; w ‘ 33.31" 3» 92 . m3. 3 r; ”333:." » - 31'? 33"‘13" h. , . ‘33? 33.1} 3.3 L131:.3;”L!w"'“f.“7’;'-7-“ ’ q — '1. q 4?? Y |* Q ‘3 ._ f 3 3731.131" -; I- ‘ L..\¢V , _ . ('3‘ 4 u “Jug”, I L.-L “(33?” . *3 2.15.13" . c _; 3:313 *L'Ifiu 3 4- 33 13v \x ‘ g3:- 3153.‘ ' .- .. - 3- 32.1.".3-23 a" m -" I- y» u_nhl . ~_.$ fl ‘;‘.‘ ~ a $51.1. . I :5: - -45. >- . .I‘ I _ =7 “32%;; 51:4” ‘ ' 1 '., 'F- “5“" . I L 9?” ‘ 1 .1 393323313“ 43' 11%” 3 «u 11...}; t. tum '3‘1‘1 3.2.39 . 3 n 1? ”351,; 1 F ‘.‘ “FILE. .'4 ‘LV""‘ ' r .L’I’353‘l—it— ‘Ax «:33- 3,1333 .3“ng 4 >..‘ ' . ‘ L‘ V I. "sif'yfifi‘" - 4 ; I J” A” ‘ ‘1’! ! .-.- .; ~_ \“o‘l’ ‘2‘. J 555'; a; 3 r id L. g.» ~17 “2 - .. w“ - a" .. :12?“ 79!. 4.. 3". -.‘. V? 5.2 L. a L g L , ‘ ‘3‘” L :7'\. 3:53}? W ""m—Laza: .3235 , V...” . .~\p—‘ "4.50.. '41" 'IxfipI — . 3 4414;“, - .\ 53f.) “-".-.L .7 :‘V’UL ' 4. . U-vw . ., .T .v ”V “3&2. ‘ 2‘ .‘123 3 fing '1'.‘ :93 iii -.’,3..,\'*11.':~33 3' '3" ’LL“ 3:2"? ' LI"... ‘2} . L‘;.-. ,2 ‘ ' r " :3 >151 “'3”! ‘53”; L LN"; 'L.L' LL.“ I 3 33' .n h. 33'1- .1. - n‘. ‘,. L .ILI ‘Ilr .LL ‘ 31' n' 1. '."'LL _.. bu 3:33.333)" I. l ‘ .. 3 |‘ L L\\'I .' .3" :3 “.3 ." l , ‘4 ‘| . '1'" “3’. “."31'\|'L:":‘3' .2333. 3 NJ 'I. I’ll I 1. 'fl L .- ‘3‘“ '3'. I ‘ 3‘. _V,‘ ‘-.|" 3‘.. -LI' | I L.‘ .L -L1LLL“"ITA\L. In“: ‘ 3| LLju .L 3;], I it?“ t'. ' 3, . ’ ‘ .1 ' .' 3: 3 L3..‘ 5:" n L; 1"” ‘IL '3 ., ' :I\"3'\.,1LI‘.L|fo ' ‘qu‘lL 4'“, 1‘. I I . #3. ; . 3 ‘ 33 ”UL 3: W,‘ 3'14 "3'“. 1.?'u\".‘..‘131'n‘L3'3?‘.‘ouLL3' .' I :3, ..1. {#1.}33’,“ 3. 3~ n - . .' 1 I. 3 . 1 I 3 "13 L. 3 ‘33 L . 3216‘ 3 L3. 3 ' -H' H :‘LLD III' -LL P .3“ H . I . 3. 13"“ L. "L . L’ _" .'\:I ha." .L ”('IK l3 - ' l 3 u (L . 3' 3 ' ' 3 ' 3 LLIL“ 6?.vale '4" 3'.\|33£'I'I [a 3'13“” ’3. 3" fin 3-\ H'“K:3‘ 3,‘ ‘Ll't'fl-L 3333’ LLLJ’yl 3. L‘jlll3"l[ LvL' L3 9"”. ‘ W ““3 .3163. ' Mint 33.". « 3;. m”: 33.33”“ 3 3 a ' 3LL "3 L, I LL;L 31L ‘ ‘ IDA-'13": 3' ‘73:» 3333" 3th 34“ ’1 \ . '3‘". “LLI'L'. |L33‘1"‘\L"‘ ' " 1" EL. .‘I'V'LL-L- MWWWWWWWWW WWWWWWW 1293 006427 ‘ 5.5"; n', .35 3 i 4—Eaii'22aw‘ ”- . u; ‘~ 0: *fiflgmhfi'fiu it W‘s-die \U‘QL.“ g6 {469185-735 . “A; \uu- Raga» ‘ This is to certify that the dissertation entitled NITROGEN CYCLING IN SOILS: SIMULTANEOUS ESTIMATION OF TRANSFORMATION RATES, DIFFUSIONAL CONTROL OF DENITRIFICATION, AND ESTABLISHMENT OF DENITRIFICATION CAPACITY presented by David Douglas Myro 1d has been accepted towards fulfillment of the requirements for Ph.D. Soil Microbiology degree in ’r professor Ma' Date 6 February 1984 MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771 RETURNING MATERIALS: MSU Place in book drop to LJBRARJES remove this checkout from .anlncjl-IL. your record. FINES will be charged if book is returned after the date stamped below. flUi 2 4 3991 1 £25; :1. (,7 NITROGEN CYCLING IN SOILS: SIMULTANEOUS ESTIMATION OF TRANSFORMATION RATES, DIFFUSIONAL CONTROL OF DENITRIFICATION, AND ESTABLISHMENT OF DENITRIFICATION CAPACITY By David Douglas Myrold A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSPOPHY Department of Microbiology and Public Health 1984 ABSTRACT NITROGEN CYCLING IN SOILS: SIMULTANEOUS ESTIMATION OF TRANSFORMATION RATES, DIFFUSIONAL CONTROL OF DENITRIFICATION, AND ESTABLISHMENT OF DENITRIFICATION CAPACITY By David Douglas Myrold Rates of mineralization, immobilization, nitrification, and denitrification were simultaneously estimated in three dissimilar soils. The estimation process included elements of mathematical modeling, nonlinear parameter estimation, and the use of 15nH4+ as a tracer. Analysis of the sensitivity coefficients showed that good estimates of the mineralization, immobilization, and rdtmification parameters could be obtained, but denitrification parameters could not be as well defined. The results suggested that biomass N estimated by the CHC13 fumigation method is the major component of the active organic N pool over short time periods (< 3 weeks). N cycling in forest soil studied was best fit with zero order kinetics. In this soil, mineralization and immobilization were the dominant processes, with rates of 1.0 and 0.67 pg N g‘1 d‘l, respectively. Two agricultural soils were used, one high in organic matter, the other low in organic matter. In both agricultural soils, N cycling was best described with the first order model. nitrification, which had a rate constant of 1.3 d’l. Nitrification was also rapid in the low organic matter soil (1.6 d‘l), however, this soil also had a high rate of immobilization (1.7 d‘l). David Douglas Myrold A model of N03” reduction and diffusion was deve10ped and used along with the Thiele modulus (a dimensionless parameter) to determine the conditions under which denitrification.would and would not be limited by N03" diffusion. Results from this exercise suggest that, under anaerobic conditions, only aggregates greater than 0.2 cm will experience a N03‘ diffusional limitatnnu In aerobic soils, only large aggregates have anaerobic centers, and under these conditions N03 diffusional limitations are more likely. Experimental results with a clay loam soil showed no effect of a N03‘ diffusional limitation which was in agreement with model predictions. Experiments were conducted to determine the effect of carbon, water, and N03‘ additions on the deve10pment of denitrification capacity in soil. There was no effect of either N03"or water additions on denitrification capacity. However, added carbon caused a significant increase in denitrification capacity. This response to added carbon was paralleled by a similar increase in microbial ATP. These results suggest that the increase in denitrification capacity was due to a proportionate increase in denitrifier and non-denitrifier biomass. To Jackie and Kirk and in memory of Herman Edwin Myrold ii ACKNOWLEDGEME NTS Many individuals have added depth and richness to my four and one half year experience at Michigan State. I thank the members of my graduate committee for their support: John Breznak and Frank Dazzo for helping me learn some microbiology; Erik Goodman for introducing me to the mathematical aspects of biology; Peter Vitousek for giving me an ecosystem perspective of microbial activity. The interactions with the other members of Jim Tiedje's lab are treasured experiences. I learned and benefited from all those I worked with and hope that they may have gained some things from me. In particular, I thank Joe Robinson, who was an excellent sounding board for my mathematical ideas and whose work encouraged me to continue pursuing parameter estimation problems. I also thank Alan Sexstone, who was never too busy to talk about 15N, denitrification, or other less scientific (but perhaps more important) subjects. I hOpe that the friendships started here will last a lifetime. My deepest gratitude goes to Jim Tiedje, who I admire greatly. I thank him for allowing me to devise and pursue my own research plans--even when they may not have been particularly germane to the denitrification project and especially when the initial results did not look very promising. His dedication, incisive thinking, concern for others, and relationship with his family will always serve as a model for me. Perhaps my greatest thanks goes to my family for their love and support. I am grateful to my mother and father who taught me that all adversities in life can be overcome through love, faith, and patience. iii I thank Jackie, my wife, for her unfailing support, encouragement, and love through the rigors, delays, and long hours of graduate study; for sharing the joys and sorrows of this experience with me. I thank my son, Kirk, whose cheerful countenance and laughter helped me remember the truly important things in life. Lastly, I think it is fitting to acknowledge the inspiration and guidance of the Lord in this work. After reflecting upon the few months it took to bring this dissertation to fruition, I think that my feelings are best summarized by the following passage: ”...it is by faith that miracles are wrought..." (Moroni 7:37). iv TABLE OF CONTENTS L IST OF TABLES O O O O O O O O O O O O O O O O O O 0 LIST OF FIGURES O O O O O O O O O O I O O O O O O 0 INTRODUCTION 0 I O O O O O O O O O O O O O O O O O 0 CHAPTER 1. CHAPTER 2. CHAPTER 3. REFERENCES 0 O I O O O O O O I O O O O O SIMULTANEOUS ESTIMATION OF SEVERAL NITROGEN CYCLE RATES: THEORY AND APPLICATION . . MODEL DESCRIPTION . . . . . . . . . . . MATERIALS AND METHODS . . . . . . . . . RESULTS AND DISCUSSION . . . . . . . . . SUMMARY . . . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . DIFFUSIONAL CONSTRAINTS OF DENITRIFICATION IN SOIL O O O O I O O O O O I O O O O 0 MATERIALS AND METHODS . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . CONCLUSIONS . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . EFFECTS OF CARBON, NO3-, AND MOISTURE ON THE ESTABLISHMENT OF DENITRIFICATION CAPACITY IN SOIL O O O O O O O O O O O 0 MATERIALS AND METHODS . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . Page viii . 16 . 21 . 27 . 70 . 72 . 85 O 94 108 109 112 116 119 128 131 LIST OF TABLES Table Page CHAPTER 1 1 Soil characteristics . . . . . . . . . . . . . . . . . 22 2 Range of active organic N fraction estimated by two different methods . . . . . . . . . . . . . . . 46 3 Rates and rate constants of N cycle processes in three dissimilar 80118 0 o o o o o o o o o o o o o o 55 4 Zero order N cycle rates of Onaway loam with and without nitrapyrin . . . . . . . . . . . . . . 66 5 Effect of carbon additions on the partitioning of 15NH4+ in two soils after a seven day incubation period 0 O O O O O O O O O O O O O O O O O O 68 CHAPTER 2 1 Soil characteristics . . . . . . . . . . . . . . . . . 79 2 Effect of N03' and glucose additions on denitrification rates in anaerobic slurries . . . . . . 88 3 Denitrification rates of anaerobic cores and anaerobic slurries . . . . . . . . . . . . . . . . . . 91 4 Denitrification rate of Capac soil amended with N03- and succinate . . . . . . . . . . . . . . . . 92 5 Soil properties and calculated Thiele moduli . . . . . 96 6 Comparison of Km values for N03“ reduction obtained in various assay systems . . . . . . . . . . . 104 vi Table Page CHAPTER 3 1 Cumulative C02 evolution over a seven day incubation period . . . . . . . . . . . . . . . . . . . 120 2 Changes in microbial ATP . . . . . . . . . . . . . . . 124 vii LIST OF FIGURES Figure INTRODUCTION 1 The terrestrial nitrogen cycle. Modified from Jansson and Persson (1982) . . . . . . . . . . . . CHAPTER 1 1 Compartmental model of the N cycle . . . . . . . . 2 Normalized sensitivity coefficients for the N114+ response in a zero order model of N cycling. C] - mineralization rate (f1) 43 '- immobilization rate (f2) 0 - nitrification rate (f3) <> - denitrification rate (f4) V7 - initial active organic N pool size (Y3(O)) 3 Normalized sensitivity coefficients for the atom Z 15N114+ response in a zero order model of N cycling. Symbols as in Figure 2 . . . . . . . . 4 Normalized sensitivity coefficients for the N03“ response in a zero order model of N cycling. SymbOIS as in Figure 2 O O O O O O O O O O I O O O 5 Normalized sensitivity coefficients for the atom Z 15NO3' response in a zero order model of N cycling. Symbols as in Figure 2 . . . . . . . . 6 Normalized sensitivity coefficients for the N114+ response in a first order model of N cycling. symbOISasinFigur820 0 so. so. 0 o oo o o 7 Normalized sensitivity coefficients for the atom Z NH4+ response in a first order model of N cycling. Symbols as in Figure 2 . . . . . . . . 8 Normalized sensitivity coefficients for the N03‘ response in a first order model of N cycling. symb018381nFigureZO000000000000 viii Page 18 29 31 33 35 38 40 42 Figure 10 11 12 13 14 15 16 Page Normalized sensitivity coefficients for the atom Z N03“ response in a first order model of N cyling. Symbols as in Figure 2 . . . . . . . . . . 44 Reduction in the residual sum of squares for experiment 1 with Onaway loam as a function of the initial active organic N pool size . . . . . . . . . . 49 Experimental and simulated pool size data for Onaway loam soil 0 - NH4+ A — N03- -—- - simulated using estimated parameters . . . . . 51 Experimental and simulated atom Z 15N data for Onaway loam soil. Symbols as in Figure 11 . . . . . . . . . 53 Experimental and simulated pool size data for Capac Clay loam. symb018 as in Figure 11 o o o o o o o o o 57 Experimental and simulated atom Z 15N data for Capac clay loam. [3 - atom Z 15N114+ from 1-'5NH4+treatment E; -atom Z 15N03- from 15NH4+ treatment atom Z 15N02‘ from 15N03' treatment simulated using estimated parameters . . . . . 59 Experimental and simulated pool size data for Spinks sandy loam. Symbols as in Figure 11 . . . . . . . . . 62 Experimental and simulated atom Z 15N data for Spinks sandy loam. Symbols as in Figure 11 . . . . . . . . . 64 CHAPTER 2 Simulated N03” concentration profiles in an anaerobic 0.4 cm aggregate under conditions with (a) and without (A) a N03“ limitation . . . . . . . 83 Effect of N03- and glucose additions on the denitrification rate of anaerobic slurries . . . . . . 87 Effect of N03” concentration on the denitrification rate of anaerobic slurries of Capac soil . . . . . . . 89 Normalized reaction rate as a function of the dimensionless bulk concentration (So - Co/Km) for different values of the Thiele modulus,¢ , shown for values from 0.1 to 500 . . . . . . . . . . . 99 ix Figure Page Relationship between aggregate size distribution and the extent of diffusion limitation in aerobic and anaerobic soils . . . . . . . . . . . . . . . . . 103 CHAPTER 3 Hypothetical soil biomass composition and anticipated response from two different mechanisms for increasing denitrification capacity . . . . . . . 115 Changes in active denitrifier biomass over time of incubation. D - 23Z, no straw; I -23Z H20, 1 mg straw-C g‘l; 13 -28Z H20, no straw; ‘ -2878 H20, 1 mg straw-C 8-1 0 0 o o o o o o o o o 12]. Changes in the ratio of denitrification capacity to microbial ATP over time of incubation. Symbols as in Figure 2 . . . . . . . . . . . . . . . . 127 INTRODUCTION Nitrogen is a constituent of the nucleic acids which serve as the blueprints for living cells; nitrogen is a component of the enzymes which construct living cells; and nitrogen is a part of the polymers which form the structure of living cells. Indeed, with the exception of carbon and water, nitrogen constitutes the largest fraction of living cells. It is no small wonder that several volumes have been devoted to the physics, chemistry, and biology of nitrogen. In terrestrial ecosystems, nitrogen is often found to be limiting for plant growth. Consequently, much work in the biological sciences, particularly in agriculture, has been focused on the transformations of nitrogen in nature. In the past few years several books have been published to review what is known about the nitrogen cycle and the dynamics of its transformations (Nielsen and MacDonald, 1978; Clark and Rosswall, 1981; Stevenson, 1982). One of the most thought-provoking representations of the nitrogen cycle is that given by Jansson and Persson (1982), in which they divide the nitrogen cycle into three subcycles: the elemental, autotrophic, and heterotroPhic cycles (Figure 1). The elemental cycle connects the large reservoir of atmospheric NZ to living organisms through the microbially mediated processes of N2 fixation and denitrification. The autotrOphic cycle involves plant photosynthesis and concomitant assimilation of inorganic nitrogen from the soil solution and the subsequent return of organic nitrogen to the soil. The heterotrophic Cycle is dominated by the activities of microorganisms performing the Processes of mineralization, immobilization, and nitrification. From 1 Figure 1. The terrestrial nitrogen cycle. Modified from Jansson and Persson (1982). coaumuamatuaz GOHUUDfiwM oz oxMuaD zu0umafiafimm< seeds 2 wmeOfim usmaa coaumuamauuacma coaumxam z H musmwm coautnaaaaoaeH gamma sumac mmMEDam uncanaam muoxam Nz GOHUNNHHWHGCfl: z z m>auo< mmmEOHm Hmanouoaz mmmaowm Iaoz z emua -Hanmum this conceptual framework it is obvious that the activities of microorganisms are central to the cycling of nitrogen. There are three distinct pools of nitrogen in soil-organic N, NH4+, and NO3'--all of which are interconnected by microbially mediated processes (Figure 1). This interconnectedness not only provides the redistribution of nitrogen necessary for life, but also contributes to the stability of the system through feedback mechanisms. In conjunction with this feedback phenomenon, two branch points exist which are susceptible to regulation--the NH4+ and N03“ pools. With this in mind, it is not surprising that the concentrations of NH4+ and N03“ in soil are usually quite low, at least in comparison to the organic N pool- Regulation of nitrogen cycling in soil is affected primarily through microbial competition for the various forms of nitrogen, within the constraints of the environment (Rosswall, 1982). The final outcome of this competition is determined by the biomass of the competing populations, their affinities for the substrate competed for, and the amount of substrate available (Tiedje Si 31., 1982). Each of these factors, in turn, may be modulated by physical or chemical processes. These interactions between the physics and chemistry of the environment and the biology are not likely to be static, but rather dynamic in nature. In order for the nitrogen cycle and its regulation to be studied as a whole, it is necessary to be able to measure nitrogen cycle processes as an integrated unit. This requires the use of a tracer, which allows one to follow the labelled nitrogen through the circles and along the arrows of the nitrogen cycle. Nitrogen-15 is the tracer of choice for any studies longer than a few tens of minutes and has been used as a qualitative, or semi-quantitative, indicator for several decades. The interested reader is referred to several excellent reviews which cover the historic uses of 15N in biology and agriculture (Hauck 1973; Hauck and Bremner, 1976; Faust, 1982). In addition to being able to qualitatively follow the fate and partitioning of nitrogen in the soil system, one would also like to be able to quantitatively measure the fluxes of nitrogen between the various pools of the nitrogen cycle and how rapidly these pools are turned over. Mathematical models or kinetic analyses are generally needed to address this problem. Jansson (1958) did the pioneering work along these lines, applying the simplified model of Kirkham and Bartholomew (1955) to mineralization-immobilization dynamics in soil. The next major push forward came in the late 1970's when Koike and Hattori (1978) used the principles of isotope dilution to simultaneously measure nitrification and nitrate reduction in marine sediments. Subsequently, this technique has been successfully applied to mineralization and immobilization in sediments (Blackburn, 1979) and water column studies (Gilbert 53E §_]_._., 1982) and to nitrification and nitrate reduction in rice paddy soils (Watanabe _e_t_:£l_., 1981). Thus far, the most ambitious application of isotOpe dilution to nitrogen studies has been the simultaneous measurement of rates of denitrification, dissimilatory N03- reduction to NH4+, mineralization, and immobilization in anaerobic soil slurries (Tiedje 3521., 1981). The logical extension of this work seems to be the application of more sophisticated mathematical modeling and rate estimation techniques. One such approach was used by Van Cleve and White (1980) who applied the principles of compartmental analysis (cf., Jacquez, 1972)--which has long been used in studies of animal metabolism--to nitrogen cycling in soils of the Alaskan bush. Using 15N as tracer, they were able to calculate total fluxes of nitrogen among the organic N, NH4+, and N03- pools, as well as the turnover times of these pools. Unfortunately their approach required them to assume that the system they worked with was at a steady state--a situation not likely to occur often in nature. This should not be a limitation, however, since methods of analysis are available to study non-steady state systems. Winkler and Hubner (1979) have applied the principles of nonlinear parameter estimation and 15N labelling to measure protein turnover in plants. In the first chapter of this thesis a similar approach will be used to simultaneously estimate rates of nitrogen cycle processes in soil. One way in which nitrogen processes are regulated is through the physical process of diffusion, which determines the rate of substrate supply and thus influences the concentration of the substrate that is available to a microorganism. On a very macroscOpic scale, this principle has been recognized in the large nitrogen transport and transformation models developed by soil physicists (cf., Tillotson, 1980). Indeed, Reddy 3321., (1978 and 1980) have shown the importance of N114+ and N03' diffusion in determining the rate and reaction order of denitrification in flooded soils. In well drained soils, much of the work on the interaction between the physics of diffusion and microbial activity has focused on oxygen diffusion in aggregates and its effect on microbial activity and the establishment of anaerobic microsites in otherwise aerobic soils (e.g., Greenwood and Goodman, 1964; Smith, 1980). This influence of oxygen diffusion and microbial respiration on denitrification was shown by Greenwood (1962) for glucose and N03‘ amended soil crumbs and has recently been demonstrated in natural soil aggregates (Sexstone 35.21}: 1984). It might be expected that the diffusion of N03" might also play a role in determining rates of denitrification in well drained soils- Indeed, it has often been suggested that the high Km values for N03- reduction measured in soils is an indication of N03" diffusion limitations (cf., Firestone, 1982). The second chapter of this dissertation examines the question of whether or not NO3- diffusion is a limiting factor of denitrification in aggregated soils. The relative size of competing populations is one component in determining the outcome of competition, and hence the regulation of ‘nitrogen cycling. Whether or not a given group of organisms can multiply and survive in soil depends upon their tolerance of adverse environmental conditions and their ability to obtain and utilize substrates needed for growth. Smith and Tiedje (1979) deve10ped an assay system for quantifying the denitrifying capacity of a soil, which directly reflects the amount of active denitrifier biomass. When this method was used to survey soils from a range of habitats, denitrification capacity was found to be directly related to the moisture regime and carbon content of the soils (Tiedje gt 31., 1982). Other work has shown that denitrifier populations increase when soils are amended with N03” and incubated anaerobically (Jacobson and Alexander, 1980). These observations led to the final chapter of this dissertation, which examines the effect of water, N03", and carbon additions on the establishment of denitrification capacity in soil. To summarize, the work reported in this dissertation is built upon the foundation of the interacting microbial transformations of the nitrogen cycle and the regulation of these transformations. Chapter I focuses on the problems of measuring several of the interconnected nitrogen cycle rates in a single experiment. Using 15N as a tracer and applying procedures long used in engineering and statistics, a method for simultaneously estimating mineralization, immobilization, nitrification, and denitrification rates is given. Chapter II addresses the area of environmental control of microbial activity. Specifically, the potential for N03- diffusion to limit denitrification is studied from both a theoretical and experimental perspective. Finally, Chapter III examines factors which control the magnitude of the denitrification capacity of soil and attempts to elucidate the mechanism of this control. This is an example of how the environment can affect microbial biomass size and thereby influence the cycling of nitrogen. 10. 11. 12. 13. 14. REFERENCES Blackburn, T.H. 1979. Method for measurin rates of NH4+ turnover in anoxic marine sediments, using a 15N-NHz, dilution technique. Apple Enflrono MicrObiolo 37:760-765. Clark, F.E. and T. Rosswall (ed.). 1981. Terrestrial nitrogen cycles. Ecol. Bull. (Stockholm) 33. 714 p. Faust, H. 1982. Stable isotopes in agriculture. p. 421-431. In_ H.-L. Schmidt, H. Forstel, and K. Heinzinger (ed.) Stable isotopes. Elsevier Scientific Publishing Company, Amsterdam, The Netherlands. Firestone, M.K. 1982. Biological denitrification. IB_F.J. Stevenson (ed.) Nitrogen in agricultural soils. Agronomy 22:289-326. Am. Soc. Agron., Madison, WI. Glibert, P.M., F. Lipshultz, J.J. McCarthy, and M.A. Altabet. 1982 . Isotope dilution models of uptake and remineralization of ammonium by marine plankton. Limnol. Oceanogr. 27:639-650. Greenwood, D.J. 1962. Nitrification and nitrate dissimilation in soil. II. Effect of oxygen concentration. Plant Soil 17:378-391. Greenwood, D.J. and D. Goodman. 1964. Oxygen diffusion and aerobic respiration in soil spheres. J. Sci. Fd. Agric. 15:579-588. Hauck, R.D. 1973. Nitrogen tracers in nitrogen cycle studies--past use and future needs. J. Environ. Qual. 2:317-327. Hauck, R.D. and J.M. Brmener. 1976. Use of tracers for soil and fertilizer research. Adv. Agron. 28:219-266. Jacquez, J.A. 1972. Compartmental analysis in biology and medicine Elsevier, Amsterdam, The Netherlands. 237 p. Jacobson, S.N. and M. Alexander. 1980. Nitrate loss from soil in relation to temperature, carbon source and denitrifier populations. Soil Biol. Biochem. 12:501-505. Jansson, S.L. 1958. Tracer studies on nitrogen transformations in soil with special attention to mineralization-immobilization relationships. K. Lantthogsk. Anntr. 24:101-361. Jansson, S.L. and J. Persson. 1982. Mineralization and immobilization of soil nitrogen. IE_F.J. Stevenson (ed.) Nitrogen in agricultural soils. Agronomy 22:229-252. Am. Soc. Agron., Madison, WI. Kirkham, D. and W.V. Bartholomew. 1955. Equations for following nutrient transformations in soil, utilizing tracer data: II. Soil Sci. Soc. Am. Proc. 19:189-192. 15. 16. 17. 18. 19. 20, 21. 22. 23. 24. 25. 26. 27. 10 Koike, I. and A. Hattori. 1978. Simultaneous determination of nitrification and nitrate reduction in coastal sediments by a 15N dilution technique. Appl. Environ. Microbial. 35:853-857. Nielson, D.R. and J.G. MacDonald (ed.). 1978. Nitrogen in the environment, Vol. 1 and 2. Academic Press, New York, NY. Reddy, K.R., W.H. Patrick, Jr., and R.E. Phillips. 1978. The role of nitrate diffusion in determining the order and rate of denitrification in flooded soil: I. Experimental results. Soil Sci. Soc. Am. J. 42:268-272. Reddy, K.R., W.H. Patrick, Jr., and R.E. Phillips. 1980. Evaluati on of selected processes controlling nitrogen loss in a flooded soil. Soil Sci. Soc. Am. J. 44:1241-1246. Rosswall, T. 1982. Microbiological regulation of the biogeochemical nitrogen cycle. Plant Soil 67:15-34. Sexstone, A.J., N.P. Revsbech, T.B. Parkin, and J.M. Tiedje. 1984. Direct measurement of oxygen profiles and denitrification rates in soil aggregates. Soil Sci. Soc. Am. J. (submitted) Smith, K.A. 1980. A model of the extent of anaerobic zones in aggregated soils and its potential application to estimates of denitrification. J. Soil Sci. 31:263-277. Smith, M.S. and J.M. Tiedje. 1979. Phases of denitrification following oxygen depletion in soils. Soil Biol. Biochem. 11:261-267. Stevenson, F.J. (ed.) 1982. Nitrogen in agricultural soils. Agronomy 22. Am. Soc. Agron., Madison, WI. 940 p. Tiedje, J.M., J. Sorensen, and Y.-Y.L. Chang. 1981. Assimilatory and dissimilatory nitrate reduction: perspectives and methodology for simultaneous measurement of several nitrogen cycle processes. IE_F.E. Clark and T. Rosswall (ed.). Terrestrial nitrogen cycles. Ecol. Bull. (Stockholm) 33:331-342. Tiedje, J.M., A.J. Sexstone, D.D. Myrold, and J.A. Robinson. 1982. Denitrification: ecological niches, competition and survival. Ant. van Leeuwen. J. Microbiol. 48:569-583. Tillotson, W.R., C.W. Robbins, R.J. Wagenet, and R.J. Hanks. 1980. Soil water, solute and plant growth simulation. Utah Agric. Expt. Stu. 31.111. 5020 53 D. Van Cleve, K. and R. White. 1980. Forest-floor nitrogen dynamics in a 60-year-old paper birch system in interior Alaska. Plant Soil 54:359-381. 28. 29. ll Watanabe, I., B.C. Padre, Jr., and S.T. Santiago. 1981. Quantitat ive study and nitrification in flooded rice soil. Soil Sci. Plant Nutr. 27:373-382. Winkler, E. and G. Hubner. 1977. Concepts for the interpretation of tracer experiments and their application in the investigation of nitrogen metabolism. p. 303-310. In_Stable isotopes in the life sciences. IAEA, Vienna, Austria. CHAPTER 1 SIMULTANEOUS ESTIMATION OF SEVERAL NITROGEN CYCLE RATES: THEORY AND APPLICATION Nitrogen is constantly replenished throughout the biosphere by the interconnected transformations which constitute the N cycle. It is this cyclic quality of N transformations which make them both ecologically beneficial and difficult to investigate. Most of the research on N in soils has focused on the activity of a single process and factors which affect its activity. Considerably less effort has been expended to examine the interactions among several N cycle processes and effects of environmental perturbations on these interacting transformations. In order to examine N transformations as an interacting unit--or even to measure gross rates of N cycle processes--it is necessary to trace N through the various compartments of the N cycle. This can be done by using 15N. Research using 15N to measure the dynamics of N cycling can be partitioned into three, somewhat overlapping categories: (1) 15N as a tracer, (2) isotOpe dilution experiments using 15N, and (3) the application of mathematical models to 15N dynamics. Most frequently 15N has been used as a tracer. This application generally involves the addition of 15N labeled N114+ or N03“ (or 15N2 in N2 fixation work) and subsequent measurement of the 15N content of the soil organic and inorganic N and plant N. In fertilizer recovery and N balance 12 13 experiments this change in 15N is usually measured over the course of one growing season (e.g. Carter £31., 1967). It can also be used to measure N cycling over shorter time periods in the laboratory (Jansson, 1958; Ross 3531., 1964; Jones and Richards, 1977 and 1978). These types of experiments are useful in determining the relative fates and partitioning of added 15N, but provide only a qualitative estimate of process rates. IsotOpe dilution experiments involve the addition of 15N into a product pool. The subsequent dilution of the atom Z 15N in this pool by natural abundance N from a precursor pool is monitored over time. This principle, with numerous modifications, has been successfully used to study nitrification and nitrate reduction in sediments (Koike and Hattori, 1978; Nishio g£_§}3, 1983) and mineralization and immobilization in sediments (Blackburn, 1979) and water columns (Glibert £5 21., 1982). Nitrification and N03' reduction have been measured by isotOpe dilution in rice paddy soils (Watanabe 3521., 1981) and Tiedje _t _l_. (1981) used 15NH4+ and 15NO3- in a double labeling experiment to simultaneously measure rates of denitrification, dissimilatory N03- reduction to NH4+, mineralization, and immobilization in anaerobic soil slurries. The isotOpe dilution method is good for simultaneously measuring short term rates of N cycle processes, but assumes that the process rates are constant over each time interval. The rate estimates are also quite sensitive to the data variability since differences in the data are taken, which is an error-amplifying process. Models of the N cycle vary greatly in complexity. Different models have been used to fit experimental data (e.g., Mehran and Tanji, 1974) and to estimated rate constants for N cycle processes (Cameron and l4 Kowalenko, 1976), illustrating that no uniquely correct model for N cycling exists. Incorporating a 15}: label into experiments of N cycle dynamics should greatly enhance the estimation of N transformation rates and their corresponding kinetic rate constants by enabling gross rates of opposing reactions to be measured and also by allowing the separation of the organic N pool into reactive and unreactive fractions (Jansson, 1958; Juma and Paul, 1981). Using 15N along with total pool sizes in mathematical N cycle models was initiated by Kirkham and Bartholomew (1955) who examined mineralization and immobilization in a closed, two compartment system under steady state conditions. Jansson (1958) successfullyapplied their technique to mineralization-immobilization dynamics in soils receiving various organic amendments. The steady state condition has also been assumed by Van Cleve and White (1980) for a field study on N dynamics in a forest ecosystem. They applied the principles of compartmental analysis (cf., Jacquez, 1972) to their data and were able to determine total fluxes of N between the NH4+, NO3’, and organic N pools and the turnover times of these pools. Under many (perhaps most) circumstances in nature the N cycle is not at a steady state; pool sizes are constantly changing, negating the usefulness of the steady state approach. Analysis of tracer data under non-steady state conditions is more difficult, but can be done by means of nonlinear parameter estimation. Winkler and Hubner (1977) have applied this method, along with 15N labeling, to measure protein turnover in bean plants. The application of mathematical modeling and nonlinear parameter Estimation techniques to N cycling in soils would allow rates of several N cycle processes to be estimated simultaneously. Such an approach, 15 however, requires the assumption of an underlying kinetic mechanism for the N transformations occurring in the N cycle. In this paper we use 15N as a tracer in several soils and test the usefulness of nonlinear parameter estimation and mathematical modeling to simultaneously estimate mineralization, immobilization, nitrification, and denitrification rates in soil. We also examine the importance of heterotrOphic nitrification in a forest soil and the effect of a C addition on the relative rates of immobilization and nitrification. MODEL DE SCRI PT ION The structure of the nitrogen cycle makes it amenable to description as a compartmental system (Figure 1). The compartments are the pools of chemically or biologically distinct forms of nitrogen and the flows among these pools are the rates of the various nitrogen cycle processes. In our work we were primarily interested in N mineralization, immobilization, nitrification, and denitrification, since these are generally the dominant processes in unvegetated soils. The process of heterotrOphic nitrification is included in Figure 1, since it appeared to be comparatively large in one soil studied. The organic N pool was divided into two components--the passive and active fractions--according to the work of Jansson (1958). It was assumed that flow between these two organic fractions, or between the passive fraction and inorganic N pools, would be insignificant over the relatively short time span of our experiments (less than three weeks). The compartmental model shown in Figure 1 is described by the following differential equations, which can be derived from the mass balance of total N (”N + 15N) and 15N for each pool. le EE— = fl ' (£2 + £3) [1] de Ti?=f3+f5'f4 [2] dY3 dt = f2 " (£1 + f5) [3] l6 17 Figure 1. Compartmental model of the N cycle. 18 a H muswfim 3 828:3; assess: Ame 828:3; in: 2 gas $12 is 5223321. is 522.33.. is 5:00:32... is? 2,28; l9 dyl (Tc— = A3f1 ‘ A1(fz+ f3) [4] dy2 dt = A1f3 + A3f5 ' A2f4 [5] dy3 —dt = Alfz - A3(f1 + £5) [6] In these equations, Y1 is total N, yi is 15N, and A1 is the atom Z 15N in the 1th pool, where i = 1 for NH4+-N, i = 2 for N03“-N, and i - 3 for active organic N. The reaction rates of the processes, or flows, are designated by f3 where j - 1,2,3,4,5 for mineralization, immobilization, nitrification, denitrification, and heterotrophic nitrification, respectively. No kinetic interpretation has been given here to the reaction rates. However, zero order, first order, and Michaelis-Menten kinetics can all be implemented by simply inserting the appropriate rate equation for the fj terms. Since Y1 and A1 are the experimentally measured variables, Equations [1-6] can be combined to form the following: dA f1(A3 - A ) l 1 —- = [7] dt Yl dt Y2 Y2 dA3 = f2(A1 - A3) [9] dt Y 20 Equations [1-3] and [7-9] were solved with a Runge-Kutta integration scheme and used to model the dynamics of nitrogen cycling on the soils used in this experiment. For the work reported in this paper, we examined first and zero order models, which can be thought of as two subsets of Michaelis-Menten kinetics. Conceptually, Michaelis-Menten kinetics should best describe the microbial N transformations, under non-growth conditions. However, incorporation of Michaelis-Menten kinetics has at least two practical limitations: (1) doubling the number of parameters to be estimatednand good estimates of Km and Vmax for a single microbial reaction in pure culture are difficult to obtain because they are inherently correlated (Robinson, 1984)--and (2) many other processes, like diffusion, influence reaction rates in soil (Reddy §_t__a_1_., 1978). MATERIALS AND METHODS _S_<_)_i_1_s_. The feasibility of estimating the rates of several N cycle processes simultaneously was tested using three soils from different habitats, with different physical and chemical prOperties (Table 1). These soils were collected from the field, sieved to < 2mm, and stored at 4°C until used. Experiment 1. Onaway loam was amended with 3.4 pg 15NH4+-N g‘lsoil as 99 atom Z (ISNH4)ZSO4. The 15NH4'+ was applied to the soil as a fine spray using a syringe with a 22 gauge needle and mixed into the soil to promote even label distribution. The labeled soil was adjusted to a water content of 0.28 g g'1 and incubated in Parafilm covered beakers at 20°C. (Several holes were punched in the Parafilm covers to insure aerobic conditions, while minimizing water loss.) Five replicates were sampled at O, 1/2, 1, 2, 4, 7, 10, 14, and 21 days. Total concentrations and atom Z 15N of the N114+ and N03” pools were determined at each sampling time. Experiment 2. Capac clay loam was amended with either 15N labeled N114+ or N03”. Treatment 1 received approximately 7 ug 15NH4+-N g"1 soil (99 atxnn Z (15NH4)ZSO4) and a corresponding amount of natural abundance NO3'-N, while treatment 2 received about 7 pg 15 NO3'-N g"1 3011 (99.4 atom Z K15N03) and the same amount of natural abundance NH4+-N. The label was added as described in experiment 1 and the water content was adjusted to 0.20 g g‘l. Soil was packed into plastic cylinders to a bulk density of 1.5 g cm53, covered with plastic wrap, and incubated at 21 22 mmo.o n.o amoa modem cofiumuou amonmom\auou mamoaamm: ufiuooaaomm mxcfiom mN.o w.o amoa zmao :oHuwuou :monmom\auoo «Hmsumunuo ofiuom ommmo 0H.o m.n smog moooavum: subcuuoz wonuuoaomn cama< zmaoao z Hmuoa N ma ousuxoa coauouowo> coauwoawfimmmao mofiuom Hwom .moauwwuouoouoso Haom .H manna 23 25°C. Four cores from each treatment were sampled at 0, 1/2, 1, 3, 4, 8, 12, 15, 19, and 22 days for 15N and pool size analysis. Experiment 3. Preincubated Spinks sandy loam was amended with about 2.5 pg 15NH4"'-N g'1 soil (99 atom Z (15NH4)2804) and the moisture was adjusted to 0.13 g g'l. In experiments 3, 4, and 5, the 15N label was added using a chromatographic sprayer, which produced a very fine spray which promoted uniformity of label addition, and was then mixed into the soil. The soil was incubated in closed, Erlenmeyer flasks which were aerated daily to prevent anaerobiosis. Five replicate samples were taken at 0, 1/2, 1, 2, 3, 5, and 7 days, extracted for NH4+ and NO3', and measured for 15N and inorganic N concentration. Biomass C and N were also measured at each time point by the CHCl3 fumigation method. W Experiment 4. Three treatments of preincubated Onaway loam were set up to evaluate the presence and magnitude of heterotrophic nitrification and the effect of added carbon on the rates of immobilization and nitrification. All treatments received 2.5 pg 15NHz,"'-N g"'1 soil as 99 atom Z (15NH4)2804 and were adjusted to a water content of 0.28 g g‘l. Treatment 1 served as a control while treatment 2 received finely ground maple leaves at a rate of 3 mg g"1 soil--this is equivalent to typical litterfall values for Northern hardwood forests (Nadelhoffer 31:31., 1983). Treatment 3 received 10 ug nitrapyrin g‘l soil; prepared according to the procedure of Bremner _e_t_ 21., (1978). These treatments were incubated and sampled according to the schedule given for experiment 3. Experiment 5. Preincubated Capac clay loam was used to examine the effect of straw addition on the relative rates immobilization and 24 nitrification. Both treatments received about 2.5 ug 15NHz,"'-N g"'1 soil (99 atom Z (15NH4)ZSO4). Treatment 2 also received finely ground alfalfa straw at the rate of 1 mg straw-C g'1 soil. These treatments were incubated and sampled as described in experiment 3. Analytical procedures. NH4+ and N03’ were extracted from soil with 2 N KCl at either a 10:1 (experiments 1 and 2) or 5:1 (experiments 3-5) extractant to soil ratio. The 5:1 ratio provided greater sensitivity for the 15N ratio analysis. Biomass N and C were determined for experiments 3, 4, and 5 on separate subsamples of soil by the CHC13 fumigation method (Jenkinson and Powlson, 1976). The N flush (Nf) after 10 days of incubation was measured by extracting the accumulated NH4+ with 2 N KCl using a 5:1 extractant to soil ratio. The C02 flush (Cf) was measured by analyzing the headspace gas on a microthermister equipped GC. A conversion factor, kc, of 0.41 (Anderson and Domsch, 1978) was used to estimate biomass C and the nitrogen conversion factor, kn - 0.39 - 0.014(Cf/Nf), of Voroney and Paul (1984) was used to estimate biomass N. NH4'I, NO3", and biomass N (as NH4+) were prepared for mass spectrometer analysis using steam distillation as described by Bremner (1965).. 15N ratio measurements were made using a Micromass 602 isotope ratio mass spectrometer. Concentrations of NH4+ and N03” (after conversion to NH4+) were measured using the Solorzano (1969) method (experiments 1 and 2) or using a Technicon autoanalyzer (experiments 3, 4, and 5). Model Evaluation and Parameter Estimation. To obtain unique parameter estimates it is necessary that the sensitivity coefficients be linearly independent (Beck and Arnold, 1977). A sensitivity coefficient 25 is defined as the first derivative of a measured variable with respect to a model parameter (e.g., 3Y1/3f1 or 3A1/3f1). Sensitivity coefficients are linearly dependent when they are a constant multiple of one another. The degree of linear independence among the sensitivity coefficients can be assessed by plotting them and examining their relationships. Plotting sensitivity coefficients also yields information about Optimal experimental design, which parameters the model is most sensitive to, and how well the parameters will be determined. When a sensitivity coefficient is large in absolute value, the respective measured response contains much information about the parameter, while a sensitivity coefficient of zero contains no information about the parameter. Consequently, the best experimental design concentrates measurements during the time when sensitivity coefficients are large in absolute value. Sensitivity coefficients for both a zero order and first order model were calculated using parameter values close to those expected for the soils used in this study. In addition to the sensitivity coefficients for the rate parameters, the sensitivity coefficient for Y3(0)--the initial concentration of N in the active fraction--was calculated. These sensitivity coefficients were determined for the four normally measured responses: the concentration and atom Z 15N of the NH4+ and NO3" pools. Rates or rate constants of the N cycle processes were estimated using a nonlinear regression technique. A Gauss minimization method (Beck and Arnold, 1977) was used in conjunction with the interpolation-extrapolation step size routine of Bard (1974). Inequality constraints in the form of penalty functions were used to 26 insure that only reasonable (non-negative) parameter estimates were obtained (Bard, 1974). The sensitivity coefficients were obtained using the finite difference method suggested by Beck and Arnold (1977), and the model equations were integrated using a Runge-Kutta technique. The objective function (i.e., residual sum of squares) was minimized according to the least squares criterion. The nonlinear parameter estimation program was written in BASIC and implemented on a microcomputer. In addition to calculating parameter estimates, the program also calculated the parameter correlation matrix and provided approximate 952 confidence intervals for the parameters. RESULTS AND DISCUSSION Model Evaluation. Sensitivity coefficients for the zero order model are shown in Figures 2-5. The sensitivity coefficients for the mineralization, immobilization, and nitrification rates change linearly with time and are therefore linearly dependent, while all other sensitivity coefficients are zero when measuring the NH4+ pool (Figure 2). If the NH4+ pool is the only response measured, the mineralization, immobilization, and nitrification rates could not be uniquely identified. As one would intuitively expect, only a net rate could be obtained. When the sensitivity coefficients for the atom Z 15NH4+ response is examined, the mineralization rate is no longer linearly dependent with respect to the immobilization and nitrification rates, thus it is uniquely determined (Figure 3). However, the immobilization and nitrification rates are still almost linearly related. Adding the N03- response allows one to uniquely estimate the immobilization and nitrification rates, since the sensitivity coefficient for immobilization is zero (Figure 4) rather that approximately twice that of the nitrification sensitivity coefficient as it was in the previously described responses (Figures 2 and 3). The final response--atom Z 15N enhances the linear independency among the immobilization, mineralization, and nitrification rates (Figure 5). Close examination of Figures 2-5 will show that the denitrification rate is also uniquely determined. However, since its sensitivity coefficients are all_small (< SZ) compared to the others, the denitrification rate will be poorly determined. This is caused, at 27 Figure 2. 28 Normalized sensitivity coefficients for the NH4+ response in a zero order model of N cycling. Cl - mineralization rate (f1) [3 O immobilization rate (f2) nitrification rate (f3) denitrification rate (f4) initial active organic N pool size (Y3(0)) 29 .83??? Time (d) 30 Figure 3. Normalized sensitivity coefficients for the atom Z NH4 response in a zero order model of N cycling. Symbols as in Figure 2. 31 II“- '.‘-'.];‘ V ... -“““€€€€\ ...... ,~ ~ .-.-..'.:::.p OOOQOQ'.!.(_‘-‘u‘v,0- _ . II”- 'Y.Yar‘{h{o(.{alolauona . I 7:4 5:131:551'51'51';'a'a'4'4n ‘ A '... .. ..’A'A'A'A'A'A'A'A'A'A'A'A‘A'A'A'A'A'A' ’ Tp‘x‘vV'VVVV'V'VV'V'VVVV" 42 . . . . 0 2 4 6 8 w TiMe (d) Figure 3 32 Figure 4. Normalized sensitivity coefficients for the N03- response in a zero order model of N cycling. Symbols as in Figure 2. m 33 Time (d) Figure 4 34 Figure 5. Normalized sensitivity coefficients for the atom Z 5NO3'response in a zero order model of N cycling. Symbols as in Figure 2. 35 Q . . . 3.. . Ct .‘. Ct ‘. .v."...‘. ."' fft.‘~ . ‘_~_~ o o-O.Q.0-0.0-0.0-Q.0.0.03.0,..{0.0-9 o o o o o 0,. o 0, -" ' " ‘ . :;:'::O;O" Q'Q'Q’O‘O'O‘O‘Q O'Ogo':::;§;..o;|.: AV 7 l' " . . . . . §'§'§';';':~..... «as. 5‘-“ v . ‘ o'l . ' " I.I.I.I.u.- ' ' ' ‘ _ ‘ ' 'A"""' VVVVV A'AVAVAvAvAv VVVVV' A'A'A'AVA'AV,' VVVVVV‘ _r_ -r.- a alll'. {Jib Time (d) Figure 5 26 insure that only reasonable (non-negative) parameter estimates were obtained (Bard, 1974). The sensitivity coefficients were obtained using the finite difference method suggested by Beck and Arnold (1977), and the model equations were integrated using a Runge-Kutta technique. The objective function (i.e., residual sum of squares) was minimized according to the least squares criterion. The nonlinear parameter estimation program was written in BASIC and implemented on a microcomputer. In addition to calculating parameter estimates, the program also calculated the parameter correlation matrix and provided approximate 95Z confidence intervals for the parameters. RESULTS AND DISCUSSION Model Evaluation. Sensitivity coefficients for the zero order model are shown in Figures 2-5. The sensitivity coefficients for the mineralization, immobilization, and nitrification rates change linearly with time and are therefore linearly dependent, while all other sensitivity coefficients are zero when measuring the NH4+ pool (Figure 2). If the NH4+ pool is the only response measured, the mineralization, immobilization, and nitrification rates could not be uniquely identified. As one would intuitively expect, only a net rate could be obtained. When the sensitivity coefficients for the atom Z 15NH4+ response is examined, the mineralization rate is no longer linearly dependent with respect to the immobilization and nitrification rates, thus it is uniquely determined (Figure 3). However, the immobilization and nitrification rates are still almost linearly related. Adding the N03- response allows one to uniquely estimate the immobilization and nitrification rates, since the sensitivity coefficient for immobilization is zero (Figure 4) rather that approximately twice that of the nitrification sensitivity coefficient as it was in the previously described responses (Figures 2 and 3). The final response-~atom Z 15N enhances the linear independency among the immobilization, mineralization, and nitrification rates (Figure 5). Close examination of Figures 2-5 will show that the denitrification rate is also uniquely determined. However, since its sensitivity coefficients are all,small (< SZ) compared to the others, the denitrification rate will be poorly determined. This is caused, at 27 28 Figure 2. Normalized sensitivity coefficients for the NH4+ response in a zero order model of N cycling. C] - mineralization rate (f1) immobilization rate (f2) nitrification rate (f3) denitrification rate (f4) initial active organic N pool size (Y3(0)) GOOD I 29 OQOOQQ'OOOQ'QOQ099.09.999.90...9.99.0999... epitome (d) Time 30 Figure 3. Normalized sensitivity coefficients for the atom Z 5N114+ response in a zero order model of N cycling. Symbols as in Figure 2. 31 -““‘4A£€Au ._-Vo.l.o,c.’ u‘. -------- - 9 l i t O O u 1 a"... _ .Ygfc‘rofaloiola(a40101auo L L' L" 'L'L'l" 'L'l '8 '4 ' . ...3Avivav.v.v.vA'stvxv‘v‘v.v.vxv.v17.!.y .. ’1‘v V'VV'VV'V'VY'VVV'VV‘ '18 . . . 9 3 4 6 8 10 Time (d) W. Figure 3 32 Figure 4. Normalized sensitivity coefficients for the N03‘ response in a zero order model of N cycling. Symbols as in Figure 2. .a. w M R w 0:: 0'3 33 Time (d) Figure 4 34 Figure 5. Normalized sensitivity coefficients for the atom Z 15NO3’response in a zero order model of N cycling. Symbols as in Figure 2. '7 F 35 - . H qt“: bib - .‘.9_.O>.0.6-o.o.o.o.o o o o o o 0.. . o - . a V, ,_ '* Q'.,QAO‘O o o o‘o‘o'o‘o‘o'o'9;.‘.'rg3...,_ .~. ' I'l ' . . . . V n. u ‘ W ..,._. I . I.u.-.... . ' . ‘ VVVV;"';';V;.;,;“"'I _;|!!:'..".|IIIIIIII o . _ .. . .' .1... .v.". C ‘. ' " o I .v',/ 3.4 . 4‘, . ...7‘AAA‘, 7‘. "..I AV 'V'VVVVVV A'A'A'A'A"'> ' 4 6 8 10 Figure 5 36 least in part, by the fact that the denitrification rate is small relative to the rates of other N cycle processes. Similarly, a good estimate of the:initial size of the active organic N fraction cannot be obtained because of the small magnitude of its sensitivity coefficient. The results of examining the sensitivity coefficients:fin:the first order model (Figures 6-9) are quite similar to those of the zero order model. Linear dependence among mineralization, immobilization, and nitrification rate constants is strong when only the NH4+ pool is measured (Figure 6). Adding the atom Z 15NH4+response separates out the udneralization rate constant from the other rate constants (Figure 7), and the N03" response allows the immobilization and nitrification rate constants to be uniquely identified (Figure 8). Once again, the denitrification rate constant will not be well estimated. The behavior of the active organic N pool is quite different in the first order model. The magnitude of the active organic N sensitivity coefficient is large, however, it is now linearly dependent with the mineralization rate constant. Thus, these two parameters cannot beIHUquely determined; one must be measured by an independent method for the other' to be determined. Plotting the sensitivity coefficients also reveals the optimal sampling strategies. With either model, the best information is obtained during the:first five days (Figures 3, 5, 7, and 9), at least for the atom Z 15N measurements. Although the sensitivity coefficients generally continue to increase with time for the total pool size responses (Figures 2, 4, 5, and 8), information is only provided for the nitrification rate since the immobilization and mineralization rates are approximately linearly dependent for these responses. 37 Figure 6. Normalized sensitivity coefficients for the NH4+ response in a first order model of N cycling. Symbols as in Figure 2. 38 4 6 Time (d) Figure 6 w 39 Figure 7. Normalized sensitivity coefficients for the atom Z 5NH4+ response in a first order model of N cycling. Symbols as in Figure 2. 40 .. _’ v 1 ,W s Azitltaaza'L'a'a'b'l'b'l't'A ' 5 "v‘vA‘A'AVAV‘VA'AVAVIV‘V‘VL'AV ' ‘ " VVVVVVVVVVVVVV‘ 'VV' l l l 4 6 8 10 Time (d) Figure 7 41 Figure 8. Normalized sensitivity coefficients for the N03“ response in a first order model of N cycling. Symbols as in Figure 2. 42 . . .v-"‘ R v nus V *'~’$‘o's' .v iv.‘.9‘9;9: Vs‘s‘o‘o‘s'y 6 w (d) 4 Time Figure 8 43 Figure 9. Normalized sensitivity coefficients for the atom Z 15NO3' response in a first order model of N cyling. Symbols as in Figure 2. 44 l J l l 2 4 6 8 Time (d) Figure 9 45 Estimation of Active Organic N. In order to estimate N cycle rates by fitting data with our model it was necessary to measure or estimate the initial size and atom Z 15N of the active fraction of the organic N pool. In all cases this pool was assumed to have an atom Z 15N of 0.3663 Z, or natural abundance. The small departures (< ilZ) from natural abundance found in soil organic N (Hauck, 1973) do not significantly affect the parameter estimation process. The size of the active fraction is more difficult to estimate. We used two different approximations to obtain our estimates. In experiments where we used the CHC13 fumigation method, we estimated biomass N and assumed that it reflected the size of the active organic N over the short length of our experimental incubations. We also employed the isotOpe dilution principle, calculating the active fraction as did .Iansson (1958). The active fraction was set equal to the 15N lost from the NH4+ and N03“ pools during the incubation period (losses of 15N via denitrification were minor) divided by the atom Z 15N of either the biomass N at the end of the incubation (when this was measured) or of the NH4+ pool. When the atom Z 15NH4+ was used, it was assumed that the atom Z 15N of the active organic N and NH4+ pools was close to equilibrium. In most cases these two methods agreed quite well (Table 2), although the isotOpe dilution method did give a slightly higher estimate of active organic N in the Capac soil. This suggests that biomass N represents most of the active organic N, at least over one to three week time span. Juma and Paul (1981) found biomass N made up 40Z of the active organic N, but their experiments were of several months duration, during which a larger pool of organic N would be expected to become involved in mineralization-immobilization dynamics. 46 Table 2. Range of active organic N fraction estimated by two different methods. Soil Active organic N Biomass N Isotope dilution ------ pg N g"1 3011 - - - - - - Onaway loam 170-190 150-170 Capac clay loam 145-165 19o+ Spinks sandy loam 25-35 25-40 +Single observation. 47 The sample of Onaway loam used in experiment 1 was collected a year earlier than that used to estimate the active N fraction in this soil (Table 2). Since the NH4+ and active organic N pools were still far from equilibrium in experiment 1 (after 21 days the N114+ pool was still highly labeled), and the size of the active N could change due to fluctuations in environmental conditions, we fit the data using a wide range of organic N pool size estimates (Figure 10). The best fit was obtained with an active N pool size of about 120 pg N g‘1 3011, which is lower than the estimates obtained for other samples of the same soil (Table 2). Estimation of N Process Rates. The N cycle model presented in this paper (Figure 1) fit the measured dynamics of N cycling in the three soils studied in several experiments. However, neither the zero order nor the first order model fit all the data in some of the experiments, particularly when a carbon source or nitrapyrin was added. Presumably in these cases other processes were functioning in the soil that were not accounted for in the model. In these cases, a portion of the data could sometimes be fit when the entire data set was not well described by the model. The following examples illustrate the usefulness, as well as the limitations, of the parameter estimation approach used in this study. Nitrogen dynamics of the Onaway loam soil in experiment 1 were well described by a zero order kinetic model which included the heterotrophic nitrification process (Figures 11 and 12). Without the inclusion of the heterotrOphic process the residual sum of squares (RSS) increased markedly from 17.3 to 399. The zero order model fit the data better than the first order model, which had residual sums of squares of 49 or 48 Figure 10. Reduction in the residual sum of squares for experiment 1 with Onaway loam as a function of the initial active organic N pool size. 49 38- 28- 3 up Residual Sum of Squaoes 8 48 88 188 188 288 Ac+ive OPQOHic N (p9 N/g) Figure 10 50 Figure 11. Experimental and simulated pool size data for Onaway loam soil. [:1 - NH4+ A - N03" -'- simulated using estimated parameters a H Time (d) Figure 11 52 Figure 12. Experimental and simulated atom Z 15N data for Onaway loam soil. Symbols as in Figure 11. 53 (N91 WOiV) .L23€ M /\ 'lW'JW HAP :W H Time (d) Figure 12 54 1660, with or without the heterotrOphic nitrification process, respectively. Rate estimates for the Onaway loam range from 17 ng N g'1 d'1 for denitrification to over 1 ug N g”1 d'1 for mineralization (Table 3). The rates of mineralization and immobilization are comparable to those given by Jansson (1958). However, the mineralization rate obtained was generally higher than those measured by the standard incubation procedure without the 15N label (Tabatabai and Al-Khafaji, 1980; Addiscott, 1983), probably because the standard technique only measures net mineralization. Approximate 95Z confidence intervals have been calculated for these N cycle rates, but it should be noted that the method used gives very optimistic values (narrower confidence intervals) and the actual variation is probably greater (Robinson, 1984). With this caveat in mind, it is apparent that the rates of mineralization, immobilization, and nitrification are estimated much more precisely than the denitrification and heterotrOphic nitrification rates. This is a reflection of the earlier discussion about sensitivity coefficients. Nonetheless, the rates obtained seem quite reasonable compared to nitrification and denitrification rates typically found in forest soils (Robertson, 1982: Robertson and Tiedje, 1984). First order kinetics best fit the data from the double labeling experiment with the Capac soil (Figures 13 and 14). This first order behavior is evident from the curvilinear trends of the NH4+ and N03" pools over time. Nitrification was initially very rapid in this soil, with rates as high as 16 to 35 ug N g"1 d‘1 during the first day of incubation. Denitrification rates varied from 450 to 890 ng N 3‘1 d'1 55 Table 3. Rates and rate constants of N cycle processes in three dissimilar soils. Process Onaway loam+ Capac clay loam§ Spinks sandy loam§ “g N g-l d-l _________ d-l ________ Mineralization 1.01 i 0.06fl 0.0117 31 0.0030 0.00745 10.00320 Immobilization 0.674 1 0.054 0.146 1- 0.076 1.70 i 0.39 Nitrification 0.113 1 0.008 1.30 i 0.14 1.60 _-_I-_ 0.28 Denitrification 0.0166 1 0.42 0.0114 i 0.0047 0.00259 :1; 0.0184 Heterotrophic nitrification 0.371 : 0.384 -- - Active or anic N 120 150 40 (us N 8‘ 8011) RSS 17.3 147 72.8 + §Mean _-_t-_ 95Z confidence interval. Zero order kinetics. First order kinetics. 56 Figure 13. Experimental and simulated pool size data for Capac clay loam. Symbols as in Figure 11. 1% 57 (B/N 8d) -80N iw H n Time (d) Figure 13 58 Figure 14. Experimental and simulated atom Z 15N data for Capac clay loam. C] - atom Z 15NH4+ from 15NH4'l'treatment ZS - atom Z 15N03‘ from 1SNHz,+ treatment <> - atom Z 15N02- from 1 N03- treatment --- simulated using estimated parameters 59 (N31 % w0+vn CS7 U-II U—IO - . - ass -80N 1 W H Time (d) m Figure 14 60 over the course of the experiment, about four times the rate estimated by Parkin _e_t_il_. (1984) for this soil using the C2H2 block method. The nitrification rate constant is comparable to those previously reported for agricultural soils (Cameron and Kowalenko, 1976; McLaren, 1976), while the denitrification rate constant is lower than those given by McLaren (1976). However, the difference in denitrification rate constants can likely be attributed to the higher water content (more anaerobic conditions) of the solution flow methods employed in the studies reported by McLaren (1976). The mineralization rate constant is high compared to those obtained using the method for determining potentially mineralizable N (Stanford and Smith, 1972; Campbell £21., 1981). As with the zero order rates obtained by this method, the lower rate constants obtained by others could be a reflection of net mineralization. However, the mineralization rate constant for the Capac soil is also twice as large as the biomass N decay rates determined by Paul and Juma (1983) using 15N tracer methods. Few other measurements of this kind have been made, so this difference may simply be an expression of the inherent variability of net mineralization constants. The Spinks sandy loam was also fit best by the first order model (Figures 15 and 16). Nitrification rates ranged from 10 to 15 pg N g‘1 d"1 in this soil, somewhat lower than those found in the Capac clay loam. Immobilization rates were considerably higher in the Spinks soil (3'13 113 N 8’1 d'l) compared with either the Capac (0.1-4 pg N g’1 d'l) or Onaway soil (0.7 pg N g’1 d'l). This higher rate may reflect the lower N content of this soil (Table 1), suggesting that the Spinks sandy loam may be more N limited than the other soils studied. The mineralization rate constant for the Spinks soil is comparable to those 61 Figure 15. Experimental and simulated pool size data for Spinks sandy loam. Symbols as in Figure 11. <_to [3] where Co is the N03' concentration at the aerobic-anaerobic interface within the aggregate (ran). This upper boundary was chosen to approximate a steady state of nitrification in the aerobic portion 0f the aggregate. Equation [1] was solved for both transient and steady state conditions. The transient case was solved by using a finite difference approximation of Equation [1], incorporating time averaging of the diffusion terms. The sink term was not time averaged because of its inherent nonlinearity with respect to concentration. In the steady state, Equation [1] becomes a two point boundary value problem, which was solved using a finite difference method described by Keller (1968). Both solutions were programmed in BASIC and implemented on a microcomputer. Aerobic and anaerobic denitrification rates and N03" flux across the aerobic-anaerobic interface in aggregates were calculated along with the expected anaerobic slurry rate. Examples of N03“ concentration profiles within a diffusion limited aggregate and within an aggregate not limited by N03” diffusion are shown in Figure 1. The marked difference in these two profiles is solely the result of a ten-fold difference in the maximum denitrification rate, V, which is Figure 1. 82 Simulated N03” concentration profiles in an anaerobic 0.4 cm aggregate under conditions with ( [j ) and without (A ) a N03’ diffusion limitation. Parameter values used (Table 5): D - 5 x 10"6 cm? 3‘1; F 8 1; Km - 0.04 pg N g'l; Co - 0.15 pg N g’l; and V - 2.0 or 0.2 ug N g'1 h"1 for the diffusion limited and non-diffusion limited cases, respectively. N03" Conceh+ho+ion (pg N/g) 83 Aggregate Radios (cm) Figure l 84 within the range found in soils (Tiedje 2321., 1982). In both cases the mass balance of N03“ is conserved since the denitrification rate is equal to the flux of N03- across the aerobic-anaerobic interface. RESULTS The effect of supplemental N03” and glucose on denitrification rates in anaerobic slurries differed between the two soils (Figure 2). Neither addition significantly altered the denitrification rate in the Media soil, while both NO3- and glucose additions increased the denitrification rates in anaerobic slurries of all treatments of the Capac 8011 (Table 2). Anaerobic slurry rates of the Capac soil increased asymptotically as N03” concentration increased (Figure 3). When these data were fit by the Michaelis-Menten relationship, a Km of 17 pg N g"1 soil (450 pM) was obtained. The addition of glucose further increased denitrification rates in all treatments of the Capac soil. Individual soil cores did not always follow the response to N03" and glucose described above so the maximum anaerobic slurry rate obtained during the course of the slurry measurement was used for all other comparisons. Anaerobic slurry rates were much greater than anaerobic core rates for both soils (Table 3). The core to slurry ratio was significantly lower in the Media soil (0.3Z) than for the treatments of Capac soil (1.1-3.4Z). A similar comparison of anaerobic core to anaerobic slurry rates was made for packed cores of Capac soil which were equilibrated With N03- and succinate solutions (Table 4). The slurry rates of the control and N03” amended soil in the cold room experiment were similar to those obtained in the long term incubation experiment (Table 3), while the anaerobic core rates were 4 to 17 times higher. Increased N03“ availability was not the cause of this difference because N03“ 85 86 Figure 2. Effect of N03- and glucose additions on the denitrification rate of anaerobic slurries. N03’ added at first arrow, glucose added at second arrow. [3 - Medio soil, A - Capac soil. 87 WW llw . 1 IIV t . 0 m m Am\z mco no>ao>m.omz Time (d) Figure 2 88 Table 2. Effect of N03- and glucose additions on denitrification rates in anaerobic slurries. Soil + Unamended + N03‘ (100 pg-N g‘l) + Glucose (1 mg-C g‘l) pretreatment --------- ngNZO-Ng‘lh’1------------- + Medio 268 i 143 282 i 190 265 i 144 Capac 256 i 53 401 i 72 540 j; 125 + N03" 405 i 105 439 i 91 528 i 100 + straw 300 i 43 406 i 72 487 i 86 + Mean 11; 95Z confidence interval. 89 Figure 3. Effect of N03“ concentration on the denitrification rate of anaerobic slurries of Capac soil. 90 188 150 288 258 N03" Concean‘ol'im gig-79) 58 m a a a. n a $02.05 as 50.8.0303 E Figure 3 91 Table 3. Denitrification rates of anaerobic cores and anaerobic slurries. Soil Anaerobic core Anaerobic slurry Core:Slurry ratio pretreatment ----- ngNzo—Ng'lh‘1----- ---z--- Medio 0.9 i 0.53+ 340 i 180 0.32 i 0.12 Capac 17.4 i 5.4 588 i 174 3.43 i 1.01 + NO3‘ 5.0 i 2.3 500 i 158 1.10 i 0.40 + straw 15.7 i 8.0 506 i 62 3.01 i 1.37 +Mean : 95Z confidence interval. 92 Table 4. Denitrification rate of Capac soil amended with N03” and succinate. Treatment Anaerobic core Anaerobic slurry Core: Slurry ratio 1120 67.1 + 50.6+ N03- 87.0 f 11.7 Succinate 5 4 i 108 N03- + succinate 400 i 43.2 __--z......_ U'ILnHH UINUIN o oo sowow |+l+l+l+ O‘HHV 0 Ho 0 t-‘o com U! .1. Mean : 95Z confidence interval. 93 concentrations were the same in each experiment, about 20 pg N03“-N g'“1 soil. The difference may have been due to greater carbon availability' in the cold room experiment. Physical disturbance of soil by mixing and packing often results in an increase in available carbon (Rovira and Greacen, 1957), and one would expect more of this physically released carbon to be available for denitrification after three days incubation at 40C than after 21 days at 20°C. There were no significant differences in anaerobic core rates, anaerobic slurry rates, or core to slurry ratios between the N03- amended and control soil (Table 4). However, the addition of succinate or N03” plus succinate resulted in significant increases in all three of these measurements when compared to the control. .Anaerobic core rates and anaerobic slurry rates were both significantly greater in the succinate treatment than in the N03” plus succinate treatment, however, the core to slurry ratios were the same. DISCUSSION Differences in denitrification rates between anaerobic cores and slurries reflect the effect of the native soil structure on denitrification rates in nature. Slurried soils always gave markedly higher denitrification rates (Table 3 and 4), presumably because of enhanced distribution of denitrification substrates and denitrifying organisms. With the Capac soil N03" supply shown was not to be limiting, since there was no difference between the core:slurry ratios of soil cores incubated with or without N03" (Table 4). However, the supply of available carbon in the Capac soil was shown to limit denitrification since succinate additions greatly decreased core:slurry ratios (Table 4). The Medio soil had fairly high concentrations of N03- ('05 pg N g'1 soil), so the low core:slurry ratio of this soil is also likely indicative of a carbon supply limitation. Even with the addition of succinate, a core:slurry ratio less than 100Z was obtained. This could be due to a phase transfer limitation of N20 diffusion out of (or 02112 diffusion into) the liquid phase of the anaerobic cores. Such limitations can be found even in well stirred solutions if the biological activity is large enough (Robinson and Tiedje, 1982). Whether substrate diffusion is, or is not, important in determining the rate of a reaction is a function of several factors including: substrate concentration of the bulk solution, substrate diffusion coefficient, the path length for diffusion, the system geometry, and the biological kinetic parameters. One of the simplest ways to evaluate 94 95 these interacting factors is to calculate the Thiele modulus, which is a dimensionless parameter constructed from these interacting factors (Goldstein, 1976). For spherical particles, like soil aggregates, the Thiele modulus for a single Michaelis-Menten reaction is defined as: _ ran[ V ]—;’ [4] where ¢> is the Thiele modulus. The other constants have been previously defined, except that V is now assumed to be independent of the aggregate radius. The Thiele modulus indicates the degree of any diffusional limitation. Reactions are not limited by diffusion when ¢‘S_1, while diffusion becomes increasingly limiting as ¢ becomes larger (Goldstein, 1976). The actual reaction rate is a function of both ¢ and the bulk concentration of substrate (Figure 4). This figure illustrates that even diffusion limited reactions can proceed at maximal velocity if the substrate concentration in bulk solution is sufficiently high. In soils, the values of each of the variables in Equation [4] for denitrification span one to over two orders of magnitude (Table 5). This variability, of course, leads to an even wider range of possible values of ‘1. Thus, whether or not N03' diffusion is rate limiting to denitrification depends upon the parameter values of the particular soil of interest. For most soils, D and Km are probably quite close to the typical values given in Table 5, so ¢ will primarily be determined by V and ran. It should be noted that (b is most sensitive to ran: which it is directly proportional to, while varying with the square root of the other parameters. 96 .ouwwouwmo 0Hnouoocm mHouonaou moasmm< .ucouaou nouns Him w 0N.o moaomm< m + 0000 .nwm.wm 0o0=o00> 00000 .amm.wm 00000 0-0 z 00 0.00 n 0.0 0.0 A000 ao0ooaooooooo 1002 u u 000 a 000.0 0.0 0000 0000003 o0o000 Aamv uaouoaoo 0000 .o00o0a 0:0 0000000 +0.0 2 0a 0.0 n 00.0 00.0 000000002 -002 0000 .ammmmmnm00o00 0>0 mono 00000 ..00 no 0000 0:0 0:0 z 0: 0.0 I 00.0 0.0 ao0omo0m0pa0ooo a=a0xnz 000 ooo0o000ooo 0000 .00050 0..0 Nao 0-0000 1 0-0000 0-00 0 0 oo00=0000 1002 0000 .ao0a0 momma .uoaouoo Eu o.m I «0.0 N.o Auv maHomu oumwouwwm ovam> mucouowom mafia: owcmm Hwoaama umuoaouom .Hfiaooa oHoHLH voumflsoamo can moHuuonoum HHom .m oHan 97 Since N03" concentrations in soil can vary from about Km to 6000 x Km (Table 5), it is important to examine the relation between and 80 presented in Figure 4. Under the low N03" concentrations (< 1 pg g":l soil) typically found in forest soils (Vitousek £91., 1982), there can be a response to N03' additions, even when N03- diffusion is not limiting. However, in agricultural soils, which generally have higher N03" concentrations (> 5 pg N g”1 soil), a response to N03" addition would be noticed only under diffusion limited conditions (¢ > 10; Figure 4). If the typical parameter values of ran V, Km, and D from Table 5 represent median values for aggregated soils, then NO3- diffusion should limit denitrification about half of the time under anaerobic conditions when carbon is not limiting, since ¢ is 1.4 for the median soil. Often, however, denitrification occurs within anaerobic microsites of a generally aerobic soil and under carbon limitation. The effect of carbon limitation on N03’ diffusion can be roughly approximated by multiplying V by the carbon limitation factor F, which varies from 0 to 1. This has the effect of decreasing <1) by the square root of F. Thus, a carbon limitation effectively decreases the magnitude of a N03" diffusion limitation. The Capac soil used in this study was carbon limited with an F of approximately 0.24. F was calculated from the quotient of the core:slurry ratios of non-carbon amended to carbon amended soil (Table 4). (This F value roughly corresponds to an available carbon concentration of Km/3 for carbon, if Michaelis-Menten kinetics are assumed). Applying this F to a V of 0.53 pg N g"1 h‘1 (Table 3) and assuming typical values for ran: D, and Km (Table 5), a d) of 0.89 for anaerobic Capac soil can be calculated. Thus, no N03- 98 Figure 4. NOrmalized reaction rate as a function of the dimensionless bulk concentration (So - Co/Km) for different values of the Thiele modulus, ¢ , shown for values from 0.1 to 500. 99 00 a a a. as a a a a 8m _ 8.0. 8— 0000 100 diffusional limitation should exist in this soil, a contention supported by the lack of response to NO3‘ in the cold room experiment (Table 4). The effect of anaerobic microsites within aerobic soil, on the other hand, would usually increase the likelihood of NO3" diffusional limitation. This effect can be studied using the zero order reaction-diffusion equation for oxygen deve10ped by Greenwood (1962) and further extended by Smith (1980). A critical radius of 1.0 cm can be calculated at an oxygen concentration of 0.17 cm3 cm‘3, an 02 diffusion coefficient of l x 10’6 cm2 3'1, and respiratory activity of l x 10‘6 cm3 C02 cm.3 3‘1. Consequently, denitrification could only occur in aggregates larger than 1.0 cm. Aggregates of this size would usually have 0 values in excess of 1.0, since ran increases very rapidly as the aggregate radius exceeds the critical radius (Smith, 1978). Consequently aggregates larger than 1 cm will often be N03" diffusion limited, even under most carbon limited situations. However, it should be noted that a diffusional limitation of N03“ may not be evident if the bulk solution N03" concentration is several fold higher than the Km for N03” reduction (Figure 4). Aggregate size is not uniform in soils, so whether or not denitrification is diffusion limited also depends upon the aggregate size distribution. The log normal distribution generally describes soil aggregate distributions (Gardner, 1956) and has been used to model soil anaerobic microsites (Smith, 1980). The volume fraction of soil that experiences a diffusional limitation may be calculated using the equations presented by Smith (1980), by simply redefining the critical radius as the radius of the largest aggregate that is not diffusion limited. This critical radius for diffusion limitation can be lOl calculated from Equation [4] by setting (8 equal to one and solving for ran. The only limitation is that ran must be as large or larger than the critical radius of anaerobiosis. This analysis was done with typical values of V, D, and Km (Table 5) for a totally anaerobic soil and for an aerobic soil with a critical radius of 1.0 cm (Figure 5). Two points can be made from the results shown in this figure: First, the fraction of anaerobic soil that is diffusion limited for N03‘ is much greater under aerobic conditions, because only larger aggregates have anaerobic centers. Second, under anaerobic conditions, even when a soil has a mean aggregate radius at which denitrification is not diffusion limited (e.g., 0.1 cm), there is a certain fraction of the soil which does experience diffusion limitations (e.g., 30Z). Most previous work which suggests that N03” diffusion limits denitrification rates is based upon the higher Km estimates obtained in soil systems compared to pure culture work (Table 6). Km values for soils are 13 to 1300 times those found in pure culture, although they are similar to the Km values obtained from cell free extracts. The high Km values found by Ryzhova (1979) and Kohl _e_t_ 9.]: (1976) for the 2Z organic-C soil can probably be attributed to limiting N03" diffusion from the water layer above the soil (Phillips 5521., 1978), however, a relatively low Km was found for the 2.2Z organic-C soil incubated in the same manner (Kohl 3521., 1976). In studies which monitored N20 evolution (Klemmedtson, 1977; Yoshinari, 2221., 1977), mass transfer of C232 into the soil and N20 out of the soil may have inflated the estimated Km values (Robinson and Tiedje, 1984), however, only the glucose amended soil of Yoshinari £3 21. had a high Km. Internal diffusion of N03" within the soil matrix is another factor which could Figure 5. 102 Relationship between aggregate size distribution and the extent of diffusion limitation in aerobic and anaerobic soils. Parameter values used: V - 0.3113 N g"1 h‘l, DNO3' - 5 x 10"6 002 8‘1, Km - 0.04 pg N g’l, and ran - 1.01 cm. Diffusion Lini’red Fraction 103 8.1 i .8 Loqlg (Hem Aqg‘egote Radius) Figure 5 10 104 n «N uo>o mums 0:00:000o 00.0 ONH omxozm uon .oovomam 0 you as N v oaoon uHHm .u.... 00: 0000 .000 o00oooooo 0:00:000o 00.0 0000 ..00 no 0000 0000 noonaooaaon0o -002 as 0 v .aoo0 0000 a new uo>o mums :oxmnm uo: .movomaw 0 Own N000 .oHpouomcm Hogans oHomwuo NHH.o «~00 .usoom one amazon One oucouwomqwmwo Imoz econ 000m me0 .ofivowa can moumsmm 001m oucmumoqommwo Imoz mooomouosaw mmoosowsmom maoomouosam mmooaooaomm .om abfiuouoopo>mam mea .oneoHH cam nomHuom mHv oosmumoaqwmfiv Imoz .Qm mocuwHHmoH< mHHoo 000:3 cowuoovoua INoz Nwma .ummzmm.oomaumu oom coonOH> Hanson “peace to wmocstpom mmcoaoesomm Il.nlu coauosoouo Imoz mcoofimwuuficoooamn mmma ..00 um owmom oomH cowoaoa> amazon "peace to maooooouon oowuosooua INoz Humfi .uowuom 0mm cowoaow> Hmuoon ”peace to momoamuuuHcoo msooooouofiz :oHuoneoum INoz HomH .wmaozofiz one common 00m mo CH moofiwuno aofiuusoou Imoz you mo=Hm> am no nomauoqaoo .c wanna 105 0000 .maoen000 0000 .umm;ww 0000020o0 0000 .oono0naao00 oon coma omH 00H 5 ea uo>o oumu omanm uoa .oooamam 0 ac: 002 0000 .000 000000000 oesoumoQQMmHo Imoz n H uo>o mums smxmnm no: 002 000 .o00ooooao 0200 2003 :o0ooooopa 002 a n0 v uo>o mums coxmcm .movomam uo: omm NOON .ownouomom mmmu £003.:00uoaeoum omz nouuma 0Hoowuo NN.m as H v owoo=Hw+ .3000 xenon omooofiwl .awOH momma Ulowoowuo N~.m as N v 0.0.00000 0 o0ona 106 have contributed to the inflated Km values found in soil. This influence can be estimated by calculating 43 values for the soils using the maximum velocity values reported in the studies, a Km of 15 HM, and typical values of 0.2 cm for ran and 5 x 10"6 cm2 s'1 for D. These parameters give a range of¢ from 0.53 to 2.68 or, in other terms, Km would have been overestimated by three fold at most. Taking into account, the soil Km values would range from 55 pM to 4 mM; still greater than whole cell values, but comparable to Km values of cell free extracts. A possible explanation for the discrepancy between the Km for N03- reduction of soils and cells is that more than one N03" uptake system exists. Recent work with assimilatory N03- reduction suggests that there may be both low affinity (1 mM) and a high affinity (5 pM) uptake mechanism (Thayer and Huffaker, 1981). Evidence for the similar situation with the dissimilatory system is lacking because whole cell studies have never been done at high enough NO3- concentrations to detect a low affinity system, while the N03" concentrations used in soils have not been low enough to measure a high affinity system. From the organism's standpoint, it would be wasteful for a denitrifier in a carbon limited system like soil to (presumably) use ATP to scavenge N03', when it already exists in the soil solution at concentrations at or above the Km of the N03' reductase enzyme. It would probably be a competitive advantage for denitrifiers to have two uptake systems; one with a relatively high Km which is energy independent, the second with a low Km powered by cellular energy. The existence of a low affinity system would decrease the importance of N03“ diffusion, since the higher Km would lower 4). However, a higher Km would make denitrification 107 more responsive to N03“ additions, since natural N03- concentrations would be in the first order region. The discussion has thus far centered on N03” diffusion and denitrification in aggregated systems. Different conceptual and mathematical models would have to be used for non-aggregated soils. Presumably a non-aggregated soil would have "hot spots" of microbial activity more or less randomly distributed throughout the profile. These centers of microbial activity would likely be associated with organic carlxni. In such a situation, one might envision the impact of NO3" diffusixni to be greater since the path length for diffusion would probably be longer. This area of reaction-diffusion in non-aggregated soils has yet to be explored either experimentally or theoretically. It should also be noted that hot spots of microbial activity are likely to exist in aggregated soils, as well. Their presence would also tend to increase the importance of diffusive limitations, in this case primarily through an increase in the maximum velocity parameter. CONCLUSIONS 1. .Aggregate size, followed by the Vmax for denitrification, are the prime determinants of whether or not denitrification is limited by N03“ diffusion. However, a formal N03“ diffusional limitation may be ameliorated by high solution concentrations of NO3‘. 2. Aerobic soils with anaerobic microsites are more likely to experience the effects of N03- diffusion limitations. 3. Carbon limitation decreases the magnitude of any potential N03- diffusion limitation by effectively decreasiiu; the maximum denitrification rate. 4. The relatively high Km values for N03“ reduction in soil may be evidence for the presence of two N03” uptake systems, since N03- diffusion was not great enough to markedly inflate the Km values that have been reported. 108 10. 11. 12. REFERENCES Betlach, M.R. and J.M. Tiedje. 1981. Kinetic explanation for accumulation of nitrate, nitric oxide, and nitrous oxide during bacterial denitrification. Appl. Environ. Microbiol. 42:1074-1084. Bouldin, D.R. 1968. Models for describing the diffusion of oxygen and other mobile constituents across the mud-water interface. J. Ecol. 56:77-87. Bowman, R.A. and D.D. Focht. 1974. The influence of glucose and nitrate concentrations upon denitrification rates in sandy soils. Soil Biol. Biochem. 6:297-301. Burford, J.R. and D.J. Greenland. 1970. Denitrification under an annual pasture. p. 458-461. IB_XI International Grassland Conference, Surfers Paradise. University of Queensland Press, St. Lucia, Australia. Carlson, C.A., L.P. Ferguson, and J.L. Ingraham. 1982. Properties of dissimilatory nitrate reductase purified from the denitrifier Pseudomonas aeriginosa. J. Bacteriol. 151:162-171. Edwards, R.M. and J.M. Tiedje. 1981. Determination of Km for dissimilatory nitrate reduction by cells of Pseudomonas fluorescens. Abstr. Ann. Meet. Am. Soc. Microbiol. N48 p. 181. Fewson, C.A. and D.J.D. Nicholas. 1961. Nitrate reductase from Pseudomonas aeruginosa. Biochem. BiOphys. Acta. 49:335-349. Firestone, M.R. 1982. Biological denitrification. In F. J. Stevenson (ed.) Nitrogen in agricultural soils. Agronomy 22:289-326. Am. Soc. Agron., Madison, WI. Forget, P. 1971. Les nitrate-reductases bacteriennes, solubilisation, purification et prOprietes de l'enzyme A de Micrococcus denitrificans Eur. J. Biochem. 18:442-450. Gardner, W. R. 1956. Representation of soil aggregate-size distribution by a logarithmic-normal distribution. Soil Sci. Soc. Am. Proc. 20:151-153. Goldstein, L. 1976. Kinetic behavior of immobilized enzyme systems. Methods Enzymol. 44:397-443. Gosz, J.R. 1978. Nitrogen inputs to stream water from forests along an elevational gradient in New Mexico. Water Res. 12:725-734. 109 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 110 Greenwood, D.J. 1962. Nitrification and nitrate dissimlation in soil II. Effect of oxygen concentration. Plant Soil 17:378-391. Greenwood, D.J. and D. Goodman. 1964. Oxygen diffusion and aerobic respiration in soil spheres. J. Sci. Fd. Agric. 15:579-588. Jahnke, R.A., S.R. Emerson, and J.W. Murray. 1982. A model of oxygen reduction, denitrification, and organic matter mineralization in marine sediments. Limnol. Oceanogr. 27:610‘623. Jorgensen, B.B. 1978. A comparison of methods for the quantification of bacterial sulfate reduction in coastal marine sediments I. Measurement with radiotracer techniques. Geomicrobiol. J. 1:11-27. Keller, H.B. 1968. Numerical methods for two-point boundary-value problems. Blaisdell Publishing Company, Waltham, MA. 184 p. Klemmedtson, L., B.H. Svensson, T. Lindberg, and T. Rosswall. 1977. The use of acetylene inhibition of nitrous oxide reductase in quantifying denitrification in soils. Swedish J. Agric. Res. 7:179-185. Kohl, D.R., F. Vithayathil, P. Whitlow, G. Sheaver, and S.H. Chien. 1976. Denitrification kinetics in soil systems: the significance of good fits of data to mathematical forms. Soil Sci. Soc. Am. J. 40:249-253. La Motta, E.J. and W.K. Shieh. 1979. Diffusion and reaction in biological nitrification. J. Environ. Engin. Div., ASCE. 105: 655-673. Parkin, T.B., E.F. Kaspar, A.J. Sexstone, and J.M. Tiedje. 1983. Use of a gas-flow soil core method for measuring field denitrification rates. Soil Biol. Biochem. (in press) Phillips, R.E., K.R. Reddy, and W.H. Patrick, Jr. 1978. The role of nitrate diffusion in determining the order and rate of denitrification in flooded soil: II. Theoretical analysis and interpretation. Soil Sci. Soc. Am. J. 42:272-278. Reddy, K.R., W.H. Patrick, Jr., and R.E. Phillips. 1978. The role of nitrate diffusion in determining the order and rate of denitrification in flooded soil: I. Experimental results. Soil Sci. Soc. Am. J. 42:268-272. Reddy, K.R., W.H. Patrick, Jr., and R.E. Phillips. 1980. Evaluation of selected processes controlling nitrogen loss in a flooded soil. Soil Sci. Soc. Am. J. 44:1241-1246. Robinson, J.A. and J.M. Tiedje. 1982. Kinetics of hydrogen consumption by rumen fluid, anaerobic digestor sludge, and sediment. Appl. Environ. Microbiol. 44:1374-1384. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 111 Rosso, J.-P., P. Forget, and F. Pichinoty. 1973. Les nitrate-reductases bacteriennes. Solubilisation, purification et proprietes de l'enzyme A de Micrococcus halodenitrificans. Biochim. BiOphys. Acta. 321:443-455. Rovira, A.D. and B.L. Greacen. 1957. The effect of aggregate disruption on the activity of microorganisms in the soil. Austral. Jo Agric. Res. 8:659-6730 Ryzhova, I.M. 1979. Effect of nitrate concentration on the rate of soil denitrification. Sov. Soil Sci. 11:168-171. Sexstone, A.J., T.B. Parkin, and J.M. Tiedje. 1984a. Interactive control of soil denitrification rates by oxygen and moisture. (prepared for publication) Sexstone, A.J., N.P. Revsbech, T.B. Parkin, and J.M. Tiedje. 1984b. Direct measurment of oxygen profiles and denitrification rates in soil aggregates. Soil Sci. Soc. Am. J. (submitted) Smith, K.A. 1978. Soil aeration. Soil Sci. 123:284-291. Smith, K.A. 1980. A model of the extent of anaerobic zones in aggregated soils, and its potential application to estimates of denitrification. J. Soil Sci. 31:263-277. Smith, M.S. and J.M. Tiedje. 1979. Phases of denitrification following oxygen depletion in soil. Soil Biol. Biochem. 11:261-267. Thayer, J.R. and R.C. Huffaker. 1981. Use of 13N-nitrate to study nitrate transport in Klebsiella pneumoniae. p. 341-351. In J.W. Root and K.A. Krohn (ed.) Short-lived radionuclides in chesttry and biology. Advances in Chemistry Series 197. American Chemical Society, Washington, D.C. Tiedje, J.M., A.J. Sexstone, D.D. Myrold, and J.A. Robinson. 1982. Denitrification: ecological niches, competition and survival. Ant. van Leeuwen. J. Microbiol. 48:569-583. Vitousek, P.M., J.R. Gosz, C.G. Grier, J.M.Melillo, and W.A. Reiners. 1982. A comparative analysis of potential nitrification and nitrate mobility in forest ecosystems. Ecol. Monogr. 52:155-177. Yoshinari, T., R. Hynes, and R. Knowles. 1977. Acetylene inhibition of nitrous oxide reduction and measurement of denitrification and nitrogen fixation in soil. Soil Biol. Biochem. 9:177-183. CHAPTER 3 EFFECTS OF CARBON, NO3", AND MOISTURE ON THE ESTABLISHMENT OF DENITRIFICATION CAPACITY IN SOIL Denitrification is a component of nitrogen cycling in soils of virtually all terrestrial ecosystems. However, the magnitude of this process varies greatly, both among and within ecosystems. This variability is undoubtedly a function of the previous environmental histories of the various habitats. Environmental factors affect both the expression of denitrification by the existing p0pulation of denitrifying bacteria and the size of the active denitrifier biomass itself. Much work has been devoted to examining the effect of such factors as available carbon, N03- concentration, pH, and aeration on denitrification rates. This work has recently been summarized in several excellent reviews (Firestone, 1982; Knowles, 1982). The effects of these variables on the establishment and maintenance of the denitrification capacity has been less well studied. A previous survey of soils from a variety of habitats suggested that active denitrifier biomass, or denitrification capacity, was directly related to moisture and organic carbon, while pH had no consistent effect (Tiedje 3521., 1982). Smith and Tiedje (1979) observed a similar result when soils received either irrigation or a glucose amendment. Flooding soil with N03” solution (Doner £21., 1975; Volz SE. 21., 1975) or anaerobically incubating N03’ amended, saturated soil (Jacobson and Alexander, 1980) resulted in increased 112 113 denitrifier populations. However, no attempt was made in these studies ‘to determine whether the pOpulation increase was due to the N03‘ addition or to a change in aeration status. Denitrification capacity could be controlled by environmental factors primarily by two mechanisms. The first involves the repression-derepression of the denitrification enyzmes by 02 and the possible induction of these enzymes by N03" (Firestone, 1982). This mechanism would not require microbial growth but simply the expression of the denitrifying potential of already existing, but inactive denitrifiers. The second mechanism involves the increase in denitrification capacity through cell division. These two mechanisms, which are illustrated in Figure 1, have been designated Phase IIa and Phase IIb by Smith and Tiedje (1979). I The purpose of this study was two-fold: (1) to examine the relative importance of carbon, moisture, and N03‘ in controlling denitrification capacity, and (2) to attempt to elucidate the mechanism(s) functioning to control denitrification capacity. 114 Figure l. Hypothetical soil biomass composition and anticipated response from two different medhanisms for increasing denitrification capacity. 115 00000003 00000000000a 0:0 £02000 0020: H ouowfim 00000001 cpgoco 00:0000m 00000000Q 000000000000 00000000000o 0000000 -I 0000000 00000000000 0>00000000 000005 0000020000: 0>000000m 000000m.00000000000-002 MATERIALS AND METHODS Capac clay loam (Aeric ochraqualf; pH 6.8, 0.28% N) was collected frouaza field previously planted to corn, sieved to < 2mm, and stored at 4°C at field moisture until used. The sieved soil was preincubated in a polyethylene bag at room temperature for a few days prior to its experimental use. Finely milled, dried alfalfa straw (2.81% N) was used as the carbon source in this study. .A 2x2x2x4-way factorial arrangement of treatments was used in this experiment. Two moisture contents (23 and 28%; 0.2 and 0.01 MPa), two levels of carbon addition (0 and 1 mg C g‘1 3011), and two levels of N03- addition (0 and 100 ug NO3'-N g"1 soil) were used. Estimates of denitrification capacity and total microbial biomass were measured at four time points (1,2,4 and 7 days). The experiment was initiated by adjusting preincubated, moist soil to the desired water content with either distilled water or a N03“ solution. Straw was thoroughly mixed into the treatments receiving C at this time. Approximately 175 g of soil, on a dry weight basis was transferred to 250 m1 Erlenmeyer flasks and packed to a bulk density of about 1.2 gm cm'3. These incubation vessels were capped with a serum stopper and incubated at 250C in the dark. Daily respiration rates were measured by analyzing headspace gas samples for C02 using a ndcrothermistor equipped GC. The head space of the flasks was replenished daily, or as needed, to maintain aerobiosis. At each time point in the experiment, one flask from each treatment was sacrificed for analysis. Six 10 g (dry wt.) subsamples were taken 116 117 for denitrification capacity measurements, seven 10 g (dry wt.) subsamples were removed for microbial biomass C measurements, and three 10 g (dry wt.) subsamples were taken for NH4+ and N03" determinations. A Technicon autoanalyzer was used to colorimetrically determine NH4+ and N03- concentrations. The remaining soil was used to determine gravimetric water content. Denitrification capacity was determined using an anaerobic slurry technique similar to the Phase I assay of Smith and Tiedje (1979). Slurries were made by adding 25 m1 of a solution containing glucose (1 mg C g'"1 3011), N03- (100 ug N g'l) soil and chloramphenicol (500 ug g"1 soil) to 10 g of soil in a 160 m1 serum bottle. The bottles were sealed with a Balch stOpper, evacuated and flushed several times with Ar to remove any traces of 02. The soil slurries were shaken on a rotary shaker (250 rpm) and N20 production in the presence of 10% C2112 was measured over the course of a one hour incubation. N20 was quanitified using a CC equipped with a 63Ni-electron capture detector (Parkin 33 g” 1984). Total microbial biomass was measured by the CHC13 fumigation method (Jenkinson and Powlson, 1976). The amount of C02 produced by fumigated samples and an unfumigated control, after a 10 day incubation, was measured by gas chromatography. Microbial biomass was also estimated in a separate experiment by measuring soil ATP. An incubation experiment similar to that described above was set up at 282 moisture with or without a 1 mg C g’1 soil straw amendment, without N03- addition. Respiration was measured daily, as described above, and ATP was extracted from four subsamples of soil for each treatment by the method of Paul and Johnson (1977) and quantified 118 by bioluminescence with a Chem-Clo photometer equipped with an Aminco integrator-timer. RESULTS Microbial respiration was increased about ten-fold by the straw addition (Table l). The effects of water content and N03' addition were insignificant in the straw amended treatments. Respiration was 20% greater in the wetter, unamended soil, while the N03” addition caused a 20% decrease in respiration in unamended soils. Denitrification capacity measurements for the various treatments had a relatively high degree of variability, with a range of coefficients of variation from 8 to 52%. An analysis of variance of the data indicated a highly significant interaction between time of sampling and water content, as well as highly significant main effects of carbon amendment and sampling date. The N03" addition had no effect on the the denitrification capacity of the soil. The effect of the straw addition ‘was most dramatic, resulting in a 40 to 632 increase in denitrification capacity over the unamended soil (Figure 2). The interactive effect of water content and sampling date was caused by the higher denitrification capacity at the lower water content on day 1, while the denitrification capacity was higher in the wetter soil at all other sampling times. However, the water content effect was not significant at any sampling date. Increases in denitrification capacity over time were evaluated with respect to the denitrification capacity at tine beginning of the incubation period. There was a significant increase in denitrification capacity with time when carbon was added at all sampling periods except day l for the 282 moisture treatment, when carbon was added. Without 119 120 Table 1. Cumulative C02 evolution over a seven day incubation period. 237. 1120 287. 1120 Carbon addition -N03" +N03’ -N03' +N03' --------- ugCOz-Cg‘l soil--------- - straw 31 i 0.73Jr 24 i 0.98 37 i 0.34 29 i 0.37 + straw 270 i 3.6 270 i 2.0 270 i 3.0 270 i 4.1 +Mean : estimated standard deviation. 121 Figure 2. Changes in active denitrifier biomass over time of incubation. D - 237., no straw; . -23Z H20, 1 mg straw-C g'l; A -282 H20, no straw; ‘ -28Z H20, 1 mg straw-C g" . 122 lil- a a / ‘0 8 .0... u. .0 2.92 010 2.300 5.08.422... Time ((0 Figure 2 123 added carbon, the denitrification capacity did not change significantly from the start of the incubation, except at 28% moisture on day 1 (when it was lower) and day 7 (when it was higher). The flush of C02 evolution was 467. greater when straw was added, while the other factors had little effect. A kc value of 0.41 (Anderson and Domsch, 1978) was used to estimate total microbial biomass from the C02 flush data. Microbial biomass in unamended soil was 455 ug C g-1 (354 ug C g"1 when unfumigated control was subtracted) and 667 11g C g'1 of microbial biomass in straw amended soil. Because the C02 evolved in unfumigated controls of the carbon amended soil was anomalously high--in one case even greater than the fumigated samples, they could not be used to provide a biomass measurement. This anomalous behavior in carbon amended soils has been previously observed (Sparling 3521., 1981) and is apparently due to the inability of the surviving microorganisms to utilize the exogenous carbon source. It was because of this behavior that we used ATP as an additional indicator of microbial biomass. Biomass ATP remained unchanged with time in the unamended treatment (Table 2). There was an increase in microbial ATP over time in the soil which received a carbon addition, presumably due to microbial growth. The new steady stateelevel of biomass was 40% greater than that of the unamended soil. The proportion of active denitrifer biomass to total microbial biomass was expressed by the ratio of denitrification capacity to microbial ATP. We chose to use ATP instead of biomass-C from fumigation because of the difficulties encountered with the control in the CHC13 fumigation method in carbon amended soils. The similarity in the temporal response of both the C02 flush and microbial ATP for the 124 Table 2. Changes in microbial ATP. Time No straw Straw added d ------- ug ATP g'l soil ----- + 0 0.93 i 0.10 0.93 i 0.10 1 1.01 i 0.10 1.19 i 0.13 2 0.96 i 0.12 1.36 i 0.11 4 1.03 i 0.11 1.42 i 0.08 7 1.03 _-l_- 0.18 1.40 i 0.11 +Mean : 957. confidence interval. 125 unamended soil supports the use of ATP for comparative purposes. The ratio of denitrification capacity to microbial ATP was not significantly affected by any of the treatments and remained relatively constant throughout the incubation period (Figure 3). 126 Figure 3. Changes in the ratio of denitrification capacity to microbial ATP over time of incubation. Symbols as in Figure 2. 127 1 ass a: 30.821230”. 508.0303 Time (CD Figure 3 DISCUSSION There was no effect of NO3‘ on the size of either denitrification capacity or total microbial biomass. This was probably because the Capac soil had a high N03- content (> 20 pg N g'l) even without the addition of 100 pg N03'-N g-l. Denitrification would rarely be limited by N03” in a soil with a N03‘ concentration this high (Myrold and Tiedje, 1984). It would also be unlikely that N03” would have been limiting for denitrifying enzyme induction. Under very low N03’ conditions (< 1 pg N g‘l), denitrifier biomass may be restricted by N03' concentrations, since 2 to 8 ug N03'-N are needed to produce 106 denitrifiers in soil (Jacobson and Alexander, 1980). Increasing the water content from 23 to 28% slightly increased denitrification capacity--but not significantly-~and had no significant effect on total microbial biomass. This is consistant with the results of Sexs tone §_t__a_l. (1984), who found no significant change in the Phase I rate (denitrification capacity) of Capac clay loam at water contents of 19.6, 23.2, and 25.5%. However, using coarser textured soils, Smith and Tiedje (1979) and Sexstone st 31. (1984) did find that denitrification capacity increased with increasing water content. These results are consistent with those of Doner _e_£ 21., 1975), who found increases in numbers of denitrifying bacteria, but no change in the total bacterial population when sandy loam soil columns were flooded with N03“ solution. The different responses to water content changes between fine and coarse textured soils may be the result of more complete derepression of denitrifying enzymes (i.e., a greater 128 129 proportion of active to inactive denitrifiers--Figure 1) in finer textured soils because of poorer aeration or a greater number of anaerobic microsites. The potential effect of water content differences on the aeration status of the soils used in this experiment was evaluated by using a model which predicts the anaerobic volume of an aggregated soil (Smith, 1980). This model requires estimates of respiration rate and pore space 02 concentration (which were measured), intra-aggregate oxygen diffusion coefficient, and log mean aggregate radius and dispersion constant. These last three parameters were estimated by to be 5 x 10"6 cm2 8'1, 0.2 cm, and 1.0, respectively, for the Capac clay loam (Sexstone _e_t_§_1_., 1984). The anaerobic functions calculated from this model were less than 0.05% for the 23 and 287.’ water content treatments. The lack of a significant moisture effect was most likely due to the insignificant difference in anaerobiosis at the two water contents used in this study. Unlike the other two factors, the addition of straw did cause an increase in the denitrification capacity and also in total microbial biomass. The carbon addition most likely caused an increase in active denitrifier biomass through growth, since the ratio of denitrification capacity to total microbial ATP remained constant. Smith and Tiedje (1979) also observed an apparently growth related response to their glucose addition to soil. This type of effect is not unlikely, since the majority of the denitrifiers in soil are chemoheterotrophs (Firestone, 1982). Thus, the active denitrifier biomass should increase in proportion to total microbial biomass, as long as denitrifiers can effectively compete with other heterotrophs for carbon. Indeed, Smith and Tiedje (1980) have shown that some denitrifiers are more capable 130 competing as aerobic heterotrophs than they are as denitrifiers. This contentitu1zis also supported by the work of Stanford ££_§13 (1975) who found long-term denitrification activity measures to be highly correlated with extractable glucose-C, a parameter which has been correlated with total microbial biomass (Jenkinson, 1968). It could, of course, be possible for derepression of inactive denitrifier biomass to occur to the same extent that the total microbial biomass increased due to growth (see Figure 1). This combination is, however, rather unlikely. In addition, when Smith's model of anaerobiosis was used with the respiration rates and pore space 02 concentrations of straw amended soil, an anaerobic fraction of less than 0.7% was obtained. This is not much greater than those of the unamended soils and is not likely to have caused a disproportionate change in the derepression of denitrifying enzymes. These results suggest that N03’ should not be important in establishing and maintaining denitrification capacity, tn: least at the generally high NO3‘ concentrations found in agricultural soils. In this study, carbon availability, through the mechanism of microbial growth was the dominant factor controlling denitrification capacity. However, in other soils or under different conditions of carbon availability and moisture, moisture could be a potentially important controlling factor and the derepression mechanisms could predominate. 10. 11. 12. 13. REFERENCES Anderson, J.P.E. and K.M. Domsch. 1978. A physiological method for the quantitative measurement of microbial biomass in soils. 8011 31010 BiOChemo 10:215-2210 Doner, R.E., M.G. Volz, L.W. Belser, and J-P. Loken. 1975. Short term nitrate losses and associated microbial populations in soil columns. Soil Biol. Biochem. 7:261-263. Firestone, M.R. 1982. Biological denitrification. IE. F.J. Stevenson (ed.) Nitrogen in agricultural soils. Agronomy 22:289-326. Am. Soc. Agron., Madison, WI. Jacobson, S.N. and M. Alexander. 1980. Nitrate loss from soil in relation to temperature, carbon source and denitrifier p0pu1ations. Soil Biol. Biochem. 12:501-505. Jenkinson, D.S. 1968. Chemical tests for potentially available nitrogen in soil. J. Sci. Food Agr. 19:160-168. Jenkinson, D.S. and D.S. Powlson. 1976. The effects of biocidal treatments on metabolism in soil--I. Fumigation with chloroform. $011 31010 BiOChemo 83167-1770 Knowles, R. 1982. Denitrification. Microbiol. Rev. 46:43-70. Myrold, D.D. and J.M. Tiedje. 1984. Diffusional constraints of denitrification in soil. Soil Sci. Soc. Am. J. (submitted). Parkin, T.B., H.F. Kaspar, A.J. Sexstone, and J.M. Tiedje. 1983. Use of a gas-flow soil core method for measuring field denitrification rates. Soil Biol. Biochem. (in press) Paul, E.A. and R.L. Johnson. 1977. Microscopic counting and adenosine 5'-tri-phosphate measurement in determining microbial growth in soils. Appl. Environ. Microbiol. 34:263-269. Sexstone, A.J., T.B. Parkin and J.M. Tiedje. 1984. Interactive control of soil denitrification rates by oxygen and moisture. Soil Sci. Soc. Am. J. (prepared for publication) Smith, M.S. and J.M. Tiedje. 1979. Phases of denitrification following oxygen depletion in soils. Soil Biol. Biochem- 11:261-267. Smith, M.S. and J.M. Tiedje. 1980. Growth and survival of antibiotic-resistant denitrifier strains in soil. Can. J. Microbiol. 26:854-856. 131 14. 15. 16. 17. 132 Sparling, G.P., B.C. 0rd, and D. Vaughan. 1981. Microbial biomass and activity in soils amended with glucose. Soil Biol. Biochem. 13:99-104. Stanford, G., R.A. Vander Pol, and S. Dzienia. 1975. Denitrification rates in relation to total and extractable soil carbon. Soil Sci. Soc. Am. Proc. 39:284-289. Tiedje, J.M., A.J. Sexstone, D.D. Myrold, and J.A. Robinson. 1982. Denitrification: ecological niches, competition and survival. Ant. van Leeuwen. J. Microbiol. 48:569-583. Volz, M.G., L.W. Belser, M.S. Ardakani, and A.D. McLaren. 1975. Nitrate reduction and associated microbial p0pu1ations in a ponded Hanford sandy loam. J. Environ. Qual. 4:99-102. HICHIGQN STQTE UNIV. LIBRQRIES llHI"WI/HlllHllUMIIIHHIIII‘UHIHINIIHINIIHHll 31293006427185