HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (23) Citing Articles via Google Scholar Google Scholar Articles by Drury, C. F. Articles by Kay, B. D. Search for Related Content PubMed Articles by Drury, C. F. Articles by Kay, B. D. Agricola Articles by Drury, C. F. Articles by Kay, B. D. Related Collections Nutrient Cycling Nutrient Management Nitrogen

DIVISION S-3—SOIL BIOLOGY & BIOCHEMISTRY

The Non-Limiting and Least Limiting Water Ranges for Soil Nitrogen Mineralization

^a Greenhouse & Processing Crops Research Centre, Agriculture & Agri-Food Canada, Harrow, ON Canada N0R 1G0
^b Dep. of Land Resource Science, Univ. of Guelph, Guelph, ON Canada N1G 2W1

^* Corresponding author (druryc{at}agr.gc.ca).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
A better understanding of factors controlling N mineralization would improve our ability to estimate fertilizer requirements more accurately. Net N mineralization approaches a small reaction rate at low and high water contents, giving rise to lower and upper limiting water contents and the least limiting water range (LLWR). Within the LLWR, there is a range in water contents in which mineralization is largely independent of water content, that is, the non-limiting water range (NLWR). An incubation study was conducted to determine the LLWR and NLWR for five soils with different properties, and two relative compaction levels with and without the addition of a legume crop residue. These soils were incubated for 1 and 3 mo at eight water contents. Net N mineralization increased with incubation time and legume addition and varied curvilinearly with water-filled pore space (WFPS). Logistic functions were generated to establish the relationships between net N mineralization and WFPS (%) and to calculate LLWR and NLWR. The mean NLWR was 32.2% after 1 mo and decreased to 18.1% after 3 mo whereas the mean LLWR was 55% after 1 mo and increased to 70.8% after 3 mo. The LLWR increased with organic C or total N but decreased with the addition of legume residue. The NLWR decreased with clay content and with the addition of legume after 3 mo. Emissions of N₂O were greatest at water contents near the upper limit of the LLWR. Use of the NLWR to differentiate soils on the basis of the sensitivity of N mineralization to variation in water content is illustrated.

Abbreviations: LLWR, least limiting water range • NLWR, non-limiting water range • NUE, N use efficiency • OC, organic C • WFPS, water-filled pore space

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
THE MINERALIZATION of organic N from crop residues and soil organic matter can provide 30 to 100% of the crop N requirement, with the largest amount being provided when a legume is incorporated into the rotation. Since inorganic N fertilizers make up >20% of the operating expenses in corn (Zea mays L.) production (Tollenaar, 1996), it is imperative that N use efficiency (NUE) be improved to increase net income. Increased NUE will also reduce the potential for environmental contamination. Management plans that successfully increase NUE must accommodate the spatial variation in texture, organic matter content, soil compaction, and soil water content that is characteristic of many agricultural landscapes.

The identification of N management units represents one approach to increase NUE in variable landscapes. Any field may consist of two or more different units with N being managed differently in each unit to optimize the overall NUE. Criteria for identifying N management units are currently being sought, and the impact of water content on N mineralization suggests that the effects of texture, organic matter, and compaction merit further examination. These factors also impact other processes including denitrification, which may dramatically reduce NUE. Specific information is required on the ranges in water contents in which the accumulation of mineral N is maximized or limited in different soils, and the relations between the water content, which limit N mineralization and those encountered in these soils through the period of greatest crop demand for N. Although the dependence of N mineralization on water content varies with temperature (Grundmann et al., 1995; Quemada and Cabrera, 1997), the spatial variability in water content is expected to be greater than that of soil temperature during the period of greatest crop demand for N and therefore emphasis is placed on the relation between N mineralization and soil water content.

Microbiological processes in a soil with a given structure are adversely affected by both high and low water contents. Inadequate aeration affects reaction rates at high water contents. McKenney et al. (1995) found that N mineralization occurs under both aerobic and anaerobic conditions. Nitrate resulting from mineralization and nitrification under aerobic conditions may subsequently be denitrified if anaerobic conditions develop. Consequently, measurements of changes in the mineral N content of soils reflect net N mineralization, as distinguished from gross or total N mineralization. Nitrogen mineralization is affected at low water contents by diminished rates of substrate diffusion (Skopp et al., 1990), physical protection of bacteria from predation by protozoan grazers (Killham et al., 1993) and low water potential (Sommers et al., 1980). It is therefore possible to identify maximum and minimum water contents at which the rates of net N mineralization approach zero, giving rise to upper and lower limiting water contents, respectively. The difference between the upper and lower limiting water contents is referred to as the LLWR. Within the LLWR, a range of intervening water contents exist where biological processes are independent of water content, or exhibit limited variation with water content. This range of water contents represents the non-limiting water range, NLWR. The ratio NLWR/LLWR describes the fraction of the range in water contents over which net mineralization is positive and independent of water content. The overall shape of the curve relating net mineralization to water content can then be represented by three parameters: the maximum mineral N (or the maximum increase in mineral N above the initial value), the LLWR, and the NLWR. Although the concepts of limiting water contents, LLWR and NLWR, and the maximum rate of a given process (e.g., leaf growth, transpiration) have been applied to processes in plants (da Silva and Kay, 1996; Sadras and Milroy, 1996), they have not been applied to microbiological processes in general or to N mineralization in particular.

Studies of the dependence of net N mineralization on soil water content in different soils have built on the seminal work of Stanford and Epstein (1974). However, the experimental designs have generally precluded the possibility of identifying water contents at which net mineralization approached zero and therefore a LLWR has not been identified. Grundmann et al. (1995) recognized the value of defining water contents that correspond to the upper and lower limits of the LLWR for net N mineralization but arbitrarily set the lower limit equal to that at the wilting point (-1.5 MPa) and the upper limit equal to that at a WFPS of 75%. Difficulties are also encountered in determining a NLWR from these studies. The dependence of net mineralization on water content has been described with curvilinear functions and these functions are used to calculate the maximum rate of net mineralization and the corresponding water content (the optimum water content). Although this approach implies that the NLWR is small or approaches zero, a cursory examination of the data originally reported by Stanford and Epstein (1974) and later data reported by Grundmann et al. (1995) suggest a finite value of the NLWR in some soils. Although the role of texture and organic matter content on the functional relation between net N mineralization and water content was considered by Stanford and Epstein (1974), there is little information on the effects of these properties on limiting water contents and none on the LLWR and the NLWR.

The impact of compaction on the functional dependence of net mineralization on soil water contents has received little attention and consequently its influence on the LLWR and the NLWR for net N mineralization has not been assessed. In addition, studies of the impact of compaction at specific water contents or water potentials have produced inconsistent results. For example, Breland and Hansen (1996) found that an increase in bulk density from 1.1 to 1.4 kg m^-3 of a sandy loam (50 g kg^-1clay, 19.1 g kg^-1 organic C [OC]) resulted in a decrease in net mineralization of clover ¹⁵N when the soil was incubated at water potentials of -6 and -25 kPa for 98 d. However, there was no evidence of anaerobiosis and the decrease in net mineralization with increasing compaction was attributed to increased physical protection of organic materials in the compacted soil. In contrast, Rasiah and Kay (1998) found that compaction increased net N mineralization in five of their seven soils (clay contents varied from 85 to 367 g kg^-1; OC contents varied from 13 to 25 g kg^-1) when the soils were incubated at -15 kPa for 70 d. Compaction only resulted in a decrease in net mineralization in the soil with the highest clay content. De Neve and Hofman (2000) did not find N mineralization to be influenced by compacting a soil (60 g kg^-1 clay, 11.4 g kg^-1 OC) from bulk densities ranging from 1.2 to 1.6 Mg m^-3 when the incubation was performed at 75% of field capacity. Differences between the studies may reflect the influence of soil structure on net N mineralization and the differences reinforce a conclusion drawn in a recent review by Jarvis et al. (1996) that more research is needed on the influence of soil texture and structure on N mineralization.

The objectives of this study were to determine the limiting water contents, the LLWR, and the NLWR for net N mineralization in soils of different texture, organic matter content, and soil compaction, and to determine the effects of legume addition on these variables.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Soils with a range of clay and OC contents were selected from various locations across central and southwestern Ontario to cover a range in soil physical and chemical properties (Table 1). The Brookston clay loam (fine-loamy, mixed, mesic Typic Argiaquoll or Orthic Humic Gleysol, Canadian Classification System) was sampled from the Hon. Eugene F. Whelan Experimental Farm in Woodslee, ON at the Greenhouse and Processing Crops Research Centre, Agriculture and Agri-Food Canada. The Fox loamy sand (Typic Hapludalfs or Brunisolic Gray Brown Luvisol, Canadian Classification System) and the Conestogo silt loam (Typic Eutrochrepts or Gleyed Melanic Brunisol) were obtained from the Cambridge Research Station and the Elora Research Station, respectively, University of Guelph, Guelph, ON. The Brady sandy loam (Aquic Hapludalfs or Gleyed Brunisolic Gray Brown Luvisol) and the Perth silt loam (Aquic Hapludalfs or Gleyed Gray Brown Luvisol) were both obtained from a private farm near Clinton in southwestern Ontario.

Soils were packed to a relative compaction (bulk density/reference bulk density) of 0.83 or 0.91. These values of relative compaction represent values commonly found under conventional till (0.83) and no-till (0.91) across a range of soil textures, OC contents, and climates (Kay et al., 1997). Values of relative compaction have little or no dependence on texture and OC contents and are, therefore, preferred to bulk density as a treatment variable in experiments with soils of different texture and OC contents. The values of the reference bulk density were determined using a modification of the method of Häkansson (1990) in which about 100 g oven dry equivalent weight of soil was puddled, vacuum saturated, poured into a Rowe cell (7-cm i.d.), a load of 200 kPa applied until drainage had ceased, the pressure released, and bulk density determined after 15 min. The desired bulk density (Table 2) was calculated using the reference bulk density and the required level of relative compaction and then used to determine the weight of soil to pack in aluminum rings (2.5 cm height and 4.8 cm diameter). Water release curves (Klute, 1986) were determined using the packed rings. Values of the water contents at -0.01 and -1.5 MPa are reported in Table 2 for the different combinations of soil and relative compaction.

Treatments included five soils (Table 1), eight water contents (20, 35, 50, 65, 80, 85, 90, and 95% WFPS), two crop residue rates (0 and 100 mg legume N kg^-1 soil), and two levels of relative compaction (0.83 and 0.91). Water-filled pores space refers to the volume of water (V_H2O) present in the soil relative to the total volume of pores (V_pores). Hence the percentage WFPS is calculated as follows: WFPS = 100 x V_H2O/V_pores. Soil samples were packed into the aluminum rings for the incubation study and the water content adjusted using distilled and de-ionized water. Potassium nitrate was added to the soil in the rings, at a rate of 50 mg N kg^-1 soil, to simulate N fertilizer addition at planting. Each aluminum ring was placed into a 265-cm³ glass jar, the glass jar was covered with parafilm, and four holes made in the parafilm to maintain an aerobic environment, while minimizing moisture loss. The soil water content was adjusted weekly. Soil samples without legume were mixed in the same way as with legume-amended samples to eliminate any differences in structure. Samples were incubated for 1 and 3 mo in a controlled environmental room at 20 ± 2°C. The incubation was arranged using a completely randomized design with three replicates.

Nitrous oxide (N₂O) flux was determined in a subset of the treatments to provide evidence of mineral N loss through denitrification. Nitrous oxide accumulation was measured in the containers at 3 d, 1, 2, 4, 6, 9, and 12 wk in three soils (Brady, Perth, and Brookston) at the low compaction treatment, with and without red clover, and at 20, 50, 65, 85, and 95% WFPS. The containers were closed and a 5-mL gas sample was removed at 5 and 35 min. Gas samples were injected into a Varian 3800 gas chromatograph (Mississauga, Ontario) fitted with a 3.0-m Porapak Q column (Chromatographic Specialties Inc., Brockville, ON, Canada) and the oven temperature was 70°C. Argon (95%) and CH₄ (5%) were used as the carrier gas with a flow rate of 30 mL min^-1. The N₂O concentration of the samples and standards were measured using an electron capture detector. Cumulative N₂O emissions were also determined over the 12-wk period by integrating under the curve.

When the incubation was completed (i.e., after 1 or 3 mo), the soil inside the incubation ring was removed and mixed. An equivalent of 10 g oven-dry soil samples was shaken with 100 mL of KCl (2 M) for 1 h, filtered through Whatman No. 40 filter papers (Whatman Ltd., Maidstone, England), and analyzed for mineral N (NH⁺₄–N and NO^-₃–N + NO^-₂–N). Concentrations of NH⁺₄–N and NO^-₃–N + NO^-₂–N in the extract were determined using the Berthelot reaction for NH⁺₄–N and the Cd reduction reaction for NO^-₃–N + NO^-₂–N using a TRAACS 800 autoanalyzer (Bran + Luebbe Analyzing technologies, Buffalo Grove, IL) (Tel and Heseltine, 1990).

The calculation of the NLWR and LLWR requires that a mathematical function be fitted to the data to calculate the WFPS at specific values of mineral N. Preliminary studies indicated that a combined logistic function gave coefficients of determination that were marginally better than the Grundmann function (Grundmann et al., 1995) and superior to a quadratic function and was therefore selected for fitting the data in Fig. 1 and 2 . The Grundmann function was based on a linear increase in mineral N with water content up to the maximum N and subsequently in the form of exp[1/( - max)], where is the relative water content, which is an exponential decline at higher water contents. The logistic function was an S-shaped function. The functional relation between mineral N and WFPS was described by the difference between two logistic functions: one described the decline in mineral N from the maximum and the second provided the adjustment to account for the increase in net N mineralization with increasing water content to a maximum value of mineral N. The two equations could have different shapes (recognizing the different effects that WFPS have on net mineralization and denitrification), but were linked by a common value of the maximum mineral N. The combined logistic equations had the following form:

[1]

where N was the measured mineral N content (mg kg^-1), and a, b, c, d, and e were curve-fitting parameters. For each logistic function, a gives the maximum value, b and d are related to the shape of the each curve, and c and e position each S-shaped curve on the WFPS axis (i.e., c and e define the WFPS at a/2 for each function). The coefficients in Eq. [1] were estimated by applying non linear estimation procedures (Statsoft, 1999) to the mean values of mineral N and WFPS (Fig. 1 and 2). The Quasi-Newton estimation method was used. Parameter starting values were chosen from a graphical examination of the data. Although the maximum number of iterations was set at 400, convergence was normally achieved with <30 iterations. Details on the fits are provided in Tables 3 and 4 for the 1- and 3-mo incubations, respectively.

View larger version (29K): [in this window] [in a new window]   Fig. 1. Soil mineral N in the Brady sandy loam, Fox loamy sand, Conestogo loam, Perth silt loam, and Brookston clay loam after the 1-mo incubation as affected by water-filled pore space, relative compaction, and legume treatments. The dashed line represents the initial mineral N levels in the soil.

View larger version (31K): [in this window] [in a new window]   Fig. 2. Soil mineral N in the Brady sandy loam, Fox loamy sand, Conestogo loam, Perth silt loam, and Brookston clay loam after the 3-mo incubation as affected by water-filled pore space, relative compaction, and legume treatments. The dashed line represents the initial mineral N levels in the soil.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Development of Functional Relations between Mineral N and WFPS
The temporal variation in extractable soil inorganic N with compaction and legume treatments at different values of WFPS are illustrated for 1- and 3-mo incubation periods in Fig. 1 and 2, respectively. The initial N content, N_initial, includes the combination of the background and added (50 mg N kg^-1) inorganic N and was indicated by a dashed horizontal line on the figures. Mineral N values above the line represent an increase in the amount of N mineralized, whereas values below the line represent a loss of mineral N.

N_max is the maximum net mineralization rate and it was determined for the fitted functions and the limiting water contents were determined for specific values of mineral N. It should be noted that, depending on the shape of the functions, N_max was not equal to the term a in Eq. [1]. The upper and lower limits of the LLWR were taken as the WFPS at which mineral N, after a given period of incubation, was equal to the value at the start of the incubation, N_initial. The upper and lower limits of the NLWR were taken as the WFPS at which net N mineralization was equal to 0.98 N_max. The lower and upper limits for the NLWR (98% of N_max) was chosen to indicate a range of WFPS values in which net N mineralization was not affected by changes in water content, but it did allow for variation associated with obtaining data from biological assays. Values of N_max, N_max - N_initial, as well as the limiting water contents are given in Table 5 for 1- and 3-mo incubations.

N_max in the Fox loamy sand soil ranged from 62 to 71 mg N kg^-1 after 1-mo incubation (Table 5). After 3 mo, N_max was 66 to 67 mg N kg^-1 in the no residue treatment and 91 to 97 mg N kg^-1 in the residue-added treatment. There was a dramatic increase in N_max with the legume-amended soils after 3 mo as compared with 1-mo incubation. Mineral N contents were greater than N_initial in the low compaction treatments between 20 and 65% WFPS and between 20 and 85% WFPS for the high compaction treatments.

Net N mineralization in the Conestogo loam soil followed a similar pattern as the Brady sandy loam soil (Fig. 1 and 2). N_max in the no residue treatment ranged from 106 to 112 mg N kg^-1 after 1 mo and was 113 to 131 mg N kg^-1 after 3 mo as compared with the initial value of 91.9 mg N kg^-1 (Table 5). After 3 mo, the residue-added treatments had considerably more N mineralized than the no residue treatments with N_max varying from 128 to 131. There was a dramatic decrease in mineral N at 95% WFPS levels, especially for the two residue-added treatments (Fig. 2). Presumably the denitrifiers were using this soluble C source from the legume treatment to convert NO^-₃ to NO, N₂O, and N₂ and perhaps nitrate was also being immobilized and converted to organic N.

There was very little N mineralized with the Perth silt loam after 1 mo (Fig. 1). After 3 mo, mineral N only increased to 70 to 71 mg N kg^-1 with the no-residue treatments and to 82 to 83 mg N kg^-1 with the residue-added treatments from the initial amount of 54 mg N kg^-1 (Table 5). Mineral N decreased from the initial concentrations at high WFPS levels, especially for the residue-added treatments.

The Brookston clay loam soil also had reduced amounts of N mineralized at 1 mo (62–67 mg N kg^-1) compared with 3 mo (71–93 mg N kg^-1)(Fig. 1 and 2). At 3 mo, mineral N increased to a maximum of 71 to 78 mg N kg^-1 for the no residue-added treatments and to a maximum values of 93 mg N kg^-1 for both of the residue added treatments from the initial value of 57 mg N kg^-1 (Table 5). It was interesting to note that in the Brookston soil, net mineralization occurred with all treatments for the 90% WFPS and for the high compaction treatment at the highest WFPS value (95%), whereas all of other soils had either negligible mineral N values or mineral N contents that were less than the initial values at these higher WFPS levels.

The values of N_max - N_initial were positive for all combinations of soil, relative compaction, legume amendment, and incubation times (Table 5). The mean N_max - N_initial at the 3-mo incubation (27.6 mg kg^-1) was significantly greater than that at 1 mo (11.6 mg kg^-1). The addition of the red clover residue increased N_max - N_initial by two fold over the no residue treatment at 3-mo incubation, on average across all treatments, by 18 mg kg^-1. However, the addition of the clover residue had little effect on N_max - N_initial at 1 mo, indicating the N released through mineralization of the legume was supporting growth in the microbial population during this period. Multiple regression analyses using forward stepwise procedures (Statsoft, 1999) were used to identify the soil characteristics accounting for most of the variation in N_max - N_initial after the 3-mo incubation. Selection was made from the following independent variables: legume (a class variable with values of 0 and 1 for the control and legume-added treatment, respectively), relative compaction, OC, total N, clay, silt, and sand and interaction terms. Multiple regression analysis described a positive dependence of N_max - N_initial on legume and total N content and a negative dependence on legume x clay (R² = 0.90, P < 0.0001). Relative compaction was not significantly related to N_max - N_initial in a simple linear regression and was not selected in the multiple regression. However, compaction did influence the WFPS at which N_max occurred; N_max occurred at greater values of WFPS at the larger relative compaction. This observation will be explored in more detail in the discussion on limiting water contents and may account for the inconsistent conclusions arising from previous studies of the effects of soil compaction on net N mineralization.

Critical Water Contents, LLWR, and NLWR for Net N Mineralization
Values of the upper and lower limiting water contents, the LLWR and the NLWR, and the ratio NLWR/LLWR (Table 7) reflected variation in the shape of the curves describing the data. Fits of Eq. [1] to data obtained after the 3-mo incubation from three data sets were selected to illustrate variations in the shape of the curves and the impact on the limiting water contents (Fig. 3) . The fitted curve was extended beyond the available data to illustrate difficulties in estimating limiting water contents in some instances. The Fox loamy sand soil had the lowest OC and total N contents of the five soils, N_max - N_initial was small, and without the addition of red clover exhibited little sensitivity of mineral N to variation in WFPS at either level of compaction. At a relative compaction of 0.83 the NLWR extended from 39.5 to 65.6% WFPS and the LLWR extended from 17.3 to 73.0% (Fig. 3a). The ratio of NLWR/LLWR was 0.47. The lack of sensitivity of mineral N to WFPS created some difficulties in estimating the lower limit of the LLWR. In the case of the Fox soil without legume residue at a relative compaction of 0.91, the initial value of mineral N was only obtained by extrapolating the curve to 0% WFPS. The lack of data at WFPS <20% precluded accurate estimation of the lower limit of the LLWR in this soil at either 1 or 3 mo. The same situation was encountered in the Brady soil at a relative compaction of 0.83 that was amended with legume after the 3-mo incubation. The clay content of the Brady soil was similar to that of the Fox, but the OC and total N contents were larger by a factor of two and three, respectively (Table 1). At a relative compaction of 0.91, the legume-amended treatment (Fig. 3b) exhibited greater curvature in the Brady than in the Fox soil. The NLWR was lower and shifted to higher limiting values (the NLWR extended from 58.5 to 74.2% WFPS). The LLWR extended from 6.0 to 91.3 (Fig. 3b). Values of the LLWR were identical for the control and legume-amended treatments (84.8 and 85.3, respectively), but the greater curvature in the amended treatment of the Brady resulted in values of NLWR/LLWR decreasing from 0.36 for the control to 0.18 for the amended treatment, respectively. More extreme curvature was exhibited in some of the treatments of the Brookston soil. At a relative compaction of 0.91, the mineral N content of the legume-amended treatment increased continuously to 90% WFPS before beginning to decline (Fig. 3c). This resulted in the NLWR extending from 89.4 to 94.5% WFPS and the LLWR extending from 33.8 to 95.4% WFPS. The LLWR was similar in magnitude to that of the Brady soil but the sharpness of the decline in the curve for the Brookston resulted in a value of NLWR/LLWR of 0.08. These analyses demonstrate that the magnitude of the NLWR and the ratio NLWR/LLWR decrease with increasing curvature (including sharpness in the decline in N at large values of WFPS). The magnitude of the NLWR, the LLWR and their limiting water contents must reflect the response of the microbial population to available substrates and O₂ and the influence of soil structure on the fluxes of substrate and O₂ to sites of microbial activity.

View larger version (23K): [in this window] [in a new window]   Fig. 3. Illustration of variation in shape of curve relating net N mineralization to WFPS (a) Fox, relative compaction 0.83, 3-mo incubation, no red clover added; (b) Brady, relative compaction 0.91, 3-mo incubation, red clover added; (c) Brookston, relative compaction 0.91, 3-mo incubation, red clover added.

As the incubation time increased from 1 to 3 mo, the decrease in the NLWR, its shift to higher water contents and the increase in the LLWR corresponded to a change in the shape of the curves from that represented by Fig. 3a toward that represented by Fig. 3b and 3c. The decrease in the NLWR and the increase in the LLWR with the increasing incubation time resulted in a significant decline in the mean value of the ratio of NLWR/LLWR (from 0.54 to 0.24). Multiple regression analyses (not shown) did not identify any significant relations between NLWR/LLWR and soil properties at 1 mo, but at 3 mo NLWR/LLWR was negatively correlated with clay content, legume addition, and relative compaction, and positively correlated with legume x clay content (R² = 0.84).

Several generalizations can be drawn from multiple regression analyses relating the NLWR, the LLWR, and their respective limits to soil properties (Table 8). First, values of the NLWR and the LLWR were generally less strongly correlated with soil variables than were their respective limits. The stronger regression with the upper and lower limits would be expected since the ranges incorporate errors associated with estimating these limits. Second, the magnitude of the LLWR at both 1- and 3-mo incubations was positively related to OC or total N and negatively related to the addition of the legume residue. Third, the magnitude of the NLWR after 3-mo incubation was much more strongly influenced by the addition of the legume residue than at 1 mo. The NLWR decreased with the addition of the legume residue, and the decrease was because of the positive effect of residue addition on the lower limit. Fourth, the upper limit of both the NLWR and the LLWR at both 1- and 3-mo incubation were more strongly correlated with variables in the study than were the lower limits with the exception of NLWR at 3 mo where there was a similar correlation coefficient for both limits. Finally, the upper limits in the NLWR and the LLWR at both incubation times were positively correlated to the same three variables: relative compaction, clay content, and total N. The consistency among the four regression equations for the upper limits is striking, and the positive correlations unexpected. The positive correlations of the upper limits of both the NLWR and the LLWR with relative compaction and clay content are not compatible with a reduction in O₂ flux, that would be expected to be associated with a decrease in pore diameter arising from either increasing compaction or increasing clay content. The positive correlation with relative compaction is considered in more detail below. The positive correlation with clay content may reflect reduced demand for O₂ with increasing clay content. Reduced demand for O₂ would reflect diminished microbial activity and would be compatible with the negative effect of clay content on N_max - N_initial referred to above. The positive correlation with total N is not compatible with an increased demand for O₂ arising from an increased biomass in soils richer in substrate N and C.

The positive correlation of the upper limit of the NLWR at 3 mo with relative compaction was complemented by a corresponding positive correlation of the lower limit with relative compaction. The NLWR was not correlated to relative compaction implying that the main effect of compaction was to shift the range to higher water contents. This observation has important implications for interpreting what would seem to be conflicting reports in the literature on the effects of compaction on net N mineralization. Although compaction did not affect N_max - N_initial, it shifted the range in water contents in which net N mineralization varied between 0.98 N_max and N_max. This would mean that incubations of soils at different levels of compaction but similar water potentials (and therefore different WFPS) could exhibit different rates of net N mineralization depending on the impact of compaction on the WFPS at the given potential and the effects of compaction on the upper and lower limits of the NLWR. It is conceivable that incubations performed at a given potential could result in increases, decreases or no change in net N mineralization with increasing compaction.

Soil compaction reduces the volume fraction of large pores as well as total porosity (Kay et al., 1997). The use of WFPS as a measure of water content in mineralization studies follows from the work of Linn and Doran (1984) who were looking for an index of aerobic microbial activity that was independent of soil compaction. Water contents expressed on the basis of WFPS incorporate compaction-induced changes in total porosity. Consequently, there is a possibility that selection of relative compaction in multiple regression analyses of dependent variables with units of WFPS is because of the influence of compaction on total porosity. This problem could be removed by expressing water contents on a volumetric basis. Values of the NLWR, the LLWR, and their respective limits, based on units of WFPS (Table 7), were converted to values with units of volumetric water contents and the multiple regression analyses rerun. The regression equations (Table 9) generally accounted for similar variation in the dependent variables and showed the same relations with textural variables, OC, total N, and legume amendment addition as given in Table 8. Although the correlations with relative compaction remained about the same at 1-mo incubation, the contribution of relative compaction to the magnitude of the dependent variables was reduced (Table 9). At 3-mo incubation, the relative compaction term was dropped from equations for limiting values (Table 9), implying that the increase in the limits of NLWR because of compaction, when defined in units of percentage (%) of WFPS (Table 8), was largely because of the compaction-induced decrease in total porosity. The correlation for NLWR at 3 mo remained unchanged.

It is instructive to compare the upper and lower limits of the NLWR and the LLWR with the water contents at field capacity and the permanent wilting point (Fig. 4 and 5) . The upper limiting water contents of the NLWR and the LLWR at 1- and 3-mo incubations, expressed on a volumetric basis, do not approximate the water contents at field capacity, taken as -0.01 MPa (Table 2). Trends are illustrated in data from the 3-mo incubation in Fig. 4. Notwithstanding difficulties related to hysteresis that arise when data obtained during water release are compared with data obtained from the addition of water to dry soils, the data do indicate that the upper limits of both the NLWR and the LLWR lie below the 1:1 line for several soils. The number of soils falling below the 1:1 line was greater at 1-mo than at 3-mo incubation. In soils falling below the 1:1 line, the water contents would not have fallen to the upper limits after rapid drainage had ceased, that is, the water contents at -0.01 MPa are greater than the limits. The potential for denitrification is greatest when soil water content fails to reach the upper limit of the LLWR when rapid drainage ceases. At 3 mo, the water content at -0.01 MPa was greater than the upper limit of the LLWR in four treatments: the Conestogo and Perth soils at a relative compaction of 0.91, with and without legume added. While the structure of the sieved soils used in this study does not duplicate that in the field, the analyses indicate the value of defining the upper limits in relation to field capacity to identify the potential for denitrification.

View larger version (19K): [in this window] [in a new window]   Fig. 4. Values of the upper limits of the NLWR and the LLWR relative to the water content at field capacity (-0.01 MPa).

View larger version (20K): [in this window] [in a new window]   Fig. 5. Values of the lower limits of the NLWR and the LLWR relative to the water content at permanent wilting point (-1.5 MPa).

Nitrous Oxide Emissions
Emissions of N₂O confirmed the contribution of denitrification to losses of mineral N at high water contents and the value of the upper limit of the LLWR in contrasting the water contents at maximum emissions from different soils. Emissions of N₂O were very high after a 3-d and 1-wk period, especially for soils incubated at 65 to 95% WFPS (Fig. 6) . The water contents at which largest emissions occurred were in the vicinity of the upper limit of the LLWR for the low compaction treatment, indicating the importance of denitrification to this limit. The highest emissions in the Brady soil were at 65% WFPS followed by 85% WFPS (the upper limits of the LLWR after 1-mo were 73 and 72% WFPS without and with legume added, respectively). In the Perth silt loam soil, the emissions were greatest at 85% WFPS in the absence of red clover (upper limit of the LLWR of 83%) and at 65% WFPS with red clover (upper limit of LLWR of 79%). The Brookston soil had the greatest emissions among these three soils at 95% WFPS (upper limits of LLWR of 85 and 83% without and with legume added, respectively) and these emissions were over ten-fold higher than those for either the Brady or Perth soils. The 85% WFPS treatment also had high emissions, especially when the soils were incubated in the presence of red clover.

View larger version (27K): [in this window] [in a new window]   Fig. 6. Nitrous oxide emissions during the 12-wk incubation for Brady sandy loam, the Perth silt loam, and the Brookston clay loam soils at 20, 50, 65, 85, and 95% water-filled pore space (WFPS).

View larger version (17K): [in this window] [in a new window]   Fig. 7. Cumulative N2O emissions at 20, 50, 65, 85, and 95% for the Brady sandy loam, Perth silt loam and Brookston clay loam soils with and without red clover after a 12-wk incubation.

The use of the NLWR in identifying N management units would first involve selecting values of NLWR for a specific incubation period and then arbitrarily setting values of the NLWR that could be used to distinguish management units. For illustrative purposes we will use values obtained after the 3-mo incubation since this period is more representative of the duration over which net mineralization contributes significantly to the supply of N used by crops such as maize in cool temperate regions. In addition, the characteristics at this time showed the greatest dependence on soil characteristics. For simplicity, we will use values of a NLWR of 0.10 m³ m^-3 to distinguish between soils in which there is weak and strong dependence of net N mineralization on water content. Net N mineralization in soils with values of the NLWR 0.10 m³ m^-3 would have a weak dependence on water content. The net N mineralized in these soils would be largely determined by the amount of potentially mineralizable N and, providing water contents throughout the season generally fell within the NLWR, management units could be further distinguished among these soils on the basis of potentially mineralizable N. On the other hand, net N mineralization in soils with a NLWR < 0.10 m³ m^-3 would have a strong dependence on water content and net N mineralization would be more strongly influenced by weather. Nitrogen management practices that accounted for this lack of predictability in net N mineralization would be required.

The NLWR after the 3-mo incubation decreased with increasing clay content, but this sensitivity to texture was reduced in the presence of added legume residue (Table 9). The NLWR (expressed on a volumetric water content basis) decreased by a factor of three as the clay content increased from 50 to 380 g kg^-1 in the no legume treatment and the regression equation between NLWR and clay content was: NLWR (m³ m^-3) = -0.003Clay + 0.1641 (r² = 0.80). Applying the critical value of a NLWR of 0.10 m³ m^-3 to the no legume treatment would mean that three of the five soils (those with clay contents 150 g kg^-1) would be expected to exhibit a weak dependence of net N mineralization on water content. These soils might be further differentiated on the basis of potentially mineralizable N. For instance, the Fox and the Brady soils both have large and similar values of NLWR, but N_max - N_initial at 3 mo in the no legume treatment was 5 and 33 mg N kg^-1in the two soils, respectively (Table 5). The differences in N_max - N_initial were related to differences in total N (Table 1) and these two soils might be considered representative of different management units. However, when legume residue was added, the sensitivity to water content increased in all soils and the NLWR fell to <0.10 m³ m^-3 on all five soils. The NLWR of the Fox and the Brady soils fell by a factor of two when the residue was added (r² = 0.41) and the N_max - N_initial at 3 mo, averaged across the compaction treatments was 33 and 48 mg N kg^-1 in the Fox and Brady soil, respectively (Table 5). These data indicate that net N mineralization in coarser-textured soils is less sensitive to variation in water content and we speculate, therefore, that N management units could be more readily established on these soils if legume residues were not present. However, in the presence of legume residue, the value in trying to distinguish between these two soils would be reduced. Net N mineralization in finer-textured soils with a NLWR <0.10 m³ m^-3 would vary with both water content and potentially mineralizable N and would require N management practices that take both factors into account.

The upper limit of the LLWR is best interpreted in relation to the water content at field capacity, in the absence of seasonal water content data. Soils with an upper limit of the LLWR that is lower than, or approaches field capacity, may require the application of more fertilizer N to accommodate for losses because of denitrification and would be expected to have the largest emissions of nitrous oxide.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
The variation in net N mineralization with time of incubation, legume addition, and soil water content followed expected trends. The amount of N mineralized increased with incubation time and the addition of legume residue, especially for WFPS between 50 and 80%. Net N mineralization varied curvilinearly with water content. At WFPS values >80%, the increased OC inputs combined with the presence of anaerobic microsites enhanced denitrification losses.

Nitrous oxide emissions were very high after a 3-d and 1-wk period, especially for soils incubated at 65 to 95% WFPS. The cumulative losses of nitrous oxide after 3 mo were greatest in the finest textured soil and represented about 43% of the N present in that soil at the start of the incubation.

Compaction did not influence the maximum amount of mineral N, but reduced mineral N losses especially when the WFPS was >80%. The impact of compaction on net N mineralization is determined by the influence of compaction on WFPS and on the limits of the NLWR.

The shape of the functional relation between net N mineralization and soil water content was characterized using three variables: N_max, the LLWR, and the NLWR. The LLWR and the NLWR provide information on the sensitivity of net N mineralization to variable water content that is not provided by the optimum water content. Values of the LLWR and the NLWR reflect an integration of the effects of soil structure on oxygen supply and substrate availability, the effects of organic matter and crop residue on microbial activity, and temporal changes in the dynamics of the population of soil organisms participating in N mineralization. The LLWR increased with incubation time and decreased with addition of legume residue. The NLWR decreased with increasing incubation time and after 3 mo decreased with the addition of legume residue and increasing clay content. Non-limiting water range, as a fraction of LLWR varied from 0.08 to 0.47 after the 3-mo incubation and was negatively correlated with clay content, legume addition and relative compaction. Coefficients of determination of regression equations relating the limiting water contents and the ranges to other soil properties were similar when water contents were expressed in units of water-filled porosity or volumetric water content.

We speculate that the NLWR and the upper limit of the LLWR can be used to differentiate soils on the basis of sensitivity of net N mineralization to variable water content. Soils with a large NLWR (coarse-textured soils without recent addition of legume residue) will be least sensitive to variation in water content. If these soils occur within the same field, different N management units could be distinguished among them on the basis of measures of potentially mineralizable N. Net N mineralization in soils with a small NLWR (fine-textured soils and soils to which legume residue has recently been added), on the other hand, will be much more strongly influenced by water content. Identification of N management units among these soils would be more complicated and management of fertilizer N on these soils would have to take into account both the amount of potentially mineralizable N and the variation in water content. The upper limit of the LLWR provides a measure of the susceptibility of a soil to denitrification and to the emission of nitrous oxide. Denitrification would be most prevalent in soils in which the upper limit of the LLWR falls below field capacity. The functional relation between net N mineralization and water content has a profound effect on the amount of N available to plants and we have developed a method to examine this relationship more thoroughly.

	ACKNOWLEDGMENTS

 
Financial support for this project from the Ontario Research Enhancement Program, Agriculture and Agri-Food Canada is gratefully acknowledged. Appreciation is also extended to Dr. T. O. Oloya for analyzing the soil extracts for ammonium and nitrate concentrations.

Received for publication July 10, 2001.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 

• Breland, T.A., and S. Hansen. 1996. Nitrogen mineralization and microbial biomass as affected by soil compaction. Soil Biol. Biochem. 28:655–663.
• da Silva, A.P., and B.D. Kay. 1996. The sensitivity of shoot growth of corn to the least limiting water range of soils. Plant Soil 184:323–329.
• da Silva, A.P., A. Nadler, and B.D. Kay. 2001. Factors contributing to temporal stability in spatial patterns of water content in the tillage zone. Soil Tillage Res. 58:207–218.
• De Neve, S., and G. Hofman. 2000. Influence of soil compaction on carbon and nitrogen mineralization of soil organic matter and crop residues. Biol. Fertil. Soils 30:544–549.
• Drury, C.F., D.J. McKenney, and W.I. Findlay. 1992. Nitric oxide and nitrous oxide production from soil: Water and oxygen effects. Soil Sci. Soc. Am. J. 56:766–770.[Abstract/Free Full Text]
• Franzluebbers, A.J. 1999. Microbial activity in response to water-filled pore space of variably eroded southern Piedmont soils. Appl. Soil Ecol. 11:91–101.
• Gee, G.W., and J.W. Bauder. 1986. Particle-size analysis. p. 383–411. In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
• Grundmann, G.L., P. Renault, L. Rosso, and R. Bardin. 1995. Differential effects of soil water content and temperature on nitrification and aeration. Soil Sci. Soc. Am. J. 59:1342–1349.[Abstract/Free Full Text]
• Häkansson, I. 1990. A method for characterizing the state of compactness of the plough layer. Soil Tillage Res. 16:105–120.
• Hendershot, W.H., H. Lalande, and M. Duquette. 1993. Soil reaction and exchangeable acidity., p. 39–42. In Carter, M. R. (ed.), Soil Sampling and Methods of Analysis. Lewis Pub./CRC Press, Boca Raton, FL
• Herlihy, M. 1979. Nitrogen mineralization in soils of varying texture, moisture and organic matter I. Potential and experimental values in fallow soils. Plant Soil 53:255–267.
• Jarvis, S.C., E.A. Stockdale, M.A. Shepherd, and D.S. Powlson. 1996. Nitrogen mineralization in temperate agricultural soils: Processes and measurement. Adv. Agron. 57:187–235.
• Kachanoski, R.J., and E. de Jong. 1988. Scale dependence and the temporal persistence of spatial patterns of soil water storage. Water Resour. Res. 24:85–91.
• Kay, B.D., A.P. da Silva, and J.A. Baldock. 1997. Sensitivity of soil structure to changes in organic C content: Predictions using pedotransfer functions. Can. J. Soil Sci. 77:655–667.
• Killham, K., M. Amato, and J.N. Ladd. 1993. Effect of substrate location in soil and soil-water regime on carbon turnover. Soil Biol. Biochem. 25:57–62.
• Klute, A. 1986. Water retention: Laboratory methods. p. 635–662. In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
• Linn, D.M., and J.W. Doran. 1984. Effect of water-filled pore space on carbon dioxide and nitrous oxide production in tilled and nontilled soils. Soil Sci. Soc. Am. J. 48:1267–1272.[Abstract/Free Full Text]
• McGill, W.B., and C.T. Figueiredo. 1993. Total Nitrogen. p. 201–211. In M.R.Carter (ed.) Soil Sampling and Methods of Analysis. Lewis Pub./CRC Press, Boca Raton, FL
• McKenney, D.J., S.-W. Wang, C.F. Drury, and W.I. Findlay. 1995. Denitrification, immobilization, and mineralization in nitrate limited and nonlimited residue amended soil. Soil Sci. Soc. Am. J. 59:118–124.[Abstract/Free Full Text]
• Nelson, D.W., and L.E. Sommers. 1986. Total carbon, organic carbon, and organic carbon content. p. 539–579. In A.L. Page et al. (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
• Quemada, M., and M.L. Cabrera. 1997. Temperature and moisture effects on C and N mineralization from surface applied clover residue. Plant Soil 189:127–137.
• Rasiah, V., and B.D. Kay. 1998. Legume-N mineralization: Effect of aeration and size distribution of water-filled pores. Soil Biol. Biochem. 30:89–96.
• Sadras, V.O., and S.P. Milroy. 1996. Soil-water thresholds for the responses of leaf expansion and gas exchange: A review. Field Crops Res. 47:253–266.
• SAS Institute, 1996. SAS User's guide: Statistics, 6th edition, SAS Institute, Cary, North Carolina.
• Schoenau, J.J., and R.E. Karamanos. 1993. Sodium bicarbonate-extractable P, K, and N. p. 51–58. In M.R. Carter (ed.) Soil sampling and methods of analysis. Lewis Pub./CRC Press, Boca Raton, FL.
• Simard, R.R. 1993. Ammonium acetate-extractable elements sodium bicarbonate-extractable P, K, and N. p. 39–42. In M.R Carter (ed.) Soil sampling and methods of analysis. Lewis Pub./CRC Press, Boca Raton, FL.
• Skopp, J., M.D. Jawson, and J.W. Doran. 1990. Steady-state aerobic microbial activity as a function of soil water content. Soil Sci. Soc. Am. J. 54:1619–1625.[Abstract/Free Full Text]
• Sommers, L.E., C.M. Gilmour, R.E. Wildung, and S.M. Beck. 1980. The effect of water potential on decomposition processes in soils. p. 97–117. In J. F. Parr et al. (ed.) Water potential relations in soil microbiology. SSSA Spec. Pub. 9. SSSA, Madison, WI.
• Stanford, G., and E. Epstein. 1974. Nitrogen mineralization-water relations in soils. Soil Sci. Soc. Am. Proc. 38:103–107.
• Statsoft. 1999. Statistica for Windows. Statsoft Inc. Tulsa, OK, USA.
• Tel, D.A., and C. Heseltine. 1990. The analysis of KCl extracts for nitrate, nitrite and ammonium using a TRAACS 800 analyzer. Commun. Soil Sci. Plant Anal. 21:1681–1688.
• Tollenaar, M. 1996. Corn production, utilization and environmental assessment: A review. Canada's Green Plan, Agriculture and Agri-Food Canada, Ottawa, ON.
• Tomer, M.D., and J.L. Anderson. 1995. Variation of soil water storage across a sand plain hillslope. Soil Sci. Soc. Am. J. 59:1091–1100.[Abstract/Free Full Text]
• Vachaud, G., A. Passerat de Silans, P. Balabanis, and M. Vauclin. 1985. Temporal stability of spatially measured soil water probability density functions. Soil Sci. Soc. Am. J. 49:822–828.[Abstract/Free Full Text]

This article has been cited by other articles:

				M. Sharifi, B. J. Zebarth, D. L. Burton, C. A. Grant, S. Bittman, C. F. Drury, B. G. McConkey, and N. Ziadi Response of Potentially Mineralizable Soil Nitrogen and Indices of Nitrogen Availability to Tillage System Soil Sci. Soc. Am. J., June 18, 2008; 72(4): 1124 - 1131. [Abstract] [Full Text] [PDF]