JEQ Grow Your Career With ASA
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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gish, T. J.
Right arrow Articles by Kung, K.-J. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gish, T. J.
Right arrow Articles by Kung, K.-J. S.
Agricola
Right arrow Articles by Gish, T. J.
Right arrow Articles by Kung, K.-J. S.
Related Collections
Right arrow Watershed and Landscape Processes
Right arrow Preferential Flow
Right arrow Maize
Right arrow Sustainable Management of the Vadose Zone
Right arrow Vadose Zone Processes and Chemical Transport
Published in J. Environ. Qual. 34:274-286 (2005).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Landscape and Watershed Processes

Using Soil Moisture and Spatial Yield Patterns to Identify Subsurface Flow Pathways

T. J. Gisha,*, C. L. Walthallb, C. S. T. Daughtryb and K.-J. S. Kungc

a USDA-ARS Hydrology Laboratory, Natural Resources Institute, Beltsville, MD 20705
b USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, MD 20705
c Soil Science Department, University of Wisconsin, Madison, WI 53706

* Corresponding author (tgish{at}hydrolab.arsusda.gov)

Received for publication December 12, 2002.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Subsurface soil water dynamics can influence crop growth and the fate of surface-applied fertilizers and pesticides. Recently, a method was proposed using only ground-penetrating radar (GPR) and digital elevation maps (DEMs) to identify locations where subsurface water converged into discrete pathways. For this study, the GPR protocol for identifying horizontal subsurface flow pathways was extended to a 3.2-ha field, uncertainty is discussed, and soil moisture and yield patterns are presented as confirming evidence of the extent of the subsurface flow pathways. Observed soil water contents supported the existence of discrete preferential funnel flow processes occurring near the GPR-identified preferential flow pathways. Soil moisture also played a critical role in the formation of corn (Zea mays L.) grain yield patterns with yield spatial patterns being similar for mild and severe drought conditions. A buffer zone protocol was introduced that allowed the impact of subsurface flow pathways on corn grain yield to be quantified. Results indicate that when a GPR-identified subsurface clay layer was within 2 m of the soil surface, there was a beneficial impact on yield during a drought year. Furthermore, the buffer zone analysis demonstrated that corn grain yields decreased as the horizontal distance from the GPR-identified subsurface flow pathways increased during a drought year. Averaged real-time soil moisture contents at 0.1 m also decreased with increasing distance from the GPR-identified flow pathways. This research suggests that subsurface flow pathways exist and influence soil moisture and corn grain yield patterns.

Abbreviations: DEM, digital elevation map • GPR, ground-penetrating radar • OPE3, Optimizing Production Inputs for Economic and Environmental Enhancement program


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
SOIL WATER CONTENT influences both crop production and water quality. However, there is substantial uncertainty in determining the impacts of hydrology at scales larger than small plots (Simmons et al., 1979; Sharma et al., 1980; Council for Agricultural Science and Technology, 1992). This uncertainty is a result of a high degree of spatial and temporal variability resulting from heterogeneous soils and variable climatic conditions. Subsurface soil horizons affect water movement, and hence plant growth, by triggering preferential funnel flow processes, especially when the subsurface restricting layer is inclined (Kung, 1990, 1993; Ju and Kung, 1993). Protocols for determining where and when these subsurface convergent flow pathways are active on a field scale are still in their infancy. Recently, an experimental protocol using primarily GPR and DEMs to identify the location of subsurface convergent flow pathways was developed (Gish et al., 2002). Unfortunately, collection and analysis of sufficient GPR data to determine the depth of the clay lenses at the field scale is laborious, time consuming, and costly. Because complete field coverage with GPR is unrealistic, a stratified sampling scheme using GPR data at different scales of observation could provide spatial information for the entire field. Incomplete GPR coverage will lead to increased uncertainties in the prediction of the clay lens location and orientation, requiring the use of independent data sets to evaluate and confirm the extent of the GPR-identified flow pathways. Using plants to monitor the soil water status may be promising as they provide complete coverage over a field and are sensitive to soil water dynamics.

Within-field spatial variability of corn grain yields is a function of several complex interactions, including climate, plant genetics, nutrient status, and soil water dynamics. Of these factors, soil water dynamics, or more specifically, soil moisture stress, is one of the major causes of yield variability (Musick and Dusek, 1980; Eck, 1984; Johnson et al., 1987; Timlin et al., 1998). Too little or too much water can cause plant stress, which subsequently reduces photosynthesis and yields. Sinai et al. (1981) found that under arid conditions, soil moisture contents were highly correlated with the curvature of the soil surface, while Graveel et al. (1989) found that corn grain yields were more variable on steep slopes. Although landscape position, soil properties, and climate influence plant-available water, little is understood on how subsurface water flow pathways may contribute to within-field yield variability. With growth of precision farming strategies, there is an urgent need to understand how and when subsurface hydrology influences yield.

The objectives of this investigation were to (i) extend the GPR protocol for determining subsurface convergent flow pathways to the field scale, (ii) evaluate uncertainties associated with the sampling protocol, and (iii) evaluate data sets for confirming the areal extent of the GPR-identified flow pathways.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Site Description
The research site for this study used one of four fields associated with the Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3) program located at the USDA-ARS Beltsville Agricultural Research Center, Beltsville, Maryland (39° 01' 00'' N, 76° 52' 00'' W). The OPE3 site has 21 ha of arable land in four small, hydrologically bounded fields, with earthen berms that feed a wooded riparian wetland and first-order stream (Fig. 1) . Additionally, the OPE3 site contains the topographical high point for the region. Angier et al. (2001) demonstrated that the adjacent first-order stream resulted from ground water flows originating from the OPE3 site. This manuscript focuses on one of the small 3.2-ha fields (Field B) that has surface slopes ranging from 1 to 3%, and consists of the following soil types (approximate area of coverage is shown in parentheses): Downer–Muirkirk–Matawan sandy loam (49%), Bourne fine sandy loam (23%), Matawan–Hammonton loamy sand (23%), and Downer–Ingleside loamy sand (5%). These soil series consist of a sandy-textured surface with at least one finer-textured subsurface horizon (Table 1). The representative soil taxonomy is a coarse-loamy, siliceous, mesic Typic (or Aquic) Hapludults. Two of the main characteristics that differentiate these soil series are their drainage and permeability classes. The Downer and Muirkik soil series are well to excessively drained and are very permeable, while the Bourne series is moderate to poorly drained and has low permeability. The Matawan series is moderately drained and is moderately permeable.



View larger version (43K):
[in this window]
[in a new window]
  		Fig. 1. Schematic of the 21-ha Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3) research site showing four field boundaries, surface topography, first-order stream, and selected instrument locations.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Texture color and structure of dominant soil series.{dagger}

 
The research site has been placed in continuous corn production since 1998. Before 1996 the site was used for raising hogs and only a few isolated areas were used for limited crop production. In 1996 and 1997 hog houses and access roads were removed. Corn variety, planting and selected plant physiology dates, as well as some precipitation data for a 3-yr period are shown in Table 2.


View this table:
[in this window]
[in a new window]
  		Table 2. Corn development stages and rainfall.

 
To characterize the research site, 1800 soil cores were removed in 1999, representing the top 0.30 m of soil. The sampling location scheme consisted of a 30- by 30-m grid over the entire 21-ha research site plus an additional 12 transects, each 195 m long. For each of the 12 transects, a sampling location was flagged every 15 m. Six cores were randomly removed within 6 m of each sampling location and sampling locations were identified with a differential global positioning system (GPS) unit. Although a wide range of soil properties were determined, only organic matter content and elevation relative to yield will be presented here.

Ground-Penetrating Radar–Identified Subsurface Flow Pathways
A DEM was developed by analyzing real-time kinematic GPS data acquired on a 5-m grid over the entire research site (Trimble, Sunnyvale, CA). The west end of Field B contains a geographic high point for the field, about 43 m above mean sea level. The DEM also showed that 40% of Field B had surface slopes of <1%, 53% had 1 to 2% slopes, and 7% had 2 to 3% slopes.

The DEM was used with the GPR data to identify potential subsurface flow pathways. Specifically, a Subsurface Interface Radar System-2 with a 150-MHz antenna (Geophysical Survey Systems, North Salem, NH) was used to identify subsurface reflections that could represent the depth of soil layers restricting vertical water movement (i.e., clay lenses). Ground-penetrating radar data were acquired for the entire production at two scales of observation and a digital trace of the subsurface reflections was produced using RADAN software (Geophysical Survey Systems, 1999). The first GPR data consisted of parallel north–south transects, 25 m apart over the entire research site. The second GPR data set consisted of parallel north–south transects, 2 m apart over forty-four 25- by 25-m plots within the OPE3 site. The spatial autocorrelation of these subsurface reflections that could restrict water movement for the entire research site was determined using GEO-EAS (USEPA, 1991) and GS+ (Gamma Design Software, 2001) geostatistical software packages. To determine the elevation of the subsurface restricting layer the depth of these subsurface reflections was averaged over each 8- by 8-m cell and was subsequently subtracted from a DEM averaged over the corresponding 8- by 8-m cell. The uncertainty associated with the kriged estimates of the subsurface clay lens depth can subsequently be quantified.

With the spatial patterns of the subsurface restricting layer determined, the location of convergent subsurface flow pathways can be calculated and identified. The ARC-GIS (ESRI, 2002) hydrologic modeling tools FLOWDIRECTION and FLOWACCUMULATION were applied to the raster grid of the elevation-corrected subsurface topography to determine potential subsurface flow pathways. To quantify the location of the subsurface flow pathways the FLOWDIRECTION command was used to create a grid indicating the direction of flow from each cell to its steepest downslope neighboring cell. The FLOWACCUMULATION command was subsequently applied to the FLOWDIRECTION grid to create a grid of accumulated water flow from its surrounding cells for a given mass of water applied. To be designated as a cell involved in a flow pathway, at least 100 m2 of land had to be draining into that cell from upslope and neighboring cells.

Corn grain yields for 1998 through 2000 were acquired with a grain combine equipped with a Yield Monitor 2000 (AgLeader, Roswell, GA) interfaced with a differential GPS. Yield data were processed to eliminate measurement errors resulting from harvester detours around field instrumentation and other obstacles. The spatial autocorrelation of corn grain yields was determined using GEO-EAS packages. To make direct comparison of yield data to the GPR-identified flow pathways the GEO-EAS (USEPA, 1991) and GS+ (Gamma Design Software, 2001) geostatistical software yield data were then block kriged at two scales, 2 by 2 and 8 by 8 m. Generally, corn grain yield values were collected every 1.4 m along the row direction.

Soil Moisture Monitoring
Soil moisture sensors were installed to independently monitor the spatial and temporal changes in soil water content throughout the growing season. Within Field B, 64 soil moisture sensors were distributed between 12 capacitance probes (EnviroSCAN; Sentek Pty Ltd., South Australia) (Table 3). These soil water sensors monitor a soil region about 0.10 m in diameter (Paltineanu and Starr, 1997). Each sensor was calibrated before installation, and programmed to record volumetric water content every 10 min. Each sensor is recalibrated every two years to ensure accurate measurements. The soil moisture probes were configured with either three, six, or seven sensors at varying depths depending on EM-38 data and the depth to the first continuous restricting layer at each plot as determined with GPR (for details of installation see Gish et al., 2002). The three-sensor probes had sensors at depths of 0.1, 0.3, and 0.8 m and were inserted in areas where the EM-38 indicated a low infiltration capacity and the GPR data indicated that a continuous restricting layer was generally less than 1.5 m from the soil surface. The seven-sensor probes had sensors at depths of 0.1, 0.3, 0.5, 0.8, 1.2, 1.5, and 1.8 m and were inserted in regions where the EM-38 data suggested a high infiltration capacity and the depth to the first continuous subsurface restricting layer was >1.5 m. The six-sensor probes had sensors at depths of 0.1, 0.3, 0.5, 1.2, 1.5, and 1.8 m and were inserted whenever the depth to the first continuous restricting layer was between 0.9 and 1.5 m and was identified as having an intermediate infiltration capacity.


View this table:
[in this window]
[in a new window]
  		Table 3. Soil moisture probe information, Field B.

 
A drilling rig was used to install the access tubes, which in turn were used to house the electronics for monitoring soil water contents. First, a 2-m steel cylinder was inserted into a thin-walled PVC pipe fitted with a 2-cm metal tip. Then, a steel auger fitted with a 5.0-cm tungsten tip was inserted into the steel cylinder. The PVC pipe, steel cylinder, and auger were vertically leveled inside a 6-m-high tripod. The tungsten steel tip protruded about 5 cm beyond the PVC metal tip and cut through sand and gravel horizons and clay lenses. The PVC–steel cylinder–auger assemblage was simultaneously inserted into the soil. All excavated soil was brought up through the center of the PVC–steel cylinder–auger assemblage. When the bottom of the access tube reached 2.5 m, the steel cylinder and auger were detached and pulled from the assemblage leaving the PVC pipe in the soil. Double O-ring plugs were inserted into the bottom of the PVC pipe to keep water from entering the access tube. Because the steel cylinder and auger were inside the PVC during soil excavation, the thin-walled PVC pipe was ensured of having excellent contact with the native soil and avoided the need for back-filling soil around the access tubes.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Ground-Penetrating Radar–Identified Subsurface Restricting Layer
The location of the subsurface convergent flow pathways is dependent on an accurate quantification of the subsurface layers that could restrict water flow. The semivariogram for the depth of the clay lens that is restricting water movement for the entire OPE3 production area is shown in Fig. 2 . The semivariogram is best fit (least squares analysis) to a spherical model that shows a nugget of 0.02 m2, an error term encompassing both sampling error and error due to microscale variability not captured by the sampling scheme. The low nugget value of 0.02 m2 suggests that much of the uncertainty at small scales was captured. The variogram exhibits a range of 76 m, indicating that data acquired at spacings of <76 m apart are related and spatially dependent. Because the largest distance between GPR transects was 25 m, there should be ample data to predict the depth of the clay lens for locations where GPR data were not collected. The sill of 0.20 m2 was close to the population variance of 0.18 m2 (dotted line on Fig. 2). Gish et al. (2002) noted similar spatial structure for the depth to a clay lens. They noted that for a 7.5-ha region, the range was 95 m, the nugget 0.02 m2, and the sill 0.18 m2. As a result, the sampling scheme was sufficient to capture the spatial structure associated with the depth to the clay lens at the field scale.



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 2. Omnidirectional semivariogram of the depth to the restricting layer for the entire Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3) research site. The dotted line represents population variance for the depth to the subsurface restricting layer.

 
The depth to the first continuous GPR-identified clay lens is shown in Fig. 3 . About 60% of the research site has a clay lens located between 1.3 and 1.7 m while 80% of the site has a clay lens shallower than 1.9 m. As a result, the soil moisture probes that extend to 1.8 m may generate representative information for most of the site even though they cover far less than 1% of the production area. At the 10- by 10-m scale of data aggregation (block kriging), the depth to the clay lens ranges from 0.9 to over 2.7 m. As a result, there will be locations where water flow may be occurring below the soil moisture probes.



View larger version (57K):
[in this window]
[in a new window]
  		Fig. 3. Depth to the ground-penetrating radar (GPR)–identified clay lens for the Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3) research site. Dotted lines denote boundaries for Fields A, B, C, and D.

 
Clay Lens Uncertainty
One of the benefits of kriging is that we can obtain standard errors associated with each estimated clay lens depth even where no measurements were taken. Accordingly, the variogram was utilized along with our known values to determine spatial contours of the standard error associated with the GPR predictions of the depth to the first continuous restricting layer (typically a clay lens) for the 3.2-ha Field B (Fig. 4) . The uncertainty of the GPR estimates is strictly a function of the spatial distribution of sampling points. Thus, within Field B those regions with the lowest standard error, <0.15 m, are typically associated with intense GPR measurements (2- by 2-m grids). Most of the field has a standard error of <0.25 m. The depth to the clay lens on the eastern half of the field, the lowest geographic points within the production area, exhibits the most uncertainty with this sampling scheme. On the eastern side of the field, the depth to the clay lens frequently changes as much as 2 m over a 25-m horizontal distance. Because the depth of the clay lens was averaged across each 10- by 10-m cell, the averaging process alone will introduce uncertainty into the exact location of the subsurface flow pathways. Gish et al. (2002) observed that a >1-m change in the depth to the subsurface restricting layer over a 25-m transect was very common. Additionally, the GPR technique used here does not differentiate between a shallow water table and a clay lens with a high water content. Consequently, the horizontal location of the subsurface flow pathways between the 25-m transects may be off several meters and other data sets are needed to support the existence of the GPR-identified flow pathways.



View larger version (64K):
[in this window]
[in a new window]
  		Fig. 4. Spatial distribution of standard error associated with depth to the clay lens in Field B. Spatial contours represent the standard error of the depth to the restricting clay layers. Plus signs (+) denote locations where depth to the clay lens was digitally recorded, and dots represent the field boundary. Flow pathways are denoted by the blue lines and the soil moisture probe locations are denoted with an alphanumeric designation.

 
Ground-Penetrating Radar–Identified Flow Pathways
After subtracting the depth to the restricting layer from the DEM, the subsurface flow network can be calculated. The subsurface flow pathways as calculated from the DEM and GPR data are displayed in Fig. 5 . The brown-scale represents the average elevation of the subsurface layer (height above mean sea level) that could restrict vertical water movement within the field. Kung (1993) and Casper and Kung (1996) successfully demonstrated that GPR could be utilized in determining the spatial locations and continuity of subsurface soil layers that could influence water movement.



View larger version (74K):
[in this window]
[in a new window]
  		Fig. 5. Topography of the subsurface restricting clay layer and location of ground-penetrating radar (GPR)–identified flow pathways. Blue lines denote the flow pathways while the moisture probe locations are designated alphanumerically. HAMSL, height above mean sea level.

 
The GPR-identified flow pathways in Field B appear to take place in two major flow networks exporting subsurface water to the east (right) into a riparian wetland. Although more than 9000 soil moisture readings are collected daily in Field B, far less than 1% of the field area is being monitored. Additionally, funnel flow processes are more likely to be a factor in the subsurface during winter and early spring when evapotranspiration is minimal. Nonetheless, preferential funnel flow is occasionally observed in the early spring and fall if the moisture probe is located near a GPR-identified flow pathway and when the restricting layer is within 1.8 m of the soil surface (Fig. 6a and 6b) . For moisture probe BM3, the probe is located 5 m from a GPR-identified flow pathway and the clay lens is located at 1.54 m below the soil surface. At this location the soil texture above the clay lens is a coarse sand with gravel and so will have a low water holding capacity. During 1998 (Day 126), volumetric water contents changed abruptly at 1.5 m from 0.08 to 0.32 m3 m–3 (near saturation) in less than 24 h, while volumetric water contents at 1.2 m changed slowly even though both depths have the same soil texture. The rapid rise in soil water content at 1.5 m is due in part to the high permeability of the coarse-textured soil. Moisture contents at 1.5 m were nearly twice the moisture contents at 1.2 m during this funnel-flow event (Days 126–155). Thus, the soil water dynamics near a GPR-identified flow pathway support the hypothesis of horizontal water flow along the clay lens interface, a preferential funnel flow process (Kung, 1990, 1993). During the 1999 growing season a large plume of water appears (Day 259) as a consequence of the rains from Hurricane Floyd. Although water contents at 1.2 and 1.5 m rise simultaneously, they remain higher at 1.5 m, indicating a larger plume and potential horizontal water flow (Fig. 6b). Abrupt changes of soil water content like these are common with preferential flow processes (Kung and Donohue, 1990). Relative to the remaining soil matrix, these latter authors reported that solution volumes and chemical concentrations were more than 400% larger in the subsurface depression areas identified with GPR. Furthermore, Kung et al. (2000) and Gish et al. (2004) observed that as soil water contents increase, soil becomes increasingly hydrologically active and chemical transit times decrease. The rapid change in soil water content through coarse-textured sand supports the hypothesis of preferential funnel flow near the GPR-identified regions.



View larger version (31K):
[in this window]
[in a new window]
  		Fig. 6. Example of preferential funnel or matrix flow processes on subsurface soil moisture as a function of proximity to a ground-penetrating radar (GPR)–identified flow pathway. Soil moisture content was determined at 1.2-, 1.5-, and 1.8-m depths. Arrows indicate precipitation events of >5 mm. The BM3 and BM1 soil moisture probes are 5 and 15 m away from GPR-identified flow pathways, respectively.

 
Subsurface soil water contents far away from the GPR-identified flow pathways for the same time frame are shown in Fig. 6c and 6d. For this soil moisture probe, BM1, the restricting clay layer is located 1.6 m below the surface and is 15 m away from a GPR-identified flow pathway. At this location, the soil above the clay lens is a sandy loam, so relative to the BM3 probe location the BM1 location will have a higher soil water holding capacity and higher soil water contents for the same environmental conditions. However, the BM1 location still has a high permeability so rapid increases in water content should be evident if subsurface funnel flow occurs. As expected, the moisture probe at this location exhibited higher soil water contents at 1.5 m. However, it never approached saturation (here, about 0.36 m3 m–3) but instead maintained a fairly constant soil water content, regardless of major precipitation events. Several rain events occurred in August 1999, which increased the subsurface water contents by only a few percent. No significant rise in soil moisture is observed at 1.2, 1.5, or 1.8 m even though Hurricane Floyd (Day 259) generated more than 200 mm of water on the watershed during the late summer of 1999. Because there was no increase in subsurface soil water contents during large rain events, infiltrating water must have bypassed this location. Furthermore, soil water content changes at Day 242 in 1999 occurred gradually. Thus, no evidence of preferential flow is observed where the moisture probe is located 15 m from a GPR-identified flow pathway.

Confirmation of Subsurface Flow Pathways
Confirmation of subsurface water movement through preferential flow pathways can be difficult due to (i) the spatial and temporal dynamics of water movement, (ii) the relatively small volumes of soil used in preferential transport, and (iii) the relatively small volume of soil evaluated by the soil moisture sensors. Nevertheless, two independent sets, crop yield during a drought year and averaged surface soil moisture, will be presented as evidence in support of the GPR-identified subsurface flow pathway locations.

The relationship between yield, surface elevation, and organic matter is shown in Fig. 7 . On this site organic matter content or elevation are not well correlated with yield (i.e., coefficients of determination are all <0.05 for any of these relationships) regardless of year. However, there is structure to the yield patterns as the spatial correlation of corn grain yield for a mild drought (1998) and a severe drought (1999) is high, r2 = 0.8.



View larger version (51K):
[in this window]
[in a new window]
  		Fig. 7. Relationships between yield and organic matter content (left) and yield and surface elevation (right) over a 3-yr period. HAMSL, height above mean sea level.

 
Evidence from Corn Grain Yields during a Drought
The amount and timing of precipitation can influence corn grain yield. Averaged 30-yr rainfall amounts for this region are about 19 mm wk–1 during the growing season. For this 3-yr period (1998–2000) climatic conditions influencing plant growth were quite variable, as summarized in Table 2. The first year of this study (1998) was considered to be a mild drought while 1999 was a severe drought. Although soil water was abundant in 2000, rain events did not start until after the crop was well established and so had little impact on crop stand. During 2000, about 24 rain events, each greater than 4 mm of water, occurred between planting and physiological maturity (R6). Because rainfall was frequent and of low intensity in 2000, water rarely ponded on the surface. However, during 2000 there were areas within Field B that had subsurface soil water contents close to saturation, which could have had a detrimental impact on yields.

Because climatic conditions were quite variable between 1998 and 2000, it is not surprising that corresponding corn grain yields were variable (Fig. 8) . Corresponding variograms of corn grain yield are shown next to the spatial contour maps of Field B for each of the three years. The variograms were similar for all three years, generating a range of about 120 m, and a sill and population variance that are close (dotted lines denote the population variance). During 1998, yields ranged from <1.2 to more than 9.4 Mg ha–1 with a median yield of about 4.0 Mg ha–1. During 1999, yields ranged from 0 to 6.9 Mg ha–1, but the median yield was less than 1.2 Mg ha–1. The drought conditions and low soil water contents during 1998 and 1999 were probably responsible for the low yields. Eck (1984) observed that 14 and 28 d of stress during the vegetative stages of growth reduced corn yields by 23 and 46%, respectively. The yield difference between 1998 and 1999 is also supported by Musick and Dusek (1980), who observed that stress during tasseling and silking stages was more detrimental to corn grain yield than during vegetative stages. During the 2000 growing season, corn grain yields varied from 1.5 to 12.5 Mg ha–1, with >80% of the watershed having yields of >9.4 Mg ha–1. As a result, water was probably the single most critical factor limiting yield during 1998 and 1999.



View larger version (55K):
[in this window]
[in a new window]
  		Fig. 8. Spatial corn grain yield patterns and their corresponding variograms for 1998 through 2000 growing seasons. On the left, darker regions indicate higher-yielding regions within the field while dotted lines on the right denote population variances. Dashed lines represent corn grain yield variance.

 
Determining the impact of GPR-identified subsurface flow pathways on yield will be spatially complex because the subsurface flow pathways vary spatially (horizontally) as well as vertically (function of the depth to the clay lens). To tackle this spatial analysis issue, three buffer zones were identified in a GIS framework as a function of the distance from a GPR-identified flow pathway: (i) 0 to 5, (ii) 5 to 10, and (iii) 10 to 15 m. For example, an illustration of the spatial areas being evaluated for each buffer zone within Field B is shown in Fig. 9 .



View larger version (43K):
[in this window]
[in a new window]
  		Fig. 9. Schematic of regions representing three buffer zones of increasing distance from the ground-penetrating radar (GPR)–identified flow pathways. Shaded blue regions indicate soil areas for each specified buffer zone and black lines denote GPR-identified flow pathways.

 
To determine the impact of flow pathway proximity and depth to the clay lens on yield, averaged corn grain yields were calculated for 2- by 2-m cells for the entire production area within the research site. Yield monitor data were geostatistically analyzed and then block kriged. The impact of vertical distance to the clay lens on averaged yield is shown as Fig. 10 . The standard error of the mean shows the first data point from each buffer zone, representing a depth to the GPR clay lens of 1.4 to 1.6 m, as significantly different. Additionally, these same three data points represent 34% of the total crop production area (dark gray shaded area in Fig. 10). This data set also indicates a significant reduction in corn grain yield when the GPR-identified clay lenses are below 2 m from the soil surface. Typically, corn plants are believed to extract soil water from depths shallower than 2 m. However, it must be recognized that the 2-m depth represents a depth whereby subsurface water flows along a restricting layer and not the depth of active water uptake by roots. Capillary action from deep soil layers will replenish shallow soil water extracted by plants. Thus, the subsurface flow network can behave like a subsurface irrigation system.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 10. Impact of the ground-penetrating radar (GPR)–identified clay lens depth on the averaged corn grain yield from three zones representing increasing distance from GPR-identified subsurface flow pathways. The gray-scale denotes the percentage of the Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3) site covered by this analysis. Yield data from all four fields are presented in this analysis.

 
To more accurately determine the impact of horizontal distance from the GPR-identified subsurface flow pathways on yield, depths of >2 m to the clay lens were excluded. Yields decrease in almost a linear fashion with increasing distance from the GPR-identified flow pathways (Fig. 11) . These three yield averages represent more than 85% of the small watershed. Although the spatial structure of the GPR-identified flow pathways is complex, the closer the corn plants are to these flow pathways, the greater the yield during a drought year.



View larger version (16K):
[in this window]
[in a new window]
  		Fig. 11. Effect of horizontal distance from flow pathways on averaged corn grain yield. Yield data from regions where the clay lens was >2 m below the soil surface were excluded from this analysis.

 
Evidence from Soil Moisture Observations
Surface soil water contents support the hypothesis that parts of Field B may be more hydrologically active than others. Figure 12 represents averaged soil moisture values as a function of the three buffer zones shown in Fig. 9. Because yield and GPR data suggest that corn grain yields were not influenced by deep GPR-identified flow pathways, soil moisture data from three probes with clay lenses below 2.0 m as well as one probe that was >15 m from a GPR-identified flow pathway were excluded from Fig. 12 (Table 3). The first dotted vertical line represents the planting date while the second dotted vertical line denotes the onset of the reproduction stage (silking).



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 12. Averaged surface soil water content as a function of distance from ground-penetrating radar (GPR)–identified flow pathways. Surface soil moisture data (0.10 m) were averaged for three buffer zones (0–5, 5–10, and 10–15 m) around the GPR-identified flow pathways. The vertical dotted line of the left identifies the planting date while the vertical dotted line on the right denotes when reproductive stage begins.

 
In general, the averaged surface soil moisture values decreased with increasing distance from the GPR-identified flow pathway. The low rainfall, especially during the reproductive stages (Table 2), and sandy-textured soil suggest that surface soil water contents for much of the field would be low during 1998 and 1999. For example, during 1998 the soil water contents in the 0- to 5-m buffer zone averaged from 0.32 to 0.20 m3 m–3 during the vegetative stages of crop growth compared with 0.27 to 0.09 m3 m–3 in the 10- to 15-m buffer zone. After silking, the difference in soil water contents becomes even more dramatic with a difference of at least 0.10 m3 m–3 between the 0- to 5- and 10- to 15-m buffer zones. The yield and soil moisture data for 1998 and 1999 support observations by Denmead and Shaw (1960) who found that yields were especially sensitive to soil moisture stress occurring during the reproductive stages. During 2000, a difference of at least 0.10 m3 m–3 between the 0- to 5- and 10- to 15-m buffer zones was observed again. However, in 2000, the averaged soil water contents in the 0- to 5-m buffer zone are approaching saturation (about 0.38 m3 m–3). As a result, the soil water contents near the GPR-identified flow pathways may be too wet and could, therefore, have a detrimental impact on crop growth and corn grain yields.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Subsurface convergent flow pathways were identified on a field scale using GPR data and DEM. Although limited, soil moisture data support the occurrence of preferential funnel flow in a 3.2-ha field. A buffer zone protocol was introduced, which showed that during a drought year subsurface restricting layers of >2 m below the soil surface had no beneficial impact on corn grain yield. Furthermore, during a drought year corn grain yield decreased as horizontal distance from the GPR-identified flow pathways increased. Averaged surface soil water contents decreased with increasing distance from the GPR-identified flow pathways. These data suggest that funnel flow processes exist and influence corn grain yield and soil moisture contents at the field scale.


   ACKNOWLEDGMENTS
 
This work was supported in part by the CSREES-NRI Grant no. 2001-01091, "Quantification and Evaluation of Subsurface Water Dynamics for Determining Water and Chemical Fluxes on Adjacent Watersheds." The authors would also like to thank the assistance and contribution from Andy Russ, Lynn McKee, Dan Shirley, Peter Buss, Wayne Dulaney, and Rob Parry.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Trade names are included for the benefit of the reader and imply no endorsement by the USDA.


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


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2005 34: 1-6. [Full Text]  



This article has been cited by other articles:


Home page
 Crop Sci.Home page
C. L. Williams, M. Liebman, J. W. Edwards, D. E. James, J. W. Singer, R. Arritt, and D. Herzmann
Patterns of Regional Yield Stability in Association with Regional Environmental Characteristics
Crop Sci., July 1, 2008; 48(4): 1545 - 1559.
[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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gish, T. J.
Right arrow Articles by Kung, K.-J. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gish, T. J.
Right arrow Articles by Kung, K.-J. S.
Agricola
Right arrow Articles by Gish, T. J.
Right arrow Articles by Kung, K.-J. S.
Related Collections
Right arrow Watershed and Landscape Processes
Right arrow Preferential Flow
Right arrow Maize
Right arrow Sustainable Management of the Vadose Zone
Right arrow Vadose Zone Processes and Chemical Transport

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
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome