Joshua Schimel 1. Introduction 177

University of California, 2. Microbiology in Biogeochemical Models 177

Santa Barb ai ^California 3- Dealing with Microbial Diversity in Models 178

4. Kinetic Effects of Microbial Population Size 178

5. Microbial Recovery from Stress 180

6. Conclusions: Integrating Biogeochemistry and Microbiology 181

1. Introduction

Microbiology examines the smallest level of the organization of life, while biogeochemistry considers the largest. What then, is the appropriate relationship between these fields? The point could be argued several ways. On the one hand, much of biogeochemistry is simply microbial physiology writ large. On the other hand, the vast gap in scale could mean that information about microbial physiology and community dynamics has limited direct utility in large-scale biogeochemical studies. The present chapter will consider this issue and will discuss some aspects of how microbiological understanding is (or is not) currently incorporated into biogeochemistry. I will also identify some directions for future research that could enhance the linkage between the two fields. As simulation models are a primary tool for linking fields and scales, much of this discussion will be targeted at how biogeochemical models handle microbiological processes and how this might change in the future.

Microorganisms (including bacteria, fungi, single-celled algae, and protozoa; Madigan et ill, 1997) are ubiquitous on Earth. They include the greatest diversity of all living things and they are dominant players in almost all global biogeochemical processes. In the C cycle, terrestrial plants may carry out slightly more than half the total global primary productivity, but single-celled algae in the ocean account for most of the rest (Schlessinger, 1997). The vast bulk of decomposition is carried out by fungi and bacteria (though mediated in some cases by faunal food webs). In the nitrogen cycle, essentially all the important transformations are carried out by microorganisms, including mineralization, N,-fixation, nitrification, and denitrification. The same is true of the cycles of sulfur, phosphorus, and many other elements as well; the dominant biological transformations are microbial, often bacterial.

2. Microbiology in Biogeochemical Models

Although there is great diversity within microbial groups in terms of physiology and environmental responses, the populations of some microbial groups can vary dramatically over time, and microbial biomass can be an important reservoir of labile nutrients, microbial physiologists and community ecologists rarely interact directly with biogeochemists. Biogeochemical models of ecosystem C and N cycling rarely include microbiology explicitly. Models typically use simple response functions for processes and almost never include microbial population size as an active control on specific process rates. A number of soil organic matter models include a pool labeled "microbial biomass." However, to quote McGill (1996):

Further, although biomass was a frequent "pool" in these models, its treatment was often indistinguishable from active forms of SOM. One might consider inclusion of soil biomass in this way to be tokenism.

Thus, one can make an argument that biogeochemical models tend to ignore microbiology or at least simplify it beyond recognition. This conclusion would suggest that there is little real, direct application of microbiological information to larger scale biogeochemical modeling. I believe, however, that conclusion would be wrong. First, several models do go into a reasonable amount of physiological detail for some processes; e.g., denitrification in the model DNDC (Li, 1996; Leffelaar et al., 1988). More importantly and more gener ally, though, most biogeochemical models try to go directly from inputs of environmental parameters directly to outputs of process rates. Thus the microbiology, rather than being nonexistent, is "implicit." It is buried in the equation structure of a model as kinetic constants and response functions. Implicit microbiology is quite different from no microbiology. A lot of basic microbial physiology and process study went into developing the response functions in most biogeochemical models. In fact, to do implicit microbiology, incorporating important mechanistic accuracy, requires a sound understanding of the processes and their control.

While making microbiology implicit in models has the advantage of simplicity, it also, however, has limitations. A model converts assumptions into predictions. By testing those predictions against reality, it allows the investigator to test the strength and accuracy of the assumptions, and how they interact with each other. That is easier to do when the assumptions are made explicit and incorporated into a model as a mechanism. When processes are made implicit, rather than explicit, it is harder to test the validity of the assumptions about those processes. To quote McGill (1996) again:

To restrict microbial biomass to a measurable SOM component, however, renders models nonmechanistic and removes the possibility that a model might simulate changes in SOM dynamics as a result of changes in activity or characteristics of soil organisms.

This raises the question: is there a need to become more mechanistic about the role of microorganisms in biogeochemical models? If so, what aspects of microbial processes and community dynamics should be considered for more focused study? In the remainder of this chapter I would like to consider these questions.

3. Dealing with Microbial Diversity in Models

There are two common core assumptions among biogeochemical models (and therefore among biogeochemists) that are worth evaluating from the microbial perspective. The first is that microbial physiologies are global. That is, microbial processes can be modeled across a range of conditions by using a single equation such as the following (Parton et al., 1987):

In this equation, C is the size of a soil carbon pool, K is a first-order rate constant, and Md and T^ are reducing functions based on temperature and moisture. Each process has a single K value and a single reducing functions for each environmental driver. The assumption is that the fundamental response functions do not change with environmental conditions or the composition of the microbial community. To say this in another way, "microbial diversity has no discrete 'role' to play with respect to ecosystem function." (Finlay et al., 1997). But is this true? Or alternatively, can changes in microbial communities change the nature of the response functions? I

and others have written rather extensively on this question (Schimel, 1995; Schimel and Gulledge, 1998; Brussard et al., 1997; Groffman and Bohlen, 1999; Wall and Moore, 1999). The general conclusion of these reviews has been that for processes that are carried out by physiologies that are broadly distributed across the microbial world (e.g., glucose metabolism), or for processes that we measure as single processes but that are really an aggregates of many specific processes (e.g., soil respiration), the composition of the microbial community is not often a major control on process dynamics. For processes that are physiologically "narrow" (i.e., carried out by physiologically/phylogenetically limited groups of organisms), such as nitrification and CH4 production and consumption, the composition of the microbial community is sometimes a substantial control on process dynamics.

Most of the good existing case studies illustrating these points have been discussed in the review papers mentioned above. However, one new study is worth mentioning. Bodelier et al. (2000) examined the effects of rice plants and N fertilization on the dynamics of methanotrophs in rice soil. They showed that while Type II methanotrophs dominated the bulk soil, rice plants selected for populations of Type I methanotrophs in the rhizosphere. They also showed that these Type I methanotrophs are N-limited and that with fertilization, Type I populations increase dramatically, significantly reducing net CH4 efflux from the system. Type II methanotrophs, while also CH4-saturated, did not show a strong increase on N fertilization. This stimulation by added NH4* is the direct opposite of the inhibition commonly found in upland soils (Gulledge and Schimel, 1998). Thus, the types and activities of methanotrophs present became a significant control on the overall methane flux from the rice ecosystem.

The conclusion from these studies is that there are cases where microbial community composition significantly affects the nature and environmental responses of biogeochemical processes. In these cases, assumptions of global physiologies and unitary response functions fail. Evaluating how diverse and significant these effects are, and then finding effective ways to integrate them into biogeochemical models is one area where microbial ecologists and biogeochemists should collaborate.

4. Kinetic Effects of Microbial Population Size

The second key assumption that models make is that microbial processes are never limited by the size of the microbial population. This assumption is clearly stated by Chertov and Komarov (1996) in their discussion of the SOMM model:

The number and species composition of decomposing organisms is dependent on the biochemical properties of organic debris and on hydrological and thermal conditions. We postulate that there are no barriers for a rapid invasion of new organisms. Thus, it is possible to calculate the decompo-

sition coefficients for the communities as a function of the biochemical properties of litter, temperature and moisture.

Finlay ct al. (1997) state the same idea even more bluntly:

(2) Microbial diversity in an ecosystem is never so impoverished that the microbial community cannot play its full part in biogeochemical cycling. The species complement of the microbial community quickly adapts, even to momentous changes in the local environment.

The assumption that microbial communities will always rapidly adapt to the available environment and substrate supply is a fundamental assumption in using first order kinetics, as in Kq. (1). Most biogeochemical processes arc catalyzed, however, and catalyzed reactions invariably show Michaelis —Menten kinetics,

where X is the product concentration, k is a reaction constant, E is the catalyst concentration, S is substrate concentration, and Km is the half saturation constant, k X E is usually expressed as Vmax (Roberts, 1977), the maximum velocity possible for the reaction, but it is worth making clear that Vmax is a linear function of the catalyst concentration. If most microbial processes actually follow Michaelis-Menten type kinetics, how can biogeochemical models represent them as 1st order? For modeling reaction kinetics across a wide range of concentrations, no 1st order model will work adequately (Fig. la). However, if substrate concentrations are moderate Km) but do not vary over an excessively wide range, it is possible to fit a line to the kinetic response, even if it is not strictly 1st order (Fig. lb). Alternatively, if substrate concentration is very low (S « Km), Eq. (2) reduces to

FIGURE 1 Fitting 1st order curves to catalyzed reaction kinetics. The individual data points were generated in a spreadsheet model of Michaelis-Menten kinetics with a random ± 5% error introduced to the individual values. The straight lines were fit to these data. The three panels all show the same data but over different ranges: (a) over the entire range from zero to close to substrate saturation; (b) over the range from 0.2 to 2 K,„; (c) over the range from 0 to K,„.

In this equation, kinetics become 1st order with VmiJKm as the rate constant (Fig. lc). Note that this constant still includes catalyst concentration (i.e., population size of the active microbes) within the Vmax term. This raises two questions: first, which conditions that allow a 1st order approximation occur in nature, and second, if all the rate constant terms are actually linear functions of the microbial population, why do models leave population size out?

To address the first of these questions, Table 1 presents data on the ratio of basal/maximal rates of soil respiration, nitrification, and denitrification. If the ratio is very small, then one can conclude that the process is naturally occurring at very low substrate concentrations, whereas a ratio close to 1 would indicate that the process is close to becoming substrate-saturated. A ratio of 0.5 would imply that the process was occurring at a substrate concentration close to Km. Table 1 is far from exhaustive, but represents the range of behaviors that occur for these processes. Respiration usually operates at between 20 and 65% of its maximum rate. This may cover a narrow enough range of concentrations so that using a 1st order approximation for a portion of the Michaelis-Menten curve (Fig. lb) would not introduce large errors. Nitrification,

FIGURE 1 Fitting 1st order curves to catalyzed reaction kinetics. The individual data points were generated in a spreadsheet model of Michaelis-Menten kinetics with a random ± 5% error introduced to the individual values. The straight lines were fit to these data. The three panels all show the same data but over different ranges: (a) over the entire range from zero to close to substrate saturation; (b) over the range from 0.2 to 2 K,„; (c) over the range from 0 to K,„.

however, appears to operate over a wide range, from 3 to 75%. However, the higher values appear to occur in natural ecosystems, while the lower values occur in agricultural systems. Regular fertilization may produce very large populations. To cover the entire range of systems, any 1st order assumption would fail, but if systems are divided into agricultural and nonagricultural, the ranges may be small enough to allow 1st order fits. Denitrification, however, behaves still differently, commonly proceeding at less than 1% of its potential rate, a range in which a 1st order assumption should work well. I hypothesize that denitrification operates so much below potential because it is carried out by aerobic organisms that switch to denitrification as a "back up" physiology when soils go anaerobic (Zumft, 1992). Thus, it should be possible to grow a large population of organisms under aerobic conditions; this would provide overcapacity when soils go anaerobic. So it appears that with some limited reparameterizing for agricultural

TABLE 1 Basal and maximal rates for respiration, nitrification, and denitrilication across a range of ecosystems. Basal rates were measured without added substrate. Maximal rates were measured using the same technique but with saturating amounts of substrate added.


Soil Location/Type

Ratio of

Basal/Maximal Rates

Notes on Approach



German beech forest English grassland English woodland English grass Icy Alaskan lichen heath Alaskan riparian carex Alaskan tussock tundra Nitrification Alaskan aider

Swedish barley, unfertilized Swedish barley, fertilized Swedish lucerne ley Swedish grass ley Ontario alder Utah Agricultural soil Dentrification Danish fen

NK USA poorly drained forest USA vegetated filter strip Kansas tallgrass prairie Kansas cultivated




0.24 (range: 0.17-0.40) 0.57 (range: 0.51 -0.76) 0.38 (range: 0.32-0.45)




Glucose Glucose Glucose Glucose Glucose Glucose Glucose amended soils amended soils amended soils amended soils amended soils amended soils amended soils

Gross nitrification/chlorate slurry Chlorate amended core/slurry Chlorate amended core/slurry Chlorate amended core/slurry Chlorate amended core/slurry Unamended/amended chlorate slurry Gross nitrification/slurry

Amended anaerobic cores Amended anaerobic cores Amended anaerobic cores Amended aerobic cores Amended aerobic cores

Anderson and Joergensen (1997)

Lin and Brookes, 1999

Lin and Brookes, 1999

Lin and Brookes, 1999

Cheng et al., 1998

Cheng et al., 1998

Cheng et al, 1998

Clein and Schimel, 1995

Berg and Rosswali, 1987

Berg and Rosswali, 1987

Berg and Rosswali, 1987

Berg and Rosswali, 1987

Hendrickson and Chatarpaul, 1984

Shi and Norton, 2000

Ambus and Christensen, 1993 Groffman et al, 1991 Groffman et al., 1991 Groffman, 1991 Groffman, 1991

July data.

versus natural ecosystems, it might be possible to use 1st order descriptions of portions of the full Michaelis-Menten substrate response curves for most of these processes. However, biomass size would still be a part of the effective rate constant, and so should still be a measurable control on the actual rate of the process in the field. This again raises the question: how can biogeochemical models exclude biomass as a factor in kinetic responses?

The effects of biomass size on process kinetics would not necessarily be apparent under some conditions. If the microbial population size is constant, then biomass can be incorporated as part of the rate constant. If population sizes vary linearly with specific environmental conditions, then biomass can be incorporated into the appropriate response function. However, neither situation is actually true. Microbial population sizes are not constant. Total microbial biomass can vary over time by a factor of 2 or more, sometimes with little obvious correlation to season or weather (Wardle, 1998). More specific populations, such as nitrifiers, denitrifiers, and methanotrophs, can vary more than that (e.g., Acea and Car-balias, 1996; Berg and Rosswali, 1987; Both et al., 1992; Saad and Conrad, 1993). There are also stresses that can rapidly reduce biomass by as much as a factor of 2 such as rewetting a dry soil (Bot-tner, 1985; Kieft et al., 1987) and freeze-thaw (Morley et al., 1983). Thus, microbial studies suggest that biomass can vary enough to have large impacts on overall process rates. This once again raises the question: why do models almost invariably exclude biomass as an active control? It is because of the second assumption implicit in using first order kinetics: microbes grow quickly enough so that they can rapidly recover from any stress, as quoted from Finlay et al. (1997) above. For models operating at the ecosystem scale and above, as long as populations can recover from stress over periods of days to weeks that assumption would probably be valid. As Escherichia coli can double in 20 min, and Penicilium spp. can cover a piece of bread in days, this probably does not seem like an unreasonable assumption. However, it is not necessarily valid.

5. Microbial Recovery from Stress

Many microorganisms (particularly some fungi) grow slowly. Even those that regrow quickly may have to recolonize habitats after stress-induced local extinction. Recolonization and regrowth dynamics in soils are not well understood. However, there are some data that suggest that they may be important in controlling ecological processes.

Clein and Schimel (1994) found that a single one-day drying-rewetting could reduce microbial respiration in birch litter for more than 60 days in a lab assay, causing a 25% reduction in total C respired. Schimel et al. (1998) did a field study to expand on that work. They placed bags containing birch litter in the field and used watering and drought shelters to establish treat

25 20 15


I 20

10 5

FIGURE 2 Carbon respired over a 10-day incubation in the lab on litter samples that had been incubated under different treatments. Panel

(a) shows data from all treatments pooled into a single analysis, while

(b) shows only data from the rewet weekly and natural conditions treatments. The continuously moist treatment was not significantly different from the rewet weekly treatment. Data are from Schimel et al., 1999.

ments including: (a) continually moist, (b) watered weekly, and

(c) natural conditions (which was actually biweekly rain). Samples were harvested every week over a month. Respiration rates at field moisture were measured over a 10-day period in the lab to establish the potential activity of the extant community. When C respired over the 10-day lab incubation was expressed as a function of moisture, the R2 for all the data pooled was 0.74 (Fig. 2a), which is often considered quite adequate for modeling ecological data. However, when the different treatments were analyzed separately, there were tighter responses and significant differences between treatments, even after accounting for the moisture of the sample (Fig. 2b). Samples that had experienced longer drought in the natural conditions had lower respiration, biomass, and specific activity at any given water content than samples that were wet more frequently. Thus, this provided clear evidence that the stress history substantially affected the size and functioning of the microbial community and that it could not recover over at least the 10-day incubation following harvest. A study examining multiple freeze-thaw cycles showed similar results (Schimel and Clem, 1996). Each stress cycle reduced the ability of the surviving microbial community to process organic matter and respire, without evidence of recovery over the course of a one-month experiment.

Another kind of stress that is important in trace gas dynamics is shifting aeration/anaerobiosis. In wetland systems the aeration history of the system is a significant controller of CFI4 efflux and the lag in the development of microbial communities is long. Moore and Roulet (1993) equilibrated microcosms for 20 days saturated, then spent 25 days dropping the water table to 50 cm, left them drained for 15 days, and then reflooded them over an additional 25 days. Methane fluxes were between 9 and 116-fold greater as the water table was dropping than as it was rising. This difference was a combination of the release of pore-water CH4 and the inability of methanogens and methanotrophs to adapt to the changes in aeration state. Temperature variations can also produce significant hysteresis in process rates, mediated through changes in microbial communities (Updegraff et al., 1998).

Thus, I believe that episodic stress events can reduce the size of even the bulk respiring community substantially enough to have measurable effects on total process rates at the ecosystem scale. These studies and others (e.g., Dickens and Anderson, 1999; Yavitt and Lang, 1990) also suggest that recovery from such stresses may not be as rapid as many have assumed. As many ecosystems experience such episodic pulse stresses with some regularity (e.g., freezing-thawing in northern and alpine systems, drying-rewet-ting in arid and semiarid systems) it seems that these effects may have ecosystem consequences. The actual importance of such variations in specific microbial populations on larger and longer scales has really not been well explored. We know very little about historical legacy effects that are mediated through microbial communities. The few studies that have actually looked for such effects often find them (e.g., Updegraff el al., 1998 and others cited above), suggesting that the assumption in biogeochemical models that microbial population size never controls process rates, and that microbial processes therefore have no history may be wrong. 1'his all suggests that studying such legacy effects and microbial community dynamics at biogeochemistry-relevant scales is a fruitful area for collaboration among microbial ecologists and biogeo-chemists.

6. Conclusions: Integrating

Biogeochemistry and Microbiology

While there are few good case studies for the two issues that I have raised (different physiological responses and biomass limitation of process rates), I believe that that is less because the cases are rare than because few researchers have designed studies to test these possibilities. Thus, as a message to a new Institute for Biogeochemistry, I believe that the important point is that the microbiology that has been incorporated implicitly into most biogeochemical models is microbial physiology. 1 have argued here and in other papers (Schimel, 1995; Schimel and Gulledge, 1997) that these models may need to consider aspects of microbial community ecology as well. Ignoring these effects will probably only very rarely produce order-of-magnitude errors, but I believe that there will be cases where 25-50% errors may be likely. Additionally, much of the unexplained error and surprises in current biogeochemical studies may be due to unaccounted-for microbial com

FIGURE 2 Carbon respired over a 10-day incubation in the lab on litter samples that had been incubated under different treatments. Panel

(a) shows data from all treatments pooled into a single analysis, while

(b) shows only data from the rewet weekly and natural conditions treatments. The continuously moist treatment was not significantly different from the rewet weekly treatment. Data are from Schimel et al., 1999.

munity dynamics (Schimel, 2000). Incorporating microbial community effects into biogeochemical models will require efforts at both the microbial and biogeochemical modeling ends:

1. From the microbial side, we need more research targeted at understanding when and where microbial community effects, either through variations in physiology or through changes in populations sizes, have large-scale impacts. Many short-term lab studies have been done, but few have attempted to extrapolate to larger spatial and temporal scales.

2. From the biogeochemical modeling side, we need to experiment with models where the microbiology is less implicit by incorporating microbial community effects. Models should consider incorporating response functions that may vary over time with environmental conditions, as proposed by Updcgraff ct al. (1998). They should also consider incorporating the size of the active microbial population as a control on the rate of a process.

Through collaborative efforts on these two points, we can determine how best to deal with the relevant microbial dynamics in biogeochemical models, and how to predict them. From this, I hope we will be able to develop models that simply, yet effectively incorporate microbial community dynamics. These models will clearly not incorporate a lot of detailed and explicit microbiology. Rather, we must develop a solid enough understanding of the relevant microbial community dynamics so that we can model past the microbes, going directly from environmental drivers to process rates, thus making the microbiology, once again, implicit.


Acea, M. J., and Carballas T. (1996). Changes in physiological groups of microorganisms in soil following wildfire. FEMS Microbiol. Ecol. 20, 33-39.

Ambus, P., and Christensen, S. (1993). Denitrification variability and control in a riparian fen irrigated with agricultural drainage water. Soil Biol. Biochem. 25,915-923. Anderson, T. -H„ and Joergensen, R. G. (1997). Relationship between SIR and FE estimates of microbial biomass C in deciduous forest soils at different pH. Soil Biol. Biochem. 29, 1033- 1042. Berg, P., and Rosswall, T. (1987). Seasonal variations in abundance and activity of nitrifiers in four arable cropping systems. Microb. Ecol. 13, 75-87.

Bodelier, P. L. E„ Roslev, P., Henckel, T„ and Frenzel, P. (2000). Ammonium stimulates methane oxidation in rice soil. Nature 403, 421-424. Both, G. J., Gerards, S., and Laanbroek, H. J. (1992). Temporal and spatial variation in the nitrite-oxidizing bacterial community of a grassland soil. EEMS Microbiol. Ecol. 101, 99-112. Bottner, P. (1985) Response of microbial biomass to alternate moist and dry conditions in a soil incubated with l4C- and 13 N-labeled plant material. Soil Biol. Biochem. 17, 329-337. Brussaard, L., BehanPelletier, V. M., Bignell, D. E., Brown, V. K., Didden, W„ Folgarait, P., Fragoso, C„ Freckman, D. W„ Gupta, V. V. S. R„ Ilat-tori, X, Hawksworth, D. L„ Klopatek, C„ Lavelle, P., Malloch, D. YV„ Rusek, J., Soderstrom, B„ Tiedje, J. M„ and Virginia, R. A. (1997). Biodiversity and ecosystem functioning in soil. Ambio 26, 563-570.

Cheng, W, Virginia, R., Oberbauer, S., Gillespie, C., Reynolds, J., and Ten-hunen, J. (1998). Soil nitrogen, microbial biomass, and respiration along an arctic toposequence. Soil Sci. Soc. Am.}. 62, 654-662.

Chertov. O. G., and Komarov, A. S. (1996). In "Evaluation of Soil Organic Matter Models" (D. S. Powlson, P. Smith, and J. U. Smith, Eds.), pp. 231-236. Springer, Berlin.

Clein, J. S., and Schimel, J. P. (1994). Reduction in microbial activity in birch litter due to drying and rewetting events. Soil Biol. Biochem. 26, 403-406.

Clein. J. S„ and Schimel, J. P. (1995). Nitrogen turnover and availability during succession from alder to poplar in Alaskan taiga forests. Soil Biol. Biochem. 27, 742-752.

Dickens, H. E., and Anderson, J. M. (1999). Manipulation of soil microbial community structure in bog and forest soils using chloroform fumigation. Soil Biol. Biochem. 31, 2049-2058.

Finlay, B. J., Maberly, S. C., and Cooper, J. I. (1997). Microbial diversity and ecosystem function. Oikos 80, 209-213.

Groffman, P. M., (1991). Ecology of nitrification and denitrification in soil evaluated at scales relevant to atmospheric chemistry. In "Microbial Production and Consumption of Greenhouse Gases: Methane, Nitrogen Oxides, and Halomethanes" (J. E. Rogers and W. B. Whitman, Eds.), pp. 201-217. American Society for Microbiology, Washington, DC.

Groffman, P. M., Axelrod, E. A., Lemunyon, J. L., and Sullivan, W. M. (1991). Denitrification in grass and forest vegetated filter strips. /. Environ. Qual. 20,671-674.

Groffman, P. M. and Bohlen, P. J. (1999). Soil and sediment biodiversity— Cross-system comparisons and large-scale effects. Bioscience 49, 139-148.

Gulledge, J., and Schimel, J. P. (1998). Low-concentration kinetics of atmospheric CH4 oxidation in soil and the mechanism of NH4T inhibition. Appl. Environ. Microbiol. 64, 4291-4298.

Hendrickson, O., and Chatarpaul, L. (1984). Nitrification potential in an alder plantation. Can. J. For. Res. 14, 543-546.

Kieft L. T., Soroker, E., and Firestone, M. K. (1987). Microbial biomass response to a rapid increase in water potential when a dry soil is wetted. Soil Biol. Biochem. 19, 119-126.

Leffelar, P. A., and Wessel, W. W. (1988). Denitrification in a homogeneous, closed system: Experiment and simulation. Soil Sci. 146, 335-349.

Li, C. (1996). The DNDC model. In "Evaluation of Soil Organic Matter Models" (D. S. Powlson, P. Smith, and J. U. Smith, Eds.), pp. 263-268 Springer, Berlin.

Lin, Q„ and Brookes, P. C. (1999). An evaluation of the substrate induced respiration method. Soil Biol. Biochem. 31, 1969-1983.

Madigan, M. T„ Martinko, J. M„ and Parker, J. (1997). "Brock Biology of Microbiology" 8th. Prentice Hall, Upper Saddle River, New Jersey.

McGill, W. B. (1996). Review and classification of ten soil organic matter (SOM) models. In "Evaluation of Soil Organic Matter Models" (D. S. Powlson, P. Smith, and J. U. Smith, Eds.), pp. 111-132. Springer, Berlin.

Moore, T. R„ and Roulet, N. T. (1993). Methane flux: water table relations in northern wetlands. Geophys. Res. Lett. 20, 587-590.

Morley C. R., Trofymow, J. A., Coleman, D. C., and Cambardella, C. (1983). Effects of freeze-thaw stress on bacterial populations in soil microcosms. Microb. Ecol. 9, 329-340.

Parton, W. J„ Schimel, D. S„ Cole, C. V„ and Ojima, D. S. (1987). Analysis of factors controlling soil organic matter levels in Great Plains grasslands. Soil Sci. Soc. Am. J. 51, 1173-1179.

Roberts, D. V. (1977). "Enzyme Kinetics." Cambridge University Press. Cambridge.

Saad, O., and Conrad, R. (1993). Temperature dependence of nitrification, denitrification, and turnover of nitric oxide in different soils. Biol. Fertil. Soil. 15,21-27.

Schimel, 1. (1995). Ecosystem consequences of microbial diversity and community structure. In "Arctic and Alpine Biodiversity: Patterns, Causes, and Ecosystem Consequences" (F. S. Chapin and C. Korner, Eds.), pp. 239-254. Springer-Verlag, Berlin.

Schimel, J. (2000). Rice, microbes, and methane. Nature 403, 375-377.

Schimel, J. P., and Gulledge, I. (1998). Microbial community structure and global trace gases. Global Change Biol. 4, 745-758.

Schimel, J. P., Gulledge, I. M„ Clein-Curley, J. S., I.indstrom, J. F.., and Braddock, I. F. (1999). Moisture effects on microbial activity and community structure in decomposing birch litter in the Alaskan taiga. Soil Biol. Biochem. 31, 831-838.

Schimel, J. P., and Clein, J. S. (1996). Microbial response to freeze-thaw cycles in tundra and taiga soils. Soil Biol. Biochem. 28, 1061 -1066.

Schlessinger, W. H. (1997). "Biogeochemistry," 2nd ed.. Academic Press, San Diego.

Shi, W., and Norton, J. M. (2000). Microbial control of nitrate concentrations in an agricultural soil treated with dairy waste compost or ammonium fertilizer. Soil Biol. Biochem. in press.

Updegraff, K„ Bridgham, S. D„ Pastor, J., and Weishampel, P. (1998). Hysteresis in the temperature response of carbon dioxide and methane production in peat soils. Biogeochemistry 43, 253-272.

Wall, D. H., and Moore, J. C. (1999). Interactions underground—Soil biodiversity, mutualism, and ecosystem processes. Bioscience 49, 109-117.

Wardle, D. (1998). Controls of temporal variability of the soil microbial biomass: a global-scale synthesis. Soil Biol. Biochem. 30, 1627-1637.

Yavitt, J. B., and Lang, G. E. (1990). Methane production in contrasting wetland sites: Response to organo-chemical components of peat to sulfate reduction. Geomicrobiol. ]. 8, 27-46.

Zumft, W. G. (1992). The denitrifying prokaryotes. In "The Prokaryotes" (A. Balows, H. G. Trsper, M. Dworkin, W. Harder, and K. H. Schleifer, Eds.), pp. 554-582. Springer-Verlag, New York.

This Page Intentionally Left Blank

0 0

Post a comment