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


     

Published online 27 April 2005
Published in Agron J 97:772-782 (2005)
DOI: 10.2134/agronj2004.0287
© 2005 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
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 ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shahandeh, H.
Right arrow Articles by Lascano, R. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Shahandeh, H.
Right arrow Articles by Lascano, R. J.
Agricola
Right arrow Articles by Shahandeh, H.
Right arrow Articles by Lascano, R. J.
Related Collections
Right arrow Nitrogen
Right arrow Site-Specific Analysis

Spatial Variability

Spatial and Temporal Variation of Soil Nitrogen Parameters Related to Soil Texture and Corn Yield

H. Shahandeha,*, A. L. Wrighta, F. M. Honsa and R. J. Lascanob

a Dep. of Soil and Crop Sci., Texas A&M Univ., College Station, TX 77843-2474
b Texas A&M Univ. Agric. Res. and Ext. Cent., Rt. 3, Box 219, Lubbock, TX 79403-9803

* Corresponding author (h-shahandeh{at}tamu.edu)

Received for publication November 23, 2004.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The spatial variability of soil properties that affect the soil N budget and corn (Zea mays L.) grain yield were studied for 2 yr in south-central Texas to better assess the potential for variable-rate N fertilization. Residual soil NO3–N with depth and soil N mineralization (Nmin) potential and their relationships with soil total N, soil organic C, and clay content were characterized. Residual soil NO3–N to 60-cm depth was more related to corn yield than NO3–N at shallower depths. Residual soil NO3–N showed temporal variation with spatial structure existing for NO3–N in the first year when NO3–N concentrations were high, but this variation was absent in the second year when NO3–N concentrations were low. The opposite trend was observed for soil Nmin. Soil total N exhibited temporal persistence. Soil Nmin showed significant correlations with soil clay content in both years. Temporal persistence of soil texture and variation in spatial structure for N parameters would likely result in different strategies for soil N management zones in 2002 vs. 2003. This experiment demonstrated the potential importance of soil texture for modifying fertilizer N recommendations. Texture generally is more easily determined than soil N parameters at the scale and intensity necessary for site-specific N management.

Abbreviations: CV, coefficient of variation • Nmin, nitrogen mineralization • SOC, soil organic carbon • SOM, soil organic matter


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
NITROGEN FERTILIZATION has historically been an important crop input from production, economic, and environmental standpoints but currently is receiving more scrutiny because of the increased cost of N fertilizer, concerns about environmental degradation, and demands for mandatory nutrient planning (Dinnes et al., 2002; Raun et al., 2002; Schmidt et al., 2002). The current recommended method for determining appropriate amounts of N fertilization for crop production involves soil testing for residual surface soil (0 to 15 cm) nitrate (NO3) concentration and integrating with anticipated crop yield goal (Oberle and Keeney, 1990; Vanotti and Bundy, 1994; Stafford et al., 1996). The prescribed quantity of N is then normally applied uniformly across a field. However, the spatial variation in NO3–N accumulated below 15-cm depth and soil N mineralized from soil organic matter (SOM) during the growing season must be assessed to determine their contribution to crop yield variability and to evaluate their necessity for variable-rate N fertilization (Mahmoudjafari et al., 1997; Schmidt et al., 2002; Eghball et al., 2003; Katsvario et al., 2003).

Variable-rate fertilization that considers differences in residual nutrient availability should be more economically and environmentally sustainable than uniform-rate application (Machado et al., 2000; Dinnes et al., 2002; Lopez-Granados et al., 2002; Crowin et al., 2003). To implement variable-rate fertilization or site-specific nutrient management practices, databases that provide spatial distributions of soil properties are needed (Cambardella and Karlen, 1999). However, the spatial variability of residual soil N is highly variable and represents potential problems in estimating the quantity of N available in the field (Parkin et al., 1988; Cabrera and Kissel, 1988).

The value of residual soil NO3 as a guide for N fertilization in precision farming using geostatistical methods has been evaluated in long-term fallow and for corn, winter wheat (Triticum aestivum L.), corn–soybean [Glycine max (L.) Merr.] rotation, and wheat–sunflower (Helianthus annus L.) rotation (Goovaerts and Chiang, 1993; Cambardella et al., 1994; Mahmoudjafari et al., 1997; Cambardella and Karlen, 1999; Lopez-Granados et al., 2002). Geostatistical techniques have been used to explore the structure of spatial variation in agricultural soils by mapping areas of environmental concern and developing site-specific fertilizer application maps (Cambardella and Karlen, 1999; Lopez-Granados et al., 2002). Lopez-Granados et al. (2002), for example, found that subsoil NO3–N concentrations had a strong spatial dependence. Goovaerts and Chiang (1993) determined that the spatial structure of mineralizable soil N was explained by the amount of soil oxidizable C. Cambardella et al. (1994) found that soil total N was spatially dependent, whereas NO3–N and mineralized N were not. However, Bruckler et al. (1997) found that soil NO3–N was both spatially and temporally dependent, and Robertson et al. (1988) found that soil Nmin had a high degree of spatial dependence.

Variable-rate fertilizer N recommendations, however, likely should not be based only on a single criterion derived from statistical models. For example, Schmidt et al. (2002) argued that variable-rate N recommendations based only on SOM were too simplistic to reflect variability in N availability within a field. Similarly, Mahmoudjafari et al. (1997) indicated that the mineralization potential of SOM was difficult to estimate because Nmin is a function of both soil temperature and water content, which can spatially vary independently of SOM content. Furthermore, Eghball et al. (2003) showed that spatial variability of corn grain yield was much lower than for soil NO3–N, indicating the possible ineffectiveness of using spatial distribution of soil NO3–N as the sole criterion for variable-rate N application.

There are other approaches that do not use a single criterion for N recommendations to delineate soil management zones. These approaches are based on the spatial structure of N parameters and related factors, including soil characteristics and cultural practices (Oberle and Keeney, 1990; Sadler et al., 1998; Bakhsh et al., 2000a, 2000b; Frogbrook et al., 2002). For example, soil properties such as clay content with only limited sampling of NO3 might be used to estimate and infer NO3 values. Oberle and Keeney (1990) and Sadler et al. (1998) in their assessments of N fertilizer requirement indicated that soil series and soil mapping information provided a basis for N management evaluation. Tabor et al. (1985) found that leaf petiole NO3–N and soil clay content had high correlation and similar spatial structure, indicating that spatial structure and optimum sampling for petiole NO3–N could be inferred from the spatial structure of clay for a particular soil mapping unit. Machado et al. (2000) also suggested that variable-rate fertilizer N application should be based on texture and soil NO3. Stafford et al. (1996) found a high correlation between soil series and crop yield, and Cox et al. (2003) found that among the soil factors affecting crop yield variability, clay content could be used as a basis for site-specific management.

The objectives of this study were to determine the temporal and spatial dependence of soil N parameters, such as total N, Nmin potential, and residual soil NO3–N with depth, and their relationships with soil texture and corn yield. Strong correlations of soil N parameters with more easily measured soil properties, such as texture, may provide an opportunity to use these parameters to estimate and infer N values for site-specific N management.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Experimental Site
Research was conducted at the Texas A&M University Agricultural Research Farm in Burleson County near College Station, TX (30°32' N, 96°25' W), in 2002 and 2003. The alluvial soil used for the study is an intergrade of Weswood silt loam (fine-silty, mixed superactive, thermic Udifluventic Haplustepts) and Ships clay (very fine, mixed, active, thermic Chromic Hapluderts) with pH of 7.9 to 8.1. The experimental field was {approx}2.5 ha, previously cropped to cotton (Gossypium hirsutum L.), and was planted with corn in this experiment. The cotton received 125 kg N ha–1 as 32–0–0 applied preplant. Sixty-three grids (plots) were superimposed on the field, with each grid being eight 1-m rows wide and 12.2 m long. The experimental layout was nine grids long extending consecutively from west to east and seven grids wide extending from north to south. Eight unsampled rows were situated between sampling grids and ran the entire length of the field.

Soil and Plant Measurements
Four soil cores (5-cm i.d.) were collected near the center of each grid using a tractor-mounted hydraulic soil sampler on 18 and 25 April in 2002 and 2003, respectively. One core was taken directly over each grid center, and three additional cores were taken at 1-m radii from the reference point. Cores for each reference point were sectioned into depths of 0 to 15, 15 to 30, 30 to 60, and 60 to 90 cm and composited with depth. Samples were dried in a forced-draft oven at 50°C and then ground with a flail-type soil grinder (Custom Lab, Orange City, FL). The field used during this study received no N fertilizer so that soil Nmin potential and residual NO3–N could be more easily correlated with crop yield.

Corn variety Dekalb 687 (Monsanto, St. Louis, MO) was planted on 21 Feb. 2002 and 28 Feb. 2003 with a Case/IH Early Riser planter (Racine, WI) at a rate of {approx}65000 seed ha–1. Bicep herbicide {metolachlor [2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-methoxy-1-methylethyl)acetamide]/atrazine [6-chloro-N-ethyl-N'-(1-methylethyl)-1,3,5-triazine-2,4-diamine]} was used for weed control, along with in-season cultivation.

Soil Nmin potential, C mineralization, soil organic C (SOC), soil total N, and soil texture were determined on 0- to 15-cm soil samples. Residual NO3–N and other plant essential nutrients were determined on soil samples from all depths. Soil C mineralization and Nmin were determined according to Franzluebbers et al. (1994a)( 1994b). Approximately 20 g of oven-dried soil were placed in 50-mL beakers, wetted to –30 J kg–1, and incubated at 25°C in airtight containers along with a vial containing 10 mL of 1.0 M KOH and another vial containing water to maintain humidity. Vials of KOH were replaced at 1, 10, and 24 d. Mineralized C as CO2 was determined at each sampling date by titrating KOH with 1.0 M HCl to the phenolphthalein endpoint (Anderson, 1982). Ammonium and NO3–N at 0 and 24 d were extracted with 2 M KCl and determined using autoanalyzer techniques (Technicon Industrial Systems, 1977a, 1977b). Initial soil mineral NH4–N and NO3–N were subtracted from that measured at 24 d to determine net soil Nmin.

Soil organic C was determined using the modified Mebius method (Nelson and Sommers, 1982) while soil total N was determined by autoanalyzer techniques (Technicon Industrial Systems, 1977a) following Kjeldahl digestion (Nelson and Sommers, 1980). Soil particle size distribution was determined on all samples using the procedure of Day (1965), which utilizes hydrometer analysis following dispersion of soil by both chemical and physical means.

Three meters of each of the middle two rows of each grid were hand-harvested for grain yield at maturity and threshed using a stationary plot thresher. Grain moisture was determined by electrical resistance, and yields were calculated at a moisture content of 140 g kg–1.

Statistical and Spatial Variability Analyses
Statistical relations between soil Nmin potential and other measured soil properties were calculated by correlation analysis. Correlation and spatial statistics were used to relate soil Nmin potential, surface and profile residual soil NO3–N, and other soil characteristics with corn grain yield. Geostatistical methods, variography, and kriging (Isaaks and Srivastava, 1989) were used to map variability of soil and plant parameters. Variograms were constructed to characterize the structure of the spatial variation of measured soil and plant parameters. Spatial variability between samples was measured by calculating the omnidirectional semivariance, {gamma}(h):

where N(h) is the number of pairs of observations separated by lag distance h and Z(xi) and Z(xi + h) give the value of the variable Z at two positions separated by h. Geostatistical software (GS+ v5.0, Gamma Design Software, St. Plainwell, MI) was used to analyze the spatial structure of the data and to define the semivariograms. Variograms quantify the spatial variation of a range of properties by measuring the degree of correlation between sampling points at a given distance apart (Webster, 1985).

Spherical, exponential, or linear models were fitted to the semivariograms, and their selections were based on visual best fit and the corresponding coefficient of determination, i.e., r2. The parameters of the model—nugget semivariance, range, and sill, or total semivariance—were calculated. Nugget semivariance is the variance at zero distance and represents field and experimental variability, or random variability, which is undetectable at the sampling scale. Sill is the lag distance between measurements at which one value for a variable does not influence neighboring values. Range is the distance at which values of one variable become spatially independent of another. Furthermore, the ratio of the nugget to sill indicates the degree of randomness in the data's spatial variability. This ratio was used to define three classes of spatial dependence for the measured soil variables (Cambardella et al., 1994), i.e., (i) when the ratio was <0.25, the measured variable was considered strongly spatially dependent; (ii) when the ratio was between 0.25 and 0.75, the soil variable was considered moderately spatially dependent; and (iii) if the ratio was >0.75, or the slope of the semivariogram was {approx}0, the variable was considered random or nonspatially correlated (pure nugget).

Semivariogram models were cross-validated, and the punctual kriging procedure in GS+ v5.0 was used to obtain point estimates of soil parameters at unsampled locations. Kriging was performed every 2.1 m across the entire study area. Counter maps of corn yield distribution and clay content percentage were produced in Surfer (Version 8, Golden Software, Golden, CO) based on grid files created from the kriged values from GS+.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Descriptive Statistics
Univariate statistical analyses of soil and plant parameters for 2002 and 2003 are given in Tables 1 and 2. The parameters measured varied spatially and temporally. The shape of parameter distributions are described by skewness. Skewness coefficients in Tables 1 and 2 showed no shifts in distribution, except for SOC, NO3–N concentrations with depth, and initial inorganic N. Grain yield and N measurements were higher, and residual NO3–N to 60-cm depth was positively skewed and exhibited nonnormal distributions in 2002. Management and temporal effects seem to be likely reasons for nonnormal distribution. Grain yield and most of the measured parameters, however, had lower values, were less variable, and followed a normal distribution with less skewness in 2003. Soil residual NO3–N values exhibiting skewed distributions have been reported to be log normally distributed (Cambardella et al., 1994). However, a logarithmic transformation of residual NO3–N values in our study reduced this parameter's skewness, but values were still nonnormally distributed (0.05 significance level). The results in 2003 were similar to those of Parkin et al. (1988) and Cambardella et al. (1994), who showed that soil N measurements tended to become more normally distributed in years following N application.


View this table:
[in this window]
[in a new window]
  		Table 1. Descriptive statistics of parameters measured for the corn experiment in 2002.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Descriptive statistics of parameters measured for the corn experiment in 2003.

 
The coefficient of variation (CV) is a useful statistic for measuring the spatial variability of soil properties. Among the N parameters, Nmin values tended to be the most variable, with CV values of 27 and 46% in 2002 and 2003, respectively (Tables 1 and 2). The range in Nmin for the 63 samples collected across the field in 2002 was from 6 to 35 mg kg–1, with a mean value of 20 mg kg–1, and from {approx}1 to 33 mg kg–1, with a mean value of 16 mg kg–1 in 2003. The Nmin variability observed in this experiment was similar to CV values of 24 and 36.4% reported by Goovaerts and Chiang (1993) and by Cambardella et al. (1994), respectively. However, CV values for Nmin in our experiment were higher than those reported by Robertson et al. (1988) and Mahmoudjafari et al. (1997). Differences in variability may result from different incubation methods and whether disturbed or undisturbed soil samples were used. Total N, organic N, and NO3–N values with depth were characterized by CVs ranging from 12 to 22% in 2002 and 11 to 16% in 2003. Clay content had a CV of 10%.

Spatial Variability
To characterize spatial variability, spherical, exponential, and linear isotropic models were fitted to calculated semivariograms for soil parameters and yield, and their spatial dependence was categorized into three classes based on the nugget/sill ratio suggested by Cambardella et al. (1994) (Tables 3 and 4). Corn grain yield was moderately spatially distributed both years. The structure of variations for grain yield, total N, and Nmin in 2003 remained the same, with moderate and strong spatial dependence, respectively. However, the spatial dependence for residual NO3–N became weak or random in 2003, indicating spatial independence. Others have also shown the spatial dependence of NO3–N to be temporally dependent (Bruckler et al., 1997). Other parameters, such as initial soil NH4–N and NO3–N concentrations and soil C mineralized, were weakly or not spatially correlated in either 2002 or 2003. Soil organic C exhibited moderate spatial dependence in 2002 but was randomly distributed in 2003. In general, less spatial dependence was observed between soil samples in 2003 than 2002.


View this table:
[in this window]
[in a new window]
  		Table 3. Semivariance parameters for the corn experiment in 2002.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Semivariance parameters for the corn experiment in 2003.

 
Semivariograms showed strong spatial dependence for soil NO3–N with depth and clay content in 2002, mineralized N in 2003, and total N in both years (Fig. 1). Variograms with spherical and exponential models reached upper bounds, i.e., sills, suggesting that the properties varied in a "patchy" way, resulting in some areas with small values and others with large ones. The range of spatial correlation for each variogram provides an average extent of these patches. The range of spatial dependence for the exponential model is the distance at which the semivariance is 95% of the sill and is estimated as three times the fitted range parameter (Isaaks and Srivastava, 1989) (Tables 3 and 4). The range of spatial dependence for NO3–N with depth was 31 to 120 m in 2002. Similarly, Lopez-Granados et al. (2002) found that subsoil NO3–N concentrations had a strong spatial dependence. The range of the variogram for NO3–N with depth in 2003, however, increased without reaching an upper bound, indicating that the extent of the spatial variation was not realized at the sampling scale employed.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 1. Variograms for (a–d) soil NO3–N with depth, (e–f) total N, (g) mineralized N, and (h) percentage clay of the surface soil (0–15 cm) in 2002 and 2003.

 
Relationship between Soil Nitrogen Parameters and Texture and Yield
To calculate the spatial variability of crop yield and to help develop variable N rate management plans, researchers have used crop and soil properties like plant height, plant population, and clay content (Cox et al., 2003; Katsvario et al., 2003; Machado et al., 2000; Tabor et al., 1985). Spatial distributions of corn grain yield and surface soil (0–15 cm) clay content of the research field in 2002 are shown in Fig. 2. In each figure, light shading represents the highest values while darker shading is associated with the lowest values. Grain yields were higher along the southern perimeter of the field, in the transect that extended from southwest to north, and in bands toward the northeast across the field. Lower grain yields tended to be located along a southeast to northwest transect and in pockets along east–west bands near the central portion of the field. Clay content of the surface 0 to 15 cm followed the opposite trend to yield distribution and tended to be higher in northern and eastern portions and lower in the southwestern quadrant (Fig. 2b).



View larger version (76K):
[in this window]
[in a new window]
  		Fig. 2. Distribution of (a) corn yield and (b) percentage clay in the surface soil (0–15 cm) in 2002.

 
Corn grain yield and clay contents were correlated against residual soil NO3–N with depth and other soil characteristics in 2002 and 2003 (Tables 5 to 7). Total N, SOC, Nmin, soil NO3–N values for all depths, and clay content were correlated in 2002 (Table 5). Grain yield was positively correlated with residual soil NO3–N. Residual soil NO3–N to 60-cm depth had a higher correlation with yield than NO3–N to either 15- or 30-cm depths. Total N was significantly related to SOC, Nmin, clay content, and NO3–N to 15- cm depth in 2002. Goovaerts and Chiang (1993) reported similar relationships for plots with long-term fallow rotations. Mineralizable N was not related to yield but was highly related to SOC and soil total N.


View this table:
[in this window]
[in a new window]
  		Table 5. Pearson correlation coefficients for corn grain yield and soil N parameters in 2002.

 

View this table:
[in this window]
[in a new window]
  		Table 7. Pearson correlation coefficients for corn grain yield and soil N parameters in 2003.

 

View this table:
[in this window]
[in a new window]
  		Table 6. Pearson correlation coefficient mean values for corn grain yield and soil N parameters for two transects of high-yielding plots (1–9) and low-yielding plots (46–54) in 2002.

 
Corn grain yield was negatively correlated with clay content (Table 5). Soil organic C, total N, and soil N mineralized in 24 d, however, were all positively and significantly related to soil clay content (Table 5). Clay protection of adsorbed organic compounds against microbial oxidation may at least partially explain this result. Soil organic C and total N were significantly correlated with soil N mineralized in 24 d. Grain yield was negatively related to clay content, and since many of the soil characteristics (SOC, total N, etc.) were positively related to clay content, they resulted in negative correlations with grain yield. Johnson et al. (2002) showed a similar result for cotton lint yield and SOC and extractable soil P.

To further illustrate the spatial distribution of corn yield and its relation with soil parameters, two field transects in 2002 with high-yielding plots (1 to 9) and low-yielding plots (46 to 54) (Fig. 2a) were selected. Field transect mean values for yield, NO3–N, Nmin, total N, SOC, and clay content with their calculated correlation coefficients are presented in Table 6. Nitrate N and clay content were significantly related to yield and seemed to be responsible for variation in high- and low-yield plots, respectively. Residual NO3–N was positively related with yield in both high- and low-yield environments while clay content exhibited a negative relationship with grain yield in low-yielding plots. Total N in low-yielding plots was positively related to NO3–N and Nmin while clay content and Nmin were negatively related. Clay protection of adsorbed organic compounds against microbial oxidation may again at least partly explain this result. In high-yielding plots where clay was lower, clay content had a significant positive relationship with Nmin. Mineralized N, however, was not related to corn grain yield in either high- or low-yielding plots (Table 6).

Corn grain yield in 2003 was less related to soil NO3–N (Table 7) compared with 2002 (Tables 5 and 6). Part of this difference was because of the wide difference in the quantity of NO3–N between years, with N availability becoming more limiting in 2003 (Tables 1 and 2). The rate of N applied before this experiment was 125 kg N ha–1 as N solution (32–0–0) for cotton at planting in 2001. Residual NO3–N measured in 2002 was sufficient for corn yields of 6271 to 9344 kg ha–1 (Table 1) (McFarland et al., 1990). However, residual NO3–N was more limited in 2003 when NO3–N to a depth of 60 cm decreased by 50% and average yield declined by 2132 kg ha–1 (Tables 1 and 2). This temporal dependence of NO3–N availability may support the need for consideration of soil N supplied through mineralization (Raun et al., 2002; Mamo et al., 2003). In our study, Nmin was not significantly related to corn grain yield in either year. However, Nmin was significantly related to NO3–N to 90-cm depth in 2003 (Table 7), suggesting that a greater proportion of residual NO3–N originated from mineralization in 2003. Mineralizable N apparently became more significant to the overall N budget as residual NO3–N diminished due to plant uptake, leaching, or denitrification. Clay content also was not related to corn grain yield in 2003, suggesting that a different relationship may occur under limited N supply.

Contour Maps of Spatial Distribution of Variables
After constructing the variograms for soil N parameters (Fig. 1), the corresponding kriged values were plotted to form contour maps of the spatial distribution of residual NO3–N, total N, and Nmin and to estimate N values in unsampled locations in 2002 and 2003 (Fig. 3). For example, high soil NO3–N concentrations were located close to the western perimeter of the field (Fig. 3a–3d) in areas with low clay content. This is particularly evident when Fig. 2b and 3a–3d are compared.









View larger version (599K):
[in this window]
[in a new window]
  		Fig. 3. Kriged contour maps of (a–d) soil NO3–N with depth in 2002, (e–f) total N in 2002 and 2003, and (g) mineralized N in 2003.

 
A comparison of kriged maps (Fig. 3a–3d) of soil NO3–N with depth and spatial distribution of corn yield (Fig. 2a) showed a similarity, especially in the western portions of a south–north band. Tabor et al. (1985) suggested that clay content can be used for estimating NO3 concentrations throughout the growing season. Our results showed that clay content exhibited a similar variation in distribution structure (Fig. 1h) as NO3–N in 2002 (Fig. 1a–1d), suggesting that spatial structure and an optimum sampling program for soil NO3–N can be inferred from the spatial structure of clay contents for particular soil mapping units (Cambardella and Karlen, 1999). However, the large nugget semivariance and the nonspatial dependence of residual soil NO3–N in 2003 suggested that a different sampling strategy would have had to have been used to capture soil NO3–N variability in 2003 and consequently a different strategy for site-specific N application. For example, Eghball et al. (1997) found that when NO3–N was low, variable-rate N application did not significantly reduce residual soil NO3–N variability. This result could have been true for our field in 2003 where spatial structure for NO3–N was not evident, and its concentration was low, and uniform applications of N across the field would have been more justified. Bruckler et al. (1997) explained that when spatial distribution structure exists, grid sampling is more efficient than random sampling, as might have been the case for applying variable-rate N for corn in 2002.

During our experiment, N fertilizer was not applied, and based on the results presented, the distance and sampling scheme used for soil NO3–N was temporally dependent. Bruckler et al. (1997) also reported that spatial dependence of soil NO3–N was time dependent. It can be hypothesized that strongly spatially dependent soil variables may be controlled by intrinsic variation in soil characteristics such as was observed for soil texture in 2002 (Cambardella et al., 1994). Machado et al. (2000) indicated that management zones for variable-rate N application should be based on information about soil elevation, texture, and soil nitrate. Cambardella and Karlen (1999) suggested, however, that this may be true only for one or two crops after large fertilizer applications. Extrinsic factors appear to play a larger role in controlling spatial dependence when there are two or more crops after N application and low or no fertilizer input as occurred in 2003.

In contrast to residual NO3–N distribution maps, soil total N exhibited temporal persistence, i.e., high and low values generally occurred at the same locations in 2002 and 2003. These distinct zones of high and low total N may provide useful information for variable-rate N management. According to Goovaerts and Chiang (1993), temporal persistence of total N reduces the cost of sampling and analysis for N recommendations. Total N was also correlated with clay content for both 2002 and 2003. These relationships indicated that spatial relatedness may bridge several soil map units, and information derived from one set of measurements for one field may have applicability for fertilizer recommendations at other field sites within the same or similar landscapes. Cox et al. (2003) studied the effect of variability of selected soil chemical and physical properties on soybean yield. Soil fertility varied from year to year and field to field. The authors found that fertility parameters had to be considered along with other soil factors to determine their relationship to yield and reported that areas with higher clay content also had higher yield, suggesting that soil clay content might be used as a basis for site-specific soil management. In our experiment, soil total N was also related to NO3–N at 15-cm depth. However, as observed by Schmidt et al. (2002), variable-rate N recommendations based solely on soil total N are too simplistic to reflect variability in N availability within a field.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Most of the measured variables showed temporal variation except for soil total N. Temporal persistence of soil N parameters may have important implications for N fertilizer recommendations. Spatial structure existed for residual NO3–N in the first year when concentrations were higher. Spatial structure was absent, however, the following year when NO3–N concentrations were lower. Residual soil NO3–N sampled to 60-cm depth was more related to corn yield than shallower sampling. Soil Nmin was not related to residual soil NO3–N when NO3 was high in 2002 but was related to NO3 when lower in 2003.

Results showed that clay content exhibited a similar variation in distribution structure as NO3–N in 2002, suggesting that spatial structure and an optimum sampling program for soil NO3–N can be inferred from the spatial structure of clay contents for particular soil mapping units. However, the nonspatial dependence of residual soil NO3–N in 2003 suggested that a different sampling strategy would have had to be used to capture soil N variability in 2003. Uniform application of N fertilizer would have resulted in overfertilization in 2002 in some areas and underapplication in others.

This experiment suggests that variable-rate fertilizer recommendations should not be based on a single criterion. Clay content with limited sampling for NO3–N might be used to estimate and infer N requirement. Texture may be more easily determined at the scale and intensity necessary for site-specific prediction. Soil textural influences on spatial variation of soil N parameters influencing crop yield require further research in larger fields.


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





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 ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shahandeh, H.
Right arrow Articles by Lascano, R. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Shahandeh, H.
Right arrow Articles by Lascano, R. J.
Agricola
Right arrow Articles by Shahandeh, H.
Right arrow Articles by Lascano, R. J.
Related Collections
Right arrow Nitrogen
Right arrow Site-Specific Analysis

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