SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Dharmakeerthi, R. S.
Right arrow Articles by Beauchamp, E. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Dharmakeerthi, R. S.
Right arrow Articles by Beauchamp, E. G.
Agricola
Right arrow Articles by Dharmakeerthi, R. S.
Right arrow Articles by Beauchamp, E. G.
Related Collections
Right arrow Soil Organic Matter
Right arrow Spatial Variability
Right arrow Plant Nutrition
Right arrow Site-Specific Analysis
Right arrow Soil Fertility and Productivity
Published in Soil Sci. Soc. Am. J. 69:453-462 (2005).
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

Division S-4—Soil Fertility & Plant Nutrition

Factors Contributing to Changes in Plant Available Nitrogen across a Variable Landscape

R. S. Dharmakeerthia, B. D. Kayb,* and E. G. Beauchampb

a Dep. of Soils and Plant Nutrition, Rubber Research Institute of Sri Lanka, Dartonfield, Agalawatta, Sri Lanka
b Dep. of Land Resource Science, Univ. of Guelph, Guelph, ON, Canada, N1G 2W1

* Corresponding author (bkay{at}lrs.uoguelph.ca)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The maximum benefits that may be obtained from site-specific N management will not be realized until we improve our understanding of the spatial variability in plant available nitrogen (PAN) under different soil and crop management practices across the landscape. The objectives of this study were to quantify the influence of soil factors on the spatial and temporal variability of N availability to corn (Zea mays L.) grown under different tillage and legume cover crop conditions. Three management treatments were established on a variable landscape in southern Ontario in 1999, 2000, and 2001: barley (Hordeum vulgare L.) followed by corn under no-tillage, barley followed by corn under spring plowing and secondary tillage, and barley underseeded with red clover (Trifolium pratense L.) that was plowed down the next spring and followed by corn. The PAN contents were measured through the growing seasons and related to soil properties, management, and their interactions using regression analysis. The spatial patterns of PAN were temporally stable, suggesting a temporal consistency in the spatial patterns of factors influencing PAN. The variation in soil temperature among landscape positions was very small, often only 1 to 2°C, and its contribution to the spatial variation in PAN was considered negligible. Soil water made only a small contribution to the variability in PAN because the seasonal average water-filled pore space (SAWFPS) exhibited little variation across the landscape and often fell within the nonlimiting water range (NLWR) for N mineralization. Most of the variation in PAN within a season in this landscape was accounted for by variation in organic carbon (OC) content.

Abbreviations: CDD, cumulative degree days • k, mineralization rate constant • Leg, legume treatment • N0, potentially mineralizable N • Nm, cumulative plant available nitrogen • NLWR, nonlimiting water range • OC, organic carbon • PAN, plant available nitrogen • SAWC, seasonal average gravimetric water content • SAWFPS, seasonal average water-filled pore space • Till, tillage treatment • WC, gravimetric water content • WFPS, water-filled pore space


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
FERTILIZER-N that has not been taken up by the crop not only reduces the economic benefits of fertilization to the farmer, but also can pose a serious threat to the environment. The severity of these problems is spatially and temporally variable. A goal of site-specific or precision management of N is to minimize these problems by varying the application rates of fertilizer-N in response to the spatial variation in PAN in the soil. However, the maximum benefits that may be obtained from site-specific N management will not be realized until we improve our understanding of the causes of variability of PAN under different soil and crop management practices within and among growing seasons at the landscape scale.

Variation of PAN content has been recorded at different landscape positions (Jowkin and Schoenau, 1998), and this variation is mainly controlled by the spatial variation in net N mineralization (Qian and Schoenau, 1995). Net mineralization is influenced by organic matter content and the readily mineralizable N content of this organic matter, texture, water content, soil structure, temperature, pH, and the C/N ratio of added organic materials. Many of these soil properties (e.g., organic matter content, texture, and water contents) are known to vary across the landscape (Afyuni et al., 1993; Goovaerts and Chiang, 1993; Brubaker et al., 1994; Hook and Burke, 2000), whereas the spatial variability in other properties is less well understood. For instance, the variation in organic matter content will contribute to variation in the amount of mineralizable N, but will also have an important impact on soil structure (da Silva et al., 1997; Kay, 1998). Jarvis (1996) emphasized that to better understand spatial variation in N mineralization, more research is needed on the influence of soil structure on N mineralization. Information on the spatial variability of soil temperature is also rather sparse. Pierson and Wight (1991) observed a large spatial variation in near-surface soil temperature along a 12-m-long transect on a sagebrush (Artemisia tridentata tridentata) rangeland, but no reports were found on the magnitude of spatial variation in soil temperature at the landscape scale. While soil properties may vary spatially across the landscape, the impact of these properties on PAN may also vary temporally due to variation in water content and temperature. Therefore, the relative importance of soil factors that control the spatial variation of the accumulation of PAN during a growing season may vary with climatic conditions.

Soil temperature and water content vary within and between growing seasons. The amount of heat accumulated in soil at a given time, referred to as soil thermal or soil heat units and expressed as cumulative degree days (CDD) above 0°C, has been used in place of time to describe net N mineralization at different temperatures in laboratory-based incubation studies (Honeycutt et al., 1988, 1991; Doel et al., 1990). Measurements under field conditions lead Honeycutt and Potaro (1990) to conclude that soil thermal units provided a "mathematically simple, pragmatic approach for predicting crop N mineralization," although subsequent studies (Honeycutt, 1999) showed that N mineralized from soil organic matter under field conditions was overestimated by laboratory predictions. Although the net N mineralization could be predicted adequately with soil thermal units for different levels of temperature x moisture combinations at soil water potentials within the –0.03 to –0.01 MPa range, Doel et al. (1990) speculated that this relationship may not be valid for prolonged drier conditions. A further level of complexity arises when inherent soil properties exhibit spatial variability. Honeycutt et al. (1991) found that the relation between the amount of N mineralized and soil thermal units varied among soils in incubation studies. Further work is required to determine if relations between PAN and soil thermal units under field conditions could be used to help quantify the effect of soil properties on the spatial variability of PAN.

Plant available N also varies with management practices such as tillage and the use of leguminous crops. Although the response of N mineralization to different management practices has been considered by numerous researchers, little of this work has been done on variable landscapes. Characterization of spatial variability of N availability across the landscape, and quantification of factors influencing this variability is therefore essential to assess maximum benefits that could be obtained from different management practices on variable landscapes.

The objectives of this study were to quantify the influence of soil factors on the spatial and temporal variability of N availability to corn grown under different tillage and legume cover crop conditions. These objectives were achieved using a field experiment conducted on a variable landscape during a 3-yr period.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The experiment was established in a variable landscape at the Elora Research Station (43°39' N and 80°35' W), which is about 23 km northwest of Guelph, Ontario. The soil was a Typic Hapludalf with a silt loam texture in the A horizon. Plots (4.5 by 8 m) were established at five positions across the landscape (arbitrarily referred to as the summit, shoulder, backslope, footslope, and toeslope positions). The elevation of the landscape dropped by about 10 m along a 160-m traverse from the summit to the toeslope position (Fig. 1). The shoulder, backslope, and footslope of the landscape faced north. Descriptive statistics of selected soil characteristics on the plots for the five positions in the landscape are given in Table 1. The clay content was relatively uniform across the site, whereas the OC content exhibited much larger variation. Measurements were made in the 1999, 2000, and 2001 growing seasons.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 1. Digital elevation map indicating the experimental sites in the 1999 and 2001 seasons (site A) and the 2000 season (site B). The five rows of dots across the slope in a given site indicate the landscape positions (from top to bottom: summit, shoulder, backslope, footslope, and toeslope).

 

View this table:
[in this window]
[in a new window]
  		Table 1. Descriptive statistics of some soil properties measured on the experimental plots during the three growing seasons.

 
Experimental Design and Treatments
Three management systems were used in this experiment: barley followed by corn under no-tillage (Barley-NT); barley followed by corn under a spring plowing and secondary tillage operation (Barley-CT); and barley underseeded with a red clover cover crop that was plowed down in the following spring and plots planted to corn (Barley+Red Clover-CT). In all seasons, barley was grown as the preceding crop after spring cultivation. The corn variety was Northrup King N17-C4. At the time of planting, 200 kg 0–20–20 N–P–K ha–1 was banded beside the corn seed to meet the P and K requirements. Corn was planted along the slope and experimental plots were established at the five positions in the landscape. Each plot had six rows of which the four center rows were used for plant sampling while the remaining two rows were used as border rows. The planting density was approximately 60000 plants ha–1, with a row spacing of 0.75 m. Corn received either 0 or 140 kg N ha–1 in the form of NH4NO3 around the fifth visible leaf tip stage (V2 stage). Data from the zero nitrogen (0N) plots under corn were used for analyses in this paper.

Altogether, there were three crop management systems (main plot), five landscape positions (subplot), and two fertilizer N levels (sub-subplot). The experimental design was a modified split-split plot design with three replicates. The experiment was conducted in two areas (sites A and B) next to each other on the same landscape (Fig. 1) in alternate years because of lack of space. Thus, corn was grown in the 1999 and 2001 seasons in one area, whereas in the 2000 season it was grown in the other.

Measurements
Soil samples were collected from the 0- to 30-cm depth for the determination of soil mineral N content and gravimetric water content (WC). Sampling was conducted once every 2 wk throughout the growing season from early May to late September or early October. There were a total of 10 or 11 sampling dates in each year. Sampling was done in the center of the third interrow of the experimental plots. A minimum of 10 cores were taken per plot at each sampling date using a standard 25-mm-diam. soil probe. Soil samples were frozen until mineral N extraction occurred. After thawing, the composite sample was mixed and sieved through a 4-mm sieve. A subsample of 5.0 g was added to 25 mL of 2 M KCl and shaken on a rotary shaker for 1 h. Concentrations of NH4–N and NO3–N in the filtered extract were determined using a colorimetric technique (Keeny and Nelson, 1982) using a Bran and Luebbe TRAACS 800 autoanalyzer (Alfa-Laval, AB, Stockholm, Sweden). The WC was determined in fresh soil samples before processing. The water contents were averaged across all sampling times to give a seasonal average water content for each landscape position, in each treatment, and for each rep.

Soils collected on the second sampling (around late May) were used for the determination of OC and particle size distribution. Organic carbon content was determined in soil samples (air dried and ground to pass through a 0.5-mm sieve) using a Leco Carbon Analyzer (LECO SC444, Leco Corporation, St. Joseph, MI). Particle size distribution was measured using the hydrometer method in 1999 and the pipette method in 2000 as described by Gee and Bauder (1986).

Dry bulk density of each plot was estimated in early October (when the corn was at postphysiological maturity) using 4.85-cm-diam. by 2.60-cm-high aluminum cores at depths of 7.5 to 10.1 and 22.5 to 25.1 cm. Four cores for each depth were collected in each plot. The average dry bulk density of the 0- to 30-cm depth of a given plot was calculated from these measurements.

In the 2000 season, soil temperature was measured using thermistors (Model TMC6-HB, Onset Computer Corporation, Bourne, MA) connected to HOBO H8 Outdoor/Industrial 4-channel External dataloggers (Onset Computer Corporation) from early July until early October. Although the thermistors were designed for outdoor use, they were further weatherproofed before field installation. Each thermistor was shielded in a 38-mm-long piece of copper tubing (8-mm i.d.), which was filled with epoxy (2216 B/A Gray epoxy adhesive, 3M Corporation, St. Paul, MN). The thermistors were calibrated in an ice slurry bath. The weatherproofed thermistors were installed at 5- and 20-cm depths in each plot, and were connected to dataloggers. There were two thermistors at each depth, about 3 m apart from each other. Because of limited availability of thermistors and data loggers, soil temperature was monitored only in one management treatment (Barley–CT) and three landscape positions (summit, backslope, and toeslope). Preliminary studies, however, revealed that the maximum variation in soil temperature occurred between the summit and backslope positions. Because of some problems in the data loggers and rodent attacks, data were not available during some periods of the growing season.

Corn plants were harvested (tops only) once every 2 wk, beginning about 4 wk after planting at the time of soil sampling. An area equivalent to 1 m by 2 rows was sampled in the first sampling, and thereafter a 1-m by 1-row area was harvested for dry matter determinations. These samples were dried (70°C), weighed, subsampled, and ground to pass through a 1-mm sieve before determining the N concentration using a Leco N analyzer (LECO FP-428, Leco Corporation).

At any given time, PAN that is not taken up by the plant should remain in the soil if the losses from the system are negligible. Hence, the total amount of PAN was approximated by summing mineral N contained in the soil (0–30 cm) and in the plant (tops). However, it was also recognized that this estimate of available N did not include root N.

Weather data (daily rainfall and minimum, maximum, and average daily air temperature) were obtained from the weather station situated on the Elora Research Station, about 1 km away from the experimental site. The average daily temperatures were used to calculate the CDD above 0°C. The accumulation of degree days was begun in the early spring when the average daily air temperature was above 0°C for three consecutive days.

Statistical Analysis
Analyses of variance were conducted using PROC GLM and nonlinear regression analyses were conducted using PROC NLIN program of the SAS software package (SAS Institute, 1996). Multiple regression analyses were conducted using the stepwise regression procedure in the S-PLUS software package (Insightful Corporation, 2002). In any model-building process using regression analysis, it is essential to detect outliers and highly influential data points because the nature of the relationship could be masked by these points (Myers, 1990). An outlier was defined statistically as data values > 1.5 interquartile ranges from the median, and were identified using box plots produced by the S-PLUS program. Influential data points were identified using the partial regression plots generated by the PROC REG procedure of SAS. The detected outliers and influential data points were removed from the analysis. During the stepwise regression analysis, higher-order interaction effects that were not significant at P < 0.05 were removed from the list of independent variables and the stepwise procedure rerun. When an interaction effect was significant, its main effects were kept in the final model, even if they were not statistically significant.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Weather
Rainfall distribution varied markedly among the three growing seasons (Fig. 2). The amounts of rainfall received from 1 May to 30 September were 471, 538, and 394 mm for years 1999, 2000, and 2001, respectively. In 2000, the amounts of rainfall received during May, June, and July were considerably higher than the 30-yr (1971–2000) average. In the 2001 season, an extended dry period was observed during the month of July and the early part of August.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 2. Distribution of rainfall during the growing period in the 1999, 2000, and 2001 seasons. The rainfall data exceeded the y-axis maximum on 6 Sept. 1999 (70.0 mm) and on 19 Aug. 2001 (63.8 mm).

 
There was a large variation among years in CDD by the end of the growing season. The CDD by mid-June (around the 5–6 visible leaf tip stage) were 665, 570, and 489, and at the end of September (around the physiological maturity of corn) were 2614, 2348, and 2416 for the years 1999, 2000, and 2001, respectively. Therefore, the early stage of the 2001 growing season was colder than the other 2 yr, and the 1999 season was the warmest as a whole.

Spatial Variation in Plant Available Nitrogen
For simplicity and also to reduce the sampling variability, the PAN data were studied at three growth stages of corn: planting, silking, and physiological maturity (i.e., maximum kernel dry weight). The 1st and 2nd samplings were done before and after planting, and therefore an average of these samplings provided a measure of PAN at the beginning of the growing season. Plants reached 50% silking between the 6th and 7th samplings in all seasons, and therefore the average of 6th and 7th samplings gave an indication of the amount of N available for plant uptake during this critical growth stage. The averages of the last two samplings indicated the maximum amount of N that was available for plant uptake during a given growing season. These values were subjected to ANOVA using PROC GLM of SAS, treating any variable that contained a block effect as a random variable.

An ANOVA indicated that management and landscape position had significant effects on PAN at most of the stages in the different years, but there was not a significant management x landscape position effect at any stage in any year (Table 2). The magnitude of the differences in PAN among landscape positions increased as the growing season progressed (Table 2). The lowest PAN was recorded in the backslope position, whereas the highest PAN was often recorded at either the footslope or toeslope positions. The magnitude of the landscape position effect also varied from one season to the other. At physiological maturity of corn, the differences in mean PAN among landscape positions varied from 57 kg N ha–1 in 2000 to 123 kg N ha–1 in 1999. The year 2000 was very wet, especially during the early periods of the season (214 mm of rainfall was received between the first and the second samplings), and a considerable proportion of PAN was lost from the rooting zone, presumably due to leaching and denitrification. This loss of N from the rooting zone could have been the reason for the smallest variation in PAN among landscape positions observed in the year 2000.


View this table:
[in this window]
[in a new window]
  		Table 2. Variation in plant available N among landscape positions in the 0N treatment at three selected times during the 1999, 2000, and 2001 growing seasons (averaged across management treatments).

 
The data in Table 2 suggest that the spatial pattern in PAN was stable across time. Temporal stability within and among growing seasons was assessed by correlating measurements of PAN at the different positions in the landscape at one stage with corresponding values at a different stage using the approach employed by Kachanoski and de Jong (1988) to assess the temporal stability in spatial patterns in soil water content. The analyses using the averaged data for the three management treatments (Table 3) showed that the correlations were significant, indicating that the spatial patterns were temporally stable both within and among growing seasons. However, the correlation of PAN data of the year 2000 with those in the years 1999 and 2001 were often not significant. This could be because of the fact that 1999 and 2001 plots were on the same site, whereas in 2000, the plots were on an adjoining site. Repeating the analyses for each management treatment indicated that the poorest correlations were obtained for the Barley+Red Clover-CT treatment (data not shown). The legume made a significant contribution to PAN, and this contribution apparently resulted in a less-consistent spatial pattern in PAN than the other treatments. Temporal stability in the spatial patterns in PAN suggests a temporal consistency in the spatial patterns of factors influencing PAN.


View this table:
[in this window]
[in a new window]
  		Table 3. Assessment of temporal stability of plant available N in the 0N treatment (averaged across the three management treatments) as determined from correlation among values of plant available N at three stages of growing seasons (at planting, silking, and physiological maturity of corn; see the text for details). Values within parentheses indicate the P values.

 
Factors Controlling the Spatial Variability in Plant Available Nitrogen
Soil Temperature
The temporal and spatial variation in mean daily soil temperature is illustrated for periods at the beginning of July, August, September, and October 2000 (Fig. 3). Because of missing data, a statistical analysis could only be done for the data in the first week of August. The mean daily air temperature at the 5-cm depth was more responsive to air temperature than at 10 cm, as would be expected. For instance, when the mean air temperature was high (about 17°C or more), the mean soil temperature was greater at the 5-cm depth than at the 20-cm depth, and the reverse was true when the mean air temperature was low (about 10°C or less) (Fig. 3). The differences among landscape positions were also greater at higher mean air temperatures. At both depths, the lowest soil temperature was observed at the summit, whereas the highest was at the backslope position (except toward the end of the season, where the soil temperature at the 5-cm depth at the toeslope position was greater than that of backslope position). Temperatures on the backslope position responded more quickly to changes in air temperature due to smaller soil water contents and reduced plant cover. Differences in mean sun angle at the different landscape positions may also have influenced soil temperature. Although there were significant differences among landscape positions, as indicated by the statistical analysis conducted for the data of the first week of August, the variation in soil temperature across the landscape appeared to be sufficiently small to warrant ignoring it in the subsequent analysis. The magnitude of the differences were often only 1 to 2°C at the early part of the season, and <1°C toward the end of the growing season. Moreover, as noted above, the highest temperature was often recorded at the position where PAN was lowest. Higher temperatures at this position could only contribute to reduced variability of PAN among different positions in this landscape.



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 3. Variation in average daily soil temperatures (°C) at 5- and 20-cm depths in three contrasting landscape positions during first week of July, August, September, and October 2000.

 
Soil and Management Factors at Different Times through the Growing Season
Stepwise regression analyses were conducted relating PAN (at planting, silking, and physiological maturity) to management (tillage and legume), OC contents, SAWFPS, and clay content (1999 and 2000 only) and their two-way interactions. The water-filled pore space (WFPS), that is, the percentage of total volume of pores that are water-filled, was selected to represent the effect of moisture because it incorporates not only the effects of water content but also the effects of compaction (Linn and Doran, 1984). Use of WFPS rather than water content also reduced multicolinearity during the regression model building because the seasonal average water content (SAWC) was very strongly related to the OC content in all 3 yr. Further, because the spatial pattern of water content is temporally stable through the growing season (da Silva et al., 2001), the spatial pattern of WFPS was assumed also to be temporally stable. Therefore, the SAWFPS was calculated from the SAWC and bulk density to represent the spatial variability in soil water content. The results of the stepwise multiple linear regression analyses are given in Table 4.


View this table:
[in this window]
[in a new window]
  		Table 4. Results of the stepwise regression analyses relating the effect of tillage (Till), legume (Leg), organic carbon (OC, g kg–1), and the seasonal average water-filled pore space (SAWFPS, %) to the amount of accumulated plant available N (kg N ha–1) in the 0N treatment at different times of the growing season.

 
The spatial variation of PAN was strongly and positively related to OC content in all three seasons. The change in PAN per unit change in OC across the landscape increased as the growing season progressed until silking. The positive effect of OC on the amount of PAN may be related to the positive relationship between the OC and the potentially mineralizable N (N0) content (Barrett and Burke, 2000; Deng and Tabatabai, 2000). The PAN was linearly related to OC at the beginning of the growing season and at silking, whereas at maturity, the OC made a parabolic contribution to PAN. The change in PAN at maturity at the average value of OC across the site (19.5 g OC kg–1 soil) was 7.9, 2.1, and 7.7 kg N ha–1 g–1 OC kg–1 soil for 1999, 2000, and 2001, respectively, or a mean of 58.9 kg N ha–1 per 1% OC. Although additional data are required to get a better estimate of the mean, it is comparable with the rule of thumb of 58.6 kg N ha–1 that would mineralize in a 5-mo growing season in Colorado for every 1% OC that was published by Soltanpour (1979), as cited by Vigil et al. (2002).

The remaining terms in the regression models were Leg (legume treatment), Till (tillage treatment), and SAWFPS. The red clover cover crop that was plowed down before planting the corn made a positive contribution to PAN, and the Leg term appeared in all of the regression models except at planting in 1999. At maturity in 2001, the contribution of the legume was diminished with increasing OC. Tillage had a significant effect (positive) on PAN in only four of the nine models. The SAWFPS term only appeared in the regression models for silking and maturity in 2000, and in both cases the magnitude of the effect was influenced by management. It is of interest to note that significant management x landscape position interaction effects were not observed (Table 2), further emphasizing the importance of soil properties in describing the variability of N availability across a landscape.

Some of the variables in the regression models given in Table 4 relate to the amount of N0, whereas other variables may influence the rate of mineralization. The simple linear regression analysis conducted at specific times does not provide an opportunity to separate these effects. Therefore, another analysis was conducted using the data from the whole season to describe the pattern of PAN accumulation and to identify the factors controlling PAN accumulation for the entire growing season.

Integral Effect of Soil and Management Factors through the Growing Season
Thermal time, expressed as CDD, was used to describe the accumulation of PAN through the growing season. Although Honeycutt et al. (1988)( 1991) used soil temperature data to calculate soil thermal units, analyses of soil temperature at the weather station at the Elora site (5-cm depth under a sod cover) and air temperature revealed a very high correlation (R2 = 0.997 and P < 0.0001) between the soil thermal units and CDD through the entire growing season. Further, as discussed earlier, the variation in soil temperature across the landscape positions was minor. Considering this and the availability of data, air temperature data were used for the calculation of CDD in this analysis. The accumulation of PAN for three contrasting landscape positions averaged across management treatments is illustrated in Fig. 4.



View larger version (31K):
[in this window]
[in a new window]
  		Fig. 4. Pattern of plant available N accumulation throughout the 1999, 2000, and 2001 growing seasons in three contrasting landscape positions. The vertical bars indicate the standard error.

 
Four different models were evaluated to identify the pattern of the accumulation of PAN in each plot during the growing season. They were

[1]

[2]

[3]

[4]
where Nm is the accumulated PAN at a given time, N0 is potentially PAN, k is a constant related to the rate of accumulation, and a, b, and c are empirical constants. Stanford and Smith (1972) were among the first to show that the accumulation of inorganic N in an incubation study followed a first order kinetic function, with time expressed in days. Griffin and Honeycutt (2000) also observed that NO–3 accumulation in soils in which manure had been incorporated could be predicted by a first order kinetic model using thermal units based on soil temperature. Simard and N'dayegamiye (1993) studied the pattern of N accumulation of meadow soils in Quebec incubating those soils at a constant temperature under laboratory conditions and observed that cumulative N mineralization curves were best described by a Gompertz equation. The selection of best model was based on the mean square error (MSE) and R2 values.

The PAN accumulation in 25% of the plots was adequately described by a linear relationship, but the vast majority of the data had a nonlinear relationship. In cases where the Nm was described by a nonlinear relationship, 33, 36, and 6% of the plots had Nm described by the first order kinetic, Logistic, and Gompertz functions, respectively. Further, the gain in R2 by the Gompertz and Logistic functions over the first order kinetic model was often <0.01. Therefore, the first order kinetic equation was selected to describe the accumulation of PAN during the growing season. Preliminary tests were also conducted with an additional constant added to Eq. [2] to account for residual N remaining in the soil from the preceding season. However, this term either prevented convergence to a solution, or where a solution was obtained, the term was often not significant. Consequently, this term was not included in the function.

The data for each of the 3 yr were fitted to the first order kinetic model, with N0 and k being replaced by soil properties and management practices. The soil properties included in the analysis were OC, clay, and SAWFPS. Bulk density and SAWC were not included because of their strong correlation with OC. The effects of management practices, tillage, and legume effects were expressed as class variables using 0 (no-tillage or no legume) and 1 (tilled or with legume). Nonlinear regression analysis was conducted by adding one variable at a time until there was no further reduction in mean square error (MSE) and all variables in the model were significant at P < 0.05 (a procedure similar to forward stepwise regression). The final models for each of the years are given in Table 5.


View this table:
[in this window]
[in a new window]
  		Table 5. Results of the stepwise regression analyses describing the accumulation of plant available N (kg N ha–1) as a function of cumulative degree days (CDD) using the first order kinetic model {Nm = No[1 – exp(–k x CDD)]} with coefficients related to tillage (Till), legume (Leg), organic carbon (OC, g kg–1), and the seasonal average water-filled pore space (SAWFPS, %) in the 0N treatment.

 
Without any soil property or management effect included in the model, that is, only N0 and k included, CDD accounted for 50% of the variation in PAN across time. When days of year were used instead of CDD, the model failed to converge in all 3 yr.

Variables that describe N0 and k could be assumed to be factors responsible for the variation in potentially available N for mineralization and the rate at which it is mineralized, respectively, within a growing season. More soil and management factors were included in the description of N0 than k (Table 5).

From the spatially variable soil properties, quadratic terms involving OC content described most of the variability in the Nm (8%), and only appeared in the description of N0. This analysis supported the argument that the N0 content increases as the organic matter content increases across the landscape (Qian and Schoenau, 1995). The quadratic terms involving OC mean that the increase of PAN per unit increase in OC decreased as OC increased. Because a significant decrease (P < 0.01) in the C/N ratio with the increase in OC content was observed, the quadratic relationship could be due to the decrease in the proportion of N that was potentially mineralizable as the OC content increased. It could be argued that in areas where the OC content is low, the proportion of more recent OM is higher than that in areas where the OC contents are high. In this landscape, the OC was low in backslope and shoulder positions where soil had been lost due to erosion. Although the SAWFPS and clay content did affect N accumulation, their contribution to the variation in PAN was marginal (often <1%).

Of the two management factors, the legume factor increased the N0 each year. About 15% in the variation in PAN was due to the presence of the legume in the Red Clover+Barley-CT treatment. However, tillage only influenced the potentially available N in 2000.

None of the spatially variable soil properties affected k, except SAWFPS in 2000, and the magnitude of this effect was small. Thus, the rate of N mineralization was constant across the landscape in a given year. This initially seems inconsistent with observations that the rate of N mineralization varies with WC (Campbell et al., 1984) or with WFPS (Rasiah and Kay, 1998). However, the apparent inconsistency may not be real for two reasons. First, although a significant variation in both SAWC and bulk density was observed across the landscape, the SAWFPS exhibited less variation and was constant across the range in OC contents found in this landscape (Fig. 5). The SAWC was strongly positively related to OC content, whereas the bulk density was negatively related to the OC content. These observations correspond to trends found on other sites (Kay et al., 1997; da Silva et al., 2001). Second, the SAWFPS was close to the optimum water content for these soils and generally fell within NLWR.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 5. Relationship between the seasonal average water-filled pore space (SAWFPS) and organic carbon content ({diamondsuit}) with their regression equations in the 1999, 2000, and 2001 seasons. This also illustrates the distribution of SAWFPS with respect to the lower ({square}) and upper ({circ}) limits of the nonlimiting water range (NLWR) (Drury et al., 2003) in the 1999 and 2000 seasons. Because the textural data for the 2001 season were not available, the NLWR could not be calculated.

 
Drury et al. (2003) observed that there is a range in water content within which the net N mineralization is maximum and exhibits minimal variation with water content. They defined this as the NLWR. They also observed that the lower limit of the NLWR, when expressed as WFPS, was related to the clay content, compaction, and presence of legume, whereas the upper limit was related to the clay content, compaction, and total N content. When the pedotransfer functions generated in their study were applied to the soils on this landscape, the NLWR of WFPS for net N mineralization was calculated to have an average of 21%. The SAWFPS fell within the calculated NLWR in almost all plots in the 1999 season and most of the 2000 season (Fig. 5), indicating that the water content was optimum, on the average, during this growing season. These observations explain why the spatial variation in SAWFPS seldom had a significant impact on the variation in PAN accumulation across this landscape.

Although the regression models describing the accumulation of PAN through the growing season have R2 ≥ 0.71 in each year (Table 5), the magnitude of the calculated values of N0 and k exhibited considerable variation between years. This variability must be because of differences in weather between seasons that are not captured in the CDD or SAWFPS terms. The values of N0 and k for 2000 were most different from the other 2 yr and may reflect large losses in N due to leaching or denitrification in 2000, when rainfall was much greater than the 30-yr average in May, June, and July. Further information on the spatial variability in these losses is required, especially in the early part of the growing season in cool humid temperate environments where such losses may have the greatest impact on fertilizer requirements.

The variability in PAN with OC among years is small relative to the variation in PAN with legume, as indicated by the coefficients in the regression equations in Tables 4 and 5. Part of the large variation in the regression coefficients related to management effects may be related to the fact that some terms (e.g., tillage) do not appear in the equations for all years. However, it is speculated that the variability in the coefficients in the regression equations may, at least in part, be due to weather x tillage and weather x legume interaction terms. If this were the case, weather has a more important role than commonly acknowledged in site-specific management studies.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The spatial and temporal variation in N availability to corn across a variable landscape and the influence of soil factors on this variation was studied when corn was grown under different tillage and cover crop conditions. Prediction of spatial variation of PAN during the growing period is possible only if the spatial patterns of the factors influencing the PAN are temporally stable. The spatial patterns of PAN in this study were found to be stable across time, suggesting a temporal consistency in the spatial patterns of factors influencing PAN. The dominant factor which controlled the spatial variability of PAN was soil organic matter content, and this effect appeared to be due to the variability of N0 content. The quadratic dependence of N0 on OC content suggests that the proportion of N0 content in the organic matter decreased as the organic matter content increased. Although a significant spatial variation in soil temperature was observed, the magnitude of this variation was very small. However, the magnitude of the spatial variation in soil temperature and its influence on spatial variation of PAN may be different on landscapes that are oriented differently. The seasonal average WFPS was constant across the range of organic matter contents observed in this landscape. Because of this observation and the large NLWRs for net N mineralization, the impact of water content on the spatial variation of PAN was minimal across the range of organic matter contents observed in this landscape. Therefore, in landscapes where the textural variation is marginal, spatial variation in water content may not have a large impact on the spatial variation in PAN. The amount of PAN at a given growth stage and the rate at which N accumulated between growth stages was heavily dependent on weather. Although the concept of CDD or thermal time could be used to reduce some of this seasonal variation, quantification of the effect of rainfall on leaching and/or denitrification losses is essential in any effort to predict changes in PAN during a growing season.


   ACKNOWLEDGMENTS
 
The Ontario Corn Producers' Association, Agriculture and Agri-Food Canada, the Natural Sciences and Engineering Research Council of Canada, and the Ontario Ministry of Agriculture and Food provided financial support for this research. The Canadian Commonwealth Scholarship and Fellowship Program provided financial support for R.S. Dharmakeerthi to pursue studies leading to the Ph.D. degree at the University of Guelph. The assistance of Jim Ferguson with fieldwork is also gratefully acknowledged.

Received for publication June 29, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
B. D. Kay, A. A. Mahboubi, E. G. Beauchamp, and R. S. Dharmakeerthi
Integrating Soil and Weather Data to Describe Variability in Plant Available Nitrogen
Soil Sci. Soc. Am. J., May 23, 2006; 70(4): 1210 - 1221.
[Abstract] [Full Text] [PDF]

Home page
 Agron. J.Home page
R. S. Dharmakeerthi, B. D. Kay, and E. G. Beauchamp
Spatial Variability of In-Season Nitrogen Uptake by Corn Across a Variable Landscape as Affected by Management
Agron. J., February 7, 2006; 98(2): 255 - 264.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Dharmakeerthi, R. S.
Right arrow Articles by Beauchamp, E. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Dharmakeerthi, R. S.
Right arrow Articles by Beauchamp, E. G.
Agricola
Right arrow Articles by Dharmakeerthi, R. S.
Right arrow Articles by Beauchamp, E. G.
Related Collections
Right arrow Soil Organic Matter
Right arrow Spatial Variability
Right arrow Plant Nutrition
Right arrow Site-Specific Analysis
Right arrow Soil Fertility and Productivity

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome