Printer Friendly

Utilizing spatial technology as a decision-assist tool for precision grading of salt-affected soils.

Precision land leveling, or precision grading, is the systematic process of removing soil from areas of higher elevation and depositing it in areas of lower elevation in order to create a uniform slope. This practice is a sizable investment that can range from $100 to $400 per hectare ($200 to $1,000 per acre). Precision grading is an approved soil and water conservation practice by the U.S. Department of Agriculture Natural Resources Conservation Service (USDA NRCS; NRCS standard 464) that is commonly implemented on the soils used for rice (Oryza sativa L.) production in the Southeastern United States (NRCS 1996).

Precision grading provides many benefits for levee-based flood irrigation of rice and soybeans (Glydne Max. L. Merr.) grown in rotation including (Scott et al. 1989):

* Improved water management

* More efficient water use

* Reduction of the number of levees required to maintain a given flood depth

* Reduction of land area dedicated to levees and lost to optimum crop production

* Levees that are straight rather than curved to follow natural contour

* Decreased risk of prolonged flooding of soybeans, which can be detrimental to normal development and growth

While land leveling offers many benefits, there also can be risks. Depending on elevation differences, the depth of soil removed can range from 0 cm to several meters. In many cases, part or all of the surface horizons can be removed, exposing sub-surface horizons. If the required soil removal is deep enough, sub-surface material may also be removed and deposited on top of surface soil. In many fields, the chemical, biological, and physical properties of the exposed and deposited sub-surface material render it unsuitable for profitable crop production (Miller et. al 1990). Corrective action is difficult and expensive and may require several years of treatment to achieve restoration (Miller et al. 1991).

Of particular concern is exposing soil horizons that contain excessive exchangeable sodium or soluble salts. Soils that naturally contain layers with elevated sodium levels at some depth in the profile comprise an estimated 200,000 ha (494,000 ac) in Arkansas (Horn et al. 1964). The use of poor quality irrigation water on soils with poor internal drainage has contributed to other soils being adversely affected by elevated sodium and soluble salts (Gilmour and Marx 1981; Baker et al. 1996).

Huey et al. (1989) outlined precautions that can be taken to reduce this risk, such as consulting USDA NRCS soil surveys and determining soil Na content at a depth of 30 cm (12 in) below the deepest depth of soil removal. Consulting soil surveys before making decisions about land leveling is a worthwhile protective measure. However, most soil surveys are mapped at a scale of 1:24,000 and may not account for inclusions that may comprise significant portions of individual fields (USDA 1987). To account for inclusions, the USDA NRCS suggests that on-site investigation is needed to plan for intensive uses in smaller areas.

Representing soil chemical properties on a field-average basis using standard sampling protocol (composite sample collected in random zig-zag pattern) may neither adequately describe spatial variability of soluble salts and sodium across the field nor adequately relate these parameters to the removal and deposition nature of precision grading. Thus, this approach may make it difficult to delineate the aerial extent of the potential hazard. Utilizing site-specific technology such as global positioning systems (GPS) and geographic information systems (GIS) in conjunction with more intensive grid soil sampling to map both the aerial extent and depth of sodium and soluble salts could provide decision-assist information that may reduce the risk of exposing unsuitable subsoil material. Site-specific technology has been utilized with electromagnetic induction methods to remotely sense the spatial and temporal distributions of soil salinity (Lesch et al. 1998; Triantafilis et at 2001).

Several studies have examined site-specific technology for variable-rate fertilizer applications (Sawyer 1994; Wollenhaupt et al. 1994). The most cited concern for variable-rate technology is the uncertainty of costs versus benefits. This concern may be minimal in the application of this technology as a decision-assist for land leveling, since the unit cost of grid soil sampling is less than 10% of the cost of land-leveling itself.

For precision grading to be a profitable investment, it is essential not to expose or re-deposit sub-surface material that is unsuitable for normal crop production. Results from two case studies where site-specific technology was utilized to assist with land-leveling decisions in soils containing elevated soluble salts and sodium are discussed. In both cases, the spatial distribution of Na, exchangeable sodium percentage (ESP), and electrical conductivity (EC) were mapped with site-specific technology and compared with maps depicting the removal and re-deposition of soil, This was done to provide an estimate of the potential sodium hazard that might result from land leveling. The statistical distribution of these parameters was also characterized so that decisions derived from traditional, composite sampling could be compared with decisions made with site-specific technology.

Methods and Materials

Two fields under consideration for precision grading were selected for case study. Both fields are used to produce rice and soybean in rotation. Field A (30 ha, 74 ac) is located in Hot Springs County, Arkansas, and its soil is mapped as a combination of Adaton (fine silty, mixed, thermic Typic Ochraqualf), Gurdon (coarse silty, siliceous, thermic, Aquic Paleudult), and Sardis (fine silty, siliceous, thermic, Fluvaquentic Dystrochrepts) silt loams (USDA 1987). Field B (24 ha, 59 ac) is located in Cross County, Arkansas, and its soil is mapped as Hilleman silt loam (fine silty, mixed, thermic Aeric Ochraqualf) (USDA 1968). It is interesting to note that none of these soils were classified as having a natric horizon.

Soil samples were collected prior to precision grading from each field in a grid pattern. In field A, 44 grid points were spaced approximately 92 m (300 ft) apart. However, the grid was irregular in some parts of the field, as grid points were established where it was visually evident that vegetative cover was restricted. In field B, 62 grid points were spaced approximately 61 m (200 ft) apart. Field A was re-sampled the following spring after precision grading was performed. Samples were collected to a 15 cm (6 in) depth at individual locations with a random zig-zag pattern and in areas where seedling rice had died. Based on the results of the study, precision grading was not performed in field B.

At each grid point, samples were collected with a hydraulic, cylindrical [7.5 cm (3 in) diameter] soil probe in 15 cm (6 in) increments by depth. Samples were collected in field A to a depth of 60 cm (24 in) in 1997 and to 15 cm (6 in) in 1998, while samples in field B were collected to 120 cm (48 in). Prior to precision grading, sampling depths were determined so that the deepest depth increment represented the 15 cm (6 in) of soil immediately below the maximum depth of removal for the entire field.

Routine soil chemical analysis was performed by the University of Arkansas Soil Test Lab in Marianna, Arkansas. Samples were treated with a modified Mehlich 3 extractant at a soil to extractant ratio of 1:7 (Mehlich 1984). This is the official procedure of the University of Arkansas Soil Test Lab. Samples were analyzed for P, K, Ca, Mg, Na, Sulfate-S, Fe, Mn, Cu, and Zn using an inductively coupled argon plasma (ICAP) spectrophotometer. Soil pH measurements were made with a pH probe on a 1:1 ratio of soil to distilled water on a volume basis, while electrical conductivity (EC) was measured on a 1:2 ratio of soil to distilled water on volume basis. Nitrate-N was measured with an ion-specific electrode on soil extracted with 0.25 M aluminum sulfate. Exchangeable sodium percentage (ESP) was calculated as:

ESP = (meq Na / (meq Ca + meq Mg + meq K + meq Na + meq H))* 100 (1)

where cations are expressed as milli-equivalents per 100 g (0.2 lbs) of soil and ESP is expressed in percent.

The latitude and longitude were determined for each soil sampling location and field perimeter with a differential global positioning system (DGPS; Trimble Navigation Ltd., Sunnyvale, CA). The DGPS data was post-processed to obtain the manufacturer's horizontal accuracy of 2-3 m (Trimble 1996). The geographic information systems (GIS) software SSToolbox (SST Development Group, Inc. 2001) was used to develop soil nutrient surfaces with respect to the field boundary. Kriging procedures were used to interpolate between grid points. Based upon geo-statistical analysis of data that involved constructing semi-variograms and model-fitting, a linear model was selected for use in determining kriging parameters.

Elevation relative to a predetermined benchmark was determined on a 30 m (98 ft) grid in field A by the Bowls Surveying firm and by USDA NRCS in field B. Coordinates for each elevation position were recorded with a DGPS system. Elevation surfaces for the field were developed using SSToolbox as well as maps that delineate areas of soil removal and deposition (precision grading maps). These maps were overlain with soil ESP maps for each depth interval, and GIS was used to determine the area where soil removal intersected ESP values of > 10%.

Soil test data for Na, ESP, and EC were statistically summarized and were tested for normality by the Shapiro-Wilk test for each sampling depth in each field using the Proc Univariate procedure in SAS (Hatcher and Stepnaski 1994).To test for log-normality, the same Shapiro-Wilk test was performed on data transformed by the natural logarithm. The geometric mean was determined by taking the exponential of the mean of the data transformed by the natural logarithm.

Threshold values of the decision parameters, Na, ESP, and EC, were values recommended by the University of Arkansas, which, if exceeded, would warrant corrective action to avoid adverse effects on crop production. For silt loam soils in Arkansas, the ESP threshold for detrimental crop development has been determined to be between 10 and 15% (Snyder et al. 1995). For silt loam soils in Arkansas, 250 mg [kg.sup.-1] (500 lbs [ac.sup.-1]) of Mehlich III extractable Na is considered elevated relative to normal levels (Chapman et al. 1998). In Arkansas, seedling rice death has been correlated to electrical conductivity (EC) values of [greater than or equal to] 0.9 dS [m.sup.-1] as determined by the 1:2 soil to water extract method (Gilmour et al. 1977).

Field A. Soil test results indicated that field A had low pH and low fertility (Table 1). Exchangeable Na, ESP, and EC were highly variable in each depth interval according to the coefficient of variation (CV) values, which were > 50%, and the large ranges (Table 2). While transforming the data by the natural logarithm greatly reduced the CV values, the data were neither normally nor log-normally distributed.

Neither the arithmetic nor the geometric means for Na, ESP, and EC exceeded 250 mg [kg.sup.-1] (500 lbs [ac.sub.-1]), 10%, or 0.9 dS [m.sup.-1] thresholds, respectively, in the top 45 cm (18 in). Only the arithmetic mean for ESP exceeded 10% in the bottom interval (Table 2). However, each depth interval contained values of Na and ESP that exceeded the respective thresholds. Only the top 15 cm (6 in) of the soil profile had EC values above the critical threshold value of 0.9 dS [cm.sup.-1].

The soil ESP and EC were linearly correlated to Na (Figures 1 and 2). The linear correlation between ESP and Na in Figure 1 excluded three data points where the Na concentration was extremely high (> 700 mg [kg.sup.-1], 1400 lbs [ac.sup.-1]). The ESP is better correlated to the natural logarithm of Na if these three data points are included. Based on the linear relationship, values of 10 and 15% ESP correspond to 202 mg [kg.sup.-1] (404 lbs [ac.sup.-1]) and 327 mg [kg.sup.-1] (654 lbs [ac.sup.-1]) Na, respectively.

The spatial distribution of ESP within the field was similar for all depth intervals (Figure 3). However, the aerial extent of ESP > 10% increased from 23% of the field at 0-15 cm (0-6 in) to 37% at 45-60 cm (18-24 in). The aerial extent of ESP > 15% increased from 4% of the field in the top 45 cm (18 in) to 11% below 45 cm (18 in). By overlaying Na and ESP distributions at each depth interval, it is evident that areas where Na exceeds 250 mg [kg.sup.-1] (500 lbs [Acre.sup.-1]) generally corresponded to areas where ESP was > 10%. However, the percentage of the field where Na was > 250 mg [kg.sup.-1] (500 lbs [Acre.sup.-1]) was 6-12% less than where ESP was > 10% for respective depth intervals. The spatial relationship between ESP and Na thresholds improved when a Na threshold of 202 mg [kg.sup.-1] (404 lbs [ac.sup.-1]), the value that was determined from the linear regression in Figure 1, was used instead of 250 mg [kg.sub.-1] (500 lbs [ac.sup.-1]). The spatial distribution of BC was not of interest, since m aximum values below 15 cm (6 in) did not exceed the 0.9 dS [m.sup.-1] threshold. Soil BC exceeded the threshold in < 1% of the field.

Since the aerial extent of ESP > 10% was greatest at the 45-60 cm (18-24 in) depth, this distribution was chosen to be overlain with the map delineating areas of soil removal and fill for precision grading to represent a worstcase scenario of exposing layers of excessive Na. Overlay analysis revealed that 3.5 ha (9 ac), 12% of field total, of soil where ESP was > 10% would be exposed through the removal process. Again, this would represent the worst-case scenario since some removal areas would not be as deep as 45 cm (18 in). This analysis does not account for the deposition of material having ESP >10% on areas of fill, which could also increase the aerial extent of excessive Na hazard.

Soil samples taken to a depth of 15 cm (6 in) after precision grading revealed that both the arithmetic and geometric means for Na and EC increased; however these levels would not be of concern (Table 3). Sodium was neither normally nor log-normally distributed, while BC was log-normally distributed. The arithmetic mean for ESP did not change, while the geometric mean increased slightly. The ESP distribution in 1998 was log-normally distributed. Thus the geometric mean is the measure of central tendency that should be considered, and it did not exceed the 10% threshold for sodicity.

Although measures of central tendency for ESP did not change appreciably, changes were observed in the spatial distribution of ESP (Figure 4). However, these changes are convoluted by the fact that a different spatial sampling scheme was used in 1998, and they may not represent changes only due to precision grading. The spatial distribution of ESP in 1998 was generally most similar to the distribution at 45-60 cm (18-24 in) depth before precision grading. However, the area of the field where BSP was > 10% increased to 52%. Most of this increase was in the eastern quarter of the field. Since this was predominantly an area of fill, this increase might be explained by deposition of material having ESP> 10%, change in sampling scheme, or a combination of both.

The ESP was > 10% in areas where Na exceeded 250 mg [kg.sup.-1] (500 lbs [ac.sup.-1]). However, the area of the field where ESP was > 10% was much larger than the area where Na was > 25G mg [kg.sup.-1] (500 lbs [ac.sup.-1]). Areas where BC was > 0.9 dS [m.sup.-1] occurred in areas where ESP was > 10%. Seedling rice death was spatially correlated to areas where ESP was > 15% and BC was > 0.9 dS [m.sup.-1]. Thus, it is difficult to determine if death was due to saline or sodic conditions or both.

Field B. Although University of Arkansas soil test recommendations would warrant additional P for rice and additional P and K for soybean production, the general soil fertility of this field is not atypical for highly productive rice and soybean crops in Arkansas (Table 1). Magnesium generally increased with depth. Below 75 cm (30 in) in field B, magnesium values are considered elevated (Chapman et al. 1998).

Sodium and ESP generally increased with depth according to both the arithmetic and geometric mean (Tables 4 and 5). The associated range for these parameters also increased with depth. Both the arithmetic and geometric mean for Na and ESP exceeded 250 mg [kg.sup.-1] (500 lbs [ac.sup.-1]) and 10%, respectively, below the 45 cm (18 in) depth. Maximum Na values exceeded this value between 30 and 45 cm (12 and 18 in), while maximum ESP values exceeded the 10% threshold between 15 and 45 cm (6 and 18 in). The distribution of ESP with depth was similar to that reported by Horn et al. (1964)

Sodium was log-normally distributed down to 30 cm (12 in), neither normally nor log-normally distributed between 30 and 75 cm (12 and 30 in), and normally distributed below 75 cm (30 in). The ESP was log-normally distributed down to 60 cm (24 in), normally distributed between 60 and 75 cm (24 and 30 in), and generally neither normally nor log-normally distributed below 75 cm (30 in).

Unlike Na and ESP, EC was fairly uniform with depth with respect to measures of central tendency (Table 6). The maximum EC value of any of the depth intervals did not exceed 0.9 dS [m.sup.-1]. Variability of BC within depth intervals was less than that for Na and ESP.

The ESP was linearly correlated to the natural logarithm of Na in the top 60 cm (24 in) ([R.sup.2] ranged from 0.60 to 0.85 for the top four depth intervals), where Mg was below 200 mg [kg.sup.-1] (400 lbs [ac.sup.-1]). Below 60 cm (24 in), where Mg was greater than 200 mg [kg.sup.-1] (400 lbs [ac.sup.-1]), the ESP was not correlated to the natural logarithm of Na ([R.sup.2] ranged from 0.23 to 0.50 in the bottom four depth intervals). Unlike field A, there was no correlation between EC and Na ([R.sup.2] = 0.1389 for linear).

The spatial distribution of ESP revealed that the land area where ESP was> 10% progressively increased with depth (Figure 5) and was not of significance in the top 30 cm (12 in). The ESP was > 10% in over 80% of the field below 45 cm (18 in), and the ESP was greater than 15% in over 80% of the field below 60 cm (24 in).

Based on overlay analysis of the map delineating areas of removal and deposition with each ESP distribution in Figure 5, precision land leveling would expose material with ESP > 10% in 35% of the field (Figure 6). Based on the assumption that the removed material with ESP > 10% is uniformly deposited 15 cm (6 in) deep on the original soil surface, then precision grading would induce a sodium hazard on an additional 37% of the field through the deposition process (Figure 6). This assumption represents the worst-case scenario, which would render 72% of the field having ESP > 10% in the top 15 cm (6 in).

Results and Discussion

Field A. None of the soil mapping units in field A had natric horizons. Based on these field study results, field A did not meet the taxonomic definition of a natric horizon according to measures of central tendency for ESP (Table 2). The spatial distribution of ESP > 15% in field A (Figures 3 and 4) before and after land leveling was confined to areas that are smaller than the minimum size delineation of 2.3 ha (6 ac) of a 1:24,000 scale soil survey map (USDA 1993). In this case, consulting the USDA NRCS soil survey may not have provided sufficient detail to offer much precaution about potential sodium hazards related to precision grading.

Based on the spatial distribution of ESP in the top three sampling increments in Figure 3 and the measures of central tendency from grid sampling in Table 3, it is unlikely that routine soil sampling that produces a composite sample from sub-samples taken in a zig-zag or grid pattern would have delineated these areas either. Perhaps the traditional field-average sampling approach would have provided a precaution for the depth interval of 45-60 cm (18-24 in) since the arithmetic mean ESP was> 10%. However, the statistical distribution of ESP at this depth was neither normal nor log-normal. This suggests that the arithmetic median (ESP = 8.3%) may be the most appropriate choice of central tendency. If so, the field-average approach would not have provided any precaution. On the other hand, composite sampling with the random zig-zag pattern could have biased the results towards the extremely high ESP values if these small areas had been disproportionately sampled as compared to the rest of the field. In this ca se, it may have wrongly led to the decision not to implement precision grading.

Targeted soil sampling in these visually apparent areas at the surface would have confirmed the presence of ESP> 15%. However, Figure 3 clearly indicates that the aerial extent of elevated sodium in the top 45 cm (18 in) was noticeably different than below 45 cm (18 in). This implies that targeted sampling based on surface conditions may not target the appropriate locations at depths below the surface. Targeted sampling does not provide sufficient information about the aerial extent of ESP > 15% or how it relates to the removal and deposition nature of precision land leveling.

Even though field-average sampling approaches did not adequately delineate sodium hazards as compared to mapping, both approaches led to the same decision to implement precision grading. However, the mapping approach provided more assurance for protecting the landowner's investment. Visual inspection of the field after precision grading confirmed that it had minimal adverse impact on rice seedling establishment (Figure 4).

Soil sampling in the top 15 cm (6 in) after sampling indicated that precision grading had not substantially changed the spatial distribution and extent of ESP> 15%, but had sub- stantially increased the extent of ESP between 10 and 15%. The more conservative threshold of ESP> 10% was used in the GIS analysis of field A before land leveling, and only the risk of exposing areas where ESP was > 10% was considered. In retrospect, re-deposition of sodium-laden soil on the surface is an extremely important consideration, since Figure 4 shows that the increase in ESP > 10% after precision grading were in areas of re-deposition and not removal.

The follow-up GIS analysis also pointed out the importance of using research-based thresholds in the decision-making process.

The more conservative threshold of ESP > 10% was used in the GIS analysis for field A before land leveling. However, rice seedling death was better correlated to ESP > 15%. The selection of a threshold could potentially affect the decision. More research is needed to better define ESP thresholds for crops and soil conditions in Arkansas.

Figure 3 clearly indicates that ESP was > 15% in small portions of the field. Visible crop damage in these areas had led the producer to conclude that poor crop production across the entire field was due to sodicity. In fact, data in Figure 3 and Table 1 indicate that low pH, P, and K resulting from altered management for elevated sodium concerns may have been limiting crop production in areas where ESP was well below the threshold values. Using spatial technology and grid soil sampling to map soil nutrients assisted the producer in changing his management philosophy for the field. Instead of managing the field to correct highly visible and severe problems in < 10% of the field, he has now decided to focus on correcting basic fertility problems that may be limiting crop production in the other 90% of the field.

Field B. In field B, the predominant map unit, the Hilleman series, does not meet the taxonomic definition as natric; however, its pedon description clearly indicates that the soil has excessive Na levels at depths below the top 40 cm (16 in) of the argillic horizon. Use of the soil survey as a basis for land-leveling decisions would have indicated concern in field B. In this particular case, routine soil sampling and measures of central tendency for ESP to depths below the greatest proposed depth of soil removal would have also indicated excessive sodium levels and would have provided sufficient information to reject the implementation of land leveling.

Based on this study, the producer chose not to implement precision land leveling with this particular grading pattern. However, he used the results to consider alternate grading strategies, such as stockpiling the top 30 cm (12 in) of removal areas and using this material to back-fill over areas where Na-laden material would be deposited.

Summary and Conclusions

Precision grading continues to be implemented in Arkansas to provide water management benefits for rice and soybean production. In Arkansas, more than 200,000 hectares (494,000 ac) of soils have naturally occurring subsurface layers containing sufficient levels of Na to be detrimental to normal crop production (Horn et al. 1964). The case study in field B represents a common scenario with regard to precision grading. Elevated sodium is far enough below the effective root zone that it exerts little influence on crop production, and the producer may not even realize that it exists. Thus, without realizing the need to take precautionary action, he unknowingly creates a sodium hazard by exposing and re-distributing Na-laden material during the process of precision grading. Unfortunately, instead of realizing the water management benefit of his investment, he has to embark on expensive corrective measures that may take years to restore productivity.

Clearly, precautionary planning needs to be taken before implementing precision land leveling to avoid exposing and spreading subsurface material that is unsuitable for normal crop production. Consulting the soil survey is a worthwhile and important first consideration in land-leveling decisions. Routine soil sampling in depth increments may also provide useful information, but it may neither delineate small areas of concern nor relate these areas to the removal and deposition nature of precision land leveling.

Mapping ESP with grid sampling and GPS and using GIS to integrate these maps with precision grading maps provided information at sufficient detail to provide decision makers with reasonable assurance about their decisions a priori. Utilizing this approach to aid decisions associated with precision grading offered many advantages including

* Greater flexibility in implementation options

* Greater assurance about the profitability of investment

* Reduction in risks before implementation

The use of this technology with precision grading may well be cost-effective, since the cost of grid sampling mapping is generally less than 10% of the investment in precision grading itself.

While most practitioners have focused the application of spatial technology in crop production to such things as yield monitoring and variable-tate fertilizer, site-specific technology offers great potential as a decision-assist tool for many agricultural applications. Some of the general benefits of site-specific technology demonstrated in this work may also be realized in other agricultural practices including

* Assistance with soil and water conservation planning

* A powerful and efficient analysis tool for identifying, diagnosing, and solving soil-related problems

* Risk assessment tool for assisting decisions on long-term conservation investments

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]
Table 1

Mean and standard deviation of selected soil test parameters by depth
increment for field A (N = 44) and field B (N = 62).

Depth    pH                P *        K                Ca

cm                            mg [kg.sup.-1]

Field A

0-15     4.7 [+ or -] 0.5  6    33 [+ or -] 7   364 [+ or -] 134
15-30    4.8 [+ or -] 0.5  6    26 [+ or -] 4   384 [+ or -] 185
30-45    4.7 [+ or -] 0.5  6    27 [+ or -] 6   340 [+ or -] 152
45-60    4.6 [+ or -] 0.4  5    29 [+ or -] 8   338 [+ or -] 143

Field B

0-15     6.0 [+ or -] 0.5  6    85 [+ or -] 25  939 [+ or -] 194
15-30    6.3 [+ or -] 0.6  5    40 [+ or -] 15  818 [+ or -] 292
30-45    6.1 [+ or -] 0.6  5    41 [+ or -] 16  575 [+ or -] 218
45-60    6.0 [+ or -] 0.5  5    47 [+ or -] 20  439 [+ or -] 154
60-75    6.1 [+ or -] 0.6  5    55 [+ or -] 20  438 [+ or -] 175
75-90    6.2 [+ or -] 0.7  5    59 [+ or -] 15  542 [+ or -] 189
90-105   6.5 [+ or -] 0.8  5    63 [+ or -] 15  679 [+ or -] 177
105-120  6.7 [+ or -] 0.8  5    66 [+ or -] 17  739 [+ or -] 162

Depth           Mg

cm         mg[kg.sup.-1]

Field A

0-15      112 [+ or -] 43
15-30     119 [+ or -] 49
30-45     126 [+ or -] 58
45-60     152 [+ or -] 66

Field B

0-15      136 [+ or -] 26
15-30     131 [+ or -] 47
30-45     130 [+ or -] 72
45-60    160 [+ or -] 107
60-75    231 [+ or -] 146
75-90    310 [+ or -] 149
90-105   361 [+ or -] 137
105-120  368 [+ or -] 112

* 5 mg [kg.sup.-1] is minimum value of P reported by the University of
Arkansas Soil Test Lab.
Table 2

Startistical summary (N = 44) of soil sodium (Na), exchangeable sodium
percentage (ESP), and electrical conductivity (EC) for each depth
interval for A before land leveling.

               Standard                               Coefficient of
Depth  Mean    Deviation    Median  Mode  Min.  Max.  Variation (CV)

Na cm               mg [kg.sup.-1]               %    mg [kg.sup.-1]

0-15    182   [+ or -] 200     109    61    50  1015       110
15-30   165   [+ or -] 114     114   106    57   537        69
30-45   177   [+ or -] 116     124    72    72   575        66
45-60   214   [+ or -] 149     148    98    68   693        70

(ESP)                        %

0-15    8.3   [+ or -] 5.4     6.3   6.5   2.6  22.9        66
15-30   8.5   [+ or -] 5.0     6.6   5.7   2.8  22.6        59
30-45   9.1   [+ or -] 4.6     7.2   4.6   4.1  21.4        51
45-60  10.2   [+ or -] 5.2     8.3   6.1   3.6  22.1        51

EC                          dS [m.sup.-1]

0-15   0.19  [+ or -] 0.27    0.10  0.04  0.04  1.37       142
15-30  0.13  [+ or -] 0.14    0.07  0.02  0.02  0.62       108
30-45  0.13  [+ or -] 0.14    0.07  0.04  0.02  0.62       108
45-60  0.15  [+ or -] 0.15    0.09  0.04  0.03  0.68       100

                 Geometric  Coefficient of
Depth  Pr<W (1)    mean       Variation (2)  Pr<W (3)

Na cm                %

0-15    0.0001      132           15         0.0007
15-30   0.0001      137           12         0.0219 &
30-45   0.0004      150           11         0.0030
45-60   0.0001      173           12         0.0092

(ESP)                         %

0-15    0.0001      7.0           29         0.0182
15-30   0.0001      7.3           26         0.1341
30-45   0.0001      8.1           21         0.0069
45-60   0.0001      9.0           22         0.0241

EC                     dS [m.sup.-1]

0-15    0.0001     0.11           20         0.0016
15-30   0.0010     0.08           20         0.0081
30-45   0.0010     0.09           19         0.0089
45-60   0.0010     0.10           19         0.0025

(1) = Probability of Shapiro-Wilk statistics.

(2) = CoefficIent of variation associated with data transformed by the
natural logarithm.

(3) = Probability of Shapiro-Wilk statistic associated with data
transformed by the natural logarithm.

& = Log-normality at the 0.05 level of significance.
Table 3

Soil sodium (Na), exchangeable sodium percentage (ESP), and electrical
conductivity (EC) in the top 15 cm of field A (N=30) after precision
grading.

                       Na         ESP         EC

                 mg [kg.sup.-1]    %     dS [m.sup.-1]

Mean                  245        10.3       0.44
s.d. ([+ or -])       238         5.0       0.50
Geometric mean        185         9.3 &     0.29 &

& indicates log-normality at the 0.05 level of probability by the
Shapiro-Wilk statistic.
Table 4

Statistical summary (N = 62) of soil sodium (Na) for each depth interval
for Field B.

                Standard                              Coefficient of
Depth    Mean   Deviation   Median  Mode  Min.  Max.    Variation

cm                      mg [kg.sup.-1]                      %

0-15     107   [+ or -]20    104     99    74    162        18
15-30    136   [+ or -]37    133    116    71    225        27
30-45    183   [+ or -]88    162    160    86    557        48
45-60    301   [+ or -]174   239    198    89    748        58
60-75    486   [+ or -]270   368    349    81   1046        56
75-90    668   [+ or -]309   695    544   102   1298        46
90-105   804   [+ or -]333   841    937   155   1766        41
105-120  838   [+ or -]324   838    883   138   1799        39

                      Geometric    Coefficient of
Depth    Pr<W (1)       Mean       Variation (2)   Pr<W (3)

cm                 mg [kg.sup.-1]        %

0-15     0.0027         104              4         0.5504 &
15-30    0.1649 *       131              6         0.3984 &
30-45    0.0001         168              8         0.0068
45-60    0.0001         260             10         0.0242
60-75    0.0002         411             10         0.0091
75-90    0.1086 *       585              9         0.0037
90-105   0.4125 *       722              8         0.0010
105-120  0.4890 *       760              7         0.0001

(1) = Probability of Shapiro-Wilk statistics

(2) = Coefficient of variation associated with data transformed by the
natural logarithm

(3) = Probability of Shapiro-Wilk statistic associated with data
transformed by the natural logarithm

* = Normally at the 0.05 level of significance

& = Log-normality at the 0.05 level of significance
Table 5

Statistical summary (N = 62) for exchangeable sodium percantage (ESP)
for each depth interval for field B.

               Standard                               Coefficient of
Depth    Mean  Deviation    Median  Mode  Min.  Max.    Variation

cm                      mg [kg.sup.-1]

0-15      4.9  [+ or -]1.1     4.8   5.1  3.0    8.5        22
15-30     6.9  [+ or -]1.9     6.9   5.4  3.4   11.1        28
30-45     9.9  [+ or -]3.6     8.9   8.9  4.2   23.8        37
45-60    14.7  [+ or -]5.7    14.0  12.0  4.7   30.1        39
60-75    17.8  [+ or -]4.6    18.5  19.5  4.4   27.8        26
75-90    18.5  [+ or -]3.6    18.6  20.5  6.4   25.1        20
90-105   18.0  [+ or -]3.1    18.2  18.1  8.1   24.0        17
105-120  17.8  [+ or -]3.0    17.8  16.4  6.0   23.1        17

                   Geometric  Coefficient of
Depth    Pr<W (1)    mean       Variation (2)  Pr<W (3)

cm                              %

0-15     0.0280       4.8           13         0.9134 &
15-30    0.1115 *     6.6           15         0.4216 &
30-45    0.0001       9.3           15         0.5036 &
45-60    0.0563 *    13.6           16         0.3640 &
60-75    0.3154 *    17.0           11         0.0001
75-90    0.0338      18.1            8         0.0001
90-105   0.0039      17.7            7         0.0001
105-120  0.0001      17.5            7         0.0001

(1) = Probability of Shapiro-Wilk statistics

(2) = Coefficient of variation associated with data transformed by the
natural logarithm

(3) = Probability of Shapiro-Wilk statistic associated with data
transformed by the natural logarithm

* = Normality at the 0.05 level of significance

& = Log-normailty at the 0.05 level of significance

Note: flx line locations
Table 6

Summary of statistics (N = 62) of soil electrical conductivity (EC) for
each depth interval for field B.

                 Standard                              Coefficient of
Depth    Mean   Deviation    Median  Mode  Min.  Max.     Variaton

cm                       dS [m.sup.-1]                       %

0-15     0.08  [+ or -]0.02   0.08   0.06  0.05  0.14        25
15-30    0.08  [+ or -]0.04   0.07   0.06  0.05  0.34        50
30-45    0.07  [+ or -]0.02   0.06   0.05  0.03  0.13        29
45-60    0.07  [+ or -]0.02   0.06   0.06  0.03  0.14        29
60-75    0.07  [+ or -]0.02   0.07   0.06  0.03  0.12        29
75-90    0.07  [+ or -]0.03   0.07   0.07  0.03  0.13        43
90-105   0.08  [+ or -]0.03   0.08   0.05  0.03  0.17        38
105-120  0.09  [+ or -]0.03   0.08   0.06  0.04  0.18        33

                     Geometric     Coefficient of
Depth    Pr<W (1)       mean         Variation (2)  Pr<W (3)

cm                 dS [cm.sup.-1]        %

0-15     0.0902 *       0.08             5          0.9074 &
15-30    0.0001         0.07             8          0.0001
30-45    0.0112         0.06             7          0.7206 &
45-60    0.0048         0.06             7          0.8905 &
60-75    0.3207 *       0.07             8          0.0669 &
75-90    0.1713 *       0.70             9          0.0648 &
90-105   0.0530 *       0.80             9          0.5322 &
105-120  0.2045 *       0.80             9          0.2783 &

(1) = Probability of Shapiro-Wilk statistics

(2) = Coefficient of Variation associated with data transformed with the
natural logarithm

(3) = Probability of Shapiro-Walk Statistic associated with natural
logarithm

* = Normality at the 0.05 level of significance

& = Log-normality at the 0.05 level of significance


Acknowledgements

The authors express appreciation to Dr. W. Baker, J. Clemons, R. Matlock, R.Wimberly and the USDA NRCS for their assistance in this study. We would also like to thank Dr. J.T. Gilmour, Dr. D. Miller, Dr. E.M. Rutledge, and Dr. H.D. Scott for their technical suggestions. And lastly, the authors wish to express their appreciation to the cooperators, Mr. Bob Fowler and Mr. Richard Imboden.

References Cited

Baker, W.H., J.S. Yadon. S.D. Carroll, CE. Wilson, S.L. Chapman, B. Teague, and P.

Tacker. 1996. Extent of saltwater intrusion into groundwater used for irrigating rice and soybeans in Arkansas. Pp. 163-168. Is: R.J. Norman and B.R. Wells (eds). Arkansas Rice Research Studies 1995. Arkansas Agricultural Experiment Station, University of Arkansas Research Series 453.

Chapman. S.L., W.E. Sabbe, and W.H. Baker. 1998. Understanding the numbers on your soil test report. Fact Sheet 2118. University of Arkansas Cooperative Extension Service, Little Rock, AR.

Gilmour, J.T. and D. B. Marx. 1981. Soluble salt concentrations in runoff water from soils previously cropped to rice. Soil Science Society of America Journal 45:1198-1201.

Gilmour, J.T., K. Sriyotai, and L. Correa. 1977. Soil Salinity and rice seedling survival. Arkansas Farm Research 26(l):4. Arkansas Agricultural Experiment Station, University of Arkansas.

Hatcher, L. and E. J. Stepnaski. 1994.A step-by-step approach to using the SAS system for univariate and multivariate statistics. Cary, NC: SAS Institute.

Horn, M.E., E.M. Rutledge, H.C. Dean, and M. Lawson. 1964. Classification and genesis of some solentz (sodic) soils in eastern Arkansas. Soil Science Society of America Proceedings 28(5):688-692.

Huey B.A., R.J. Norman, R.S. Helms, B.R. Wells, and H.M. Chaney. 1989. Growing rice on precision leveled soils. Rice Information Article 101. University of Arkansas Cooperative Extension Service. Little Rock, AR.

Lesch, S.M., J. Herrero, and J.D. Rhoades. 1998. Monitoring for temporal changes in soil salinity using electromagnetic induction techniques. Soil Science Society of America Journal 62:232-242.

Mchlich, A. 1984. Mechlich 3 soil test extraction: a modification of Mehlich 2 extractant. Communications in Soil Science and Plant Analysis 15:1409-1416.

Miller, D.M., B.R. Wells, R.J. Norman, R.S. Helms, and W.N. Miley. 1990. Poultry litter as an amendment for precision graded soils. Soils and Fertilizer Information Article 2-90. University of Arkansas Cooperative Extension Service. Little Rock. AR.

Miller, D.M., B.R. Wells, and R.J. Norman. 1991. Fertilization of rice on graded soils using organic materials. Pp. 55-57. In W.E. Sabbe (ed). Arkansas Soil Fertility Studies 1990. Arkansas Agricultural Experiment Station Research Series 411. University of Arkansas.

Natural Resources Conservation Service. 1996. Field Office Technical Guide - Arkansas. Pp. 464-1-464-3. Natural Resources Conservation Service, Little Ruck. AR.

Sawyer, J.E. 1994. Concepts of variable rate technology with considerations for fertilizer application. Journal Production Agriculture 7:195-201.

Scott, H.D., J. DeAngulo, M.B. Daniels, and L.S. Wood. 1989. Prolonged flooding effect on soybean. Agronomy Journal 81:631-636.

Snyder, C.S., S.L. Chapman, and J.T. Gilmour. 1995. Management of soils with high soluble salts. Soil Test Note - ST003. University of Arkansas Cooperative Extension Service, Little Rock. AR.

SST Development Group, Inc. 2001. SSToolbox user guide. Stillwater, OK: SST Development Group, Inc.

Triantafilis, J., I.O.A. Odeh, and A.B. McBratney. 2001. Five geostatistical models to predict soil salinity from electro-magnetic induction data across irrigated cotton. Soil Science Society of America Journal 65:869-878.

Trimble Navigation Ltd. 1996. GPS Mapping Systems. General Reference. Sunnyvale, CA: Trimble Navigation Ltd. Surveying & Mapping Division.

United States Department of Agriculture. 1968. Soil Survey: Cross County. Arkansas.

United States Department of Agriculture. 1987. Soil Survey: Clark and Hot Springs Counties Arkansas.

United States Department of Agriculture. 1993. Soil Survey: Manual. P. 5-3.

Wollenhaupt, N.C., R.P. Wolkowski, and M.K. Clayton. 1994. Nutrient management-mapping soil test phosphorus and potassium for variable-rate fertilizer application. Journal Production Agriculture 7:441-448.

Michael B. Daniels is an environmental management specialist, Stanley L. chapman is a soils specialist, and William Teague is an extension associate with the university of Arkansas cooperative Extension Service in Little Rock, Arkansas.
COPYRIGHT 2002 Soil & Water Conservation Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

 
Article Details
Printer friendly Cite/link Email Feedback
Author:Daniels, M.B.; Chapman, S.L.; Teague, W.
Publication:Journal of Soil and Water Conservation
Geographic Code:1USA
Date:May 1, 2002
Words:6778
Previous Article:Riparian zone impact on phosphorus movement to a Coastal Plain black water stream. (Research).
Next Article:Simulating grass productivity on diverse range sites in Texas.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters