Chemosphere 63 (2006) 1892–1902

www.elsevier.com/locate/chemosphere

The impact of rice plant roots on the reducing conditions

in flooded rice soils
a,c a,c,* b,c
Gregory Doran , Philip Eberbach , Stuart Helliwell
a
CRC for Sustainable Rice Production, Farrer Centre, School of Agricultural and Veterinary Sciences,
Charles Sturt University, Locked Bag 588, Wagga Wagga 2678, NSW, Australia
b
CRC for Sustainable Rice Production, Farrer Centre, School of Science and Technology, Charles Sturt University,
Locked Bag 588, Wagga Wagga 2678, NSW, Australia
c
EH Graham Centre for Agricultural Innovation, Charles Sturt University, Locked Bag 588, Wagga Wagga 2678, NSW, Australia

Received 24 August 2005; received in revised form 7 October 2005; accepted 16 October 2005
Available online 5 December 2005

Abstract

The impact of oxygen diffusion from plant roots on the soil redox in the root zone in flooded rice bays was inves-
tigated using two Australian rice growing soils. The rates of production of Fe(II) and Mn(II) in pore water resulting
from the reduction of soil minerals was used to gauge the extent of development of anaerobic conditions and the time
taken for equilibrium to establish. Soil concentrations of readily reducible Fe were 13–28 times greater than Mn, mak-
ing Fe a more reliable indicator of redox conditions than Mn. In addition, Mn(II) concentrations reached equilibrium
far more rapidly than Fe, which made the identification of any contribution to soil redox by oxygen diffusing from rice
plant roots difficult to observe. Dissection of soil cores showed that more than 80% of the rice root mass occurred in the
top 4 cm of soil, suggesting that any contribution roots may make to the redox potential of the flooded soils would
occur in this region. However, studies conducted indicated that the diffusion of oxygen from the surface floodwater into
soil pore water in the 2.5 cm layer of soil was so substantial that it would mask any contribution made by rice plant
roots to the overall soil redox in this root zone.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Redox; Iron; Manganese; Anaerobic; Reduction; Pore water

1. Introduction

The flooding of rice bays rapidly generates anaerobic

* soil microbes during decomposition of organic matter
Corresponding author. Address: CRC for Sustainable Rice
Production, Farrer Centre, School of Agricultural and Veter-
inary Sciences, Charles Sturt University, Locked Bag 588,
Wagga 2678, NSW, Australia. Tel.: +61 2 6933 2830; fax: +61 2
6933 2924.
E-mail address: peberbach@csu.edu.au (P. Eberbach).
conditions in soil pore water within a few days of flood-
ing. Dissolved oxygen in the floodwater is consumed by
(Tsukano, 1986), and cannot be adequately replenished
by diffusion from the atmosphere, resulting in an oxygen
deficit. Soil microfauna capable of using terminal elec-
tron acceptors (TEAs) other than molecular oxygen
are able to survive under these anaerobic conditions.

0045-6535/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2005.10.027
 G. Doran et al. / Chemosphere 63 (2006) 1892–1902
Reduction of NO 3 to N2 generally occurs first, followed
by Mn(III/IV)/Mn(II), Fe(III)/Fe(II) , SO2 4 =H2 S and
organic matter/CH4, however, considerable overlap
between the different redox systems has been observed
(Gao et al., 2002).
Re-oxidation of these reduced chemical species may
occur when the reduced form encounters dissolved oxy-
gen which has diffused into soil pore water from surface
floodwater, from the roots of aquatic plants, or from the
atmosphere during sampling (Doran et al., 2005). Plants
such as rice, are able to grow under flooded conditions
due to hyper-developed intercellular air spaces within
leaves, stems and roots which allow air to diffuse from
foliar regions of the plant to the roots. Kumazawa
(1984) reported that oxygen diffusion from rice plant
roots into the soil caused the oxidation of reduced soil
components within a few millimetres of the root surface
of young roots, but this effect diminished with the age of
the root. This work demonstrated that Fe(II) in soil pore
water was oxidised to ferric oxide and formed concentric
layers around the root surface as the root aged, until the
epidermis eventually became impregnated with iron
oxide. Additionally, it demonstrated that the potential
for soil components, including applied organic chemi-
cals, to be chemically or microbially affected indirectly
as a consequence of oxygen diffusion from rice roots
exists.
Pesticides are used extensively in agriculture in an
attempt counteract the estimated $4–5 billion annual
cost (Kookana et al., 1998), of which more than $3 bil-
lion has been attributed to losses in economic produc-
tion of crops due to weeds (Combellack, 1989).
However, as most crops are grown under aerobic soil
conditions, most of our understanding of pesticide deg-
radation rates and mechanisms has evolved from studies
conducted under aerobic conditions. Given that rice is
grown under flooded conditions, and that the soil is
anaerobic for a large amount of the time, some of these
compounds may behave unexpectedly. For example,
MCPA has been shown to degrade rapidly under aero-
bic conditions, but slowly under anaerobic condi-
tions (Vink and van der Zee, 1997). Conversely, DDT
degrades rapidly under anaerobic conditions, but slowly
under aerobic conditions (Guenzi and Beard, 1968), as a
consequence of reductive co-metabolism to DDD by
anaerobic microorganisms (Aislabie et al., 1997). Such
dramatic variations in the rates of degradation of pesti-
cides may indicate a dependence upon the presence of
oxygen for chemical degradation, or a particular aerobic
or anaerobic organism for degradation (Kuhlmann and
Kaczmarczyk, 1995), and may be further complicated by
the diffusion of oxygen from the roots of aquatic plants.
The diffusion of oxygen from developing rice plant
roots, in conjunction with the secretion of carbohydrates
and amino acids, allow the survival of large aerobic
microbial populations in the rhizosphere (Yu et al.,
1893
2003). In these zones, high levels of oxygen diffusion
from roots may result in an increase in the redox poten-
tial of the soil, and thus enhance rates of degradation of
rice pesticides which degrade primarily by aerobic mech-
anisms. Therefore, the aims of the following work were
to determine the appropriateness of using Fe(II) or
Mn(II) as indicators of the extent of anaerobic condi-
tions, the time taken for equilibrium to establish, the
impact of rice plant roots on redox conditions at depth,
and to determine the effect of oxygen diffusion from the
atmosphere.
2. Methodology
Development of reducing conditions in flooded soil
were gauged using the rates of production of Fe(II)
and Mn(II) in soil pore water, as work by Gao et al.
(2002) indicated a better correlation of the soil EH with
the production of Mn(II) and Fe(II) than other redox
species, such as dissolved oxygen, NO 2
3 or SO4 .
2.1. Selection of soil type and rice variety
The two rice growing soils from south-western NSW,
Australia used for experimentation were a Birganbigil
clay loam from Yanco (3436 0 S, 14624 0 E) and a
Wunnamurra self-mulching clay from Coleambally
(3447 0 S, 14559 0 E), and were classified as a brown chro-
mosol and a grey vertosol, respectively (Isbell, 1996).
These soils were chosen as being representative of rice
soil due to their relative production levels. While the
Coleambally and Yanco soils had similar total Fe con-
centrations (4.9% and 4.8%, respectively), their readily
reducible Fe contents differed significantly (0.28% and
0.48%, respectively). Prior to sowing, rice seed was ger-
minated by placing on moist filter paper in Petri dishes
and incubated (4 days) using simulated day and night
temperatures (27 and 22 C, respectively). The cultivar
Amaroo (Oryza sativa cv Amaroo) was sown at a rate
of 150 kg ha1 and plants were thinned-out to achieve
an establishment rate of 200–300 plants m2.
2.2. Experimental setup
Soil cores were collected using lengths of PVC sewer
pipe (15 cm ID · 45 cm long) which had the outer edge
of the end centimetre of each tube tapered. Eight PVC
tubes were pushed into the ground to a depth of 20 cm
at each of the two field sites. After insertion, the tubes
were removed and carefully transported back to the
glasshouse, where the bottom end of each tube was
sealed using a PVC plate (3 mm · 20 cm · 20 cm).
Holes (22 mm) were drilled through the wall of the
pipe at 2.5 and 7.5 cm (nominally the upper and lower
rice root zone, respectively), and 12.5 cm (sub-root zone)
1894 G. Doran et al. / Chemosphere 63 (2006) 1892–1902
from the soil surface. Through each of these holes, a
20 mm hole was then drilled horizontally into the soil
to the middle of the core and one assembled sipper
was placed in each hole, with the removed soil used to
backfill the hole. Sampling frits (sippers) were used to
sample pore water for Fe(II) and Mn(II) analysis and
were prepared from air stones normally used to aerate
fish tanks. Water flow was restricted through the air
stones by coating them with a latex gap sealant, leaving
a 4–5 mm wide region free around the stoneÕs circumfer-
ence. The stones were then attached to Teflon tubing
(30 cm · 1.6 mm OD · 0.5 mm ID), and had an internal
volume of <0.7 ml. The tubing from the sipper was
threaded through a patching plate, which was then
sealed with PVC pipe cement followed by multipurpose
silicone sealant (Fig. 1A). The end of the tubing was
capped to prevent water leakage and allow water sam-
pling. The cores were then flooded to a depth of several
centimetres using Murrumbidgee River water.
2.3. Oxygen translocation and soil redox
Prior to sowing, the cores were allowed to stand over-
night to allow water percolation and then topped up to
3 cm at their shallowest point. Urea was applied at field
rate (185 kg ha1) by adding an aliquot of urea solution
to each tube (5 ml, 65 g l1), followed by the sowing of
the pre-germinated rice seed. This corresponded to 11
seeds per tube, which was reduced to 5 plants/tube
approximately 4–5 days after establishment in accor-
dance with the 50% establishment rate expected in the
field. Each treatment was replicated 4 times, with four
tubes for each soil type remaining free of rice. The exper-
iment was conducted in a glasshouse at temperatures
ranging from 15 to 35 C and a bank of five iridescent
lights (60 W) was used to extend the growing day from
glass
square
air/water
interface
A B
water/soil
interface
2.5 cm
sampling
7.5 cm
ports
12.5 cm
PVC
square
Fig. 1. Structure of soil core and soil pore water sampling system for oxygen translocation (A) and tube modified to prevent
atmospheric oxygen diffusion (B).
5 am to 8 pm to simulate summer growing conditions.
Water depth was maintained at 3 cm until 15 days after
flooding (DAF), at which time the depth was increased
to 5 cm. Water depth was increased to 10 cm at mid til-
lering (3 shoots/plant). Yellowing of plant leaves was
observed at 51 DAF and urea was applied at field
rate (62 kg ha1) by adding an aliquot of urea solution
to each tube (5 ml, 22 g l1). Ammonium sulfate
(123 kg ha1) was applied at 58 DAF to counteract the
continuing yellowing of leaves. Urea was again applied
(22 g l1) at panicle initiation, and water depth was
increased to 20 cm over the next 5 days and then
decreased to 10 cm at flowering. The surface water was
allowed to decrease via natural evaporation over the
next few weeks and was only increased if evaporation
was excessive. Surface water was allowed to evaporate
completely when the rice grains reached the late dough
stage.
2.4. Dissolved oxygen replenishment in the water
column and soil redox
Three cores of each soil type on which no rice had
been grown were selected for subsequent experimenta-
tion under conditions which prevented oxygen from dif-
fusing into the water column, thus allowing unhindered
observation of redox processes. Gas inlet and outlet
ports were added to the top of each PVC tube (two of
each soil type) to allow nitrogen to be purged through
the headspace, removing some of the gaseous products
of reduction, and any oxygen produced by photosynthe-
sis. A glass square (20 cm · 20 cm · 5 mm), silanised
with Coatasil (dimethyl dichlorosilane) on one side to
minimise water condensation, was attached to the top
of each tube with silicone sealant (Fig. 1B). This pre-
vented air diffusion into the tube headspace whilst allow-
gas inlet/
outlet ports
 G. Doran et al. / Chemosphere 63 (2006) 1892–1902
ing light penetration for photosynthesis by soil and
waterborne organisms. The cores were flooded with irri-
gation water to a depth of 3 cm using one of the
gas ports and nitrogen was purged through the tubes
(10 ml min1) to remove oxygen already present. The
remaining cores which had not been covered with glass
squares were flooded to a depth of 3 cm and served as
aerobic controls. The water and nutrient application
schedule described in the previous section was followed
for all tubes.
2.5. Water sampling and analysis
The cap was removed from the sipper and an aliquot
of water (1 ml) was drawn and discarded to flush the
sampling line. A second aliquot (3.5–4 ml) was drawn
and then passed through sequential 0.45 lm · 25 mm
and 0.22 lm · 13 mm Teflon syringe filters (Millipore,
Massachusetts) to remove particles which may interfere
with sample preparation and spectrophotometric analy-
sis. The filters were flushed with excess solution and the
remainder (1.5 ml) was expelled into an Eppendorf
tube (1.5 ml) containing HCl (50 ll, concentrated). The
Eppendorf tube was sealed, shaken and reserved for
Fe(II) and Mn(II) analysis. Mn(II) analysis was per-
formed using a Varian SpectraAA20 Plus Atomic
Absorption Spectrometer at 279.5 nm using a slit width
of 0.2 nm, and air and acetylene flow rates of 2 and
8 ml min1, respectively. Fe(II)/total Fe analysis was
performed using the ferrozine method as described in
Doran et al. (2005).
2.6. Root distribution and soil analysis
Soil cores on which rice was grown were cut trans-
versely into 2-cm thick slices. After washing the soil off
the roots, the roots were oven dried (70 C, 2 days)
and weighed. Additional Coleambally soil cores on
which rice was not grown were cut into slices trans-
versely. The red material that lined the holes which
had been occupied by decayed rice plant roots was col-
lected and dissolved in HCl (1 ml, 30% v/v), diluted with
distilled water and analysed by AAS for total Fe at
248.3 nm using a slit width of 0.2 nm, and air and acet-
ylene flow rates of 2 and 8 ml, respectively. Determi-
nation of readily reducible Fe and Mn in soil used
TammÕs Reagent (acid ammonium oxalate) and pro-
Table 1
Variation of readily reducible Fe and Mn concentrations in soils with depth (mean ± standard error, n = 4)
Soil depth (cm) Yanco
1
Fe (mg kg ) Mn (mg kg )
1.5–3.5 4530 ± 370 111 ± 3
6.5–8.5 4810 ± 280 184 ± 4
11.5–13.5 4870 ± 260 196 ± 7
1895
vided an estimate of the ‘‘active’’ iron oxides that result
from a poorly crystalline structure (Loeppert and Ins-
keep, 1996). The, ashing step in the original method
was replaced with decomposition using hot aqua regia,
followed by analysis using flame AAS (Table 1).
2.7. Statistical analysis
Statistical analysis was performed using repeated
measures ANOVA (Howell, 2002).
3. Results and discussion
The existing model of soil redox under flooded condi-
tions has been described by Ponnamperuma (1984) and
Tsukano (1986). In this model, the floodwater and top
few millimetres of surface soil of a rice bay have high
redox potentials. The soil immediately beneath this layer
may exhibit a negative redox potential, while the soil
below the plough layer remains positive. These condi-
tions are generated as a result of the continual cycling
between aerobic and anaerobic conditions under rice
agriculture from season to season, and differ from envi-
ronments which remain permanently under flood, such
as lake and river sediments (Tsukano, 1986). Thus,
weaker reducing conditions in the surface layer of the
soil and decreased reduction of Fe(III) and Mn(III/IV)
species are predicted compared to subsequent soil layers
due to a greater influence exerted by dissolved oxygen in
the water column.
3.1. Establishment of equilibrium and the effect
of sampling depth
The rates of production of Fe(II) and Mn(II) in soil
pore water were used to gauge developing redox condi-
tions in two flooded soils. In general, Fe(II) and Mn(II)
concentrations increased rapidly, plateaued and in most
instances, gradually decreased (Figs. 2 and 3). Initial
rates of Fe(II) production were similar for the first few
days regardless of depth, suggesting that microbial
action was fairly uniform throughout the soil core. As
predicted by the redox potential models of Ponnampe-
ruma (1984) and Tsukano (1986), Fe(II) and Mn(II)
concentrations in the pore water of the 2.5 cm layer
(upper root zone) plateaued at lower values than in
Coleambally
1
Fe (mg kg1) Mn (mg kg1)
2590 ± 110 142 ± 6
2690 ± 180 186 ± 6
3230 ± 160 238 ± 9
1896 G. Doran et al. / Chemosphere 63 (2006) 1892–1902

Yanco No Rice A
250 Yanco Rice
Coleambally No Rice
200 Coleambally Rice

150

100

50

B
250
Fe(II) concentration (mg l-1)

200

150

100

50

C
250

200

150

100

50

0 20 40 60 80 100 120

Time (days)

Fig. 2. Fe(II) concentration on Yanco and Coleambally soils at 2.5 cm (A), 7.5 cm (B) and 12.5 cm (C) in the presence and absence of
rice plants (mean ± standard error, n = 4).

the 7.5 (lower root zone) and 12.5 cm layers (sub-root
zone). This was attributed to the presence of photosyn-
thetic algae on the soil surface (Ponnamperuma, 1984)
or diffusion of water from the water column into the
pore water. Both events would have resulted in elevated
dissolved oxygen levels in the surface soil, causing the
oxidation of Fe(II) and Mn(II) to their insoluble forms.
Higher plateau concentrations of Fe(II) and Mn(II)
were observed for the Yanco soil compared with the
Coleambally soil. This was considered as a direct result
of the Yanco soil containing more than 1.7 times the
amount of readily reducible Fe than the Coleambally
soil, as described in Section 2. This in accordance
with work using Mexican, Spanish (Velazquez et al.,
2004) and Brazilian soils (de Mello et al., 1998), which
showed that greater amorphous iron content in soil
generally resulted in higher Fe(II) concentrations in
water.
A study by Gao et al. (2002) using Californian rice
bays compared a series of redox pairs to gauge the
 G. Doran et al. / Chemosphere 63 (2006) 1892–1902 1897

Yanco No Rice
A
20 Yanco Rice
Coleambally No Rice
Coleambally Rice
15

10

B
20
Mn(II) concentration (mg l-1)

15

10

C
20

15

10

0 20 40 60 80 100 120
Time (days)

Fig. 3. Mn(II) concentration on Yanco and Coleambally soils at 2.5 cm (A), 7.5 cm (B) and 12.5 cm (C) in the presence and absence of
rice plants (mean ± standard error, n = 4).

establishment of reducing conditions in soil pore water
under flooded rice agriculture. Their work showed that
Fe(II) concentrations in soil failed to reach a threshold
value within the 106 day period after commencement of
flooding, whilst Mn(II) concentrations peaked after
approximately 70 DAF. Conversely, the present study
indicated an equilibrium concentration occurred as
early as 27 DAF for Fe(II) and Mn(II), but this was
dependent upon the sampling depth (Table 2). The dif-
ference in establishment of equilibrium time between
the two experiments was attributed to differences in
environmental factors such as pH, temperature, soil
composition, and initial microbial population size and
diversity. Additionally, the work conducted here indi-
cated that the time taken for establishment of equilib-
rium concentrations of Fe(II) and Mn(II) increased
with sampling depth, as predicted by the existing
model (Ponnamperuma, 1984; Tsukano, 1986), and is
1898 G. Doran et al. / Chemosphere 63 (2006) 1892–1902

Table 2 3.2. Impact of plant roots and depth of influence

Approximate number of days required for Fe(II) and Mn(II)
concentrations to reach equilibrium Soil removed from the lining of old root holes in
Sampling depth Coleambally Yanco Coleambally soil was qualitatively analysed by AAS
(cm)
Rice No rice Rice No rice and found to contain Fe, which was presumed to be in
the Fe(III) state due to its precipitated state. Due to
Fe
the Fe layer extending approximately 1 mm from the
2.5 27 27 34 43
7.5 27 27 43 43
surface of the pore into the bulk soil, its presence
12.5 34 27 81 43 appeared to be the result of diffusion of oxygen from
the surface of the rice plant root which oxidised Fe(II)
Mn in the pore water, as described by Kumazawa (1984),
2.5 27 27 27 27
rather than accumulation of Fe(III) in the tissue of the
7.5 65 65 43 43
now-decayed root. Analysis of root abundances at the
12.5 65 65 50 50
conclusion of the experiment revealed that more than
80% of the root mass occurred in the top 4 cm of soil
presumably a consequence of a lesser influence of dis- (Fig. 4). Hence, a greater amount of oxygen would be
solved oxygen from the surface floodwater, which orig- expected to diffuse from rice roots, which would result
inated from the atmosphere. The study by Gao et al. in the greatest affect on soil redox at the 2.5 cm region
(2002) only provided an indication of the overall redox compared to the 7.5 cm or 12.5 cm regions of the soil
conditions in the top 10 cm of soil, as their sampling profile. Fig. 2 shows that Fe(II) concentrations were sig-
system consisted of an open-ended piezometer posi- nificantly different (P < 0.05) in Yanco cores with rice
tioned 5 cm below the soil surface on 5 cm bed of sand. than without rice in the 2.5 cm layer, whereas the
Thus, water in the 0–10 cm soil profile would be 7.5 cm layer showed no difference. Whilst the cores with-
expected to be sampled. out rice produced greater Fe(II) concentrations at
The levels of Fe(II) in pore water were higher on 12.5 cm for the first 60 DAF, the rice treatment pro-
Coleambally and Yanco soils than in the Californian duced more Fe(II) until 95 DAF, but the difference
rice soil used by Gao et al. (2002). Comparison with between the treatments was not significant. Conversely,
the Californian field data is difficult as neither readily the Coleambally soil showed no significant difference
reducible Fe or soil organic matter data were reported. between the cores with and without rice, and may be
Both Fe reduction and soil organic matter has been an artefact of the heterogenous distribution of readily
shown to be coupled (Borges et al., 2001; Velazquez reducible Fe(III) in the soil relative to the Yanco soil.
et al., 2004), and may be influential in the amount of Like Fe(II) concentrations on Yanco cores in the
Fe(III) reduced. While this study attempted to minimise 2.5 cm sampling layer, Mn(II) concentrations were
the effect of dissolved oxygen in the pore water, dissolu- lower in the presence than the absence of rice after only
tion of atmospheric oxygen by wave action on the rice a few weeks. Similarly, the 7.5 cm layer exhibited lower
bays would have occurred in the field-based Californian Mn(II) concentrations after about 40 DAF, suggesting
study, which was also expected to have an influence on that the presence of rice was influencing the redox
the amount and rate of Fe(II) produced in the surface
soil.
Both Fe(II) and Mn(II) levels in this study were 70
0-2 cm
found to increase with time and concentration levels 60 2-4 cm
generally plateaued after 30–40 days, and in some 4-6 cm
%Dry root mass

instances, Mn(II) concentrations began to decrease. 50 6-8 cm

8-10 cm
Gao et al. (2002) experienced similar trends for Mn(II)
40
and attributed this to the precipitation of Mn(II) as
MnS. Since sulfate reduction to sulfide is known to gen- 30
erally occur after the reduction of Mn and Fe minerals
as the soil redox potential becomes more negative 20
(Ponnamperuma, 1984), a delay in the formation of
10
FeS and MnS would be expected and may be the reason
for the observed decrease in the Fe(II) concentration at 0
the 2.5 cm depth on the Yanco soil when no rice was Coleambally Yanco
grown. Pockets of gas were observed during water sam-
Soil type
pling and smell indicated the presence of H2S and possi-
bly other sulfurous compounds, supporting the concept Fig. 4. Rice plant root mass in soil as a function of depth after
of sulfide precipitation. plant maturation and death after approximately 5 months.
 G. Doran et al. / Chemosphere 63 (2006) 1892–1902 1899

conditions in the soil. Mn(II) concentrations at the
12.5 cm sampling depth appeared lower in the absence
than in the presence of rice. However, the difference
between the presence and absence of rice was only signif-
icantly different (P < 0.05) on the Yanco soil at 12.5 cm
and not at all on the Coleambally soil. It is expected that
considerable time would be required before developing
roots reached the 7.5 cm layer of soil, and no difference
between the treatments would therefore be expected.
Additionally, as no difference between the treatments
would be expected at 12.5 cm as rice roots tend not pen-
etrate the soil beyond a depth of 10 cm, no logical expla-
nation for these observation can be reached from these
data.
3.3. Effect of O2 diffusion
When the tops of the tubes were sealed with glass to
prevent oxygen diffusing into the water column and soil
pore water, the Fe(II) and Mn(II) concentrations in soil

Yanco Open A
Yanco Sealed
80
Coleambally Open
Coleambally Sealed
60

40

20

B
80
Fe(II) concentration (mg l-1)

60

40

20

C
80

60

40

20

0 20 40 60 80 100 120
Time (days)

Fig. 5. Fe(II) concentration on Yanco and Coleambally soils at 2.5 cm (A), 7.5 cm (B) and 12.5 cm (C) open to the air and sealed
under N2 purge (mean ± standard error, n = 3).
1900 G. Doran et al. / Chemosphere 63 (2006) 1892–1902

pore water for both soils increased with time (Figs. 5
and 6). The threshold Fe(II) and Mn(II) concentrations
observed for each soil type at soil depths of 2.5, 7.5 and
12.5 cm were lower than for the initial redox potential
experiments conducted when oxygen was able to diffuse
into the water column, and may have been due to the
removal of some of the readily reducible Fe and Mn in
the previous experiment. The Fe(II) concentration in
the 2.5 cm layer showed threshold values similar to those
in the 7.5 and 12.5 cm layers in this experiment. The
anaerobic gases which are generated in flooded environ-
ments normally react with dissolved oxygen in the water
column, or diffuse from the water column into the atmo-
sphere. As the placement of glass squares on top of the
PVC tube (Fig. 1B) prevented oxygen from re-entering
the water column, it also prevented the trapped anaero-
bic gases above the water column from being removed.
Consequently, the upper 1 cm of soil which is normally
oxic, was replaced with a more anaerobic environment
resulting in more extensive and rapid reduction of read-

Yanco Open A
10 Yanco Sealed
Coleambally Open
Coleambally Sealed
8

B
10
Mn(II) concentration (mg l-1)

C
10

0 20 40 60 80 100 120
Time (days)

Fig. 6. Mn(II) concentration on Yanco and Coleambally soils at 2.5 cm (A), 7.5 cm (B) and 12.5 cm (C) open to the air and sealed
under N2 purge (mean ± standard error, n = 3).
 G. Doran et al. / Chemosphere 63 (2006) 1892–1902
ily reducible Fe in the soil. For this reason, the Fe(II)
concentration in the 2.5 cm layer was hypothesised to
be greater on both soils in sealed tubes. In fact, the dif-
ference in the Fe(II) concentration in the 2.5 cm layer
was found to differ significantly (P < 0.05) between
sealed and unsealed treatments, suggesting that diffusion
of atmospheric oxygen into the water column was able
to affect the concentration of Fe(II) in the pore water
of the surface layer of soil.
The Mn(II) concentration in the 2.5 cm layer in the
Yanco soil was lower than the 7.5 and 12.5 cm layers,
and may have been due to the loss of Mn(II) into the
water column, or depletion of Mn(II) in the previous
experiment. However, the latter seems unlikely due to
the abundance of Mn in soil and the small pore water
volumes removed in the previous experiment. Since the
Fe(II) concentrations were similar throughout the
Yanco soil profile but Mn(II) concentrations was lower
in the surface layer, re-oxidation of Mn(II) may have
been possible. However, based on the recognised E0 val-
ues for the Fe(II) and Mn(II) redox couples, it is unlikely
that Mn(II) would be oxidised before Fe(II). Thus, sim-
ilar Fe(II) concentrations throughout the Yanco soil
profile indicate that Mn(II) re-oxidation may have
occurred.
As with the Fe(II) concentration, the difference in the
Mn(II) concentration in the 2.5 cm layer differed signif-
icantly (P < 0.05) between sealed and unsealed treat-
ments, indicating that diffusing atmospheric oxygen
into the water column in the unsealed tubes was able
to re-oxidise Mn(II) to insoluble forms of Mn. Addition-
ally, the significant difference between Fe(II) and Mn(II)
concentrations in sealed and unsealed tubes indicates
that the lower Fe(II) and Mn(II) concentrations
observed in 2.5 cm layer of both soils in the previous
experiment was due to the diffusion of atmospheric oxy-
gen into the water column, which re-oxidised Fe(II) and
Mn(II) in pore water in the surface layer.
4. Conclusion
The reduction of Fe and Mn minerals was used as an
indicator of the ability of rice plants to control soil redox
under flooded conditions. Readily reducible Fe was
present in soils at 13 and 28 times the Mn concentration
in Coleambally and Yanco soils, respectively, and was
thus a more reliable indicator of redox conditions than
Mn. Furthermore, experiments showed that the Fe(II)
concentration did not plateau as rapidly as the Mn(II)
concentration, and that the Mn(II) concentration did
not increase as dramatically with time as the Fe(II) val-
ues. The use of Fe(II) as a redox indicator also had the
advantage of being more readily determined with fewer
assumptions than Mn(II) because Fe(II) could be deter-
mined directly using spectrophotometric methods,
1901
whereas determination of Mn(II) was by atomic absorp-
tion spectrometry and relied on the assumption that all
soluble Mn was in the Mn(II) form. In fact, this may
not be the case as water samples were preserved for anal-
ysis using HCl. This may have promoted dissolution of
Mn(III) and (IV) particles which were sufficiently small
to pass through the 0.22 lm filters. However, monitor-
ing of Mn(II) concentrations in addition to Fe(II) con-
centrations may still be beneficial in that sampling
errors or other anomalous events may be identifiable
if spikes in the concentrations for both elements are
observed concurrently.
Pesticides adsorbed to the top few centimetres of
flooded rice bay soils will most likely be retained in an
oxic environment as a consequence of diffusion of oxy-
gen from the surface floodwater into soil pore water.
Data suggested that oxygen diffusion from Amaroo rice
plant roots in a flooded environment did not have a sig-
nificant impact on the overall soil redox, and for this
reason, the fate of adsorbed pesticides would be influ-
enced more by the oxic environment of the surface soil,
and only influenced by the anoxic conditions if lodged
deeper in the soil profile. However, more mobile pesti-
cides may be subject to more localised effects in the rhi-
zosphere in subsequent soil layers. Additionally, the
effect of oxygen diffusing into the water column from
the atmosphere appeared to mask the effect that rice
plant roots had in the top few centimetres of soil. Thus,
whilst the oxygen diffusing from the rice plant root may
not be substantial enough to influence the overall fate of
applied pesticides, the exudation of proteins and carbo-
hydrates into the rhizosphere by rice plant roots may
contribute to biotic degradation of pesticides (Yu
et al., 2003).
Acknowledgements
The authors would like to thank the CRC for Sus-
tainable Rice Production providing the funding to
undertake the work reported, and the EH Graham Cen-
tre for Agricultural Innovation for financially support-
ing the senior author during the writing up of this work.
References
Aislabie, J.M., Richards, N.K., Boul, H.L., 1997. Microbial
degradation of DDT and its residues—a review. New
Zealand J. Agric. Res. 40, 269–282.
Borges, M., De Mello, J.W.V., Abrahao, W.A.P., Jordao, C.P.,
Simas, F.N.B., 2001. Methods for evaluation of easily-
reducible iron and manganese in paddy soils. Commun. Soil
Sci. Plant Anal. 32, 3009–3022.
Combellack, J.H., 1989. The importance of weeds and the
advantages and disadvantages of herbicide use. Plant Prot.
Quart. 4, 14–32.
1902 G. Doran et al. / Chemosphere 63 (2006) 1892–1902
de Mello, J.W.V., Barron, V., Torrent, J., 1998. Phosphorus
and iron mobilization in flooded soils from Brazil. Soil Sci.
163, 122–132.
Doran, G., Helliwell, S., Eberbach, P., 2005. Extraction of
fipronil and thiobencarb from anaerobic water samples
using solid-phase extraction. J. AOAC Int. 88, 854–859.
Gao, S., Tanji, K.K., Scardaci, S.C., Chow, A.T., 2002.
Comparison of redox indicators in a paddy soil during
rice-growing season. Soil Sci. Soc. Am. J. 66, 805–817.
Guenzi, W.D., Beard, W.E., 1968. Anaerobic conversion of
DDT to DDD and aerobic stability of DDT in soil. Soil Sci.
Soc. Am. Pro. 32, 522–524.
Howell, D.C., 2002. Statistical Methods for Psychology, fifth
ed. Wadsworth, California, USA.
Isbell, R.F., 1996. The Australian Soil Classification. CSIRO
Publishing, Melbourne, Australia.
Kookana, R.S., Baskaran, S., Naidu, R., 1998. Pesticide fate
and behaviour in Australian soils in relation to contamina-
tion and management of soil and water: a review. Aust. J.
Soil Res. 36, 715–764.
Kuhlmann, B., Kaczmarczyk, B., 1995. Biodegradation of the
herbicides 2,4-dichlorophenoxyacetic acid, 2,4,5-trichloro-
phenoxyacetic acid, and 2-methyl-4-chlorophenoxyacetic
acid in a sulfate-reducing aquifer. Environ. Toxicol. Water
10, 119–125.
Kumazawa, K., 1984. Physiological specificity of rice root in
relation to oxidizing power and nutrient uptake. In: Biology
of Rice. Elsevier Science Publishers, Amsterdam.
Loeppert, R.L., Inskeep, W.P., 1996. Chapter 23—Iron. In:
Sparks, D.L. (Ed.), Methods of Soil Analysis. Part 3:
Chemical Methods. Soil Science Society of America, Wisc-
osin, USA, pp. 639–664.
Ponnamperuma, F.N., 1984. Effects of flooding on soils. In:
Flooding and Plant Growth, Orlando, FL, pp. 9–45.
Tsukano, Y., 1986. Transformations of selected pesticides in
flooded rice-field soil—a review. J. Contam. Hydrol. 1, 47–63.
Velazquez, M., Del Campillo, M.C., Torrent, J., 2004. Tempo-
rary flooding increases iron phytoavailability in calcareous
vertisols and inceptisols. Plant Soil 266, 195–203.
Vink, J.P.M., van der Zee, S.E.A.T.M., 1997. Effect of oxygen
status on pesticide transformation and sorption in undis-
turbed soil and lake sediment. Environ. Toxicol. Chem. 16,
608–616.
Yu, Y.L., Chen, Y.X., Luo, Y.M., Pan, X.D., He, Y.F., Wong,
M.H., 2003. Rapid degradation of butachlor in wheat
rhizosphere soil. Chemosphere 50, 771–774.