield of the beans, profitability of the wheat–bean system as
a whole was not reduced by intercropping and increased with
decreasing risk of N leaching (Fig. 5). The results do suggest
that N leaching is reduced in intercropped systems, probably by
increased efficiency of use of mineral N, but other explanations
for a reduction in leaching are possible: for example losses other
than leaching or enhanced storage in soil organic matter or other
retention in the system.
4.1. Are there losses of N from the system?
There is no evidence that more N is being lost under intercropping than sole cropping. The model calculates the amount of
Hauugaard-Nielsen et al. (2001) found greater and deeper
exploitation of soil by the roots of intercropped barley than by
its companion pea crop. Kage (1997) has used a model to argue
that the root density of faba beans is sparse enough to restrict
uptake relative to oats. Hauugaard-Nielsen et al. (2001) found
an increase in N fixation from 40% to 80% in pea intercropped
with barley compared with sole crops, and in different work the
same authors (2003) found an increase from 70% to 90%. Other
work has suggested lesser reliance on N-fixation (HauugaardNielsen et al., 2006), but this more recent work did not control
weeds and attributed the different behaviour to competition with
the pea crop for resources other than N. The assumption we
make in the model (Section 2.1.4) seems to be justified, that
the benefits of intercropping, say, a cereal with a legume derive
from the ability of the cereal to out-compete the legume for N
(Stern, 1993), probably because the cereal has a more efficient
root system (Kage, 1997). As a result, the legume usually derives
more of its N from the atmosphere, thus increasing the apparent
efficiency with which other N is used. Compared with sole crops
at a similar total density, the intercrops yield more and leave less
N for leaching than a legume sole crop but more than a cereal sole
crop (e.g. Hauugaard-Nielsen et al., 2003), most likely because
legume N only becomes available in soil after senescence or
harvest (Stern, 1993) and the legume does not appear to transfer
N to the wheat crop during the growing season.
4.3. Does an intercropped system enhance storage of N in
soil organic matter?
If N does not leach, more N is introduced through fixation
and the N already present is used more effectively then what
becomes of N under intercropping? Schröder et al. (1996) commented on the large difference in leaching between cropped and
fallow treatments at high N rates, attributing the difference to
an unknown extra sink in the grass and rye soils. A possible
candidate for this sink is the dry matter developed by the grass
A.P. Whitmore, J.J. Schröder / Europ. J. Agronomy 27 (2007) 81–88
and rye crops in autumn that does not survive the winter. The
N in this tissue will eventually re-mineralise, but perhaps not
within the winter and so not be at risk of leaching. This phenomenon is simulated in the model, but data is not available
to allow us to deduce the particular cold resistances of the rye
and the grass. The simulations of the N-uptake by both winter
crops were probably the least satisfactory outputs from the model
(data not shown). Schröder et al. (1996) also remarked that the
grass crop appeared to have a detrimental effect on maize yields
in 1988, 1991 and 1993 but not the other years. In agreement
with this, the difference between the modelled N uptake in the
maize grown after fallow and after grass crop falls in the order
1991 > 1988 > 1990 > 1993 > other 3 years, although these modelled differences were small. The model suggests that in addition
to competition between crops for N, there was competition in
these years between the growing maize crop and soil microbes
that were immobilising N as a result of decomposing the grass
residues (e.g. Thorup-Kristensen, 1994). Both Schröder et al.
(1996) and Hauugaard-Nielsen et al. (2003) found decreases in
N leaching in intercropped plots relative to other plots but also
deduced or found greater mineralisation of N after winter. Van
Dam (2006) has also found that N taken up by catch crops may
still be lost by leaching if it re-mineralises before the main crop
can take it up. Intercropping probably increases the input of
residues into the soil and understanding the fate of the N these
retain and then the timing with which they release N is essential
to managing intercrops to reduce N loss but maintain yield and
profit. If subsequent fertiliser applications are not reduced, the
leaching may increase in the longer term. Intercropped systems
may enrich soil organic matter that turns over slowly, but this
is difficult to measure and simulations over short timescales are
also not reliable.
4.4. Yield and profitability
Where both crops have a value, intercropping can benefit
both environment and profit (Fig. 5). Where the legume has no
cash advantage (e.g. Clements et al., 1997; Thorsted et al., 2006)
the benefits of intercropping may be more ambiguous because
a certain amount of the yield of the commercial crop must be
sacrificed in order to grow the companion crop. Profitability
depended more heavily on wheat in these mixtures, which was
the more valuable component (Bulson et al., 1997). A more
valuable second crop, however, might increase the profit of an
intercropped system. Even though the environmental benefits of
intercropping appear to be clear in Fig. 5, the argument in Section
4.3 still applies: that N released as a result of decomposition of
legume residues will be at risk of leaching during the following
winter and therefore requires control.
5. Conclusions
This modelling study suggests that intercropping holds the
potential to reduce nitrate leaching while maintaining yield
and possibly profitability, provided appropriate combinations of
crops are grown. However such a system probably needs careful
management in order to derive these benefits. If less nitrogen is
87
leached from under intercrops or if substantial amounts of N are
fixed, inputs to the system as a whole may need to be reduced in
proportion to any remineralisation of soil organic N, otherwise
subsequent losses are likely to be larger. Nitrogen mineralising
after harvest of a legume, intercropped or otherwise, may be at
risk of leaching unless a winter crop is planted.
Acknowledgements
Rothamsted Research is grant-aided by the Biotechnology
and Biological Sciences Research Council. We thank Lawrence
Clark for commenting on the manuscript. APW thanks Landcare
Research, New Zealand for hosting, and the Lawes Agricultural
Trust for funding, a period of study leave in 2005 and 2006.
References
Addiscott, T.M., 1977. A simple computer model for leaching in structured soils.
J. Soil Sci. 28, 554–563.
Addiscott, T.M., Whitmore, A.P., 1987. Computer simulation of changes in soil
mineral nitrogen and crop nitrogen during autumn, winter and spring. J.
Agric. Sci. (Cambridge) 109, 141–157.
Anon., 2000. Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for Community action in
the field of water policy. Official Journal L 327 22/12/2000, 1–73.
Ayisi, K.K., Putnam, D.H., Vance, C.P., Russelle, M.P., Allan, D.L., 1997. Strip
intercropping and nitrogen effects on seed, oil, and protein yields of canola
and soybean. Agron. J. 89, 23–29.
Berntsen, J., Hauugaard-Nielsen, H., Olesen, J.E., Petersen, B.M., Jensen, E.S.,
Thomsen, A., 2004. Modelling dry matter production and resource use in
intercrops of pea and barley. Field Crops Res. 88, 69–83.
Bouniols, A., Cabelguenne, M., Jones, C.A., Chalamet, A., Charpenteau, J.L.,
Marty, J.R., 1991. Simulation of soybean nitrogen nutrition for a silty clay
soil in southern France. Field Crops Res. 26, 19–34.
Bradbury, N.J., Whitmore, A.P., Hart, P.B.S., Jenkinson, D.S., 1993. Modelling
the fate of nitrogen in crop and soil in the years following application of 15 Nlabelled fertilizer to winter wheat. J. Agric. Sci. (Cambridge) 121, 363–379.
Brisson, N., Bussiere, F., Ozier-Lafoontaine, H., Tournebize, R., Sinoquet, H.,
2004. Adaption of the crop model STICS to intercropping. Theoretical basis
and parameterization. Agronomie 24, 409–421.
Bulson, B.A.J., Snaydon, R.W., Stopes, C.E., 1997. Effects of plant density on
intercropped wheat and field beans in an organic farming system. J. Agric.
Sci. (Cambridge) 128, 59–71.
Clements, R.O., Donaldson, G., Purvis, G., Burke, J., 1997. Clover: cereal
bicropping. Optimising cereal inputs: its scientific basis. Aspects Appl. Biol.
50, 467–469.
De Wit, C.T., Van Den Bergh, J.P., 1965. Competition between herbage plants.
Neth. J. Agric. Sci. 13, 212–221.
den Hollander, N.G., Bastiaans, L., Kropff, M.J., 2007. Clover as a cover crop
for weed suppression in an intercropping design. I. Characteristics of several
clover species. Eur. J. Agron. 26, 92–103.
Fowler, D.B., Limin, A.E., Rtichie, J.T., 1999. Low-temperature tolerance in
cereals: model and genetic interpretation. Crop Sci. 39, 626–633.
Gerwitz, A., Page, E.R., 1974. An empirical model to describe plant root systems.
J. Appl. Ecol. 11, 773–781.
Ghaffarzadeh, M., Prechac, F.G., Cruse, R.M., 1997. Strip intercropping effects
on yield and yield components of corn, grain sorgh um, and soybean. Agron.
J. 89, 893–899.
Giller, K.E. Day, J.M., 1985. Nitrogen fixation in the rhizosphere: significance
of natural and agricultural systems. In: Fitter, A.H. et al. (Eds.), Ecological
Interactions in Soil: Plants, Microbes and Animals: papers presented at a
British Ecological Society Meeting, York, 1984, British Ecological Society:
Special Publication 4, Blackwell Scientific, Oxford, pp. 127–147.
Hassink, J., Whitmore, A.P., 1997. A model of the physical protection of organic
matter in soils. Soil Sci. Soc. Am. J. 61, 131–139.
88
A.P. Whitmore, J.J. Schröder / Europ. J. Agronomy 27 (2007) 81–88
Hauugaard-Nielsen, H., Andersen, M.K., Jørnsgaard, B., Jensen, E.S., 2006.
Density and relative frequency effects on competitive interactions and
resource use in pea–barley intercrops. Field Crops Res. 95, 256–267.
Hauugaard-Nielsen, H., Ambus, P., Jensen, E.S., 2003. The comparison of nitrogen use and leaching in sole cropped versus intercropped pea and barley.
Nutr. Cycl. Agroecosyst. 65, 289–300.
Hauugaard-Nielsen, H., Ambus, P., Jensen, E.S., 2001. Temporal and spatial
distribution of roots and competition for nitrogen in pea-barley intercrops—a
field study employing 32 P technique. Plant Soil 236, 63–74.
Kage, H., 1997. Is low rooting density of faba beans a cause of high residual
nitrate content of soil at harvest? Plant Soil 190, 47–60.
Lesoing, G.W., Francis, C.A., 1999. Strip intercropping effects on yield and yield
components of corn, grain sorghum, and soybean. Agron. J. 91, 807–813.
Pomeroy, M.K., Andrews, C.J., Fedak, G., 1975. Cold hardening and dehardening responses in winter wheat and winter barley. Can. J. Plant Sci. 55,
529–535.
Schjoerring, J.K., Kyllingsbaek, A., Mortensen, J.V., Byskovnielsen, S., 1993.
Field investigations of ammonia exchange between barley plants and the
atmosphere. 1. Concentration profiles and flux densities of ammonia. Plant
Cell Environ. 16, 161–167.
Schröder, J.J., van Dijk, W., de Groot, W.J.M., 1996. Effects of catch crops on
nitrogen fluxes in a silage maize production system. Neth. J. Agric. Sci. 44,
295–315.
Schröder, J.J, ten Holte, L., van Dijk, W., de Groot, W.J.M., de Boer, W.A.,
Jansen, E.J. 1992. Effecten van wintergewassen op de uitspoeling van stikstof
bij de teelt van snijmaı̈s. PAGV report 148, 105 pp.
Stern, W.R., 1993. Nitrogen fixation and transfer in intercrop systems. Field
Crops Res. 34, 335–356.
Thorsted, M.D., Olesen, J.E., Weiner, J., 2006. Width of clover strips and wheat
rows influence grain yield in winter wheat/white clover intercropping. Field
Crops Res. 95, 280–290.
Thorup-Kristensen, K., 1994. The effect of nitrogen catch crop species on the
nitrogen nutrition of succeeding crops. Fert. Res. 37, 227–234.
Trenbath, B.R., 1999. Multispecies cropping systems in India—predictions of
their productivity, stability, resilience and ecological sustainability. Agroforestry Syst. 45, 81–107.
Trenbath, B.R., 1993. Intercropping for management of pests and diseases. Field
Crops Res. 34, 381–405.
Tsubo, M., Walker, S., Ogindo, H.O., 2005. A simulation model of cereal-legume
intercropping systems for semi-arid regions. I. Model development. Field
Crops Res. 93, 10–22.
Van Dam, A., 2006. Understanding the reduction of nitrogen leaching by catch
crops. Published PhD Thesis, Wageningen University, Wageningen, 171 pp.
Van Diepen, C.A., Wolf, J., Van Keulen, H., Rappoldt, C., 1989. WOFOST: a
simulation model of crop production. Soil Use Manage. 5, 16–24.
van Dijk, W., Schröder, J.J., de Groot, W.J.M., 1995. Effecten van wintergewassen op verliezen en benutting can stikstof bij de teelt van snijmaı̈s. PAGV
report 201, 97 pp.
Vinten, A.J.A., Whitmore, A.P., Bloem, J., Howard, R., Wright, F., 2002. Factors
affecting N immobilisation/mineralisation kinetics for cellulose, glucose and
straw amended sandy soils. Biol. Fert. Soils 36, 190–199.
Weir, A.H., Bragg, P.L., Porter, J.R., Rayner, J.H., 1984. A winter wheat crop
simulation model without water or nutrient limitations. J. Agric. Sci. (Cambridge) 102, 371–382.
Whitmore, A.P., 1995. Modelling the mineralization and immobilization, leaching and crop uptake of nitrogen during three consecutive years. Ecol. model.
81, 233–241.
Whitmore, A.P., 1996a. Describing the mineralization of carbon added to
soil in crop residues using second-order kinetics. Soil Biol. Biochem. 28,
1435–1442.
Whitmore, A.P., 1996b. Alternative kinetic laws to describe the turnover of the
microbial biomass. Plant Soil 181, 169–173.
Whitmore, A.P., 1996c. Modelling the release and loss of nitrogen after the
incorporation of vegetable crop residues. Neth. J. Agric. Sci. 44, 73–
86.
Whitmore, A.P., 2007. Describing the transformation of organic carbon and
nitrogen in soil using the MOTOR system. Comp. Electron. Agric. 55, 71–88.
Whitmore, A.P., Groot, J.J.R., 1994. The mineralization of N from finely or
coarsely chopped crop residues: measurements and modelling. Eur. J. Agron.
3, 103–109.
Whitmore, A.P., Groot, J.J.R., 1997. The decomposition of sugar beet residues:
mineralization versus immobilization in contrasting soil types. Plant Soil
192, 237–247.
Whitmore, A.P., Gunnewiek, H.K., Crocker, G.J., Klı́r, J., Körschens, M.,
Poulton, P.R., 1997. Modelling the turnover of carbon in soil using the
Verberne/MOTOR model. Geoderma 81, 137–151.
Whitmore, A.P., Handayanto, E., 1996. Simulating the effects of quality on the
decomposition of crop residues. In: Giller, K.E., Cadisch, G. (Eds.), Driven
by Nature. Wye College England, pp. 337–348.
Whitmore, A.P., Matus, F.J., 1996. The decomposition of wheat and clover
residues in soil: measurements and modelling. In: Van Cleemput, O., Hofman, G., Vermoesen, A. (Eds.), Progress in Nitrogen Cycling Studies, 68.
Kluwer Academic Publishers, pp. 465–469, Developments in Plant and Soil
Sciences.
Whitmore, A.P., Schröder, J.J., 1996. Modelling the change in soil organic C
and N in response to applications of slurry manure. Plant Soil 184, 185–
194.
Whitmore, A.P., Coleman, K.W., Bradbury, N.J., Addiscott, T.M., 1991. Simulation of nitrogen in soil and winter wheat crops: modelling nitrogen turnover
through organic matter. Fert. Res. 27, 283–291.
Whitmore, A.P., Addiscott, T.M., 1987. A function for describing nitrogen
uptake, dry matter production and rooting by wheat crops. Plant Soil 101,
51–60.
Whitmore, A.P., Addiscott, T.M., 1986. Computer simulation of winter leaching
losses of nitrate from soils cropped with winter wheat. Soil Use Manage. 2,
26–30.
Yang, L., Cai, Z., 2006. Effects of shading soybean plants on N2 O emissions
from soil. Plant Soil 283, 265–274.
Zhu, Y., Chen, H., Fan, J., Wang, Y., Li, Y., Chen, J., Fan, J., Yang, S., Hu, L.,
Leung, H., Mew, T.W., Teng, P.S., Wang, Z., Mundt, C.C., 2000. Genetic
diversity and disease control in rice. Nature 406, 718–722.