Biol Fertil Soils (1995) 19:333-342

9 Springer-Verlag 1995

J.L. G a u n t 9 H.-U. Neue 9 K.G. Cassman - D.C. Olk J. R.M. Arah 9 C. Witt 9 J.C.G. Ottow 9 I.F. Grant

Microbial biomass and organic matter turnover in wetland rice soils

Received: 2 February 1994

Abstract A decline in rice yields has been associated with intensification of rice production. In continuously irrigated systems this has been attributed to a decline in soil N supply. Nutrient mineralisation and immobilisation is constrained by the quantity and nature of the organic substrates and the physico-chemical environment of the soil system itself. A flooded soil is very different from an aerobic one; electron acceptors other than oxygen have to be used. The transition to continuously anaerobic conditions associated with the intensification of wetland rice systems affects their organic matter turnover and may adversely affect their productivity.

Key words Mineralisation 9 Immobilisation 9 Humification 9 Microbial biomass 9 Oryza sativa L. 9 Intensive production 9 Continuous flooding 9 Yield decline

J.L. Gaunt (~) ~ 9 I.F. Grant Natural Resources Institute, Central Avenue, Chatham Maritime, Kent ME4 4TB, UK J.L. Gaunt . H.-U. Neue - K.G. Cassman 9 D.C. Olk J. R. M. Arah 9C. Witt International Rice Research Institute, PO Box 933, 1099 Manila, Philippines J. R.M. Arah Scottish Agricultural College, West Mains Road, Edinburgh EH9 3 JG, UK C. Witt 9 J.C.G. Ottow Instkut fur Angewandte Mikrobiologie Justus-Liebig-Universit~it, Senckenbergstrasse 3, D-35390 Giessen, Germany Current address:

~Rothamsted Experimental Station, Harpenden, Hertfordshire, UK

Introduction Approximately 75~ of the world's rice is produced in bunded, puddled fields where irrigation is assured for at least one crop per year. O f this, 92~ is produced in Asia, primarily for domestic consumption. Rapid population growth necessitates increased production; there will have to be an increase of almost 70~ over the next 35 years if the supply of rice is to match the demand (IRRI 1993). In view of the relatively low yields (around 2 t h a - l) and fragility of rain-fed and upland rice systems, the current lack of investment in irrigation, and the pressures of urbanisation, most of this extra production will have to come from increased productivity in today's irrigated systems. Irrigation has long been used in rice cultivation. It was widespread in China by 400 B.C. Artificial flooding in traditional irrigated systems was limited to one rice crop per year. Between rice crops, soils that were not naturally waterlogged were cultivated with an upland crop or kept dry fallow. In the mid 1960s, rapid intensification of rice cultivation took place, with the introduction of highyielding varieties (HYVs), expansion of irrigation, and increased inputs of fertiliser and pesticides. These HYVs were less photoperiod sensitive and matured more quickly than traditional varieties, allowing two or even three irrigated rice crops to he grown per year. Systems under such intensive cultivation have shown a long-term (approx. 30 years) yield decline (Flinn and De Datta 1984). In longterm field experiments, plant N deficiency has been identified as a factor limiting to yield, despite constant or increasing total soil N (Cassman et al., in press). It has been suggested that the near-continuous flooding associated with year-round rice production might somehow reduce the availability of N released from soil organic matter. The predominant method of land preparation is puddling; the soil is tilled by ploughing and harrowing whilst saturated with water. Puddling breaks down soil aggregates into a uniform mass, reducing bulk density and soil strength and creating a relatively uniform soil-water me-

334

dium. Gaseous diffusion within such a medium is some 10000 times slower than in air. Oxygen diffusion is too slow to supply the microbial respiratory demand for electron-acceptors. The flooded soil develops a thin oxidised layer at the floodwater-soil interface and a reduced layer below. The oxidised region is a photic environment with a positive redox potential. The reduced region is nonphotic, with a predominantly negative Eh, usually low enough to allow the reduction of iron oxides. The microbiology of the two zones reflects these important differences. This paper focuses on the microbial decomposition of organic matter and the consequent mineralisation and/or immobilisation N in intensively cultivated continuously irrigated wetland rice ecosystems. Biologically mediated losses of C and N, especially by nitrification, denitrification and methane emission, are adequately covered elsewhere and we discuss them only briefly.

Organic matter in wetland rice soils A key factor driving biological processes in soil is the decomposition of organic matter. Decomposition converts some of the nutrients contained in organic substrates into mineral forms that may be taken up by plants and other organisms. Mineralised nutrients not taken up by plants or lost from the soil system may become incorporated into the soil microbial biomass, or they may react chemically with soil mineral and organic surfaces. These processes, collectively termed "immobilisation", render nutrients at least temporarily inaccessible.

Organic matter turnover The total organic carbon content of a soil changes only slowly, because a large fraction is inert, with residence times measured in thousands of years. Jenkinson and Rayner (1977), and many others since, have modelled the turnover of soil organic matter in terms of five discrete pools defined by their rates of decomposition. The pools were identified as decomposable plant material, resistant plant material, microbial biomass, and chemically and physically stabilised fractions of inert organic matter. However, the association between conceptual pools (defined by their turnover rates) and organic matter fractions (separable by physical and chemical procedures) is not always apparent. Understanding the interaction between soil micro-organisms and soil organic matter calls for the identification and characterisation of biologically relevant labile and inert organic components.

Organic matter input Organic matter inputs to the flooded ricefield are controlled by the net primary production of the system, the

allocation of assimilated carbon into roots, and the amount of recycled material (residues of the previous crop, etc.). Intensifying rice production affects the nature and increases the quantity of the inputs from the aquatic system. Inputs of plant material, both above and below ground, as root exudates and tissues shed during growth, increase with intensified production. The quantity of standing material returned to the soil at the end of a crop cycle varies according to management, depending on cutting height, straw removal, and whether or not the stubble is burnt. The primary productivity of the photosynthetic biomass in rice field floodwater is between 500 and 1000 kg C ha -~ per crop cycle (Yamagichi et al. 1980; Vaquer 1984), approximately 10~ of the standing rice crop biomass. Plant residues and detritus from the photosynthetic aquatic biomass fall on the soil as fresh litter. Litter is not generally considered part of soil organic matter if it constitutes a distinct layer on the soil's surface. However, like other agricultural soils, rice soils are ploughed at least once a year, thereby incorporating litter directly and preventing the formation of a separate layer. Studies of 13C and 14C abundance have shown a uniform distribution of organic matter in the Ap horizon of rice soils, presumably due to puddling (Becker-Heidmann and Scharpenseel 1986). Estimates of carbon release to the soil during plant growth range from 5~ to 33~ of the total assimilated (Keith et al. 1986; Sauerbeck and Johnen 1977; Warembourg and Paul 1973; Shamoot et al. 1968). Moreover, the distribution of photosynthates differs according to plant species, phenological stage and/or growth conditions (van Veen et al. 1989). Biological N fixation in the soil aquatic system can be significant. Fixation occures both in the floodwater (photodependent N fixation) and in the rhizosphere (associative N fixation). Estimates of N inputs from biological N fixation range from 15 to 50 kg ha -~ (Koyama and App 1979). The contribution of associative biological nitrogen fixation in the root rhizosphere to biological N fixation is small. Estimates based on acetylene reduction assays were summarised by Roger and Watanabe (1986) as 0 . 8 - 6 k g N ha -1 per crop cycle, and as 1.3-7.2kg N ha -~ per crop based on ~SN incorporation.

Organic matter composition The decomposition and recycling of organic matter sustains a microbial biomass which under steady-state conditions is determined largely by carbon availability in the soil, modulated by interactions with mineral particles. Under aerobic conditions the soil microbial biomass generally forms a relatively constant proportion of total soil C (Theng et al. 1989). Where soils are accumulating or losing C, the biomass will deviate from this relationship (Anderson and Domsch 1986). The soil microbial biomass is thus sensitive to management-related changes in soil organic matter.

335 Constituents of organic matter obtained from soils by chemical extraction have traditionally been characterised as humic and non-humic compounds (Schnitzer 1978). Non-humic compounds include carbohydrates, protein, fats, waxes, alkanes and low molecular weight organic acids. In general they are attacked readily by micro-organisms and have a short survival span. Humic compounds comprise the major portion of organic matter in soils and waters. They are amorphous, dark coloured, hydrophilic, acidic, partly aromatic, complex polymers that are heterogeneous in structure. Although humic compounds are generally thought of as recalcitrant, some are less so than others. Humic acid fractions that are more accessible to micro-organisms and more active in nutrient cycling are termed mobile humic acids or MHA. Olk (1993) found an M H A fraction less than 40 years old in an aerobic calcareous soil planted to irrigated cotton in California, which acounted for about 15~ of the extractable humic compounds (the remainder being calcium-bound humic acids, or Ca-HA) and 3~ of total soil organic C. A similar extraction procedure applied to two irrigated rice soils in the Philippines, one calcium-dominated and triple-cropped since 1963, the other double-cropped since 1983, produced the same two fractions (MHA and Ca-HA). As in the Californian soil, The M H A fraction appeared more labile than the Ca-HA, with a higher N content, lower C / N ratio, and a higher ratio of light absorbed at 465 nm to that at 665 nm (Table 1); this ratio decreases with increasing molecular weight and condensation. However, in the Philippine soils M H A accounted for about half of the total extractable humic compounds and 12O7oof total C. This fourfold difference between the aerobic and the flooded soils suggests that the younger humic acids may play a more important role in the latter. Macroscopic soil organic matter, a significant component of the non-humic fraction, comprises plant residues and detritus from the photosynthetic aquatic biomass.

The chemical structure of organic molecules alone is not sufficient to account for the enormous range of ages and turnover times of organic matter encountered in soils. According to Duxbury et al. (1989) old humic fractions with turnover times of thousands of years have a half-life in the order of weeks when extracted and re-added to soils. This emphasises the importance of the physical stabilisation of organic matter associated with inorganic components. Francois et al. (1991) showed that organic matter included with the sand and silt fractions consisted predominantly of decaying plant material, that fine silt fractions consisted of aggregates of silts, clays, bacteria, fungi and plant fragments and that the clay fraction contained approximately 50% of the total soil organic C. Organic C associated with clay was predominantly amorphous and contained little identifiable plant or microbial residue. Studies with 13C indicated that the organic matter associated with the fine clay fraction was heterogeneous with regards to decomposability, containing materials with a high turnover rate alongside stable organic matter (Balesdent et al. 1987). Van Veen et al. (1984) suggested that clay in soil can protect soil microbial biomass, slowing its turnover rate. Gregorich et al. (1991) concluded that soil texture did not affect the eventual decomposition of a readily decomposed substrate, but did control the rate

Table 1 Carbon and nitrogen content and light absorption ratio (465 nm/665 nm) for mobile humic acids (MHA) and calcium

humates (Ca-HA) extracted from soils with different cropping histories (data from Olk and Cassman 1993, Olk 1993)

Carbon (C)

Single-cropped aerobic cotton (California) a Double-cropped flooded rice (Philippines) b Triple-cropped flooded rice (Philippines) c

These substrates are difficult to quantify, especially given their transitory nature in soils. In aerobic soils organic debris can be isolated as a light fraction by densiometric separation procedures. Flooded rice soils receive large inputs of such biologically active and relatively undegraded organic matter, added to the system in consequence of its high primary production.

Organic matter-mineral interactions

Nitrogen (N) Total gkg -1 soil

E465/E665

Total gkg-lsoil

MHA gkg -1 soil (% total org. C)

Ca-HA gkg -1 soil (% total org. C)

11.3

0.32 (28)

2.18 (19)

22.5

2.61 (12)

2.72 (12)

1.90

0.23 (12)

28,8

3.74 (13)

2.82 (10)

2.44

0.30 (12)

a 160kg N ha -1 applied to each crop b 116 kg N ha -~ applied in dry season (DS); 58 kg N ha -1 in wet season (WS) c 150kg N ha -1 applied in DS, 90kg N ha -1 in early and late WS

MHA gkg -1 soil (% total org. C)

Ca-HA gkg -1 soil (~ total org. C)

MHA

Ca-HA

6.0

4.4

0.19 (10)

5.1

4.0

0.19 (8)

6.0

4.3

336 and turnover of C through the microbial biomass during short periods. Van Gestel et al. (1991), discussing the interaction between soil organic matter, texture and micro-aggregate stability, suggested that organic matter held within microaggregates was not accessible for microbial decomposition. Upon drying and rewetting of soils, mechanical disruption occurred rendering this organic matter more susceptible to decomposition. Christensen (1987), however, found no release of organic matter upon disruption of aggregates by sonication. Puddling of rice soils largely disrupts aggregates and disperses soil particles. This may release substrates physically enclosed in aggregates, but increase other protective organo-mineral interactions.

Soil m i c r o b i a l p o p u l a t i o n s

Bacterial flora predominate over fungi and actinomycetes in the reduced zone of flooded soils. In a Japanese ricefield, Bacillus, coryneform bacteria, Acinetobacter, Flavobacterium and Erwina were the major aerobic isolates, while Clostridium, Propionibacterium, Actinomyces and related organisms were the main species isolated in anaerobic incubations (Hayashi et al. 1978). Most strict anaerobes belong to the group Clostridium, of which only one isolate (Clostridium tertium) was able to assimilate carbohydrate vigorously. Bacterial populations in anaerobic paddy soils have been divided into two groups analogous to autochthonous and zymogenous bacteria on the basis of their growth characteristics in different concentrations of nutrient broth (Suwa and Hattori 1987). Autochthonous bacteria are always numerous in soil, with relatively constant populations requiring few external nutrient inputs. Zymogenous bacteria are opportunists, including more active organisms which respond rapidly to added nutrients.

precise than manual procedures. Observation of bacteria in intact soil samples is hindered by the presence of soil particles. More recently this problem has been overcome by the application of confocal laser microscopy (CLSM), which can scan a number of successive planes to build images with a greater depth of focus (Bloem, personal communication). Soil microbial biomass in three Japanese paddy soils was measured by direct counting using fluorescence microscopy. Bacterial numbers were high, ranging from 1.1 to 7.6• 10~~g-~ as compared with the values of the order of 109g -1 more usually encountered in aerobic soils (Hasebe et al. 1984). Soil samples taken from a waterlogged Philippine rice field and examined using CLSM gave cell numbers (2.1-9.8x109g -1) comparable to those above. Of the various recently developed indirect methods for measuring soil microbial biomass, the most promising in wetland rice soils is the chloroform-fumigation extraction procedure. In this procedure soils are exposed to chloroform, which causes cell lysis. Cell contents released are extracted and C or N quantified. Earlier attempts to measure biomass by chloroform-fumigation incubation were hindered by the long incubation time required. The chloroform-fumigation direct extraction method (Amato and Ladd 1988) was simplified by Gaunt (1993) for use with waterlogged field samples. A one-step fumigation with chloroform liquid in closed bottles was found as effective as fumigation in a desiccator under vacuum. Sieving of field samples prior to fumigation strongly affected the biomass determination. Biomass estimates for aerobically incubated wetland rice soils from the Philippines were similar to those from Indian and Nigerian soils measured by fumigation-incubation (Fig. 1).

Microbial processes

All else being equal, the maximum rates of decomposition and mineralisation processes are determined by the

Microbial biomass

1.6 O United Kingdom, Temperate

[ ] Philippines, Tropical (rice soils) It is important to be able to quantify the biomass as an 9 -I- India, Tropical organic fraction predicted by organic matter turnover ~ 1.2 Nigeria, Tropical models and a nutrient pool. There has been much [] research on the measurement of soil microbial biomass in d 0.8 aerobic soils, while similar studies for flooded rice soils, particularly under field conditions, are limited. The most o direct method for quantifying the microbial biomass is tfl 0.4 ++ microscopic observation, which has been used as a reference in the development of more indirect methods. Direct 0 ' 9 o lo 2'0 30 1o 5o 60 counting is generally too time consuming and tedious for Total soil organic C, g kg "1 routine application, but improvements are in hand. The use of computers in counting and analysing data and the application of image analysis to the microscopic environ- Fig. 1 Relationship between soil microbial biomass carbon and soil organic carbon for temperate and tropical soils. Redrawn from ment have advanced direct microscopy techniques. Bjorn- Theng et al. (1989). Biomass of Philippine rice soils measured unsen (1986) reported the use of video image analysis for der aerobic conditions by fumigationextraction as ninhydrin-reacepifluorescence microscopy, which is considerably more tive N, other values measured by fumigation incubation O

,m

337

resource quality of the substrate and the physical environment. Nutrient deficiencies and toxins can only reduce the rate of these processes. These environmentally determined maximum rates thus provide a yardstick against which to assess the function of the microbial community.

Decomposition of reduced carbon Cell metabolism occurs via oxidative catabolic and reductive anabolic pathways. The reduction of intracellular electron acceptors, such as nicotinamide adenine dinucleotide (NAD) or nicotinamide adenine dinucleotide phosphate (NADP) to NADH and NADPH respectively, facilitates the catabolic oxidation of carbon compounds. The energy produced is utilised in phosphorylation and stored as adenosine triphosphate (ATP). Anabolic processes are endergonic, utilising the energy stored in ATP to synthesise cell compounds from simple substrates (Lynch and Hobbie 1979). Decomposition of organic compounds in the absence of oxygen, as in wetland rice soils, relies on the re-oxidation of intracellular electron acceptors produced by catabolic reactions in excess of the requirements by anabolic respiration. The two general mechanisms of re-oxidation are anaerobic respiration and fermentation.

Anaerobic respiration If metabolic oxidation of reduced organic carbon compounds is the dominant decomposition process in soils, the supply of oxidised substrates which can act as electron acceptors is an important limiting factor. The sequence of reduction is generally as predicted by thermodynamics. The redox couple that has the highest affinity for electrons is reduced first, the energy released from this reaction being greatest. Although soil micro-organisms provide enzymes or catalysts to increase the rate of reaction (Rowell 1981) they do not affect the quantity of energy released. In the absence of 02 the next-best electron acceptor commonly available in soils is nitrate (NO3), which is reduced to nitrous oxide (N20) and/or dinitrogen (N2) in the process of denitrification. Since the products are both gases they are eventually lost to the atmosphere. The ratelimiting step is diffusion of ammonium from the reduced to the oxidised surface zone, where nitrification produces NO;- which may subsequently be denitrified anaerobically. Nitrification-denitrification in flooded soils has been extensively covered in the literature (e.g. Reddy and Patrick 1986; Katyal et al. 1988); we do not propose to reproduce that coverage here. The major product is N2, which is environmentally neutral; losses of the greenhouse gas N20 may, however, be greater during and immediately after periods of dry fallow, when NO 3-levels are higher (Bronson et al. 1993). Losses as high as 90kg NO3--Nha -~ have been reported during such periods (Buresh et al. 1989).

Where neither 0 2 nor NO 3 is available, oxidised manganese (Mn4+ , iron (Fe3+) or sulphate (SO]-) may be employed (e.g. Munch and Ottow 1980; Ponnamperuma 1984). Patrick and Reddy (1978) discussed the possibility of overlap between the individual reduction systems, but concluded that SO 2+ reduction was unlikely to occur in the presence of O2, NO~ or NO~-. However, soil heterogeneity may result in different reduction processes occurring at separate locations. The large quantity of iron in rice soils is generally considered to provide the main redox buffering system (Patrick and Reddy 1978). In the absence of any energetically preferable electron acceptor, some microbial consortia are able effectively to disproportionate reduced organic compounds, oxidising part to CO2 and reducing part to methane (CH4). There are two main pathways, one involving acetate, the other H +. The specific mechanisms of, and constraints affecting, these pathways are complex and poorly understood (Conrad et al. 1989). Methane produced at depth must pass through an oxidised zone (either at the floodwatersoil interface or within the rhizosphere) where it may be converted to CO2 by methanotrophic bacteria (Anthony 1986; de Bont et al. 1978); the aerenchymous roots of rice plants simultaneously provide the O 2 required for this oxidation, and a means of escape from the soil for that CH 4 which is not oxidised (e.g. Holzapfel-Pschorn et al. 1986; Nouchi and Mariko 1993). There is an extensive literature on the subject (Cicerone and Shetter 1981; Khalil and Rasmussen 1983; Holzapfel-Pschorn et al. 1985; Scht~tz et al. 1989; Neue et al. 1990; Conrad 1993) arising from concern over the role of CH4 as a greenhouse gas.

Fermentation Fermentation alone does not result in mineralisation of carbon (except in the case of the conversion of acetate to methane and carbon dioxide) because intracellularly generated organic electron acceptors are used instead of external ones. Thus, one or more products of fermentation are partially oxidised compounds. Anaerobic fermentation, however, does play an important role in the decomposition of reduced carbon in breaking down complex substrates prior to oxidation, resulting in an array of substances, many of them transitory and not found in wellaerated soils. Interactions between micro-organisms influence the processes of fermentation. Of particular importance is interspecies hydrogen transfer, whereby hydrogen-utilising bacteria use H + ions produced by fermentation as a source of energy and thereby maintain a low concentration of H + . The low H § concentration facilitates what would otherwise be energetically unfavourable reactions such as the fermentation of ethanol to acetate. Obligate anaerobic organisms are relatively limited in their nutritional and metabolic activities. Anaerobic species are generally dependent upon the activities of other organisms to supply nutrients and establish the required physico-chemical conditions. Lovely and Phillips (1989)

338 showed that decomposition of glucose at reduced 0 2 supply involved fermentation to fatty acids which were then oxidised to CO 2 by a consortium of fermentative, fatty-acid-oxidising and Fe3+-reducing bacteria. Such consortia require favourable ultra-micro-environments on a scale greater than the micro-environments associated with each individual organism. These ultra-micro-environments may be associated with organic or inorganic surfaces or, as in the case of biofilms, other living organisms.

Nitrogen immobilisation and mineralisation A decreasing soil N supply has been hypothesised as a cause of declining yield in double- and triple-cropped continuous rice systems, in which the soil remains submerged throughout most of the year. It is thus especially important to examine net N mineralisation or immobilisation in these systems. At any one time the net rate of mineralisation or immobilisation is the resultant of the gross rates of the two individual processes. In order to predict net N mineralisation it is necessary therefore to consider the N requirements of the microbial population, and their ability to immobilise that N, as well as the nature of substrates being decomposed. In aerobic soil systems, the soil microbial biomass is generally C limited (Jenkinson 1988). Under these conditions the biomass has a small capacity to immobilise nutrients, and net N mineralisation occurs at a rate that reflects the nature of the organic substrates being decomposed. The N requirement of anaerobic metabolism is lower than that of aerobic metabolism (Acharya 1935). Therefore net N mineralisation is higher and residues with a wider C/N ratio can be utilised than under aerobic conditions. Fresh inputs of organic matter (which constitute the macro organic, light and non-humic fractions described earlier) probably provide the major source of energy for soil micro-organisms. The "light" fraction is predominantly associated with sand and silt particles and is not well protected. Intensification of cultivation in aerobic systems has been shown to deplete this fraction more rapidly than heavy fractions (Dalai and Mayer 1987). Its role in N mineralisation is less clear. Janzen et al. (1992) found that soil respiration rate and microbial N content were strongly correlated with a light fraction separated by such a procedure, although N mineralisation was less strongly related. Tian et al. (i992) found that the decomposition of added plant residues in an aerobic soil depended on their C/N ratio and lignin and polyphenol content; residues with low C/N ratios frequently could not be decomposed without additional inorganic N which was thereby immobilised. Using aerobic soils in a pot study, Appel and Mengel (1993) found that N mineralisation and plant N uptake was related to an organic N fraction extractable by CaC12. Although Wenzl (1990) associated this fraction with non-humic substances, it was observed that the fraction was relatively constant with time, indi-

cating that it may represent a more stable fraction. Subsequent analysis showed that the CaCl2-extractable material contained a high molecular weight fraction as well as non-humic materials. Total N content was shown to be highly related to mineralisable N (Sahrawat 1983). However, the relationship was less clear for soils from China (Zhu et al. 1984) and throughout south and Southeast Asia (Kawaguchi and Kyuma 1977), and it breaks down completely in continuously irrigated systems where the N supply has declined despite constant or accumulating N. Where C is not limiting, inorganic N has been shown to determine the size of the soil microbial biomass (Merckx et al. 1987; van Veen et al. 1989). Bremer and van Kessel (1992) investigated seasonal microbial biomass dynamics after the incorporation of lentil and wheat residues in aerobic soils, and suggested that the soil microbial biomass may reduce losses of nutrients during periods of low crop demand and may act as a source of nutrients during active crop growth. Net microbial N immobilisation occurred in field soils where no fertiliser N was added (Wheatley et al. 1991). Observations in a Philippine rice field showed that the soil microbial biomass may rapidly immobilise inorganic N added at transplanting (Fig. 2). A similar immobitisation was observed by Schimel and Firestone (1989), who found that nitrogen dynamics in the forest floor were characterised by rapid microbial uptake of NH~ after which transformations were slower. The initial turnover time of the ammonium pool was 1 day. The mechanism of this is not clear; however, biomass N was related to the solution NH~- concentration (Fig. 3). Microbial assimilation of most ions and large molecules occurs via specific carriers, whilst lipophilic and small, uncharged compounds rapidly pass membranes by diffusion. At high external concentrations of NH~, the diffusion of NH3 may be sufficient to supply microbial N requirements (Kleiner 1985). In wetland rice soils, where N is added in the form of urea, temporarily high pHs and therefore NH 3 concentrations arise near urea particles. This suggests that the rapid immobilisation of N by the biomass may reflect a temporary 11o lOO

.~

+120 kg N

ha-1

I-Fertilised, 180 kg N ha"1 ~ Unfertilised

9o

"

80 70 60 50 40 30 20 lO

o

i.

40

60

80

Days after transplanting

Fig. 2 lmmobilisationof N within soil microbial biomass at two fertilisation rates: ON (unfertilised control) and t80N (180kg N ha -1 split: 120 kg N at transplanting and 60 kg N at panicle initiation, 42 days after transplanting)

339

Fig. 3 Relationshipbetween soil microbial biomass ninhydrin-reactiveN and solution ammonium concentration in an N-fertilised flooded rice soil

Biomass

ninhydrin-reactive-N,

m g kg "1

60

60

4 84

40

30

20

: I

o

Solution

passive influx of ammonia rather than immobilisation of NH~ that requires energy. The fate of N apparently re-mineralised from the soil microbial biomass is not clear. Bremer and van Kessel (1992) found that microbial 15N declined 5 times faster than lSN mineralised in an aerobic agricultural system. Studies based on acid fractionation of organic N suggest that a significant proportion of this N may be immobilised within the hydrolysable organic fractions and is not re-mineralised during a growing season. Up to two-thirds of N added with straw remained immobilised after 1.5 years of continuous cropping (Broadbent and Nakashima 1970). The apparent stability of these pools may reflect processes of humification and re-synthesis of organic compounds. Likewise, rapid isotopic exchange between added labelled N and N in more recalcitrant organic compounds may contribute to observed immobilisation.

--

L

4

2

0

6

ammonium,

8

m g N L "1

duced gross or net mineralisation; and greater physical or chemical protection of some fraction of previously accessible N. We shall examine each of these possibilities in turn.

Reduced N inputs The nature and quantity of inputs from the soil aquatic system are affected by intensification of rice production.

10000

8000

o

Yield decline in continuous irrigated systems Declining yields have been seen in several long-term experiments where continuous irrigated rice is grown in the Philippines (Cassman et al., in press). Various hypothetical explanations were discussed by these authors, a significant observation being that all crops were apparently N deficient. Plotting N response curves at various periods (Fig. 4) showed that the efficiency of the grain yield response to applied N (the gradient) changed little while the intercept (the yield with no input of N) decreased significantly. This decrease reflects a reduction of N supplied by the soil-floodwater system and/or the ability of the crop to take up N. It occurred in spite of the fact that total soil organic C and total soil organic N remained constant or even increased. Possible microbiological contributors to the observed yield decline factors include: a decrease in N inputs; re-

=

z

6000

4000 c

l

-

- 1970-72 Y = 6 3 0 0 + 3 6 X - O . 1 0 X 2 , r 2 = 0 . 8 6 1989-91 Y = 3800 + 34X - O.11X2 , r 2 = 0.94

2000

f

I

60

r

I

100

f

I

150

FERTILIZER-N i N P U T (kg ha "1)

Fig. 4 Yield r e s p o n s e to fertiliser N rates in a l o n g - t e r m e x p e r i m e n t at I R R I at t i m e periods 1 9 7 0 - 1 9 7 2 a n d 1 9 8 9 - 1 9 9 1

340 Inputs of above and below ground plant material have been discussed above. Photodependent N fixation in the floodwater has been shown to be inhibited by broadcast application of fertiliser. Although the amount of N derived from biological nitrogen fixation and taken up directly by a crop may be small Eskew et al. (1981), the process may nevertheless be important in maintaining soil organic N.

Reduced mineralisation/increased immobilisation

Remediation strategies The yield decline in continuously irrigated rice systems may be explained by a shift in the system equilibrium under new management and/or degradation of the soil resource. Classifications that distinguish between nearcontinuously waterlogged rice soils developed in natural wetlands and those created solely through irrigation may be helpful in this context. It will also be useful to know whether there are similarities between the organic matter fractions presently accumulating in continuously irrigated rice soils and those already present in natural wetlands. It may be possible, where ground water conditions permit, to manage the organic N pools altered during the shift from an aerobic to an anaerobic system. Introduction of aerobic rotations to manipulate the existing N pools will not however, recover N actually lost by degradation. A decline observed in the rice-wheat system, which is not continuously flooded, has yet to be attributed to a change in the soil resource base. If it is so attributed, however, then the decline in that system may only be addressed through control of organic matter outputs and inputs.

Aromatic compounds are generally more resistant to decomposition under anaerobic conditions. Degradation of lignin, the major plant-derived organic compound found in soil, may then be slowed. Lignin is generally decomposed into simple phenolic substances which then polymerise through various reactions to form humic substances. If sufficient quantities of organic material are added to soils that remain continually anaerobic, these phenolic intermediate products of lignin degradation may accumulate. Thus humic acids extracted from aquatic sediments more closely resemble partially degraded lignin than the humic acids found in aerobic soils (Yonebashi and Hattori 1988). Similarly, humic acids from several temperate and subhumid water systems were found to contain high levels of phenolic C as determined by ~3C- Conclusions NMR (Malcolm 1990). Phenolic compounds can immobilise N abiotically It is clear that to address the issues of declining soil fertili(Verma et al. 1975). Inorganic ammonium mineralised ty requires a mechanistic understanding of the soil prowithin the soil system is vulnerable, as also are amino ac- cesses involved in organic matter decomposition, nutrient ids released by the turnover of soil microbial populations. mineralisation, and immobilisation. To understand net N Accumulation of aromatic compounds and the associated mineralisation in any depth it will be necessary to identify abiotic immobilisation of N might go some way to ex- the organic fractions involved. Every effort should be plaining the yield decline in intensively cultivated irrigat- made to determine the biological relevance of the organic ed rice grown under continuously flooded conditions matter fractions identified by extraction procedures: we with no aerobic period between crops (Cassman et al., in need to know not just where the N is, but whether or not press). As the system becomes anaerobic the proportion is is accessible to the soil biomass. This is particularly imof total N that is labile reduces (whilst N in recalcitrant portant in the area of recent inputs. The macro organic fractions increases), resulting in a decrease in yield. This or light fractions are transient and therefore difficult to decrease may therefore reflect the change to a new equi- quantify, whilst other substrates such as root exudates librium between total and labile N fractions, rather than may contribute significantly to microbial activity without being a consequence of resource base degradation (a de- ever being identified as a separable fraction. The integracline in labile N occurring without reversible internal pool tion of the information obtained using mechanistic modtransfer - i.e. a net loss of N). Preliminary results con- els will provide an essential step to validate our underfirm the accumulation of aromatic compounds, whilst to- standing of the soil system. tal N seems constant. It is important to understand and partition those components of the change in the resource base (total soil N and C) caused by the transition from a predominantly References aerobic system to one dominated by anaerobic conditions Acharya C (1935) Comparison of the course of decomposition of and those that result from other management practices rice straw under anaerobic, aerobic and partially aerobic condiassociated with intensification. Prior to intensification, tion. Biochem J 29:1116-1 i 20 soils were generally flooded only for one season or were Amato M, Ladd J (1988) Assay for microbial biomass based on ninhydrin-reactive nitrogen in extracts of fumigated soils. Soil Biol cultivated with upland crops. These crops were mainBiochem 20:107-114 tained with much lower inputs of fertilisers and pestiAnderson J, Domsch K (1986) Carbon link between microbial biocides. Alterations in these areas of management may also mass and soil organic matter. In: Megusar F, Gunter M (eds) perhaps have contributed to the decline. Proceedings of the Fourth International Symposium on Microbial Ecology, 24-29 August 1986, Ljubljana, Yugoslavia

341 Anthony C (1986) Bacterial oxidation of methane and methanol. Adv Microb Physiol 27:113 - 210 Appel T, Mengel K (1993) Nitrogen fractions in sandy soils in relation to plant nitrogen uptake and organic matter incorporation. Soil Biol Biochem 25:685-691 Balesdent J, Mariotti A, Guillet B (1987) Natural 13C abundance as a tracer for soil organic matter dynamics studies. Soil Biol Biochem 19:25- 30 Becker-Heidmann P, Scharpenseel H (1986) Thin layer 13C and Algc monitoring of "Lessive" soil profiles. Radiocarbon 28: 383-390 Bjornsen P (1986) Automatic determination of bacterioplankton biomass by image analysis. Appl Environ Microbiol 51:1199-1204 Bremer E, van Kessel C (1992) Seasonal microbial biomass dynamics after addition of lentil and wheat residues. Soil Sci Soc Am J 56:1141-1146 Broadbent F, Nakashima T (1970) Nitrogen immobilization in flooded soils. Soil Sci Soc Am Proc 34:218-221 Bronson K, Neue H, Singh U (1993) Nitrous oxide and methane fluxes in rainfed fallow-irrigated rice systems. In: Agronomy abstracts. American Society of Agronomy, Madison, pp 242 Buresh R, Woodhead T, Shepherd K, Flordelis E, Cabangon R (1989) Nitrate accumulation and loss in a mungbean/lowland rice cropping system. Soil Sci Soc Am J 53:477-482 Cassman K, De Datta S, Olk D, Alcantara J, Samson M (in press) Yield decline and nitrogen balance in long-term experiments on continuous, irrigated rice systems in the tropics. Adv Soil Sci Christensen B (1987) Decomposability of organic matter in particle size fractions from field soils with straw incorporation. Soil Biol Biochem 19:429-435 Cicerone R, Shetter J (1981) Sources of atmospheric methane: measurements in rice paddies and a discussion. J Geophys Res 88:11022-11024 Conrad R (1989) Control of methane production in terrestrial ecosystems. In: Andreae MO, Schimel D (eds) Exchange of trace gases between terrestrial ecosystems and the atmosphere. Dahlem Konferenzen. Wiley, Chichester, pp 39-58 Conrad R (1993) Mechanisms controlling methane emission from wetland rice fields. In: Oremland R (ed) Biogeochemistry of global change: Radiatively active trace gases. Chapman and Hail, New York, pp 317-335 Conrad R, Mayer H, Wust M (1989) Temporal change of gas metabolism by hydrogen-syntropic methanogenic bacterial associations in anoxic paddy soil FEMS Microbiol Ecol 62:265- 274 Dalal R, Mayer R (1987) Long term trends in fertility of soils under continuous cultivation and cereal cropping in southern Queensland. VI. Loss of total nitrogen from different particlesize and density fractions. Aust J Soil Res 25:83-93 de Bont J, Lee K, Bouldin D (I978) Bacterial oxidation of methane in a rice paddy. Ecol Bull (Stockholm) 26:91-96 Duxbury J, Scott-Smith M, Doran J, Jordan C, Szott L, Vance E (1989) Soil organic matter as a source and sink of plant nutrients. In: Coleman DC, Oades JM, Uehara G (eds) Dynamics of soil organic matter in tropical ecosystems. NifTAL, University of Hawaii, pp 32-57 Eskew D, Eaglesham A, App A (1981) Heterotrophic N 2 fixation and distribution of newly fixed nitrogen in a rice-flooded soil system. Plant Physiol 68:48-52 Flinn J, De Datta S (t984) Trends in irrigated-rice yields under intensive cropping at Philippine research stations. Field Crops Res 9:1-15 Francois C, Feller C, Guiraud G, Loury J, Boudot J (1991) Immobilization of nitrogen from urea and plant residues in a ferrallitic soil: Laboratory experiments and study by size-fractionation. Biol Fertil Soils 12:182-188 Gaunt J (1993) Soil microbial biomass in wetland rice soils. Ph D thesis, Wye College, University of London Gregorich E, Voroney R, Kachanoski R (1991) Turnover of carbon through the microbial biomass in soils with different textures. Soil Biol Biochem 23:799-805

Hasebe A, Kanazawa S, Takai Y (1984) Microbial biomass in paddy soil: I. Microbial biomass calculated from direct count using fluorescence microscope. Soil Sci Plant Nutr 30:175-187 Hayashi S, Asatsuma S, Nagatsuka T, Furusaka C (1978) Studies on bacteria in paddy soil. Rep Inst Agr Res Tohoku Univ 29:19-38 Holzapfel-Pschorn A, Conrad R, Seller W (1985) Production, oxidation and emission of methan in rice paddies. FEMS Microbio Ecol 31:343-351 Holzapfel-Pschorn A, Conrad R, Seiler W (1986) Effects of vegetation on the emission of methane from submerged paddy soil. Plant and Soil 92:223-233 IRRI (1993) 1993-1995 IRRI rice almanac. International Rice Research Institute, Manila, Philippines Janzen H, Campbell C, Brandt S, Lafond G, Townley-Smith L (1992) Light-fraction organic matter in soils from long term crop rotations. Soil Sci Soc Am J 56:1799-1806 Jenkinson D (1988) Determination of microbial biomass carbon and nitrogen in soil. In: Wilson J (ed) Advances in nitrogen cycling in agricultural ecosystems. CAB, Wallingford, UK, pp 368-386 Jenkinson D, Rayner J (1977) The turnover of soil organic matter in some of the Rothamsted Classical Experiments. Soil Sci 123:298-305 Katyal J Carter M, Vlek P (1988) Nitrification activity in submerged soils and its relation to denitrification loss. Biol Fertil Soils 7:16-22 Kawaguchi K, Kyuma K (1977) Paddy soils in tropical Asia. Univ. Hawaii Press, Honolulu Keith H, Oades J, Martin J (1986) Input of carbon to soil from wheat plants. Soil Biol Biochem 18:445-449 Khalil M, Rasmussen R (1983) Sources, sinks and seasonal cycles of atmospheric methane. J Geophys Res 88:5131-5144 Kleiner D (1985) Bacterial ammonium transport. FEMS Microbiol Rev 32:87-100 Koyama T, App A (1979) Nitrogen balance in flooded rice soils. In: Nitrogen and rice. International Rice Research Institute, Los Bafios, Philippines, pp 95-103 Lovely D, Phillips E (1989) Requirement for a microbial consortium to completely oxidise glucose in Fe(III)-reducing sediments. Appl Environ Microbiol 55:3234-3236 Lynch J, Hobbie J (1988) Micro-organisms in action: concepts and applications in microbial ecology. Blackwell, Oxford Malcolm R (1990) Variations between humic substances isolated from soils, stream waters and groundwaters as revealed by t3CNMR spectroscopy. In: MacCarthy P, Clapp C, Malcolm R, Bloom P (eds) Humic substances in soil and crop sciences: Selected readings. American Society of Agronomy, Madison, pp 13-35 Merckx R, Dijkstra A, den Hartog A, van Veen J (1987) Production of root-derived material and associated microbial growth in soil at different nutrient levels. Biol Fertil Soils 5:126-132 Munch J, Ottow J (1980) Preferential reduction of amorphous to crystalline iron oxides by bacterial activity. Soil Sci 129:15-21 Neue HU, Becker-Heidmann P, Scharpenseel HW (1990) Organic matter dynamics, soil properties, and cultural practices in rice lands and their relationship to methane production. In: Bouwman AF (ed) Soils and the greenhouse effect. Wiley, Chichester, pp 457-466 Nouchi I, Mariko S (1993) Mechanism of methane transport by rice plants. In: Oremland RS (ed) Biogeochemistry of global change: Radiatively active trace gases. Chapman and Hall, New York, pp 336-352 Olk D (1993) Effects of soil moisture and soil organic matter on potassium availability in a calcareous vermiculitic soil. PhD thesis, University of California, Davis Olk D, Cassman K (1993) Reduction of potassium fixation by organic matter in vermiculitic soils. In: Mulongoy K, Merckx R (eds) Soil organic matter dynamics and sustainability of tropical agriculture. Wiley, New York, pp 307-316 Patrick W, Reddy C (1978) Chemical changes in rice soils. In: IRRI

342 soils and rice. International Rice Research Institute, Los Bafaos, Laguna, Philippines, pp 361-380 Ponnamperuma F (1984) Effects of flooding on soils. In: Kozlowski T (ed) Flooding and plant growth. Academic, New York, pp 9-45 Reddy K, Patrick W Jr (1986) Fate of fertilizer nitrogen in the rice root zone. Soil Sci Soc Am J 50:649-651 Roger P, Watanabe I (1986) Technologies for utilizing biological nitrogen fixation in wetland rice: Potentialities, current usage, and limiting factors. Fert Res 9:39-77 Rowell D (1981) Oxidation and reduction. In: Greenland D, Hayes M (eds) The chemistry of soil processes. Wiley, New York, pp 401-453 Sahrawat K (1983) Mineralization of soil organic nitrogen under waterlogged conditions in relation to other properties of tropical rice soils. Aust J Soil Res 21:133-138 Sauerbeck D, Johnen B (1977) Root formation and decomposition during plant growth. In: Soil organic matter studies. Proceedings of symposium IAEA/FAO/GSF, Vienna, Austria, pp 141-148 Schirnel J, Firestone M (1989) Inorganic nitrogen incorporation by forest floor material. Soil Biol Biochern 21:41-46 Schlesinger W (1991) Biogeochemistry: An analysis of global change. Academic, San Diego Schnitzer M (1978) Humic substances: Chemistry and reactions. In: Schnitzer M, Khan S (eds) Soil organic matter. Elsevier, New York, pp 1-64 Schutz H, Seller W, Conrad R (1989) Processes involved in formation and emission of methane in rice paddies. Biogeochem 7:33-53 Shamoot S, McDonald L, Bartholomew W (1968) Rhizodeposition of organic debris in soil. Soil Sci Soc Am Proc 32:817-820 Suwa Y, Hattori T (I 987) Population dynamics of soil bacteria cultured under different nutrient conditions. Soil Sci Plant Nutr 33:235-244 Theng B, "Fate K, Sollins P (1989) Constituents of organic matter in temperate and tropical soils. In: Coleman D, Oades M, Uehara G (eds) Dynamics of soil organic matter in tropical ecosystems. NifTAL Project, Department of Agronomy, College of Tropical Agriculture and Human Resources, University of Hawaii, pp 5 - 31

Tian G, Kang B, Brussard L (1992) Biological effects of plant residues with contrasting chemical compositions under humid tropical conditions - decomposition and nutrient release. Soil Biol Biochem 24:1051 - 1060 Van Gestel M, Ladd J, Amato M (1991) Carbon and nitrogen mineralization from two soils of contrasting texture and microaggregate stability: Influence of sequential fumigation, drying and storage. Soil Biol Biochem 23:313-322 Van Veen 3, Ladd J, Frissel M (1984) Modelling carbon, nitrogen turnover microbial biomass soil. Plant and Soil 76:257-274 Van Veen J, Merckx R, Van der Geijn (1989) Plant and soil related controls on the flow of carbon from roots through the soil microbial biomass. Plant and Soil 115:179-188 Vaquer A (1984) La production algae dans les rizieres de Carargue pendant la periode de submersion. Vet Internat Verein Limnol 22:1651 - 1654 Verma L, Martin J, Haider K (1975) Decomposition of ~4C labeled proteins, peptides and amino acids: free and complexed with humic polymers. Soil Sci Soc Proc 39:279-284 Warembourg F, Paul E (1973) The use of 14CO2canopy techniques for measuring carbon transfers through the plant-soil system. Plant and Soil 38:331-345 Wenzl W (1990) Extraktion and Charakterisierung yon stickstoffreichen Nichthuminstoffen einer Braunerde. Kongrel3band 1990. VDLUFA-Schriftenreihe 32:337-344 Wheatley R, Griffiths B, Ritz K (1991) Variations in the rates of nitrification and denitrification during the growth of potatoes (Solanum tuberosum) in soil with different carbon inputs and the effect of these inputs on soil nitrogen and plant yield. Biol Fertil Soils 11:157-162 Yamagichi T, Okada K, Murata Y (1980) Cycling of carbon in a paddy field. Jpn J Crop Sci 49:135-145 Yonebayashi K, Hattori T (1988) Chemical and biological studies on environmentalhumic acids: I. Composition of elemental and functional groups of humic acids. Soil Sci Soc Plant Nutr 34:571-584 Zhu Z, Cai G, Yu Y, Zhang S (1984) Nitrogen mineralization of paddy soils in Tai-lake region and the prediction of soil nitrogen supply. Acta Pedologica Sinica 21:29-36