vvEPA
EPA600/R-22/070 | May 2022 | www.epa.gov/research
United States
Environmental
Protection Agency
Using Vadose Zone
Instruments and Post-
Harvest Soil Nitrogen
To Evaluate Nutrient Transport
at an Agricultural Field Site
-------
EPA/600/R-22/070
May 2022
www.epa.gov/research
Using Vadose Zone Instruments and Post-Harvest
Soil Nitrogen to Evaluate Nutrient Transport at an
Agricultural Field Site
by
Stephen R. Hutchins1^, Julie N. Weitzman2, J. Renee Brooks3, Jana E. Compton3, Barton R. Faulkner1,
Susanna L. Pearlstein4, Ronald E. Peachey5, Mark V. White1, and Robert A. Coulombe6
1US Environmental Protection Agency
Office of Research and Development
Center for Environmental Solutions and Emergency Response
Groundwater Characterization and Remediation Division
Ada, OK 74820 (^deceased)
2ORISE Postdoctoral Fellow
US Environmental Protection Agency
Office of Research and Development
Center for Public Health and Environmental Assessment
Pacific Ecological Systems Division
Corvallis, OR 97333
3US Environmental Protection Agency
Office of Research and Development
Center for Public Health and Environmental Assessment
Pacific Ecological Systems Division
Corvallis, OR 97333
4FormerORISE Research Fellow
US Environmental Protection Agency
Corvallis, OR
5Oregon State University
College of Agricultural Sciences
Department of Horticulture
Corvallis, OR 97331
6CSS
Corvallis, OR 97333
-------
Notice/Disclaimer
The US Environmental Protection Agency, through its Office of Research and Development, funded and
conducted the research described herein under an approved Quality Assurance Project Plan (Quality
Assurance Identification Number G-GWERD-0030111-QP-1-7). It has been subjected to the Agency's peer
and administrative review and has been approved for publication as an EPA document. Approval does not
signify that the contents reflect the views of the Agency, nor does mention of trade names, commercial
products, or services constitute endorsement or recommendation for use.
Questions concerning this document, or its application should be addressed to:
Dr. Ann Keeley, Division Director
US Environmental Protection Agency
Office of Research and Development
Center for Environmental Solutions and Emergency Response
Groundwater Characterization and Remediation Division
Robert S. Kerr Environmental Research Center
Ada, OK 74820
Phone: 580-436-8890
Email: keeley.ann@epa.gov
11
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the Nation's land,
air, and water resources. Under a mandate of national environmental laws, the Agency strives to formulate
and implement actions leading to a compatible balance between human activities and the ability of natural
systems to support and nurture life. To meet this mandate, EPA's research program is providing data and
technical support for solving environmental problems today and building a science knowledge base
necessary to manage our ecological resources wisely, understand how contaminants affect our health, and
prevent or reduce environmental risks in the future.
The Center for Environmental Solutions and Emergency Response (CESER) within the Office of Research and
Development (ORD) conducts applied, stakeholder-driven research and provides responsive technical
support to help solve the Nation's environmental challenges. The Center's research focuses on innovative
approaches to address environmental challenges associated with the built environment. We develop
technologies and decision-support tools to help safeguard public water systems and groundwater, guide
sustainable materials management, remediate sites from traditional contamination sources and emerging
environmental stressors, and address potential threats from terrorism and natural disasters. CESER
collaborates with both public and private sector partners to foster technologies that improve the
effectiveness and reduce the cost of compliance, while anticipating emerging problems. We provide
technical support to EPA regions and programs, states, tribal nations, and federal partners, and serve as the
interagency liaison for EPA in homeland security research and technology. The Center is a leader in
providing scientific solutions to protect human health and the environment.
Gregory Sayles, Ph.D., Director
Center for Environmental Solutions and Emergency Response
111
-------
This report is dedicated to the memory of Stephen R. Hutchins for his thoughtful leadership in
designing, conducting, and analyzing this study, andfor his broader contributions to groundwater
science and management.
iv
-------
Table of Contents
Notice/Disclaimer ii
Foreword iii
List of Figures vi
List of Tables viii
Abbreviations ix
Executive Summary x
1.0 Introduction 1
2.0 Materials and Methods 2
2.1 Site Description and Monitoring Network 2
2.2 Instrumentation Installation 4
2.3 Core Sampling and Analyses 6
2.4 Groundwater and Precipitation Sampling and Analyses 6
2.5 Lysimeter Pore Water Sampling and Analyses 7
2.6 Metrics for N budget 8
2.7 Bromide Tracer Study 9
2.8 Additional Data Collection 10
2.9 Vadose Zone and Groundwater Modeling 11
2.10 Monitoring and Field Operations Schedule 16
2.11 Subfield Comparison Considerations 18
3.0 Results 19
3.1 Crop Performance 19
3.2 Water Transport Through the Vadose Zone 21
3.3 Water Quality Trends 27
3.4 Post-Harvest Soil Data 37
4.0 Quality Assurance 40
5.0 Literature Cited 42
Appendix A. Soil Water Characteristic Curves 47
v
-------
List of Figures
Figure 1. Aerial view of the two subfields at OSU's Vegetable Research Farm field site, showing well
transects with interior and perimeter wells 3
Figure 2. Schematic of monitoring network at the two subfields at OSU's Vegetable Research Farm field site,
showing locations of wells, soil moisture probes, tensiometers, and lysimeters 3
Figure 3. Schematic of monitoring network at the two subfields at OSU's Vegetable Farm Research site,
showing location and orientation of vadose zone devices 4
Figure 4. HYDRUS calibration results for north subfield 12
Figure 5. HYDRUS calibration results for south subfield 13
Figure 6. Plan view of MODFLOW-USG conceptual model showing boundaries and internal features 14
Figure 7. Stratigraphic column for the study site showing range of thicknesses and assigned hydrogeologic
units (not to scale) 15
Figure 8. Active three-dimensional MODFLOW-USG model discretization. Vertical exaggeration: 5x 16
Figure 9. OSU Vegetable Research Farm a) irrigation and rainfall data during field study, and b) water table
response during same time interval. Data show 10/18/16-12/23/19 21
Figure 10. Location of top of gravel/rock matrix and relation to well screened interval 22
Figure 11. Lithology relative to sampling depth for center wells and lysimeters 22
Figure 12. Flux profiles for the north and south subfields for three wet season and two dry season dates,
corresponding to lysimeter sampling dates. Water table depth is represented by inverted triangles... 23
Figure 13. Dynamics of water isotopes (6180) over three water years for precipitation (top, light blue
circles) and lysimeters at 2.5-ft (yellow symbols), 5-ft (red symbols) and 10-ft (brown symbols) depths.
Lines represent input waters with the blue line representing cumulative precipitation over the water
year (starting Oct 1). The red line represents irrigation 6180 values, the green line is the 6-month
running mean weighted by precipitation amount, and the black line is the well water 6180 values
measured in ORN 24
Figure 14. HYDRUS predictions of bromide tracer breakthroughs in a) north subfield and b) south subfield.
25
Figure 15. Lysimeter bromide concentrations at a) north plot north lysimeter, b) north plot east lysimeter,
c) north plot south lysimeter, d) south plot north lysimeter, e) south plot east lysimeter, and f) south
plot south lysimeter. Yellow triangle indicates date of bromide application 26
Figure 16. Groundwater flow contours based on MODFLOW results 26
Figure 17. a) Nitrate and b) phospate concentrations in two center wells. ORN is the well in the north
subfield which was later interseeded with cover crops starting in Jul WY17. Dashed boxes indicate
intervals where it is hypothesized that denitrification is occurring in ORS 27
Figure 18. a) ORS well nitrate versus water table elevation (the 04/17/18 sample date is highlighted) b)
change in water level gradient across site during 04/17/18 event 27
Figure 19. Center well a) nitrate, b) 6180-H20, and c) 615N-N03 profiles 28
Figure 20. Effect of sample depth on a) nitrate, b) 615N-N03, c) chloride, and d) 6180-H20 profiles in ORN
and ORS wells 29
Figure 21. Stable nitrate nitrogen isotope data for the two center wells, ORN and ORS. The first panel shows
the time series of nitrate (mg L-l) and 615N-nitrate (%o) from WY17-WY19 (and beginning of WY20).
Black arrows represent timing of fertilizer additions. Rectangles capture time periods identified as
showing strong negative trends between nitrate concentrations and 615N-nitrate values, which are
plotted linearly in the second panel. The dual isotope graph of 615N-nitrate (%o) vs 6180-nitrate (%o) in
vi
-------
the third panel distinguishes whether identified periods of decreasing nitrate concentrations with
increasing 615N-nitrate values are due to distinct nitrate source signatures (and thus potentially
represents times of source-mixing) or microbial processing (i.e., times in which denitrification
dominates) 30
Figure 22. Lysimeter nitrate concentrations at a) north plot north lysimeter, b) north plot east lysimeter, c)
north plot south lysimeter, d) south plot north lysimeter, e) south plot east lysimeter, and d) south plot
south lysimeter 31
Figure 23. Mean lysimeter nitrate profiles in a) north and b) south subfields. The north subfield has the
cover crop. Triangles indicate dates of fertilizer application 31
Figure 24. Lysimeter water isotope values (6180) and nitrate concentrations during the wet winter/spring
periods (Feb-May) in WY17 and WY19. Triangles and circles represent north and south subfields,
respectively, at 2.5-ft (yellow symbols), 5-ft (red symbols) and 10-ft (brown symbols) depths 32
Figure 25. Stable nitrate nitrogen isotope data for all 2.5-ft and 10-ft lysimeters in both the north (left) and
south (right) subfields. The first panel shows irrigation (gray bars) and rainfall (black bars) values (in)
over time from WY17-WY19 (and beginning of WY20). Black arrows represent timing of fertilizer
additions. In the second and third panels changes in 615N-nitrate (%o) over time in the three replicate
2.5-ft lysimeters and 10-ft lysimeters, respectively, are shown for the two subfields. Like Figure 18,
rectangles capture time periods identified as showing strong negative trends between nitrate
concentrations and 615N-nitrate values. Dual isotope data of 615N-nitrate (%o) vs 6180-nitrate (%o)
for each lysimeter (not shown) was used to determine whether periods of decreasing nitrate
concentrations with increasing 615N-nitrate values were attributable to distinct nitrate source
signatures (and thus potentially representing times of source-mixing) or microbial processing (i.e., times
in which denitrification dominated) 33
Figure 26. Lysimeter phosphate concentrations at a) north plot north lysimeter, b) north plot east lysimeter,
c) north plot south lysimeter, d) south plot north lysimeter, e) south plot east lysimeter, and d) south
plot south lysimeter 34
Figure 27. Mean lysimeter TKN profiles in a) north and b) south subfields. The north subfield has the cover
crop. Triangles indicate dates of fertilizer application 35
Figure 28. Mean lysimeter total phosphate profiles in a) north and b) south subfields. The north subfield has
the cover crop. Triangles indicate dates of phosphorus fertilizer application 35
Figure 29. Average daily nitrate loss with depth for each subfield 36
Figure 30. Annual nitrate-N mass loss with depth for each subfield 37
Figure 31. Post-harvest data for soil nitrate (a,d,g,j), soil ammonium (b,e,h,k), and soil total nitrogen (c,f,i,l)
during each year of the study for the north (interseeded) and south (conventional) subfields. Note that
the north subfield was not interseeded in WY16 38
Figure 32. Post-harvest data for soil phosphorus (a,c,e,f) and soil organic carbon (b,d,f,h) during each year
of the study for the north (interseeded) and south (conventional) subfields. Note that the north
subfield was not interseeded in WY16 39
Figure A-l. Observed water content versus capillary pressure head. Scatter plot color corresponds to soil
texture as the color mixture, red-green-blue, with red being sand percent, green as silt percent, and
blue as clay percent. The continuous blue line represents the van Genuchten equation curve used for
the HYDRUS soil layer, and black dots along the curve represent points used during the HYDRUS
simulation. Also shown are the parameters used for the equation 48
-------
List of Tables
Table 1. Derived dispersivity values 14
Table 2. Field operations in the north and south subfields at the OSU Vegetable Research Farm 17
Table 3. Cover crop dry matter on north subfield that was planted to a cover crop 19
Table 4. Annual nitrogen (N) fluxes (inputs and exports) across fields (north vs south) and soil depths (2.5,
5.0, and 10.0 ft) by water year (WY) 20
Table 5. Sweet corn ear yield in north (cover cropped) and south subfields 20
Table 6. Texture analysis for cores taken at perimeter well locations 23
viii
-------
Abbreviations
BGS Below Ground Surface
CAFO Concentrated Animal Feeding Operation
CL Chloride
COND Conductivity
DO Dissolved Oxygen
EQBLK Equipment Blank
FLDBLK Field Blank
FPLSD Fisher's Protected Least Significant Difference
FT Foot
GCRD Groundwater Characterization and Remediation Division
ICP-OES Inductively-Coupled Plasma-Optical Emission Spectrometry
ID Inner Diameter
IN Inch
ISIRF Integrated Stable Isotope Research Facility
MCPA Dimethylamine salt of 2-methyl-4-chlorophenoxyacetic acid
MSL Mean Sea Level
MWL Meteoric Water Line
NH4-N Ammonium Nitrogen
N02-N Nitrite Nitrogen
N03-N Nitrate Nitrogen
N03&N02-N Nitrate plus Nitrite Nitrogen
OD Outside Diameter
0-P04-P Phosphate Phosphorus
ORP Oxidation-Reduction Potential
OSU Oregon State University
PESD Pacific Ecological Systems Division
RO Reverse Osmosis
RSKERC Robert S. Kerr Environmental Research Center
S04 Sulfate
TKN Total Kjeldahl Nitrogen
TOC Total Organic Carbon
TP Total Phosphorus
WY Water Year (October-September)
IX
-------
Executive Summary
The original objective of this project was to test different methods for measuring water and nutrient
transport throughout the vadose zone underlying a sweet corn crop grown in Corvallis, Oregon, from 2016-
2020. A secondary objective was to evaluate the impact of different management strategies on water and
nutrient fluxes within the field. Specifically, a pilot study was implemented to determine the feasibility of
approach for comparing the impact of interseeding a cover crop into the sweet corn crop versus
conventionally managing the sweet corn, without a cover crop. However, large spatial variation at the field
site, and lack of replication among the two management strategies prevented us from evaluating the
influence of interseeded cover crops. Overall, our study led to a deeper understanding of how nutrients
and water are transported through the vadose zone, below the roots to the groundwater (Weitzman et al.,
2022). This report presents detailed results from different measurement methods (monitoring wells,
vadose zone lysimeters, post-harvest soil tests) that can be utilized when trying to understand hydrologic
nitrate transport in the vadose zone and groundwater.
A two-acre field was uniformly fertilized and planted with a sweet corn crop and monitoring networks were
then installed in two subfields (north and south) consisting of groundwater wells and vadose zone
instrumentation installed near the center of each. The monitoring system for both subfields consisted of a
center groundwater well and three replicate lysimeters (n = 3) at three depths each - 2.5, 5, and 10 ft
below ground surface (BGS). Replicate soil moisture probes (n = 3) and tensiometers (n = 3) were also
included at the three depths in the north subfield only. A precipitation collector device was constructed
and installed in the south subfield to collect precipitation with minimal evaporative losses for stable isotope
analyses of water (6H2O). In addition to the center monitoring system within each subfield, eight total
perimeter wells (four in each subfield) were later installed around the field; all wells were screened at 15-
35 ft BGS. Groundwater and lysimeter pore water samples were collected every two weeks and analyzed
for a suite of parameters, including standard water quality measures, nutrients, cations, and stable isotopes
of water (6H2O) and nitrate (6NO3). Post-harvest soil cores were obtained each year in September 2016-
2019 across incremental depths to 10 ft at six locations around the central monitoring system of each
subfield and analyzed for nutrients. The field was irrigated throughout the summer growing season through
fixed irrigation lines equipped with sprinklers. In June 2019 a bromide tracer test was conducted for one of
these irrigation events through the established irrigation system to track the movement of irrigation water
through the vadose zone.
Sweet corn was planted in either June or July each year from WY16-WY20 and two different crop
management strategies (cover crop versus conventional) were implemented on the subfields in June 2017
and repeated through 2020. Briefly, the north subfield was interseeded with a cover crop (barley, tall
fescue, or triticale depending on year) in late summer as a pilot study for feasibility of approach, while the
south subfield continued to be conventionally managed for sweet corn only, without a cover crop. Both
subfields were fertilized with nitrogen (N) two times during the growing season and spray-irrigated with
well-collected groundwater approximately every week throughout the summer months. Rates of urea-N
application were determined based on yearly pre-sidedress soil nitrate tests (PSNT), and in 2019 and 2020
were increased 25% above the calculated optimum to assess whether increased fertilizer N inputs might
lead to increased nitrate leaching losses. Cover crops were drilled (interseeded) between corn rows in the
north subfield when corn growth reached the V6 stage. Corn was harvested each fall of the study when
kernel moisture was 72-75% of total kernel weight using a commercial picker to mimic corn ear removal
x
-------
from the field as would occur in production fields. Cover crops were terminated in the spring by spraying
with herbicide ~1-1.5 months before new planting, with remaining dead residue plowed down every year.
Corn yield varied across the study years, ranging from 8-16 ton/acre ("as-harvested" fresh weight). Overall,
the corn yields for most years of the study were above average for western Oregon where grower yields
tend to have a mean of ~10-12 ton/acre ("as-harvested" fresh weight) (Sullivan et al., 2020b). Cover crop
production also varied across the study years, ranging from 17-72 lb N/acre in dry matter (biomass) prior to
termination in the spring (Table 3). When the cover crops established successfully (WY2018 and WY2019),
they performed as expected for western Oregon where interseeded cereal/grass cover crops usually
contain ~40-80 lb N/acre in dry matter (biomass) when established successfully in the summer/fall and
killed in early April (Sullivan et al., 2020a).
The field site was unexpectedly complex belowground, with extensive site heterogeneities throughout the
vadose zone as evidenced by the variance in nutrient concentrations between replicate lysimeters and
wells, and indications of significant impacts from preferential flow paths as shown by the bromide tracer
test and 6H2O data. Such variance complicated data interpretation regarding the effects of water and
nutrient transport at the site. However, while such variability did exist, a clear trend emerged with respect
to nitrate leaching - annual nitrate losses generally decreased with depth across the field. Losses averaged
~162 lbs N/acre/yr near the surface (2.5 ft BGS) versus ~51 lbs N/acre/yr at depth (10 ft BGS), a 68%
reduction in leaching between 2.5 and 10 ft BGS. Greatest leaching losses occurred during the wet fall and
winter, and on average the two seasons accounted for ~75% of total nitrate losses at all three depths.
When all nitrogen inputs and outputs were taken into account, our study revealed that N was lost in excess
of inputs at the 2.5-ft BGS depth, but ultimately accumulated above the 10 ft BGS depth, indicating
nitrogen storage or removal between the surface and deeper horizons, as the 6NO3 data revealed
denitrification is not a major loss pathway of nitrate at the study site. Most nutrient leaching studies focus
sampling efforts on relatively shallow depths (3 ft BGS or less), but here we see that only approximately
30% of surface nitrate leaching moves below 10 ft BGS into the deeper soil and groundwater. A better
understanding of N storage and removal mechanisms is needed to assess the effectiveness of management
practices that aim to reduce its delivery to groundwater.
Overall, this project shows that long-term research to intensively monitor water and nutrient transport
across the vadose zone can be complicated by site heterogeneities. The use of several different
methodologies can help tease apart potential interactions, but it is clear that an abundance of within-field
and across-field replication is necessary to properly evaluate the complex relationship of water and
nutrient transport in the vadose zone. A long-term, replicated study of the impact of cover crops in sweet
corn systems, as well as within other cropping systems, could greatly benefit from the utilization of the
methods presented in this study. Intensive monitoring of the vadose zone across field sites will ultimately
help land managers to more effectively evaluate their efforts to improve water quality.
XI
-------
1.0 Introduction
In U.S. public drinking water supplies, nitrate has been a dominant contaminant responsible for violations
of the Safe Drinking Water Act (Pennino et al., 2017). Fertilizer application can cause nitrogen accumulation
in agricultural soils, eventually moving nitrate below the rooting zone and into groundwater systems via
rainfall and irrigation. Groundwater nitrate concentrations tend to be highest in aquifers underlying
agricultural areas across the U.S. (Burow et al., 2010). Residents in rural areas often rely on private wells,
where drinking water safety is not ensured by the Safe Drinking Water Act. In Oregon, this is particularly
the case in agricultural valleys where state agencies established Groundwater Management Areas to
address such issues (Kite-Powell and Harding, 2006; Hoppe et al., 2014).
Agricultural practices aiming to improve nitrogen management and reduce nitrate leaching to groundwater
have received a great deal of attention in crop and soil science (Raun and Johnson, 1999; Dinnes et al.,
2002; Sela and van Es, 2018). Recommended practices for western US agriculture include 1) better
management of irrigation timing and amount, 2) appropriate timing, placement, rate, and source of
fertilizers, 3) improvements to soil health, and 4) establishment of cover crops during the non-growing
season to ensure better nitrogen uptake and minimize leaching.
Cover crops are grown to reduce soil erosion and nutrient losses through runoff and leaching, as well as
other benefits to soils and agroecosystems and can reduce N leaching in many situations (Tonitto et al.,
2006; Delgado et al., 2007; Dabney et al., 2010). Cover crops or relay/interseeded crops may be particularly
effective in the Pacific Northwest because much of the rainfall, and in turn hydrologic N export, occurs
during the fall and winter (Burket et al., 1997; Lin et al., 2019; Compton et al., 2021). In Oregon's
Willamette Valley, Feaga et al. (2010) studied the influence of winter cover crops (cereal rye, triticale, and a
vetch/triticale mix) on nitrate leaching for 11 years in summer vegetable crops. Cover crops significantly
reduced nitrate leaching and led to soil N accumulation that could be used by the next vegetable crop. In
their study, early establishment of the cover crop increased the overall nitrogen removal, and precipitation
timing and amount controlled the rates of nitrogen leaching.
Feaga et al. (2010) noted that despite the reductions in nitrogen leaching associated with cover crops, soil
solution nitrate concentrations could still exceed the drinking water standard at their sampling depth of 4 ft
below ground surface (BGS) when a high fertilizer rate for corn (200 lb N ac"1) was used. However, leaching
studies often do not sample below a 3-ft BGS depth (Valkama et al., 2015). Thus, a better understanding of
nitrogen dynamics through the vadose zone is needed to assess the effectiveness of management practices
aiming to reduce nitrate leaching to groundwater and drinking water supplies.
In addition to the use of cover crops as a land management strategy, some states are using post-harvest
soil nitrate values as a screening tool to identify fields at highest risk of nitrate loss under Concentrated
Animal Feeding Operations (CAFOs) management. The idea is that any residual nitrate within and below the
root zone after harvest will be subject to leaching, and if the post-harvest nitrate values are excessive (e.g.,
>20 ppm nitrate-nitrogen for silage corn), then there is a greater risk of groundwater contamination. The
post-harvest test tends to be most useful in identifying nitrogen excess that originates from a) excess
nitrogen application (especially as manure or other organic amendments), b) late season application of
nitrogen fertilizer, c) poor crop growing conditions, or d) water limitations during the late summer which
can favor soil nitrate accumulation near the soil surface (Sullivan et al., 2021). While the post-harvest
1
-------
nitrate test is not intended to predict groundwater nitrate concentrations, it may offer insight into fields at
higher risk of nitrate leaching (Carey and Harrison, 2014; Carey et al., 2017). University extension services
often promote the use of post-harvest soil nitrate sampling as a voluntary effort to help reduce the costs
associated with fertilizer and minimize groundwater impacts. However, there has been little rigorous
testing to determine whether these practices are truly predictive of nitrate leaching rates under agricultural
management regimes that rely on commercial fertilizers and summer irrigation, like that needed for corn
growth in Western Oregon.
As discussed, leaching losses of fertilizer nitrogen represents a cost for farmers and has consequences for
human health and the environment. This is especially the case in the southern Willamette Valley, Oregon,
where groundwater nitrate contamination is prevalent. To better understand issues regarding seasonal and
annual nutrient loss in the region, a four-year field study was undertaken from 2016 to 2020 in Corvallis, OR
to evaluate water and nutrient transport throughout the vadose zone underlying a sweet corn crop
subjected to varying land management strategies. The study utilized an array of monitoring techniques,
including the use of groundwater monitoring wells, vadose zone instrumentation (e.g., lysimeters, soil
moisture probes, and tensiometers), and the collection of post-harvest soil cores.
This report focuses on presenting the details of the many methodologies employed throughout the study
and the corresponding results, while interpretations and recommendations are kept to a minimum. Our
hope is that the report be used as a resource from which other researchers can draw when designing their
own network-based vadose zone study. Mining data from the collected results can also benefit other
researchers as they pursue similar projects to understand the complexity of vadose zone soil-water
interactions in the future.
2.0 Materials and Methods
2.1 Site Description and Monitoring Network
The approximately 2 ac field site is located at Oregon State University's (OSU) Vegetable Research Farm in
Corvallis, Oregon. Using a site that was part of the OSU Vegetable Farm allowed for more control over the
study design, as well as over management decisions. Having guaranteed access to the site was especially
important given the long-term (four years) nature of the study. Given the site was on the experimental
farm also meant the field could be fully instrumented in a manner that fit the needs of the project. The
specific field within the OSU Vegetable Farm (one of multiple fields within the facility) was chosen based on
recommendations from collaborators with extensive experience working at the university-owned location.
Located ~400 m east of the Willamette River in Linn County, the field site is typical of the agriculturally
productive Willamette Valley, with a relatively flat-lying, low-relief terrain that is largely comprised of the
fine-textured alluvial-derived Chehalis (Cumulic Ultic Haploxerolls) silty clay loam soil series in the upper
soils (most especially within the top 1.5 m). Below the silty clay loam layer, silt loam and loam layers
intermix, with the layers about twice as thick in the south subfield as compared to the north subfield. Sandy
loam and loamy sand make up the bottom of the soil profile, below which a gravel/rock matrix dominates,
which appears between depths of 4.3 to 6.7 m below the ground surface across the site.
2
-------
The field was planted in sweet corn in June
2016. The field was again planted in sweet corn
in June 2017 but was then divided into two
subfields (north and south) (Figure 1) so one
subfield could serve as a pilot study for
feasibility of approach in regard to interseeding
with a cover crop. One subfield (south)
remained planted in corn as before, and the
other subfield (north) was planted with corn
interseeded with a cover crop (additional
details are provided in Section 2.9). The
monitoring network for this site includes
monitoring of groundwater and precipitation as
well as vadose zone water and properties.
*9
*
*ff | *
X
*
Transect #3
Ja
Transect tf2
Transect #1
North SuMUld
(Intwiccdcd)
:.7.
S
-------
with three sets of triplicate lysimeters installed at those same different depths (2.5, 5, and 10 ft). Thus, a
total of six lysimeters were established at each soil depth across the entire field site. Tensiometers and soil
moisture probes were not installed in both subfields due to cost restraints. A precipitation collection device
was mounted in the south subfield to collect precipitation for isotopic analysis. Eight additional monitoring
wells were installed in September 2017 around the field and were included in the monitoring program to
better determine groundwater flow direction and potential additional nutrient inputs (Figure 1).
2.2 Instrumentation Installation
Monitoring wells were installed using the GCRD 8140 Sonic Geoprobe which created 6-in OD boreholes.
Cores were continuously collected during this process using 3-in OD sleeves. The completed wells consisted
of 2-in ID PVC risers and a series of 5-ft screens equipped with 2-in polyester filter socks (250 nm) to
provide 20 ft of screened interval, with the screened intervals at 15-35 ft BGS at all locations. Sand was
added to provide a sand pack extending at least 2 ft above the top of the screen, and then bentonite pellets
were added to within 2 ft of the surface. For the two center wells, concrete was poured to provide a
surface seal tied into a 3-ft x 3-ft concrete pad encompassing a lockable steel protector, and steel pipe
bollards were installed around the wells. However, for the perimeter wells, the well tops were flush-
mounted to allow surface passage of vehicles and agricultural equipment. These wells were completed
BGS, and traffic-rated manholes were installed within a concrete base. These flush-mount wells were not
surrounded by bollards. Each well was developed using a stainless-steel submersible Grundfos pump to
surge and pump the well in an attempt to clear out accumulated solids and help seal the sand pack around
the well.
Vadose zone devices were installed at three different depths for each subfield. The goal was to install the
vadose zone devices at 5 ft, 10 ft, and 15 ft BGS. However, a cobble matrix was discovered 14 ft BGS during
construction of the monitoring well in the north subfield, and it was therefore decided to install all vadose
zone devices at 2.5 ft, 5 ft, and 10 ft BGS at the site. The footprint was 23 ft x 23 ft in each subfield, allowing
the vadose zone devices to be installed perpendicular to the footprint boundary (Figure 3). The vadose
North Subfield
South Subfield
V
DRIVEWAY orn^-IoJ —= TN10E
¦ | ]¦ • * SN10E
~
DRIVEWAY ors—»| ~ 6^
"t
LN2.5E
0—° TN2.5E
# SN2.5E
I LS2.5E
[p] Well
O Lysimeter
Q Tensiometer
# Soil Moisture Probe
Q Lysimeter Manhole Cover
Q Data Logger
¦ Bollard
0 Precipitation Collector
Figure 3. Schematic of monitoring network at the two subfields at OSU's Vegetable Farm
Research site, showing location and orientation of vadose zone devices.
-------
zone devices were installed using the GCRD 6610 Geoprobe to create 3.25-in OD boreholes constructed at
45-degree angles within the immediate vicinity of the monitoring wells. Continuous cores were again
collected, but not sampled, during creation of these boreholes to try to minimize disturbances within the
native matrices and to provide material for backfill. These cores were collected using 1.625-in OD sleeves.
Once the desired depth was attained, a 1.25-in steel rod was used to push out the bottom soil plug and
create a small-diameter hole for the lysimeter or sensor. A 2-in ID PVC tube was then inserted within the
drill rod to provide a flush fit so that the inserted device did not hang up on the lip of the cutting shoe.
The lysimeters consisted of 7.5-in long x 1.2-in OD Prenart® Super Quartz MAXI lysimeters with 0.25-in OD
Teflon tubing. Prior to installation, the Teflon tubing was routed through PA11SF plasticized polyamide
tubing to provide more protection. Each lysimeter was set at the specified depths in the borehole using
1.25-in OD PVC tubing with specially-designed end caps allowing for loose contact with the top of the
lysimeter with the tubing running through the center up to the surface. For each lysimeter, a silica slurry
was prepared by adding 600 g silica to 300 mL reverse-osmosis (RO) deionized water and shaken
vigorously. Part of the slurry was added through the 2-in PVC tubing and then the lysimeter was slowly
pushed into place. The remaining silica slurry was then added using a funnel, and the 1.25-in PVC push rod
was slowly withdrawn, while ensuring that the lysimeter remained in place. A slurry of the native material
was prepared and poured through the 2-in PVC tubing to provide a thin barrier, and then the 2-in PVC
tubing and the drill pipe was removed. The remaining core material was then broken up and pushed down
the borehole to create at least 2 ft of backfill above the installed lysimeter, and then bentonite pellets were
added up to 2 ft BGS. Tap water was used to hydrate the pellets, and then the remaining core material was
used to fill the borehole up to the surface. Shallow trenches 1 ft in depth were excavated so that the
protected lysimeter tubing could be routed from the emergence points into buried steel manholes. Each
steel manhole had two holes drilled into its side to allow entry of the lysimeter tubing and was set to
include two 1-liter lysimeter bottles.
The soil moisture probes were 4-in long x 1.5-in wide x 0.5-in deep Decagon Devices® 5TE soil moisture,
temperature, and electrical conductivity sensors with 50-ft cables. These were installed the same way as
the lysimeters, except that no silica slurry was used. The tensiometers were 3-in long x 0.9-in OD Campbell
Scientific® 257-L Soil Matric Potential Blocks with 50-ft cables. These were permanently affixed to 0.5-in OD
PVC tubing which was used as push rods to set these devices in place. The electrical leads for these devices
were taped to the outside of the push rod, and the top of the push rod was cut off at just BGS. The soil
moisture probe and tensiometer electrical leads were routed through 1-ft deep trenches into a central
large hard plastic water meter manhole box to accommodate the excess electrical line. The ends of these
lines were then routed through a 2-in ID PVC conduit that was connected to a steel electrical box mounted
on a steel pipe set in concrete. Within this box these electrical leads were connected to a Campbell
Scientific® CR1000-ST-SW-NC data logger that was powered by D-sized batteries. All lysimeter lines and
electrical wiring outside of the meter box and manholes were then buried to a depth of at least 6 in BGS.
The precipitation collector was constructed in-house according to guidelines described by Groning et al.
(2012) to prevent isotopic fractionation related to evaporation. Briefly, such a device is designed to collect
precipitation over an extended period of time for stable isotope analysis, with precipitation collecting in a
removable 2-liter bottle with an outlet connected to 30-ft of plastic tubing wound around the outer casing
to minimize kinetic fractionation related to sample evaporation, which would invalidate the isotope
signatures. The collector was constructed of PVC and was mounted on a steel post. Well protectors, data
5
-------
logger boxes, and precipitation collectors were all equipped with locks to help secure access and provide
data integrity. The manhole covers for the flush-mount wells and lysimeters were not lockable, but access
was limited due to the need for specialized large hexagon wrenches for removal of the inset lug bolts.
2.3 Core Sampling and Analyses
As stated earlier, continuous cores were obtained in 5-ft sleeves during construction of the within-field
monitoring wells at the field site. These sleeves were transported back to the EPA's Robert S. Kerr
Environmental Research Center (RSKERC) in Ada, Oklahoma, and subsampled for textural classification
(sand, silt, clay). Near-surface cores were collected in 2016 as the sampling network was installed, and then
annually after each crop harvest for the first three years of the study (2017-2019). For each subfield, six
locations were selected at least 15 ft away from the subfield sampling infrastructure to avoid taking cores
directly above the lysimeter and sensor locations, which could have potentially introduced preferential flow
path interferences. Near surface core samples were taken using the GCRD 8140 Sonic Geoprobe which
created a 4.5-in OD borehole and allowed for the acquisition of a 3-in ID core which was 5 ft in length. An
initial 6-in deep bore hole was created and then the 5-ft core sample was acquired. The 5-ft cores were
subsampled in situ at the field site. A clean hacksaw blade was used to cut out sections corresponding to
0.8-1.2 ft, 2.8-3.2 ft, and 4.8-5.2 ft BGS. These represent subcores taken at 1, 3, and 5 ft BGS. This was done
for the 2016 and 2017 post-harvest soil cores. For the 2018 post-harvest soil cores, an additional core
sample was taken at 2.5 ft BGS to better represent the 2.5-ft lysimeter depths. For the 2019 post-harvest
soil cores, the 3-ft depth sample was eliminated and instead a 10-ft core was obtained at each location to
provide core samples taken at 1 ft, 2.5 ft, 5 ft, and 10 ft BGS. These subcore sections were capped and
transported back to RSKERC for lithological and nutrient analyses. The boreholes were then backfilled
initially with bentonite, followed by native material taken from the unused core sections.
At RSKERC, a clean hacksaw blade was used to cut out sections of interest from the 5-foot cores obtained
during installation of the monitoring wells. These sections, along with the sections obtained during post-
harvest sampling, were emptied into cloth soil sample bags provided by the Soil, Water, and Forage
Analytical Laboratory (SWFAL) located at Oklahoma State University. These samples were then shipped to
SWFAL, without refrigeration or other preservation, for textural classification (% sand, % silt, % clay) as well
as soil pH, nitrate-nitrogen, ammonium-nitrogen, Mehlich 3 phosphate, organic carbon, and total nitrogen
analyses.
2.4 Groundwater and Precipitation Sampling and Analyses
Monitoring wells (as well as lysimeters and precipitation collectors) were sampled every two weeks. Prior
to sampling, all wells were gauged for water levels relative to tops of casings using a 100-ft water level
meter and measurements were recorded to the nearest one-hundredth of a foot. A depth-calibrated 0.25-
in polyethylene sample line attached to a 1-ft stainless steel spacer was lowered into the well so that the
sample line inlet was 2 ft below the measured depth to water. Groundwater was pumped using an Alexis®
battery-powered peristaltic pump with size 24 PharMed® tubing at a rate of 200-300 mL/min. Groundwater
was pumped through a sample filter bypass system prior to an in-line YSI low-volume flow-through cell for
15 min prior to sample collection to flush out the sample system and flow-through cell. The flow stream
was analyzed in-line for pH, Eh, conductivity, temperature, and dissolved oxygen every 5 min using the YSI
field electrodes associated with the flow-through cell. After 15 min, unfiltered samples were obtained and
6
-------
analyzed on-site for turbidity using an Oakton®T-100 turbidimeter according to the manufacturer's
instructions. Unfiltered samples were also analyzed on-site for alkalinity using Chemetrics®9810/9820
alkalinity kits according to the manufacturer's instructions. Afterwards, unfiltered samples were obtained
for laboratory analysis as follows. First, approximately 20 mL was dispensed into a 20-mL plastic scintillation
vial with a conical cap and sealed without headspace for stable water isotope (6180-H20 and 62H-H20)
analysis. This sample was not acidified. Second, approximately 60 mL was dispensed into a 60-mL plastic
bottle for nutrient analysis. This sample was not acidified. Third, approximately 60 mL was dispensed into a
60-mL plastic bottle for acidified nutrient analysis. This sample was acidified to pH < 2 with sulfuric acid.
Fourth, approximately 40 mL was dispensed into each of two 40-ml glass volatile organic analysis vials for
total organic carbon (TOC). These samples were not acidified. After this sampling was completed, filtered
samples were collected. The flow stream was switched to pass through a 0.45-|am groundwater capsule
filter and the first 500 mL of filtrate was discarded. Then approximately 15 mL of filtrate was dispensed into
a 20-mL plastic scintillation vial for stable nitrate nitrogen and oxygen isotope (615N-N03 and 6180-N03)
analysis. This sample was not acidified. Finally, approximately 125 mL of filtrate was dispensed into a 125-
mL plastic bottle for inductively-coupled plasma-optical emission spectrometry (ICP-OES) analysis. This
sample was acidified to pH < 2 with nitric acid. The only samples to be collected and analyzed from the
precipitation collector were the samples for water isotope, nitrate isotope, and acidified nutrients analysis.
Samples were transported back to EPA's nearby Pacific Ecological Systems Division (PESD) lab, where the
isotope samples were separated out for analysis at PESD's Integrated Stable Isotope Research Facility
(ISIRF). The water isotope samples were stored at room temperature and the nitrate isotope samples were
frozen prior to analysis. The other samples were refrigerated at 6°C or less prior to being shipped on ice to
RSKERC. Nutrient and general parameter samples were analyzed by Standard Methods (American Public
Health Association, 2017), whereas metals/cations were analyzed by inductively-coupled plasma-optical
emission spectrometry. The ISIRF lab used an LGR liquid-water isotope analyzer for water isotope analysis,
and a Thermo Electron Gas Bench II Inlet with cryogenic trapping and a Thermo Electron Delta V Isotope
Ration Mass Spectrometer (IRMS) for nitrate isotope analysis following bacterial denitrification (Sigman et
al., 2001; Casciotti et al., 2002).
Beginning in June 2019, this sampling procedure was adjusted for wells ORN and ORS to assess differences
in groundwater quality with depth. For each of these two wells, the well was first sampled 2 ft below the
measured depth to water as described above, and then immediately afterwards the sample inlet was
withdrawn from the well and the water level was again recorded. The sample inlet was lowered to the
bottom of the well and a second sample was obtained as described above. Analysis of recorded water
levels showed insignificant changes between the two sample events for each well, indicating that the water
obtained was representative of the aquifer matrix at that depth.
2.5 Lysimeter Pore Water Sampling and Analyses
The lysimeters were primed (i.e., had a vacuum applied) before each scheduled sample retrieval date.
Priming was done by attaching a line from a battery-powered vacuum pump to the lysimeter bottle outlet
tubing and applying a vacuum of -60 KPa. The lysimeters were primed one week before sample retrieval.
Sample retrieval consisted of disconnecting and capping the partially-filled lysimeter bottles and
transporting them to an adjacent location for subsampling. The partially-filled lysimeter bottles were
replaced with clean lysimeter bottles dedicated to those particular lysimeter locations and the pinch-
7
-------
clamps were closed to ensure that any rainfall that seeped into the manholes would not enter into the
lysimeter bottles. The lysimeter pore water was analyzed for the same series of analytes as was done for
groundwater. However, in most cases there was insufficient volumes collected for the full range of
analyses, and therefore a prioritization protocol was put into place based on the individual volume.
Analytes were prioritized as follows: 1) acidified nutrients (60 mL), 2) nitrate isotopes (20 mL), 3) water
isotopes (20 mL), 4) TOC (80 mL), and 5) ICP-OES (125 mL). Each lysimeter bottle had been previously
calibrated in 50-mL increments, and estimates were made of total volume for each lysimeter bottle to
determine which subsamples would be collected for which analytes. Subsamples were collected and
preserved as described for groundwater sampling, with one exception - because of the low volumes, 0.45-
|am syringe filters were used with 30-mL and 60-mL plastic syringes instead of high-capacity groundwater
filters to provide filtered water for nitrate isotope and ICP-OES analyses. An initial aliquot of 5 mL of sample
was used to flush the syringe filter prior to collecting sample for analyses. Once subsampling was done, the
lysimeter pore water subsamples were handled and analyzed the same way as described for the
groundwater samples.
2.6 Metrics for N budget
A comprehensive N budget for the field site was constructed in which the relationships between different N
input and export components could be evaluated as distinctive indicators of performance across the study
years and soil depths. Inputs included fertilizer N, irrigation water N, and modeled total atmospheric N
deposition, as represented by the equation:
Total N inputs = Fertilizer N + Irrigation water N + Atmospheric N deposition
Irrigation water N was determined by quantifying nitrate concentrations in the spray-irrigated well water-
collected groundwater. Total (wet + dry) N deposition for the field site was estimated to be 4.5 lb ac 1 yr1
from a 4 km x 4 km Community Multiscale Air Quality Model (CMAQ version 4.7.1, Schwede and Lear,
2014) as presented in Compton et al. (2020). Exports were presented as leached N and crop harvest N.
Fractional leaching export, or the total leached fraction of N inputs (i.e. leached from the deepest soil depth
measured), serves as a useful metric for understanding how much N may be exported to groundwater,
similar to fractional watershed export (c.f. Howarth et al., 2006; Compton et al., 2020), and was calculated
as follows:
Fractional leaching export rate (%) = [(Leached N at 10 ft depth) / (Total N inputs)] *100
Crop N uptake by the aboveground portion of plants across the field site was determined by multiplying the
aboveground whole plant biomass and N content for each specific crop (i.e., whole corn crop and cover
crops), and then adding all the crop specific N uptake values together, as shown below:
Crop N uptake = Corn crop N uptake + Cover crop N uptake
= [Crop biomasscorn * %Ncorn] + [Crop biomasscover crop * %Ncover crop\
Crop N harvest is distinct from crop N uptake in that it represents all N removal from the field and
movement off site (Compton et al., 2021). In this study, only corn ears were harvested from the site, and
thus make up the crop N harvest value (i.e., corn ear yield x N content); corn stalks and leaves (and roots),
8
-------
as well as cover crops were left on the field, and therefore any N taken up remained behind in residues.
However, the N content in corn ears was not measured separately from the whole corn aboveground
biomass, and so, crop N harvest could not be calculated directly. Rather, we found that corn ear yield
varied systematically in relation to the whole aboveground corn plant biomass by ~39%. Therefore, we
applied this same ratio to the aboveground N uptake values to determine corn ear N uptake (i.e., crop N
harvest). Hart et al. (2010) reported that ~40% of the total aboveground N was in the ears, and so our
estimate seemed reasonable. Accordingly, for the purposes of this study:
Crop N harvest = N removed in corn ears = [Crop biomasscorn * %Ncorn] * 0.39
Indicators of field-level performance were also evaluated and included N use efficiency (NUE), which is a
benchmark for N management effectiveness, and N surplus and N remainder, proxies for N loss to the
environment (Lin et al., 2019; Quemada et al., 2020). Estimated for each fertilizer year, we calculated NUE
as the ratio of crop N harvest to total N inputs, while N surplus was calculated as the total N inputs minus
crop N harvest (Zhang et al., 2015) and N remainder was calculated as the total N inputs minus crop N
harvest and leaching N loss (Lin et al., 2019) for the upper soil layer (2.5 ft). For each of the two deeper soil
layers, N remainder was calculated as N leached from the soil depth immediately above (representing N
inputs to the below soil layer) minus leaching N loss from the respective soil layer (e.g., at 10 ft depth N
remainder would equal the N leached from the 5 ft depth minus the leaching loss from the 10 ft depth).
These components are represented by the following:
Nitrogen Use Efficiency (NUE) (%) = [(Crop N harvest)/(Total N inputs)] * 100
Surplus N = Total N inputs — Crop N harvest
N Remainder at 2.5 ft = Total N inputssurface — (Leached N2.^ft + Crop N harvest)
N Remainder at 5 ft = Leached N2.sft — Leached N5ft
N Remainder at 10ft = Leached N5ft — Leached N10ft
2.7 Bromide Tracer Study
A field tracer study using sodium bromide as a conservative tracer was conducted in June 2019. The study
was designed to deliver the tracer through the irrigation system allowing it to be sprayed onto the field
during a standard irrigation event. A concentrated solution was prepared in the field using site
groundwater and solid sodium bromide. The concentrated sodium bromide stock solution was pumped into
the main irrigation line at a rate to provide a final application concentration of 50 mg/L bromide. Tracer
injections were conducted using an Agri-lnject G55 metering pump. The pump drew solution from a 330-gal
tote with graduations and equipped with a 2-in quick connect valve. Pump outflow was measured using a
Watts model IWTG pressure gauge and an Assured Automation DM-P-050 digital flow meter/totalizer. An
inline check valve, downstream of the flow meter and pressure gauge, prevented irrigation system pressure
from backflowing into the injection system.
9
-------
Due to the pulsating action of the metering pump, the flow rate function of the flow meter was not
functioning, therefore, manual flow rates were calculated using the totalizer volumes and a timer. Totalizer
accuracy was checked against a graduated container. One week prior to tracer injections, injection system
settings for the desired flow rates were determined with a trial run of the injection system installed onto
the farm irrigation system and using clean water. There are ten smaller irrigation lines equipped with
above-ground sprayers that cover the entire field. For each irrigation event, six lines are operated for three
hours, and then the remaining four lines are operated for another three hours. The irrigation pump was
turned on and allowed to reach operating pressure, then the injection system was turned on and adjusted
to the desired flow rates (0.80 GPM for the six-line set and 0.53 GPM for the four-line set) by adjusting the
pump stroke and making manual rate calculations. Desired settings were recorded, and the injection
system was uninstalled from the irrigation system.
The bromide tracer solution was prepared by adding 200 gal of water to the 330-gal tote/tank one day
prior to the tracer injections. A total of 57.2 lbs of Acros Organics extra pure, anhydrous sodium bromide
(NaBr) was added to the tote and a circulation/mixing pump added to aid in dissolution. After
approximately 2 hr the mixing pump was turned off as all NaBr salt was dissolved. Irrigation times for each
line set (6 and 4 lines) was 3 hr. Tracer volume for each set (120 gal for 6 lines and 80 gal for 4 lines) was
injected into the system over the first 2.5 hr of the irrigation time to allow residual tracer in the system to
be flushed out prior to switch over. A data table containing time, initial totalizer reading, ending totalizer
reading, elapsed time, calculated flow rate and injection system pressures was created, and recordings
were made every 15 min to ensure injection rate was stable and total desired volume was injected. Injected
volumes were checked against graduations on the 330-gal tote. The tank mixer pump was turned on for the
duration of the tracer injections to ensure no NaBr precipitation or tank stratification. This bromide
injection was conducted during only one irrigation event.
Prior to irrigation, three replicate pans were placed within the plots for each of the two subfields to collect
irrigation water during the entire irrigation event. A 30-mL plastic syringe was used to transfer 20 mL from
each pan into a separate 20-mL plastic vial that was transported back to RSKERC for bromide analysis along
with the regular biweekly samples. This was repeated for the next irrigation event to ensure that remaining
bromide concentrations were negligible. Bromide analyses via ion chromatography were incorporated into
the biweekly monitoring schedule for all lysimeter and groundwater samples immediately prior to the
bromide tracer study and for all biweekly sampling events after the bromide tracer study.
2.8 Additional Data Collection
In addition to the physical samples being collected, supplementary data was collected at various times
throughout the project period, including electronic data and written records. Electronic data was recorded
from data loggers placed in the monitoring wells. Information on the water levels was collected every 20
min and the data was downloaded every 6 months. Electronic data was also recorded from the soil
moisture probes and tensiometers through separate data loggers. Information on soil moisture and soil
capillary pressure was collected every 30 min and the data was downloaded monthly.
Rainfall was automatically collected by a Washington State University Weather Net station (Corvallis East)
located two fields away from the study site, and this information was downloaded from the Internet
10
-------
(http://weather.wsu.edu/). Other site information such as planting times, fertilizer application, irrigation
frequency, and similar land management information was recorded manually by site personnel.
Topographic Lidar data for the site was collected from the State of Oregon Department of Geology and
Mineral Industries using their Lidar Data Viewer (https://gis.dogami.oregon.gov/maps/lidarviewer/). The
Lidar was flown between 31 August - 14 September 2008. Its resolution is 1 yard (0.9144 m) on a uniform
rectangular grid. Its accuracy is 0.13 ft (root mean square error) with an average vertical error of 0.10 ft
(DOGAMI, 2009).
The monitoring wells were surveyed using a GPS receiver (Model 4700, micro-centered L1/L2 geodetic
ground plane antenna, Trimble Navigation Limited, Sunnyvale, CA, USA). The average elevation of the site
was 215.94 ft with an average standard deviation of 0.10 ft.
We used data on Willamette River bathymetry which were surveyed as part of a USGS study by Sullivan and
Rounds (2004). The portion of the survey which borders the OSU Vegetable Research Farm covers
approximately river miles 130-133. The file containing this data is "willamette_elevations.zip" and can be
downloaded at https://or.water.usgs.gov/prois dir/will tmdl/main stem bth.html.
2.9 Vadose Zone and Groundwater Modeling
The HYDRUS-1D model (Simunek et al., 2013) uses the linear finite element method to solve the Richards
equation for fluid flow in variably saturated, heterogeneous, one-dimensional porous media. It can also use
the flow solutions to solve the advection-dispersion equation for solute transport. We used the
tensiometer and soil moisture data to calibrate the model. The modeling was done: 1) to support
inferences on nitrate behavior in the soil profile as water moves from the soil surface to the saturated zone,
2) to estimate nitrogen loss rates from lysimeter-sampled depths, and 3) to inform the sampling schedule
for bromide following tracer application.
The raw data from the soil moisture probes was used to obtain volumetric water content. The raw data
from the tensiometers showed enough temperature dependence that it needed to be corrected using the
method described in Thomson and Armstrong (1987). This method is referenced by the manufacturer
(Campbell Scientific). To use this method, the temperature data from the soil moisture sensors installed at
the same depth was used to obtain the corrected values of soil matric potential. These data were
converted to capillary pressure head (in H2O), the unit which is reported for the remainder of this
document.
In order to facilitate use in vadose zone model calibration, data from the sensors was examined for values
which seemed unrealistic, given knowledge of the soil system at hand. In many instances, data from the
eastern soil matric potential sensor at the 5-ft depth had values much less than the approximate minimum
of the other two. This was found to be related to zero values of resistivity in the probe related to very wet
conditions. When these values occurred and appeared as sudden drops contrary to adjacent value trends,
they were assumed to result from temporary instrument failure to record actual resistivity. In that sensor,
these conditions were so pervasive throughout the multiyear monitoring period, that its data were omitted
in the final calibration data for the HYDRUS-1D model. From 09/16/16 to 10/14/16, sensor one of the 2.5-ft
soil matric potential probes exhibited anomalously high values. These were assumed to be related to lack of
11
-------
equilibration of the sensor with ambient soil conditions and were therefore omitted in the final calibration
data. Also, data from the southern 2.5-ft soil moisture probe was deleted, due to lack of realistic variability
(persistent high levels, never less than about 0.45), as compared to the other two. During two periods of
time between 10/05/18 through 10/16/18, and 11/27/19 through 12/03/19 the data logging batteries
failed, leading to missing data for these dates.
Following these modifications, data from the sensors was averaged for each depth, and averaged to obtain
daily values, yielding a single set of calibration data for soil matric potential, moisture, and probe
temperature, allowing trial and error fitting of functional relationships for in-situ soil characteristic curves
(Appendix A). Figures 4 and 5 show the HYDRUS-1D model calibration results for the north and south
25
0
-25 -
o
CM
X
-50 -
c
-75 -
-C
-100 -
-125 -
-150 L
25 r
0 -
-25 -
O
-
CM
X
-50 -
c
-75 -
-c
-100 -
-125 -
-150 L
25 r
0 -
-25 :
O
X
-50 -
c;
-75 -
-c
-100 -
-125 -
-150 L
measured
modeled
2017
2018
2019
2020
5 ft
1111
2017
2018
2019
2020
10 ft
i
2017
2018 2019
Date
2020
c
ro
C
0.48
0.46
0.44
0.42
0.4
0.38
0.36
0.34
0.32
0.48
0.46
0.44
0.42
0.4
0.38
0.36
0.34
0.32
0.48
0.46
0.44
0.42
0.4
0.38
0.36
0.34
0.32
U — >
1 "2L
2017
2018
2019
2020
2017
2018
2019
2020
2017
2018 2019
Date
2020
Figure 4. HYDRUS calibration results for north subfield.
12
-------
25
0
-25
o
I
-50
c
-75
-c
-100
-125
-150
25
0
-25
O
I
O
LO
c
-75
-c
-100
-125
-150
25
0
-25
O
Ol
I
-50
c
-75
-c
-100
-125
-150
measured ORN
modeled ORS
-1
2.5 ft
i
2017 2018 2019 2020
5 ft
2017 2018 2019 2020
10 ft
2017
2020
0.48
0.46
0.44
0.42
0.4
0.38
0.36
0.34
0.32
0.48
0.46
0.44
0.42
0.4
0.38
0.36
0.34
0.32
0.48
0.46
0.44
0.42
0.4
0.38
0.36
0.34
0.32
»c>
1 "-"2.
2017 2018 2019 2020
2017 2018 2019 2020
2018 2019
Date
Figure 5. HYDRUS calibration results for south subfield
2017
2018 2019
Date
2020
subfields, respectively. Fitting the calibration with HYDRUS-1D is a complex undertaking due to the
numerous parameters that it requires, so a trial-and-error approach was used. Solute transport modeling
was done for bromide transport using dispersivity parameters given in a study by Vanderborght and
Vereecken (2007). Heat transfer modeling was done using empirical formulation for thermal conductivity
with their given parameters for loam and sand (Chung and Horton, 1987). HYDRUS-1D provides an
implementation of the (Feddes et al., 1978) water stress response as a function of leaf area index. Seasonal
root growth was estimated using the model of Flenet et al. (1996), with corresponding leaf area estimated
using the allometric model of Colaizzi et al (2017). In their Eq. 4, Vo was set to 33 cm, do to 1, y to 0.542,
and canopy height to 223.9 cm. The latter value was given in Peachey and Hart (2009). The root water
uptake parameters were selected as those of Wesseling et al. (1991) in HYDRUS-1D. Bromide ion transport
13
-------
Table 1. Derived dispersivity values
National Agriculture Imagery Program
MODFLOW Conceptual Model
River Package
Study Site
Wells
Sloughs
Ponds
(Specified Head)
'Driller's Well Logs
Computational Grid
Soil logio Dispersivity (cm)
Type Mean Std. Dev. n_
Silty day 0 244
loam
Silt loam 0.595
Loam 0.691
Sandy loam 0.368
Loamy sand 0.68
Silty clay 1.448
Sand/cobble 0.882
0.876 10
0.488 106
0.382 108
0.401 144
0.505 106
0.232 2
0.06 2
was modeled using the dispersivity values
given in Vanderborght and Vereecken (2007),
which listed values for all seven soil types
present at the site (Table 1). Bromide
transport simulation was continued beyond
9/30/20 by assuming precipitation, potential
evapotranspiration, irrigation, and crop
growth were all identical to those of the years
2016-2019.
Although the HYDRUS-1D model can handle
vertical saturated flow below the vadose zone
soil columns under study, it is not well-suited
to estimate groundwater flow under and
beyond the main study site. Instead, the
MODFLOW-USG model (Panday et al., 2013)
was used to better understand how
groundwater movement affects results.
MODFLOW-USG uses a control volume finite difference method adapted to accommodate a variety of
unstructured grids. This is useful because it allows arbitrary refinement of the solution grid in areas of
interest, while relaxing the grid scale at distance (miles) away from the study area (Figure 6). This reduces
the computational burden, making large scale models tractable. The boundary conditions for the model
included the Willamette River along the north side of the area, several sloughs and ponds, and specified
head along the more permanent, but smaller watercourses surrounding the south side of the modeled area
(Figure 6).
USGS Bathymetry
Lake
(Specified Head)
(MODFLOW Drains)// Specified Head
Figure 6. Plan view of MODFLOW-USG conceptual model showing boundaries and internal features.
14
-------
The conceptual model for the MODFLOW-USG model was developed based on a combination of driller's
well logs, descriptions in Conlon et al. (2005), and observations during drilling and coring at the site. The
driller's logs were obtained using the Oregon Water Resources Department database. The data were
obtained using the Well Report Query (https://apps.wrd.state.or.us/apps/gw/well log/Default.aspx). The
driller's well logs were for wells
scattered about the active
model area (Figure 6). During
drilling, we noted at least three
distinct stratigraphic layers.
The uppermost layer consisted
primarily of silty clay loam
(SCL) of the Chehalis soil series,
underlain by a cobbly upper
sedimentary unit (USU). At
greater depths drillers typically
noted a "blue clay" indicating
the top of the deeper basin-fill
sediments (Conlon et al.,
2005). This lower sedimentary
unit (LSU) extends at depth
until Tertiary marine
sedimentary bedrock is
encountered and forms the
bottom no-flow boundary
(Figure 7). The layers were
built into the MODFLOW-USG
model.
Each layer had 24,779 cells (74,337 computational nodes, total). The easting, northing, and depth of each
layer at the wells was used to create interpolated surfaces using the regularized spline method. The layers
in the MODFLOW-USG model were thus created as the ground surface (from the Lidar surface), and each of
the deeper layers using the spline surfaces (Figure 8). The top of the bedrock surface was created by
subtracting the data on LSU thickness interpreted using the Conlon et al. (2005) maps for the LSU surface.
Parameters for the drain conductance, river conductance, and hydraulic conductivity of each layer were
estimated using the PEST algorithms of Doherty (2010). This is an optimization program that runs
MODFLOW-USG iteratively; varying the unknown parameters until the sum of squared weighted residuals is
close to zero. We used the measured water levels in wells ORN and ORS, and in the perimeter wells. The
residual is the difference between the measured water level (head) and computed water level for the cell in
which the well resides. The measured water level consisted of the averaged water level for the time period
of interest, with seasonal variation. The weights were determined as l/(2xVAR) where VAR is the variance
of the water level measurements. This makes the weights proportional to the inverse of the observation
error.
0 - 6 m
0-26 m
0 -17 m
Chehalis silty clay loam (SCL)
'••V-:'o'-'VA-ro
Upper Sedimentary Unit (USU)
-cobbles, alluvial floodplain deposits
Lower Sedimentary Unit (LSU)
-basin-fill sediments
7777777777777777777777777777
Bedrock (no-flow)
Figure 7. Stratigraphic column for the study site showing range of thicknesses
and assigned hydrogeologic units (not to scale).
15
-------
Figure 8. Active three-dimensional MODFLOW-USG model discretization. Vertical exaggeration: 5x.
In addition to the conductance, River Package cells require a river stage and bottom. The bottom elevations
were taken from the bathymetry data (Section 2.7), and the stages were taken from the Lidar raster, since
topographic Lidar does not penetrate the surface of open water. The Drain Package requires only the
elevation of the slough bottom ("drain") which was assigned to the drain cells from the Lidar raster. When
computed water levels rise above the bottom of drain cells, water is removed from groundwater,
simulating the effect of contribution to the larger stream or river system. Gross initial estimates of
hydraulic conductivity were taken from the driller's well logs that documented pumping tests (ranging from
46-4,055 ft/day) and these were refined in the PEST estimation process.
2.10 Monitoring and Field Operations Schedule
The entire field was planted in sweet corn in July 2016 prior to field installations and the initiation of the
monitoring program (Table 2). Field installations were completed, in part, in September 2016 and the
biweekly monitoring program began in October 2016. The corn field was flail mowed in October once all
instrumentation was installed. The perimeter wells were installed in September 2017 and incorporated into
the biweekly monitoring program. Post-harvest core samples were taken in September 2016-2019 and
analyzed as described in Section 2.3.
In 2017, sweet corn was planted in June and pre-herbicides applied to the entire field. At about the V6
stage of corn growth, tall fescue was drilled (interseeded) between corn rows in the north subfield.
Herbicide carryover prevented proper establishment of the fescue during the summer, so Steptoe barley
16
-------
Table 2. Field operations in the north and south subfields at the OSU Vegetable Research Farm
Date
Activity
Details
30-Jun-18
Irrigation
Begin periodic irrigation
l-Jut-18
Planting
planted Devotion super sweet 28,000 plants per acre
l-Jut-18
Herbicide
Outlook (16 oz/acre) and Atrazine {16 oz/acre) herbicide
l-Jut-16
Fertilizer
3 25 lbs 12-10-10(38 lbs as N) banded at planting
26-JuI-16
Fertilizer
Sidedressed 200 lbs/acre urea (90 lbs as N}
15-Sep-16
frrigatiori
Finish periodic irrigation
12-Oct-16
Harvest
Corn mowed, left down on field
2-Jun-17
Planting
Planted Devotion super sweet corn, 12-inch spacing in row
l-Jun-17
Fertilizer
325 lbs 12-10-10 (38 lbs as N) banded at planting
5-Jun-17
Herbicide
Outlook (16 oz/acre) and Atrazine i 18 oz/acre) herbicide
fr-Jurv-17
Irrigation
Begin periodic irrigation
fr-Juf-17
Fertilizer
Sidedressed 156 lbs/acre urea (70 lbs as NJ
fr-Juf-17
Planting
Planted fescue at 40 lbs/acre in north subfield
4-Sep-17
Irrigation
Finish periodic irrigation
10-Sep-17
Harvest
Harvested corn
27-Oct-l 7
Planting
Direct-seeded 80 lbs/acre Steptoe barley in north subfield
27-Oct-l 7
Herbicide
Roundup applied to south subfield
2-May-18
Herbicide
Roundup (4% Glystar Plus) applied to both subfields
6-Jur>-18
Planting
Plant Glacier Sh2 sweet corn, 26,000 seeds/acre
6-iun-18
Fertilizer
325 lbs 12-10-10 (38 lbs as N) banded at planting
6-iun-18
Herbicide
Dual Magnum (1 pt/acre) and atrazine (1 pt/acre) herbicide on south subfield
friun-18
Herbicide
torsban 15G (1.3 lbs/acre) T-banded over seed row with planter to entire field
7~iun-18
Irrigation
Begin periodic irrigation
2-Jut-18
Herbicide
Impact (1 oz/acre) and Basgran (1 qt/acre) herbicide on north subfield
10-Jul-18
Fertilizer
Sidedressed 278 lbs/acre urea (125 lbs as NJ
ll-Jul-18
Planting
Interseeded triticale on north subfield at vS-6 corn stage
27-Aug-18
Irrigation
Finish periodic irrigation
13-Sep-18
Harvest
Harvested both subfields
lO-Oct-18
Herbicide
DualGlyphosate (2 qt/acrej and MCPA (1 pt/acre) herbicide on south subfield
20-Apr-19
Herbicide
Applied glyphosate 2 qts/acre 3 lb ai formulation
31 May 19
Planting
Plant Driver Sh2 sweet corn, 28,000 seeds/acre
31 May 19
Fertilizer
325 lbs 12-10-10 (38 lbs as N) banded at planting
31 May 19
Herbicide
Outlook (1 pt/acre) and atrazine (1 pt/acre) herbicide on south subfield
31 May 19
Herbicide
torsban 15G (1,3 lbs/acre) T-banded over seed row with planter to entire field
2-Jun-13
Irrigation
Begin periodic irrigation
22 Jun 29
Herbicide
1 oz Impact (topramezone) and 0.25% MSO+2.5%UAN solution 32 - 3 lb/acre on north subfield
26-Jun-19
Fertilizer
Sidedressed 360 lbs/acre urea (162 lbs as N)
28-Jun-19
Planting
Cultivated entire field; interseeded Variety 093 (Bolt) triticale 110 lbs/acre on north subfield
26-Aug-19
Irrigation
Finish periodic irrigation
7-Sep-19
Harvest
Harvested both subfields
ll-Jun-20
Planting
Plant Devotion Sh2 sweet corn, 28,000 seeds/acre
ll-Jun-20
fertilizer
240 lbs/acre 16-16-16 (38 lb/acre as N) banded at planting
ll-Jun-20
Herbicide
Outlook (1 pt/acre) and atrazine (1 pt/acre) herbicide on south subfield
ll-Jun-20
Herbicide
torsban 15G (8 oz/acre) T-banded over seed row with planter to entire field
08-Jul-20
Planting
Broadcast Fayette Trail Tall fescue on north subfield
08-Jul-20
Irrrigation
Begin periodic irrigation
15-Ju!-20
Fertilizer
Side-dressed urea (165 lbs N/acrej
17-Juf-20
Herbicide
Laudis on entire field; atrazine on south subfield
16-5ep-20
Irrigation
Finish periodic irrigation
21-5ep-20
Harvest
Harvested both subfields
17
-------
was seeded in late October to supplement the fescue cover crop. In 2018, sweet corn was planted in June
with pre-herbicides applied only to the south subfield that would not have cover crops. Triticale was drilled
between corn rows at the V6 stage of corn growth in the north subfield and established well. A similar plan
was followed in 2019, whereas in 2020 tall fescue was used as the cover crop instead of triticale. Corn was
harvested with a commercial picker in 2017-2020 to remove ears from the field as would occur in
production fields.
Fertilizer application was conducted twice during each growing season. A commercial 20-10-10 (NPK)
fertilizer was side banded when the corn was first planted and was applied at the same rate for each year.
Four to five weeks later a soil nitrate pre-sidedress test was conducted to determine how much additional
nitrogen was needed and urea was sidedressed at the prescribed rate. Urea was sidedressed with a hand-
push spreader about 4 to 6 in from the seed row at the V6 stage of corn growth. For the 2019 and 2020
growing seasons, the last two seasons of the study, an additional 25% urea was included in the sidedress
application to determine whether this additional application would contribute to nitrate leaching. Weeds
were managed on the cover cropped north subfield with cultivation at V4, followed by an application of
Impact (topramezone, 4-HHPD, Gr 27) and Basagran (bentazon, Gr 5) herbicides in 2018, Impact herbicide
only in 2019, and Laudis herbicide in 2020. Weeds were managed on the south subfield by applying
Outlook and Atrazine after planting corn and cultivation at V4. Roundup (glyphosate) with or without MCPA
was applied to the south subfield (without cover crop) in the fall and spring to keep the plot vegetation-free
prior to corn planting.
Corn was harvested when kernel moisture was 72-75% of total kernel weight. Corn yield was assessed by
harvesting ears from six 20-ft rows in each half of the field in 2018-2020. In 2019, nitrogen uptake by the
above-ground portion of corn plants was measured by harvesting entire plants from 10 ft of row at six sites
in each half of the field, grinding plants, drying at 70°C, and then extracting a small subsample for N
analysis. Cover crops were harvested from four 6.5-ft2 areas immediately after harvest and again in spring,
from four 25-ft2 areas just before desiccation with glyphosate. Cover crop biomass was dried at 70°C and a
subsample sent for N analysis. Additional details regarding field operations and the planting of corn and
cover crops are provided in Table 2.
2.11 Subfield Comparison Considerations
While the field site utilized for this study was split into two halves, which were managed differently (i.e.,
cover crops versus conventional methods), the north and south subfields simply served as replicate
sampling areas within the entirety of the field. Thus, it is not appropriate to interpret the differences
between the two subfields as an evaluation of management effects, as the treatments were not
randomized nor replicated. Cover crops were planted on half of the field to capture preliminary
observations for a future study concerning the effects of cover crops on nutrient movement, and thus,
comparing crop management was not a main objective of this study. Cover crop establishment was poor
during much of the study period, resulting in low biomass N uptake through the rainy fall and winter
seasons, and as such, the cover crop versus conventionally managed subfields did not behave differently.
Data in the report is presented based on subfield, but as stated, direct comparison of the two subfields is
not appropriate.
18
-------
3.0 Results
The framework for the following results is set around water years (WY) rather than calendar years. A WY
begins on Oct 1st of one year through Sept 30th of the following year and is dated with the ending year
number. Thus, for example, WY16 refers to the interval from 10/01/15 to 09/30/16.
3.1 Crop Performance
Rates of fertilizer application changed significantly from year to year, especially the final two years when
25% extra N was added over what was required based on the soil nitrate pre-sidedress test. The total
nitrogen applied for each year was 128 lbs/acre N in WY16, 108 lbs/acre N in WY17, 163 lbs/acre N in
WY18, 200 lbs/acre N in WY19, and 203 lbs/acre in WY20.
Cover crop production and sweet corn yield varied from year to year. Cover crop dry matter accumulation
following the WY17 growing season only averaged 688 lb/acre because of very poor fescue survival and late
drilling of the barley (Table 3). The interseeded triticale cover crop established well under the corn canopy
after interseeding in July of WY18 and WY19 and easily survived the damage caused by the corn-picker at
harvest. Avoiding pre-herbicides but cultivating and applying Impact plus Basagran herbicides at V4 to V6
sufficiently controlled weeds on the north subfield and allowed the cover crop to establish without damage
from herbicide carryover. Pre-herbicides and cultivation were sufficient to control weeds on the south
subfield without a cover crop. Cover crop (triticale) dry matter accumulation in the spring of WY19
(interseeded in July WY18) was more than 4000 lbs/acre and captured 72 lb/acre N (Table 3). Even though
an excess of 25% N was added for WY19, the cover crop dry matter accumulation in the spring of WY20 was
only 3200 lbs/acre and captured only 40 lb/acre N (Table 3). The tall fescue that was broadcast in early July
of WY20 did poorly, and weeds proliferated due to the later application of Laudis herbicide in late July,
which was designed to avoid the impact on the tall fescue.
Table 3. Cover crop dry matter on north subfield that was planted to a cover crop
Cover Crop Dry Matter and Nitrogen Content
Water
Year
# Obs
Crop
After Corn Harvest (Fall)
Before Dessication (Spring]
Dry Matter
Nitrogen Date
Dry Matter
Nitrogen
Date
(lb/acre)
(lb/acre) Sampled
(lb/acre)
(lb/acre)
Sampled
WY17
Fescue, Interseeded and Steptoe
688
17
4/30/2018
Barley (direct seeded after harvest)
WY18
4
T ritica le, interseeded
317
9 3/17/2018
4100
72
4/13/2019
WY19
4
Triticale, interseeded
284
8 9/10/2019
3200
40
4/7/2020
WY20
4
Tall Fescue, interseeded (broadcast,
rather than drilled)
467"
10*» 9/24/2020
MS
NS
NS
* Poor survival of fescue because of herbicide injury
** Very little if any fescue biomass was collected after harvest; mostly weed biomass
NS: not sampled
19
-------
Table 4. Annual nitrogen (N) fluxes (inputs and exports) across fields (north vs south) and soil depths (2.5, 5.0,
and 10.0 ft) by water year (WY) (adapted from Weitzman et al., 2022]
Year
Field
N Inputs
(lb ac"1 yr1)
NtV-N Leached by Depth*
(lb ac1 yr"1)
Fractional
Leaching
Export
Rate+ (%)
Crop N
Uptake*
(lb ac1 yr"1)
Crop N
Harvest5
(lb ac"1 yr1)
Surplus N11
(lb ac"1 yr"1)
N Remainder
by Depth11
(lb ac"1 yr"1)
NUE#
(%)
Fertilizer
Irrigation
Water
Deposition
(Wet+Dry)
2.5 ft
5.0 ft
10.0 ft
Cover
Crops
Corn
2.5ft
5ft 10ft
WY17
North
South
108
34.8
4.50
225
104
171
172
57.2
74.4
38.8
50.5
NA
168
189
66.2
71.3
81.1
76.0
-144
-28.0
54.0 114
-68.0 97.6
45
48
WY18
North
South
163
31.1
4.50
113
141
99.3
95.9
55.2
83.5
27.8
42.0
25.1
206
256
81.1
96.5
118
102
4.50
-38.9
13.7 44.1
45.1 12.4
41
49
WY19
North
South
200
19.9
4.50
87.7
66.9
36.4
100
29.8
66.1
13.3
29.5
80.1
207
252
81.7
95.2
143
129
55.0
62.3
51.3 6.60
-33.1 33.9
36
42
WY20
North
South
203
11.3
4.50
238
317
52.7
45.5
7.60
35.2
3.47
16.1
50.9
121
145
47.9
54.7
170
164
-67.1
-153
185 45.1
272 10.3
22
25
*Leached nitrate (N03-N) fluxes are expressed as depth averaged mean values
""Fractional leaching export rate (%) is expressed as the amount of N leached at the 10 ft depth, divided by the total N inputs (fertilizer + irrigation water + deposition)
*Crop N uptake includes both cover crop N and aboveground corn crop N
§Crop N harvest for this study is equal to N uptake in corn ears
i.e., the only plant material removed from the field
"Surplus N gives an indication of what is lost or retained by the environment and is estimated as the total N inputs minus crop N harvest
^Nitrogen remainder is defined as the total N exports from a particular depth subtracted from the total N inputs of the above depth
#Nutrient use efficiency (NUE) is calculated as the percent of crop N harvest divided by total N inputs
In WY19, sweet corn accumulated 207 and 252 lbs/acre N in the north (cover cropped) and south (without
cover crop) subfields, respectively (Table 4). It is of interest to note that neither the fall harvest cover crop
dry matter (and N) (Table 4) nor the sweet corn yield, ear diameter, average ear length, and tip fill (Table 5)
increased in WY19 versus WY18, even though the total nitrogen applied was 163 lbs/acre N in WY18 and
200 lbs/acre N in WY19, and the WY19 urea N application was 25% greater than what was required based
on the soil nitrate pre-
sided ress test. In WY20,
fertilizer application was 203
lbs/acre N, and the WY20
urea N application was again
25% greater than what was
required based on the soil
nitrate pre-sidedress test, but
in this year the sweet corn
yield dropped substantially.
The cause for this discrepancy
is not apparent but was
probably due to a
combination of poor soil
conditions at planting,
proliferation of weeds, and
the fact that WY20
represented five years of
continuous corn planting
without rotation.
Table 5. Sweet corn ear yield in north (cover cropped) and south subfields
WY18 (Triticale Interseed)
Glacial Sh2 White, 6/6/2018
Subfield Cover Crop
Herbicide
# Obs
Ear Yield
(t/acre)
Ear Diameter
(in)
Average Ear Length
(in)
Tip Fill (%)
North Triticale
Topramezone, bentazon POSTV4
6
14.9
2.1
10.2
96
South None
S-metolachlor +atrazine PRE
6
16.7
2.1
10.1
97
FPLSD (0.15) p-value
ns*
ns
ns
ns
WY19 (Triticale Interseed)
Glacial Sh2 Yellow, 5/31/2019
Subfield Cover Crop
Herbicide
# Obs
Ear Yield
(t/acre)
Ear Diameter
(in)
Average Ear Length
(in)
Tip Fill (%)
North Triticale
Topramezone POST V4
6
13.2
2.0
8.6
91
South None
Dimeth-P, atrazine PRE
6
13.5
2.0
8.8
93
FPLSD (0.15) p-value
ns
ns
ns
ns
WY20 (Fescue Interseed)
Devotion Sh2 Sweet Corn, 6/11/2020
Subfield Cover Crop
Herbicide
# Obs
Ear Yield
(t/acre)
Ear Diameter
(in)
Average Ear Length
(in)
Tip Fill (%)
North Fescue
Tembotrione POST V4 (directed spray
to avoid emerging fescue cover crop)
6
6.6
2.0
8.1
96
South None
Dimeth-P, atrazine PRE and
Tembotrione, atrazine POST V4
6
8.5
2.1
8.3
98
FPLSD (0.15) p-value
ns
ns
ns
ns
* Not significant statistically at alpha 0.15
20
-------
Crop N uptake for the corn ranged from 121-256 lbs/acre across all years, with an additional 25-80 lbs/acre
taken up by the cover crop in the north subfield (Table 4). Total inputs of N, which included fertilizer N plus
contributions of N from irrigation water and atmospheric deposition, ranged from 147-224 lbs/acre.
Nutrient use efficiency (NUE) for the corn crop varied across years, from 22-49%. In WY20 NUE was merely
~25%, driven by poor stand growth of the corn crop even though N inputs were similar to those in WY18
and WY19. Late spring rains may have contributed to the poor development of the corn crop in both
subfields in WY20, as wet soil conditions delayed planting (Table 2) and inhibited initial growth. Further, as
mentioned above, WY20 represented the fifth consecutive year of corn planting without rotation, which
may have also impacted corn yield. It should also be noted that different sweet corn varieties were grown
across the study years (Table 2), potentially affecting performance. However, it is likely that differences in
field conditions and cultivation operations in WY20, and not N fertilizer amounts or corn variety, played a
greater role in impacting corn yield, resulting in the poor NUE observed in the two subfields.
3.2 Water Transport Through the Vadose Zone
Precipitation in the Willamette Valley, OR is
greatest during the Fall and Winter months,
requiring the use of irrigation to supplement the
water needs of most crops (Figure 9a). The
annual rainfall was 49.4 in for WY17, 28.6 in for
WY18, 35.5 in for WY19, and 28.3 in for WY20.
The annual irrigation input was relatively
consistent (11.9-17.6 in) during the four-year
study. Depths to groundwater generally ranged
from 10-22 ft BGS during the study, and the
water table rose and fell over 8-11 ft annually
(Figure 9b). Heavy rains during mid-April WY19
caused flooding of the adjacent Willamette River
and briefly inundated the fields, resulting in a
brief rise in the water table up to 2-4 ft BGS. A
comparison between rainfall and irrigation
events with water table response indicate that
irrigation had minimal effect on water table
elevations (Figure 9a and 9b). Data from the
USGS 14171600 Willamette River Gauge Station,
located about 5000 ft from the field, show that
these water table responses appear to be better
correlated with the flood stages of the adjacent
Willamette River, which is located only about
1000 ft from the corner of the field (Figure 9b).
This would help to explain why the water table
elevation does not remain relatively high when
consistent rainfall is replaced by consistent
irrigation during the summer months.
a) OSU Vegetable Research Farm Irrigation/Rainfall
I Irrigation
¦ Rainfall
Oct-16 Jan-17 Apr-17 Jul-17 Oct-17 Jan-18 Apr-18 Jul-18 Oct-18 Jan-19 Apr-19 Jii-19 Oct-19
b) Well Water Table Elevationsand River Gauge
i.
K
A A
A
- it
HS
J
it
H' VVV
J
7
IN
J
—ORN
—ORNA
—ORNB
—ORNC
—ORND
—ORS
—ORSA
—ORSB
—ORSC
—ORSD
River Gauge
Oct-16 Jan-17 Apr-17 Jul-17 Oct-17 Jan-18 Apr-18 Jul-18 Oct-18 Jan-19 Apr-19 Jul-19 Oct-19
Figure 9. OSU Vegetable Research Farm a) irrigation and
rainfall data during field study, and b) water table response
during same time interval. Data show 10/18/16-12/23/19.
21
-------
Transect #1
Transect #2
Transect #3
Transect #4
The upper soil across
the field site is
predominately a
Chehalis silty clay
loam, especially within
the first 4-5 ft BGS.
Below this depth the
soil type varies
considerably and
includes a gravel/rock
matrix in the lower
profile whose top
ranges from 14 ft BGS
on the northern edge
of the field to 22 ft
BGS on the southern
edge (Figure 10). Figure 11 shows the lithological profiles for the cores taken within each of the two plots
located in the center of the subfields relative to the lysimeter depths and well sample depths. In the south
plot (ORS location), the silt loam and loam layers underlying the silty clay loam are about twice as thick as
those in the north plot (ORN location), and the top of the gravel/rock matrix is much lower. Cores were also
taken during installation of the perimeter wells, but only for the first 10 ft BGS. The top of the gravel/rock
matrix was estimated during drilling by listening to the sound of the Geoprobe rod as it reached and then
advanced into the gravel/rock matrix. As observed for the center well locations, the soil matrix was
predominantly silty clay loam for the first 5 ft BGS and then varied considerably below that for the
perimeter well locations (Table 6).
Figure 10. Location of top of gravel/rock matrix and relation to well screened interval.
ORS ORS
¦ Lys
ORS Location
Sample Depth
(ft BGS)
0
1
1 2
I 3
4
1 5
6
7
Texture
Class
Silt Loan
Silt Loam
Loam
SiH-Leara--
SiltLoam
Loam
Silt Loam
San
m_
15.0
16.3
8.8
16.0""
6.3
8.8
27.5
27.5
35.0
-18.-8-
213
42.5
36.3
Silt
ff
56.3
Clay
m
275
27.5
32.5
58.8
57.5
52:5--
50.0
52.5
52.5
48.8
65:0-
63.8
43.8
51.3
31.3
36.3
-48.-0
41.3
20.0
20.0
16.3
-16.-3
15.0
13.8
12 5
ORN Location
Sample Depth
(ft BGS)
Texture
Class
Sand
J*L
8.8
8.8
100
"giit
I*)
625
60.0
58.8
Clay
3
4
5-
6
7
8
9
-K>
11
12
13
Sthiroam-
SiltLoam
Silt Loam
Loam
Sandy Loam
Sandy-Leam--
Loamy Sand
Loamy Sand
Lo^ny Sand
7.5
8.8
--18:8-
18.8
213
413
70.0
-7«-
77.5
80.0
76.3
613
61.3
---62.5-
65.0
62.5
47.5
22.5
-225-
18.8
17.5
20.0
31.3
30.0
-18.8
16.3
16.3
113
7.5
—6:3
3.8
2.5
3.8
ORN
Lys
ORN
Sample
Inlet
Range
14
15
lb
17
18
19
20
21
22
"23"
24
25
26
27
28
29
30
31
32
33
34
35
Loam
Sandy Loam
Sandy Loam
Loamy Sand
Loamy Sand
Loamy Sand
Loamy Sand
Sand
(Gravel/Rock)
51.3
55.0
70.0
825
78.8
80.0
875
87.5
36.3
32.5
17 5
6.3
13.8
12.5
6.3
8.8
ravel/Rock)
(Gravel/Rock)
(Gravel/Rock)
(Gravel/Rock)
(Gravel/Rock)
(GravelJRock)
(GraveWRock)
(Gravel/Rock)
(Gravel/Rock)
(core lost)
(core lost)
(core lost)
(core lost)
14
15
S
17
18
19
20
21
22
"23"
24
25
26
27
28
29
30
31
32
33
34
35
(GraveWRock)
(Gravel/Rock)
(Gravel/Rock)
(core lost)
(core lost)
(core lost)
(core lost)
(core lost)
(core lost)
(core lost)
(core lost)
(GravelJRock)
(Gravel/Rock I
(core lost)
(core lost)
(core lost)
(Gravel/Rock)
(Gravel/Rock)
(GraveVRock)
(core lost)
(core lost)
(core lost)
Figure 11. Lithology relative to sampling depth for center wells and lysimeters.
22
-------
Table 6. Texture analysis for cores taken at perimeter well locations
Location ORNA
Location ORNB
Location ORNC
Location ORND
Sample Depth
(ft BGS)
(%)
Silly Clay U
>il .
Silly Clay Lt
>il .
Silly Clay Lt
Silty Clay Li
Silt Loam
Silt Loa m
Silt Loa m
Silt Loa m
(no sample)
30.0
27.5
62.5
62.5
62.5
60.0
62.5
61.3
56.3
52.5
67.5
66.3
Clay
{%)
30.0
32.5
32.5
35.0
32.5
31.3
13.8
20.0
15.0
Top of Rock/Gravel
Clay
m
30.0
37.5
37.5
35.0
30.0
Sample Depth (ft
BGS)
Top of Rock/Gravel
(%)
20.0
20.0
35.0
45.0
55.0
55.0
60.0
60.0
60.0
55.0
60.0
60.0
60.0
52.5
42.5
35.0
35.0
Clay
m
32.5
35.0
35.0
40.0
35.0
20.0
20.0
12.5
12.5
10.0
10.0
Sample Depth (ft
BGS)
(%)
Clay Lc
Clay Lc
Clay Lc
Top of Rock/Gravel
47.5
30.0
65.0
60.0
65.0
61.3
62.5
65.0
70.0
72.5
72.5
37.5
55.0
22.5
Clay
(«)
25.0
27.5
36.3
35.0
32.5
25.0
20.0
20.0
15.0
15.0
12.5
Location ORSA
Location ORSB
Location ORSC
Location ORSD
Sample Depth
(ft BGS)
Texture
Class
Sample Depth (ft
BGS)
Texture
Class
Sample Depth (ft
BGS)
Texture
Class
Sample Depth (ft
BGS)
0
Silty clay Loam
10.0
1
Sjlty clay Loam
7.5
2
Silty clay Loam
5.0
3
sjlty clay Loam
5.0
4
5
Silty clay Loam
Silty Clay Loam
7.5
7.5
6
Silt Loam
15.0
7
Silt Loam
27.5
8
Loam
37.5
9
Silt Loam
30.0
10
Loam
37.5
8.5
Top of Rock/Gravel
60.0
61.3
62.5
60.0
30.0
31.3
32.5
35.0
33.8
32.5
26.3
Silty Clay Loam
Silt Loam
Sandy Loam
Sandy Loam
Sandy Loam
Top of Rock/Gravel
32.5
20.0
62.5
70.0
77.5
60.0
60.0
60.0
60.0
61.3
62.5
46.3
55.0
25.0
30.0
32.5
32.5
32.5
33.8
30.0
21.3
25.0
(nc
Silt Loam
Silt Loam
Sandy Loam
Top of Rock/Gravel
57.5
61.3
57.5
42.5
55.0
31.3
32.5
31.3
35.0
27.5
26.3
17.5
25.0
Top of Rock/Gravel
65.0
62.5
52.5
47.5
48.8
42.5
40.0
45.0
27.5
30.0
15.0
12.5
12.5
The soil texture data for the two center wells was
used with HYDRUS-1D to generate infiltration flux
profiles for the north and south subfields (Figure
12). This was done for five separate lysimeter
sampling dates corresponding to three wet
seasons and two dry seasons. As expected, the
infiltration fluxes were higher during the wet
season sampling dates as compared to the dry
season sampling dates. However, there was
significant variability in infiltration flux rates even
at given depth profiles in each subfield for both
the wet season and dry season sampling events
(Figure 12). The trapezoidal rule for integration
with depth was used to calculate the average flux
for each sample event from ground surface to the
water table. Based on this, it takes anywhere from
2 years to 27 years for infiltrating water to reach
the water table. Although this is a very wide range,
it is not surprising given that the infiltration fluxes
vary by almost an order of magnitude between the
wet and dry seasons, and the fact that the water
table elevation changes by over 10 ft from year to
year. According to heat transfer modeling, the
average time to peak for each of the three
seasonal temperature cycles at each depth were
21.3, 36.7, and 72.3 days for the 2.5-ft, 5-ft, and
10-ft depths, respectively.
North Subfield
0.00
-5.00
— -10.00
-C
+->
Q.
0.30
0.20
0.10
n
o
0.40 <
0.30
n
o
3
Ct-
rl)
3
-0.02 -0.01 0.00
Water Flux (ft/day)
Figure 12. Flux profiles for the north and south subfields
for three wet season and two dry season dates,
corresponding to lysimeter sampling dates. Water table
depth is represented by inverted triangles.
23
-------
-10
-12
-11
-7
O
->P
O
00
CO
-10
-11
-7
-10
Apr Oct
WY17
Apr
WY18
Apr
WY19
Apr Oct
WY20
Precipitation 0
5 ft Ly si meters
Irrigation line
Precip 1 yr running mean
Cumulative Precip
Figure 13. Dynamics of water isotopes (6180) over three water years for
precipitation (top, light blue circles) and lysimeters at 2.5-ft (yellow
symbols), 5-ft (red symbols) and 10-ft (brown symbols) depths. Lines
represent input waters with the blue line representing cumulative
precipitation over the water year (starting Oct 1). The red line
represents irrigation 6180 values, the green line is the 6-month running
mean weighted by precipitation amount, and the black line is the well
water 6180 values measured in ORN.
small fraction of fall precipitation
reached the deeper depth. Water in the 5-ft BGS lysimeters was even less responsive isotopically to the fall
2016 precipitation input indicating that much of the fall precipitation seemed to bypass these lysimeters in
early WY17. in the summer when irrigation water with higher isotope values was applied to the field, water
Evaluation of stable isotope data was
also used to gain insight into
infiltration rates at the site based on
water sampled from lysimeters and
the precipitation collector
throughout individual water years.
We used the isotopic value of
incoming precipitation as a water
tracer because incoming precipitation
in WY17 and WY19 was isotopically
distinct and lower than irrigation
water (Figure 13). In October 2016
(WY17), lysimeter water at the 2.5-ft
and 10-ft depths had isotopic values
similar to that of irrigation water (-
7.6%o), while the cumulative
precipitation dropped rapidly to -
10%o by January (Figure 13, blue
line). Even though 2 ft of rain fell
from October-January, the isotopic
values of lysimeter water initially
stayed similar to irrigation water.
However, in January 2017, water
isotopes in several lysimeters at 2.5 ft
BGS began dropping rapidly to match
that of the accumulated precipitation
by early March. This lag in response
to input precipitation can be used as
a measure of transit rate. It took
approximately 2-3 months for the fall
precipitation to reach two south 2.5-
ft BGS lysimeters, with the other 2.5-
ft BGS lysimeters showing more of a
lagged response.
Water in the 10-ft BGS lysimeters
also dropped in isotopic values in
response to the fall 2016
precipitation, but not as quickly nor
as completely as the 2.5-ft BGS
lysimeters, indicating that only a
24
-------
isotopes in the 2.5-ft BGS iysimeters slowly increased and it took until December 2017 for the 2.5-ft BGS
lysimeters to reach isotopic values similar to the irrigation water, again, indicating a 3-4 month lag. Water
isotopes in the 5-ft BGS lysimeters also began to increase during the summer of WY17, but very slowly,
while water isotopes in the 10-ft BGS lysimeters did not increase in response to irrigation inputs. The
precipitation inputs during WY18 were not as low as in WY17 and did not induce a pulse shift in water
isotopes of any lysimeters. However, precipitation in WY19 did produce declines in lysimeter water
isotopes at the 2.5-ft BGS and 5-ft BGS levels, but the 10-ft BGS lysimeters remained relatively constant
over time, never reaching the high values observed at the beginning of monitoring. In early WY19, water at
2.5 ft BGS responded within a month of some isotopically low precipitation inputs, whereas water at 5 ft
BGS responded 2-3 months after the input. No distinct temporal pattern in lysimeter water isotopes was
noted for WY20. No water found in lysimeters matched the water-year accumulated precipitation, nor a 1-
year running mean of precipitation inputs, indicating that much of the water captured by the lysimeters has
a much longer residence time within the soil. Long residence time would cause lysimeter water to gradually
shift in isotopic values after input isotope values changed, and the response would be highly muted
compared to inputs. Interestingly, the water isotopic ratios at 2.5 ft BGS were generally higher than the
cumulative mean for the entire study, as well as the 1-year running mean and water year cumulative mean.
The closest isotopic match for water in the 2.5-ft BGS lysimeters was irrigation water. At the 5-ft BGS depth,
lysimeter water isotope ratios were lower than at 2.5 ft BGS, and isotopic values at the 10-ft BGS level were
the lowest, more closely matching that of groundwater and the cumulative value over the four-year study.
This systematic difference between depths, the consistency of water isotope ratios within a depth, and
decrease in water isotope variation over time is in stark contrast to patterns found for nitrate
concentrations in the same water (see below).
It was hoped that the bromide tracer
data could be used to provide
insights into infiltration rates at the
site, but the tracer profiles were far
from what was expected. The
HYDRUS-1D model was used to
predict the bromide breakthrough at
the three lysimeter depths for both
subfields (Figure 14). The initial
bromide breakthrough, as defined by
the quantitation limit of 0.2 mg/L,
was predicted to occur at about Day
100 and Day 225 at the 2.5-ft BGS
and 5-ft BGS depths, respectively for
both subfields. The bromide
breakthrough profiles were predicted
to be relatively uniform, with peak
bromide breakthroughs occurring at
about Day 200, Day 300, and Day 600
at the 2.5-ft BGS, 5-ft BGS, and 10-ft
BGS depths, respectively. However,
instead of sharp breakthrough
25
a) North Subfield
1.0
0.9
0.8
*C
re
E
0.7
k.
0.6
CO
0.5
-—•
0.4
i—
CO
0.3
—'
0.2
0.1
0.0
23 ft
5ft
10 ft
8 8
8 § § |
8 8
b) South Subfield
1.0
0.9
0.8
| 0.7
i 06
— 0.5
.— 0.4
0.3
' 0.2
0.1
0.0
CO
I.
1
i
i
/
1 /
i
i:
• i
• i
I!
i
* i
\
* i
v
:
. -7
<
<•!
— • 2.5 ft
5ft
10ft
W ^ Vf 4* y' Ul ^ W V F- I— r—
8888888iS|§§§
Days after release
Figure 14. HYDRUS predictions of bromide tracer breakthroughs in a)
north subfield and b) south subfield.
-------
a) North Plot - North Lysimeter Br
4.7
b) North Plot - East Lysimeter Br
4.7
c) North Plot - South Lysimeter Br
4.7
Figure 15. Lysimeter bromide concentrations at a) north plot north lysimeter, b) north plot east lysimeter,
c) north plot south lysimeter, d) south plot north lysimeter, e) south plot east lysimeter, and f) south plot
south lysimeter. Yellow triangle indicates date of bromide application.
curves, the bromide breakthrough profiles were drawn out and showed substantial variability between
lysimeter locations (Figure 15). In addition, bromide began to break-through at rates faster than what was
predicted. In contrast to the prediction of 100 days needed to observe initial bromide breakthrough at the
2.5-ft BGS depth (Figure 14), in some cases bromide began to break through at that depth during the next
sampling event following injection, or within a 14-day window (Figure 15d, 15e). At two locations, bromide
began to break-through and peak at the 5-ft
BGS lysimeters prior to breaking through at the
2.5-ft BGS lysimeters (Figure 15a,d).
Collectively, these data show the large impact
that site heterogeneity and preferential flow
paths have on water transport through the
vadose zone at this site.
The groundwater flow model developed using
MODFLOW-USG shows that the groundwater
flow direction is generally to the northeast,
tracking along the Willamette River, which
flows in the same direction (Figure 16). The
area of the map where the contour lines do not
meet the river at a perpendicular angle (e.g.,
north of the study site) represents where the
potential exists for groundwater to discharge
into the river. However, such effects are
beyond the scope of this study.
Figure 16. Groundwater flow contours based on MODFLOW
results.
26
-------
3.3 Water Quality Trends
3.3.1 Monitoring Wells
a) Well Nitrate {N03&N02-N)
-ORN
-ORS
Water quality data for the wells, lysimeters, and precipitation collector can be found in the data file
corresponding to the report. Nitrate and phosphate concentrations were highly variable in the two center
wells even before initiation of the cover crops in July
WY17 (Figure 17). Nitrate concentrations in the ORN
well generally peaked from March to June in each
year of the study (Figure 17a). However, this trend
was not observed in the ORS well, which often had
much lower nitrate levels than found in the ORN well,
and yet exhibited very sharp increases in nitrate
concentrations during selected sampling events. The
well phosphate concentrations in ORN did not exhibit
the periods of increased concentration observed for
the corresponding nitrate concentrations in this well
(Figure 17b). In contrast, sharp increases in phosphate
concentrations in well ORS closely mirrored the same
sharp increases in nitrate concentration in this well.
The short duration and sharpness of these nutrient
spikes in ORS are very unusual and do not readily
correlate with changes in other water quality
parameters like chloride (see data file).
Sep-16 Feb-17 Jul-17 Dec-17 May-18 Sep-18 Feb-19 Jul-19 Dec-19 Apr-20 Sep-20
b) Well Phosphate {o-P04-P)
-ORN
-ORS
Feb-17 Jul-17 Dec-17 May-18 Sep-18 Feb-19 Jul-19 Dec-19 Apr-20 Sep-20
Figure 18. a) Nitrate and b) phospate concentrations in
two center wells. ORN is the well in the north subfield
which was later interseeded with cover crops starting
in Jul WY17. Dashed boxes indicate intervals where it
is hypothesized that denitrification is occurring in ORS.
The well data indicate that ORN and ORS are sampling
separate groundwater pools, except on certain
occasions when groundwater normally supplying ORN
also supplies ORS. The spikes in nitrate concentrations
in the ORS well generally correlated with rises or
spikes of the water table during those sampling
events (Figure 18a). This phenomenon was not
observed in the ORN well (see Figure 17a). An analysis
of the changes in water level gradient across the site
was done for five sequential sampling events
encompassing one of these water level spikes on
04/17/18, and clearly shows the water level gradient
decreasing in magnitude during the water level spike
(Figure 18b). This may have had an impact on the
source of the water being sampled in ORS during this
sample event.
27
a) ORS Well Nitrate vs Water Table Elevation
Sep-16 Feb-17
May-18 Sep-18
-ORS Well Nitrate
Feb-19 Jul-19 Dec-19
— •• - Water Table Elevation
Apr-20 Sep-20
b) Water Level Gradient Across Site During 04/17/18 Water Level Spike
-•-03/20/18
—•—04/03/18
—•—04/17/18
—•—05/01/18
—•—05/15/18
ORSC
ORSB ORSD
ORNA
ORS ORNB
Figure 17. a) ORS well nitrate versus water table
elevation (the 04/17/18 sample date is highlighted) b)
change in water level gradient across site during
04/17/18 event.
-------
a) Well Nitrate (N03&N02-N)
-ORN
-ORS
Sep-16 Feb-17 Jii-17 Dec-17 May-18 Sep-18 Feb-19 Jul-19 Dec-19 Apr-20 Sep-20
b) Well 6180-H20
c) Well 615N-N03
Sep-16 Feb-17 Jul-17 Dec-17 May-18 Sep-18
-ORN
-ORS
Sep-16 Feb-17 Jii-17 Dec-17 May-18 Sep-18 Feb-19 Jul-19 Dec-19 Apr-20 Sep-20
-ORN
-ORS
Dec-19 Apr-20 Sep-20
Figure 19. Center well a) nitrate, b) 6180-H20, and c) 615N-
N03 profiles.
The 62H-H20 and the 615N-N03 values were
generally dissimilar between ORS and ORN
and were similar only during the nitrate spikes
when the nitrate values were similar between
these two wells (Figure 19). This indicates that
the source water was changing during these
nutrient spike events. The reason for this
change is not clear. One explanation could be
that the groundwater flow direction shifts
during periods of increasing or spiking water
table elevation, but this seems unlikely given
the relatively small water level gradients
across the site. Another explanation is that
this may be an artifact caused by upwards
water flow within the well during these
periods. Because water samples were
obtained 2 ft below the current water table at
the time of sampling, the seasonal variation in
the water table indicates that sometimes
samples were obtained from screened
intervals adjacent to the gravel/rock matrix
and sometimes the samples were obtained
from screened intervals adjacent to the silt
loam or sandy loam layers above (see Figure
11). This would hold true for ORSA and ORSD
in addition to ORS. Further, at times of very
high water tables, the samples were obtained
above the screened intervals, and it is not
known at what depth within the screened
intervals most of the water was entering the
wells. Beginning in June 2019, wells ORN and
ORS were sampled at two depths during each
sampling event to address this issue.
Figure 20 shows the effect of sample depth on nitrate, chloride, 615N-N03, and 6180-H20 profiles in ORN
and ORS. The water chemistry remains consistent throughout the water column for ORN, but not always for
ORS. Although there were only a few instances of significant differences in the more conservative
parameters of chloride and 6180-H20 with depth in ORS (Figure 20c,d), there were several instances where
there were significant differences in the less conservative parameters of nitrate and 615N-N03 with depth in
this well (Figure 20a,b). From June to August 2019, 615N-N03 was higher in ORS (top and bottom) compared
to ORN, even though nitrate concentrations were generally similar, indicating differences in the nitrate
sources between wells. However, from September to November 2019, 615N-N03 was higher only for the
upper part of ORS, while the bottom of ORS matched that of ORN. The nitrate concentration in the upper
ORS was also lower than ORS bottom, indicating that water within the upper portion of the well might have
been from a different source than the water in the lower part of the well.
28
-------
May-19 Jul-19 Sep-19 Nov-19 Jan-20 Mar-20 May-20 Jul-20 Sep-20 May-19 Jul-19 Sep-19 Nov-19 Jan-20 Mar-20 May-20 Jul-20 Sep-20
Figure 20. Effect of sample depth on a) nitrate, b) 615N-N03, c) chloride, and d) 6180-H20 profiles in ORN and
ORS wells.
The well dynamics between nitrate concentrations and 615N-N03 across the entire timeseries of the
experiment illustrate that nitrate source mixing processes may play a role in explaining nitrate dynamics
(Figure 21). When denitrification occurs, nitrate concentrations decrease while both 615N-N03 and 6180-
NO3 increase, creating a negative correlation between nitrate concentration and 615N-N03, and a positive
correlation between 615N-N03 and 6180-N03. Overall, these trends of denitrification rarely occurred in ORN
or ORS (Figure 21). In ORN, while nitrate concentrations followed an annual cycle of increasing in the wet
winter, and decreasing during the dry summer, 615N-N03 values were relatively stable with an isotopic
value indicative of synthetic fertilizer, indicating one source of nitrate supplying ORN. To assess if
denitrification was an important process in ORN, periods with decreasing nitrate concentrations and
increasing 615N-N03 values were identified (color boxes Figure 21). Those periods did show the diagnostic
patterns indicating denitrification, but they were short in duration. Thus, denitrification may have
happened occasionally in ORN, but it does little to explain the observed nitrate dynamics. Increases in 615N-
NO3 overtime also happen when nitrate is increasing (instead of decreasing as expected with
denitrification), or nitrate concentrations are stable. Increases in both nitrate and 615N-N03 can indicate
mixing with a more isotopically enriched nitrogen source (Figure 18a). The high variability in 615N-N03
within ORS, with only limited periods of clear denitrification trends, indicates that mixing of nitrate sources
was likely a major factor in explaining changes in the nitrate concentrations.
Other parameters of interest include NH4-N, TOC, TKN, and TP; these are available in the corresponding
data file but are not discussed here because in most cases the observed concentrations were less than the
quantitation limits for these analytes. Quantitation limits were 0.05 mg/L for NH4-N, 0.5 mg/L for TOC, and
0.1 mg/L for both TKN and TP.
29
-------
8,5rw(%«) 8,SPW(%°)
Figure 21, Stable nitrate nitrogen isotope data for the two center wells, ORN and ORS. The first panel
shows the time series of nitrate (mg L-l) and Sl5N-nitrate (%o) from WY17-WY19 (and beginning of
WY20), Black arrows represent timing of fertilizer additions. Rectangles capture time periods
identified as showing strong negative trends between nitrate concentrations and 6l5N-nitrate values,
which are plotted linearly in the second panel. The dual isotope graph of 6l5N-nitrate (%o) vs 6180-
nitrate (%o) in the third panel distinguishes whether identified periods of decreasing nitrate
concentrations with increasing 6l5N-nitrate values are due to distinct nitrate source signatures (and
thus potentially represents times of source-mixing) or microbial processing (i.e. times in which
denitrification dominates).
3.3.2 Ly si meters
In addition to the variability of nitrate profiles in the wells, there was also quite a bit of variability in nitrate
profiles among the replicate lysimeters (Figure 22). Although nitrate concentrations were generally lower in
pore water obtained from the deepest lysimeters at 10-ft depths, this trend was not consistent, and in
some cases nitrate concentrations were greatest in pore water obtained from the mid-level lysimeters at
30
-------
a) North Plot - North Lysimeter Nitrate
160
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
d) South Plot - North Lysimeter Nitrate
160
140
120
100
b) North Plot - East Lysimeter Nitrate
40
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
e) South Plot - East Lysimeter Nitrate
c) North Plot - South Lysimeter Nitrate
£ ioo
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
f) South Plot - South Lysimeter Nitrate
ioo
90
80
"52
E
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
50
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
Figure 22. Lysimeter nitrate concentrations at a) north plot north lysimeter, b) north plot east lysimeter, c) north
plot south lysimeter, d) south plot north lysimeter, e) south plot east lysimeter, and d) south plot south lysimeter.
~ N Application Dates
40 AA
O 30
08
O 20
5-ft depths. Two of the six 2.5-ft lysimeters began to show a large nitrate breakthrough beginning in August
WY19 (Figure 22b,e) following the June WY19 application of urea nitrogen at 25% greater than what was
required based on the soil nitrate
pre-sidedress test. We have reason to
believe an application error occurred,
explaining why this trend was not
exhibited in the other four lysimeters
at the same depth. Farm managers
with knowledge of the site indicated
that the process of backing large
planting equipment into the field
could sometimes release a "hot spot"
of concentrated fertilizer. Therefore,
we speculate that the large
divergence in nitrate concentrations
during this time was from some
concentrated fertilizer "hot spots"
occuring right near the two impacted
lysimeters. When the nitrate data for
the replicate lysimeters was averaged
for the two subfields, concentration Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18 Dec-18 Jun-19 Nov-19 Apr-20 Sep-20
profiles were found to be relatively Figure 23. Mean lysimeter nitrate profiles in a) north and b) south
consistent across the three depths subfields. The north subfield has the cover crop. Triangles indicate dates
(Figure 23). of fertilizer application.
31
a) Mean North Lysimeter Nitrate (N03&N02-N)
¦¦¦ vw\nAnM''A''A.v'ri""
Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jui-18 Dec-18 Jun-19 Nov-19 Apr-20 Sep-20
b) Mean South Lysimeter Nitrate (N03&N02-N)
N Application Dates
2.5 ft
-5ft
-10 ft
-------
Most changes in lysimeter water nitrate concentrations were not related to shifting sources of surface
water (i.e., irrigation water, recent precipitation), but occasionally water source did have an effect on the
nitrate levels. Unlike the nitrate concentration data, water isotopes between lysimeters at the same depth
were relatively similar (Figure 13), and thus the nitrate variation between lysimeters was not the result of
different water sources for the most part. However, when water in lysimeters shifted to reflect recent
precipitation (Figure 13), nitrate concentrations tended to decrease in the upper soil layers (Figure 24). For
water in the 2.5-ft lysimeters (yellow symbols), nitrate concentrations declined as water isotopes declined
in both WY17 and WY19, reflecting a dilution effect with the recent precipitation inputs. However, water in
the 10-ft lysimeters (brown symbols) increased in N concentration with lower water isotope values in
WY17. In WY19, water in the 5-ft lysimeters responded to the precipitation inputs, and a dilution effect was
observed. In WY19, water in the 10-ft lysimeters did not change isotopically in response to precipitation
(Figure 13), but different lysimeters had unique water isotopic values, and those with lower water isotope
values tended to have lower nitrate concentrations. This pattern in the 10-ft lysimeters was observed from
the second half of WY18 through WY19 when the water isotopes at 10 ft within a lysimeter were relatively
stable (no consistent increasing or decreasing trends). The water isotope values in those north 10-ft
lysimeters with low nitrate had isotope values that were also lower than what was observed in the
groundwater wells (Figure 13, black line), and matched more with the cumulative precipitation inputs in
Spring (Figure 13, blue and green lines). Thus, at the deepest depths, water source may also play a role in
the nitrate variation between lysimeters beyond the dilution effect observed in the upper profile.
50
Wet Winter/Spring WY17
Wet Winter/Spring WY19
40 -
O)
30 -
E
-10.5 -10.0 -9.5 -9.0 -8.5 -8.0 -7.5 -7.0-10.5 -10.0 -9.5 -9.0 -8.5 -8.0 -7.5 -7.0
d80 (%o)
D^O (%o)
Figure 24. Lysimeter water isotope values (6180) and nitrate concentrations during the wet
winter/spring periods (Feb-May) in WY17 and WY19. Triangles and circles represent north and south
subfields, respectively, at 2.5-ft (yellow symbols), 5-ft (red symbols) and 10-ft (brown symbols) depths.
32
-------
Like the water isotopes (Figure 13), stable nitrate nitrogen isotopes for almost all the lysimeters at the
same depth and in the same subfield were relatively similar (Figure 25). The largest variation in S15N-N03
values between replicate lysimeters occurred at the shallowest depth, i.e., 2.5 ft. in the north subfield. In
such instances, two of the three replicate lysimeters tended to have similar 515N-N03 values, with only one
lysimeter significantly diverging from the average 615N-N03 values of the other two. The difference in the
515N-N03 values of the one divergent lysimeter over distinct time periods appears to be driven by the
occurrence of hotspots and/or hot moments of denitrification within that one lysimeter or with fertilizer
pulses (Figure 25). While the north subfield was the one planted with cover crops, it should be noted that
North Subfield
Irrigation
Rainfall
vV
10' Lvsimeters
•1.12 (ION)
¦L15(10E)
A 1.18 (1 OS)
^6 «£> A ^ tfj -Jj ^ ^ A ^ 'j
/r V V/ /» /> Vi v-j % -J-) ¦/> Vi Vi
% *0, % /> °/> * '& '* '* 's> '* <£> \ *> <& )
for each lysimeter (not shown) was used to determine whether periods of decreasing nitrate
concentrations with increasing 6l5N-nitrate values were attributable to distinct nitrate source
signatures (and thus potentially representing times of source-mixing) or microbial processing (i.e.
times in which denitrification dominated).
33
-------
the few instances in which denitrification appeared to occur in the 2.5-ft lysimeters was from WY17
through the beginning of WY18 - a time period marked by low establishment of the cover crops. In August
2018, nitrate isotope ratios drop in both east lysimeters at the 2.5-ft depth, while the nitrate
concentrations spike. This isotopically lower signal is consistent with a fertilizer signal, and likely indicates
the leaching of the earlier fertilizer application. In the deeper 10-ft lysimeters of both the north and south
subfields changes in 615N-N03 values were less pronounced, yet distinct enough to suggest source mixing
could help explain periods of decreasing nitrate concentrations. Evidence of source mixing within the
lysimeters tended to occur following fertilizer applications in WY18 and WY19 where nitrate isotopic
signatures drop to values consistent with fertilizer indicating the breakthrough of fertilizer application at
the 10-ft depth. Nitrate levels in the soil and pore water would likely spike following the addition of a big
pulse of fertilizer nitrate (as plant uptake is unlikely to be 100%), and then slowly equilibrate back to lower,
natural soil levels as any residual fertilizer N was leached, immobilized, or taken up by the crops. Thus,
source mixing of fertilizer N with soil N may likely be responsible for the many time periods over which
nitrate concentrations decrease within the lysimeters.
With the exception of the north lysimeter in the north subfield (Figure 26a) there was little variability in
phosphate profiles for any given depth throughout the study. Phosphate data are derived from the
analyses of the unacidified nutrient samples, and since these were of a lower priority, there was often not
enough volume collected to allow for phosphate analyses. Therefore, no samples were analyzed for
phosphate at the 5-ft depths for two lysimeters in the south subfield (Figure 26f). Even though these
profiles are absent, it does not look like phosphate concentrations decreased significantly with depth. The
north lysimeter in the north subfield had relatively high phosphate concentrations at the 2.5-ft depth
throughout the study (Figure 26a). This lysimeter was also anomalously high in nitrate values at the
beginning of the study; the reason for this is not clear.
a) North Plot - North Lysimeter Phosphate
1
2.5N
-5N
-ION
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
b) North Plot - East Lysimeter Phosphate
c) North Plot - South Lysimeter Phosphate
"m
E
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
d) South Plot - North Lysimeter Phosphate
e) South Plot - East Lysimeter Phosphate
"m
E
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
f) South Plot - South Lysimeter Phosphate
"bJ
E
Sep-16 Jul-17 May-18 Feb-19 Dec-19 Sep-20
Figure 26. Lysimeter phosphate concentrations at a) north plot north lysimeter, b) north plot east lysimeter, c)
north plot south lysimeter, d) south plot north lysimeter, e) south plot east lysimeter, and d) south plot south
lysimeter.
34
-------
The replicate lysimeterTKN and TP
data were also averaged for each
subfield. For TKN, pore water values
decreased with depth and were often
below the quantitation limit (0.1 mg/L)
at the 10-ft depths (Figure 27). This
trend was not as obvious for the TP
data, and most concentrations were
below the quantitation limit of 0.1
mg/L (Figure 28).
The mean nitrate profiles were used
along with the HYDRUS-1D model to
estimate nitrogen leaching at each
depth for both subfields. As the
HYDRUS-1D model solves the Richards
equation, it computes the flux as
described in Simunek et al. (2013),
Section 5.4.5. Therefore, for a given
snapshot in time, we may plot flux
profiles, i.e., flux (in/day) as a function
of depth (in), but with depth on the
vertical scale. During solution, the model also computes the volumetric water content (in/in) at each time
step, which can be shown as a function of depth. The nitrate concentration (converted to lbs/ft3) times the
water flux (converted to ft/day) gives
the mass flux (Ibs/acre/day). The
average daily nitrate loss profiles are
shown for each depth within each
subfield in Figure 29.
For each depth, the nitrate loss profiles
in the north (interseeded) subfield
were surprisingly similar to those in the
south (conventional) subfield. This is in
stark contrast to the nitrate profiles
from the center wells in each subfield
and indicates that the well nitrate data
may be compromised by periodic
vertical flow within the well. Applying
the biweekly measured nitrate values
to each day over the two-week interval
between sampling events and
multiplying those nitrate
concentrations by the daily water flux Figure 28. Mean lysimeter total phosphate profiles in a) north and b)
values computed by the HYDRUS-1D south subfields. The north subfield has the cover crop. Triangles
indicate dates of phosphorus fertilizer application.
35
"55
E
a) Mean North LysimeterTKN
Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18 Dec-18 Jun-19 Nov-19 Apr-20 Sep-20
"55
E
Li |
L.O AA
).l
Jun-16
b) Mean South LysimeterTKN
-------
a) Average Daily N03&N02-N Loss (2.5 ft)
10.00
1.00
0.10
0.01
0.00
0.00
0.00
Oct-16 Oct-17 Oct-18 Oct-19 Sep-20
North Subfield South Subfield
10
1
0.1
•5. 0.01
0.001
0.0001
0.00001
Oct-16 Oct-17 Oct-18 Oct-19 Sep-20
North Subfield South Subfield
10
1
0.1
0.01
0.001
0.0001
0.00001
Oct-16 Oct-17 Oct-18 Oct-19 Sep-20
North Subfield South Subfield
Figure 29. Average daily nitrate loss with depth for each subfield.
model, the daily mass flux was integrated over time to give an estimate of the total annual mass flux,
converted to Ibs/acre/year, which is often used in agricultural studies of mass leaching. These data are in
Figure 30 for each year of the study and show there was generally not a large difference in leaching profiles
between the interseeded north subfield and the conventional south subfield.
In interpreting these data, it is important to realize that in this context nitrate mass loss refers to the nitrate
mass loss due to leaching from that particular depth. If no biodegradation of nitrate occurs as dissolved
nitrate is transported down through the soil column, then the annual nitrate mass loss should be similar
from one depth to the next. For WY17, it appears that nitrate mass loss was similar for the 2.5-ft and 5-ft
depths but was much lower at the 10-ft depth (Figure 30). For WY18 (Figure 30) nitrate mass loss was much
36
b) Average Daily N03&N02-N Loss (5 ft)
c) Average Daily N03&N02-N Loss (10 ft)
-------
VVY17
3S0
J 00
| 250
200
Lb
5fl 5 ft
¦ No-ih Subfield « South Subfield
VVY18
3IX)
250
¦ 1 ¦
25 ft 5 ft
¦ Mo-t:i Subfield m Scuih Subfield
WY19
300
2 SO
2uO
150
100
2,5 ft Sft
¦ fslortn Sufctfcld s South Subfield
WY20
» Mfi'lh Subfield « Scuth SuKHd
Figure 30. Annual nitrate-N mass loss with
depth for each subfield.
lower in the upper two depths compared to those in WY17,
and the extent of nitrate mass loss decreased throughout the
soil profile instead of just below the 5-ft depth as was
observed in WY17. For WY19, nitrate mass loss was generally
much less at each depth for both subfields compared to the
two previous years (Figure 30). However, this trend was not
observed at the 2.5-ft depth in WY20, which saw the highest
nitrate mass loss among all the depths and years (Figure 30).
As stated earlier, the total nitrogen applied for each year was
108 lbs/acre N in WY17,163 lbs/acre N in WY18, 200 lbs/acre
N in WY19, and 203 lbs/acre in WY20. One would then expect
that vadose zone water concentrations would increase if the
additional nitrogen was not being sequestered or lost.
The proportion of added N exported through leaching below
10 ft declined over time, from 39% and 51% of added N for
the north and south subfields, respectively, in WY17, to 3.5%
and 16% in WY20 (Table 4). This decline in the proportion of
inputs that were leached from the two subfields is surprising
considering that the overall fertilizer rates increased over
time. The nitrogen remainder within the subfields, i.e., the N
remaining after crop harvest and leaching are taken into
account, tended to be negative mostly at the 2.5-ft depth,
indicating substantial losses of N in excess of inputs, likely
due to mining of accumulated or legacy soil N from previous
years. Others have similarly seen such buildup of N in soil
from cumulative inputs of fertilizer, which carry over from
prior years and continue to affect N export, particularly in
agricultural areas (Van Meter et al., 2016). Current N
additions are not large enough to explain the level of crop N
uptake alone, much less the levels of leaching that are
occurring from these fields at the 2.5-ft depth. Positive N
remainder values at the 10-ft depth (where inputs equal the
N leached past 5 ft) provide evidence for potential N
accumulation at depth (Table 4). On average, across all years
of the study, only ~30% of the surface nitrate that leached at
the 2.5-ft depth moved below 10 ft into the deeper soil and
groundwater.
3.4 Post-Harvest Soil Data
Post-harvest soil cores were collected WY16-WY19 from six locations surrounding the monitoring networks
in each subfield. For WY16, WY17, and WY18, cores were taken from 1 ft, 3 ft, and 5 ft BGS. An additional
core was also taken at 2.5 ft BGS in WY18 to better correlate with the lysimeters at that depth. In WY19,
the 3-ft BGS core was excluded, and a deeper core at 10 ft BGS was added, thus sample cores in WY19 were
37
-------
taken at 1 ft, 2.5 ft, 5 ft, and 10 ft BGS. Analytical results for the individual core sections are presented in
the corresponding data file. The soil nitrogen profiles, as represented by soil nitrate, soil ammonium, and
soil total nitrogen are shown for each year in Figure 31.
WY16 Soil Nitrate
WY16 Soil Ammonium
WY16 Soil Total Nitrogen
Depth (ft)
WY17 Soil Nitrate
Depth (ft)
WY18 Soil Nitrate
-Norh
-South
-North
—South
6 9 12
Depth (ft)
WY19 Soil Nitrate
- North
-South
T I
1 r t I
— North
—South
K
\
E 8
2 6
E 8
2 6
— North
—South
Depth (ft)
WY17 Soil Ammonium
— North
—South
Depth (ft)
WY18 Soil Ammonium
— North
—South
Depth (ft)
WY19 Soil Ammonium
— North
—South
Depth (ft)
WY17 Soil Total Nitrogen
— North
—South
Depth (ft)
WY18 Soil Total Nitrogen
—North
—South
Depth (ft)
WY19 Soil Total Nitrogen
— North
— South
— North
—South
Depth (ft)
Figure 31. Post-harvest data for soil nitrate (a,d,g,j), soil ammonium (b,e,h,k), and soil total nitrogen
(c,f,i,l) during each year of the study for the north (interseeded) and south (conventional) subfields. Note
that the north subfield was not interseeded in WY16.
Soil nitrate levels were particularly high at 1 ft BGS for WY16 compared to the other years (Figure 31a).
Surprisingly, except for the WY16 data, soil nitrate levels did not drop appreciably with depth below those
observed at the 1-ft BGS depth (Figure 31d,g,j). The soil ammonium data were highly variable for the
intermediate depths for WY16 and WY19, making it difficult to discern any trends with depth (Figure
31b,e,h,k). However, a slight decrease in soil ammonium levels with depth could be observed in the WY17
and WY18 data. In contrast, soil total nitrogen levels showed consistent decreases with depth for all four
years (Figure 31c,f,i,l).
38
-------
There was little difference in soil phosphorus and soil organic carbon patterns from year to year (Figure 32).
Soil phosphorus levels dropped significantly between the 1-ft BGS depth and the 2.5-ft BGS or 3-ft BGS
depths but did not continue to drop at the 5-ft BGS depth or even at the 10-ft BGS depth (Figure 32a,c,e,g).
In contrast, the soil organic carbon profiles were similar to the soil total nitrogen profiles in that the
concentrations decreased
£
Q.
— 60
120
100
£
Q.
~ 60
WY16Soil Phosphorus
WY16Soil Organic Carbon
-North
-South
3 6 9:
Depth (ft)
WY17 Soil Phosphorous
-North
-South
Depth (ft)
WY18Soil Phosphorus
-North
-South
3 6 9:
Depth (ft)
WY19 Soil Phosphorous
L
\
\
•=
3 6 9
Depth (ft)
£ 15000
-North
-South
-North
-South
3 6 9 12
Depth (ft)
WY17 Soil Organic Carbon
-North
-South
3 6 9 12
Depth (ft)
WY18 Soil Organic Carbon
-North
-South
3 6 9 12
Depth (ft)
WY19 Soil Organic Carbon
-Nortl
-Soutl
3 6 9
Depth (ft)
Figure 32. Post-harvest data for soil phosphorus (a,c,e,f) and soil organic
carbon (b,d,f,h) during each year of the study for the north (interseeded)
and south (conventional) subfields. Note that the north subfield was not
interseeded in WY16.
39
steadily with depth and did not
change much from year to year
(Figure 32b,d,f,h).
The uncertainty in the time
required for infiltrating water to
reach groundwater, coupled with
the presence of preferential flow
paths, precludes the direct use of
post-harvest soil nitrate values in
one year to predict groundwater
nitrate concentrations in
subsequent years for complex
conditions such as those at this
field site. Post-harvest nitrate
values may be more useful in
sites with sandier soils and
higher soil nitrate values, but at
least one study also has shown
this not to be the case. Gehl et
al. (2006) evaluated variability of
post-harvest soil nitrate values
with depth for eight sites with
sandy textured soil along Kansas
rivers using multiple nitrogen
application and irrigation rates
for irrigated corn and concluded
that post-harvest soil nitrate
content was not conclusive
evidence for determining nitrate
leaching risk associated with
nitrogen application rates in
excess of plant uptake.
Furthermore, Gehl et al. (2006)
reported that post-harvest soil
nitrate distributions in the S-ft
soil profile were variable among
sites, variable within a field, and
sometimes variable between
years for the same field.
-------
4.0 Quality Assurance
This research was conducted under an approved Quality Assurance Project Plan (Analyses of Agricultural
Nutrient Sources of Impairment and Effects on Downstream Waters, Quality Assurance Identification
Number G-GWERD-0030111-QP-1-7). An initial plan was developed to allow for site characterization prior
to water quality data collection and was approved in May 2016. This plan was revised to include water
quality data collection and was approved in July 2016. A field technical systems audit was conducted in
June 2016 at a separate site in Oklahoma which had been instrumented and operated under similar
conditions as that in Oregon. One Finding requiring corrective action, four Observations with
recommendations, and one Noteworthy Best Practice were identified, primarily relating to sample date
labeling, changes in sampling procedures, and summary data table entry and management. These were all
addressed in the revised final plan which was approved in August 2017, and the audit was closed in August
2017. Results of the audit had no impact on the data that were collected prior to the audit but improved
the overall quality of the data and the communication of data analysis following revision of the Quality
Assurance Project Plan.
Field data were obtained as described in this report and met the project's data quality requirements.
Equipment used for field sampling, including meters and pumps, were tested prior to each sample event.
Those devices using rechargeable batteries were fully charged prior to each sample event, and extra
batteries were brought along for the other battery-powered devices. New PharMed tubing was installed on
the peristaltic pumps, and additional lengths of PharMed tubing were brought along to change out
between wells. Meters for turbidity, pH, specific conductance, and ORP measurements were calibrated
prior to use on a daily basis, and the calibration data was recorded in the field notebook. Calibration checks
were performed before use and after the last sample measurement of the day. Calibration checks for
turbidity were performed using the 20.0 NTU turbidity calibration standard. Calibration checks for pH,
specific conductance, and ORP measurements were performed using the YSI 5580 Confidence Solution,
which provides an acceptable range of values for those parameters. The target goals for calibration checks
were within ± 0.2 pH units for pH and within ±15% of known concentrations for all other checks. These
calibration checks were also recorded in the field notebook. If a calibration check failed, this was recorded
in the field notebook and the possible causes of the failure were investigated. Upon investigation,
corrective action(s) were taken, and the instrument was recalibrated. During the course of this project,
there were occasions where some calibration checks were out of range, and in these instances the data
were flagged in the summary data table. However, these instances were rare, and the discrepancies were
not extreme, and so the data were still considered to be useable.
Field blanks, equipment blanks, temperature blanks, and field duplicates were utilized to assess data
quality during collection and transport of water samples. Field blanks were used to assess contamination
introduced from the sample containers or the applied preservatives. Equipment blanks were used to check
whether sample cross-contamination occurred due to insufficient purging of the sampling pump, sample
filter bypass system, or flow-through cell. Equipment blanks were also used to determine whether filtration
of the samples introduced contaminants. Water for field and equipment blanks was prepared by filling a
20-L carboy with RO water. For field blanks, this water was dispensed directly into the appropriate sample
containers while in the field, and preservatives were added as required. Field blank samples were not
filtered, even though actual samples were filtered for certain analytes. For equipment blanks, RO water
from the carboy was pumped through the same sample filter bypass system and flow-through cell used for
40
-------
the site wells. The water was first pumped for 15 min to purge the system prior to sampling, and samples
were then collected and processed (including filtration) as described in the sampling protocol. The
temperature blanks were prepared by filling a 30-mL plastic bottle with tap water, and these were then
taken to the field and one blank was added to each ice chest containing samples to determine whether the
samples were properly chilled during transport back to RSKERC. Finally, field duplicates were obtained at a
rate of one duplicate for every ten samples to assess data precision. For the vast majority of samples,
conditions and analyses met the requirements of the data quality objectives. Occasionally an analyte was
detected in the field or equipment blank at concentrations above the quantitation limit and these occasions
were flagged. However, in almost all cases these concentrations were at or near the quantitation limit and
corresponding samples from previous or subsequent sampling events in which these exceedances did not
occur generally showed consistent values, and so these instances had no impact on data usability. In one
instance an ice chest of samples was sent to the wrong location and as a consequence all samples were at
room temperature upon arrival. All the impacted samples were flagged, but the only data that are
considered suspect were the analytes for samples that were not preserved with acid.
Laboratory analyses consisted of analysis of general parameters and metals/cations at RSKERC, analysis of
stable isotopes at PESD's ISIRF facility, and soil texture and analyses at OSU's SWFAL facility. RSKERC
laboratory instrumentation used for analysis of project analytes are in routine use and were tested for
acceptable performance through the analysis of standards and QC samples according to internal standard
operating procedures. The QC checks and acceptance criteria in the RSKERC SOPs were adequate to meet
the project requirements. Routine inspection and maintenance of these instruments was documented in
instrument logbooks, and the standard operating procedures were followed for instrument testing and
corrective actions. When corrective actions occurred in laboratory analysis, they were documented in the
analytical report and the samples were reanalyzed. Other QC metrics included in analytical reports were %
recovery, calibration check results, precision of laboratory duplicate analyses, and matrix spike recoveries.
Any exceedances of acceptance criteria were documented in the analytical reports and these data were
flagged. These occasions were very rare, and the data were determined to be useable. Performance
evaluation samples were routinely submitted to the RSKERC lab twice a year for the project analytes in
water and the resulting analyses met the data quality objectives during this project.
For the stable isotope analyses conducted at PESD's ISIRF facility, volume variance and drift were corrected
for in each run and the data were calibrated to the standards on the international isotope scale. Memory
effects were calculated and the first few injections that showed memory from the previous sample were
discarded. Each run was calibrated with three standards and an independent fourth standard was used as a
QC check on all the calculations. Two samples were duplicated in each run, one that was split (beginning
and end of run), and the other was in the middle of the run. All of these QC metrics were summarized in
the individual analytical reports. Overall, the data quality was excellent and always met the data quality
objectives. Measured precision from the standard deviation of 135 lab duplicates was 0.27 and 0.08 %o for
62H and 6180, respectively. Accuracy across 68 isotopic runs averaged 0.01 ± 0.23%o SD and 0.05 ± 0.08%o
SD for 62H and 6180, respectively. On one occasion the samples had to be rerun because of a few low
repeatability numbers. The ISIRF facility also participated in three IAEA Interlaboratory Comparison studies
and was scored Excellent and was ranked 15th in uncertainty of the 235 labs that participated. For the soil
analyses conducted at OSU's SWFAL facility, accuracy and precision of test results were assured through
daily analysis of quality control samples, a three-step internal data review process, and participation in
external certification and sample exchange programs. All instruments were calibrated with certified
41
-------
standards and maintained according to specifications. One soil check standard was included for every nine
samples for each soil pH, nitrate, and phosphorus, ammonium, and potassium analysis. A certified soil was
also used as a check for analysis of OC and TN. The permissible ranges were set at two times the standard
deviation. If results were outside the permissible ranges, corrective actions were taken, including re-
analysis of affected samples. If additional sample was not available for re-analysis, results were flagged. In
addition, SWFAL participates in the Agricultural Laboratory Proficiency Program by analyzing test soils three
times per year, and a performance evaluation soil sample for nitrate and phosphate was also submitted by
RSKERC, and the reported ranges were within acceptable limits. All of the sample data reported met the
data quality objectives.
Specific circumstances precluded the continuous acquisition and analysis of certain analytes during the
four-year project period. Following the 10/15/19 sampling event, RSKERC made the decision to eliminate
TKN analyses to minimize hazardous waste generation. An attempt was made to analyze samples for TN
and calculate TKN by difference, but subsequent comparison of the different analyses showed
overestimation of TN in many samples. Therefore, neither TN nor TKN analyses were conducted on samples
collected after 10/15/19. In mid-March 2020, RSKERC was shut down due to concerns over the COVID19
pandemic. Samples continued to be collected by Oregon field personnel and were archived. RSKERC
resumed laboratory operations in mid-August 2020 and began analyzing archived as well as current
samples. Because holding times had been exceeded for the unacidified samples for nitrate, nitrite,
phosphate, and TOC, these archived samples were not analyzed during this time period. The results of
analyses of unacidified samples for bromide and acidified samples for combined nitrate and nitrite
(reported as nitrate throughout this report) were considered acceptable despite the extended holding
times, and so these data are reported for the entire duration of the project.
5.0 Literature Cited
Abdulkareem, J., Abdulkadir, A., Abdu, N., 2015. A review of different types of lysimeter used in solute
transport studies. Int. J. Plant Soil Sci. 8, 1-14. https://doi.org/10.9734/iipss/2015/18Q98.
American Public Health Association, 2017. Standard methods for the examination of water and wastewater
(23rd Edition). Baird, R.B., Eaton, A.D., Rice, E.W. (Eds.). American Public Health Association, American
Water Works Association, and Water Environment Federation, Washington, D.C., p. 1504.
Burket, J.Z., Hemphill, D.D., Dick, R.P., 1997. Winter cover crops and nitrogen management in sweet corn
and broccoli rotations. HortScience 32, 664-668. https://doi.Org/10.21273/HQRTSCI.32.4.664.
Burow, K.R., Nolan, B.T., Rupert, M.G., Dubrovsky, N.M., 2010. Nitrate in groundwater of the United States,
1991-2003. Environ. Sci. Technol. 44, 4988-4997. https://doi.org/10.1021/esl00546v.
Carey, B.M., Harrison, J.H., 2014. Nitrogen dynamics at a manured grass field overlying the Sumas-Blaine
Aquifer in Whatcom County. Washington State Department of Ecology Publication No. 14-03-001, Olympia,
WA.
Carey, B.M., Pitz, C.F., Harrison, J.H., 2017. Field nitrogen budgets and post-harvest soil nitrate as indicators
of N leaching to groundwater in a Pacific Northwest dairy grass field. Nutr. Cycl. Agroecosyst. 107,107-123.
https://doi.org/10.1007/slQ705-016-9819-5.
42
-------
Casciotti, K.L., Sigman, D.M., Galanter Hastings, M., Bohlke, J.K., Hilker, A., 2002. Measurement of the
oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method. Anal.
Chem. 74, 4905-4912.
Chung, S.-O., Horton, R., 1987. Soil heat and water flow with a partial surface mulch. Water Resour. Res. 23,
2175-2186. https://doi.org/10.1029/WR023i012pQ2175.
Colaizzi, P.D., Evett, S.R., Brauer, D.K., Howell, T.A., Tolk, J.A., Copeland, K.S., 2017. Allometric method to
estimate leaf area index for row crops. Agron. J. 109, 883-894.
https://doi.org/10.2134/agroni2016.ll.0665.
Compton, J.E., Goodwin, K.E., Sobota, D.J., Lin, J., 2020. Seasonal disconnect between streamflow and
retention shapes riverine nitrogen export in the Willamette River Basin, Oregon. Ecosystems 23,1-17.
https://doi.org/10.1007/sl0021-019-0Q383-9.
Compton, J.E., Pearlstein, S.L., Erban, L., Coulombe, R.A., Hatteberg, B., Henning, A., Brooks, J.R., Selker,
J.E., 2021. Nitrogen inputs best predict farm field nitrate leaching in the Willamette Valley, Oregon. Nutr.
Cycl. Agroecosys. 120, 223-242. https://doi.org/10.1007/slQ705-021-10145-6.
Conlon, T.D., Wozniak, K.C., Woodcock, D., Herrera, N.B., Fisher, B.J., Morgan, D.S., Lee, K.K., Hinkle, S.R.,
2005. Ground-water hydology of the Willamette Basin, Oregon. U.S. Geological Survey Scientific
Investigations Report 2005-5168, Reston, VA, p. 83.
Dabney, S.M., Delgado, J.A., Meisinger, J.J., Schomberg, H.H., Liebig, M.A., Kaspar, T., Mitchell, J., Reeves,
W., 2010. Using cover crops and cropping systems for nitrogen management. In: Delgado, J.A., Follett, R.F.
(Eds.), Advances in nitrogen management for water quality. Soil and Water Conservation Society, Ankeny,
IA, pp. 230-281.
Delgado, J.A., Dillon, M.A., Sparks, R.T., Essah, S.Y., 2007. A decade of advances in cover crops. J. Soil Water
Conserv. 62,110A-117A.
Dinnes, D.L., Karlen, D.L., Jaynes, D.B., Kaspar, T.C., Hatfield, J.L., Colvin, T.S., Cambardella, C.A., 2002.
Nitrogen managment strategies to reduce nitrate leaching in tile-drained Midwestern soils. Agron. J. 94,
153-171. https://doi.org/10.2134/agroni2002.1530.
DOGAMI, 2009. LiDAR remote sensing data collection by Watershed Sciences, Inc.: Willamette Valley phase
I, Oregon. Oregon Department of Geology and Mineral Industries, Portland, OR.
Doherty, J., 2010. PEST, model-independent parameter estimation, user manual (5th edition). Watermark
Numerical Computing, Bethesda, MD.
Feaga, J.B., Selker, J.S., Dick, R.P., Hemphill, D.D., 2010. Long-term nitrate leaching under vegetable
production with cover crops in the Pacific Northwest. Soil Sci. Soc. Am. J. 74, 186-195.
https://doi.org/10.2136/sssai2008.0178.
Feddes, R.A., Kowalik, P.J., Zaradny, H., 1978. Simulation of field water use and crop yield. John Wiley &
Sons Inc., New York, NY.
43
-------
Flenet, F., Kiniry, J.R., Board, J.E., Westgate, M.E., Reicosky, D.C., 1996. Row spacing effects on light
extinction coefficients of corn, sorghum, soybean, and sunflower. Agron. J. 88,185-190.
https://doi.org/10.2134/aeronil996.00021962008800020011x.
Gehl, R.J., Schmidt, J.P., Godsey, C.B., Maddux, L.D., Gordon, W.B., 2006. Post-harvest soil nitrate in
irrigated corn. Soil Sci. Soc. Am. J. 70, 1922-1931. https://doi.org/10.2136/sssai2006.033Q.
Groning, M., Lutz, H.O., Roller-Lutz, Z., Kralik, M., Gourcy, L., Poltenstein, L., 2012. A simple rain collector
preventing water re-evaporation dedicated for 6180 and 62H analysis of cumulative precipitation samples.
J. Hydrol. 448-449, 195-200. https://doi.Org/10.1016/i.ihvdrol.2012.04.041.
Hart, J.M., Sullivan, D.M., Myers, J.R., Peachey, R.E., 2010. Nutrient management guide: Sweet corn
(Western Oregon). Oregon State University Extension Service Publication EM 9010-E, Corvallis, OR, p. 21.
Hoppe, B., White, D., Harding, A., Mueller-Warrant, G., Hope, B., Main, E., 2014. High resolution modeling
of agricultural nitrogen to identify private wells susceptible to nitrate contamination. J. Water Health 12,
702-714. https://doi.org/10.2166/wh.2014.047.
Howarth, R.W., Swaney, D.P., Boyer, E.W., Marino, R., Jaworski, N., Goodale, C., 2006. The influence of
climate on average nitrogen export from large watersheds in the Northeastern United States.
Biogeochemistry 79, 163-186. https://doi.org/10.1007/slQ533-006-9010-l.
Kite-Powell, A.C., Harding, A.K., 2006. Nitrate contamination in oregon well water: Geologic variability and
the public's perception. J. Am. Water Resour. As. 42, 975-987.
Lin, J., Compton, J.E., Leibowitz, S.G., Mueller-Warrant, G., Matthews, W., Schoenholtz, S.H., Evans, D.M.,
Coulombe, R.A., 2019. Seasonality of nitrogen balances in a Mediterranean climate watershed, Oregon, US.
Biogeochemistry 142, 247-264. https://doi.org/10.1007/slQ533-018-0532-0.
Panday, S., Langevin, C.D., Niswonger, R.G., Ibaraki, M., Hughes, J.D., 2013. MODFLOW-USG version 1: An
unstructured grid version of MODFLOW for simulating groundwater flow and tightly coupled processes
using a control volume finite-difference formulation. Techniques and Methods. U.S. Geological Survey
Techniques and Methods Report 6-A45, Reston, VA, p. 78. https://doi.org/10.3133/tm6A45.
Peachey, R.E., Hart, J., 2009. Nitrogen fertilizer use efficiency in sweet corn; effect of variety. Research
Report to the Oregon Processed Vegetable Commission and the Agricultural Research Foundation, Corvalis,
OR.
Pennino, M.J., Compton, J.E., Leibowitz, S.G., 2017. Trends in Drinking Water Nitrate Violations Across the
United States. Environ Sci Technol 51,13450-13460. https://doi.org/10.1021/acs.est.7b04269.
Quemada, M., Lassaletta, L., Jensen, L.S., Godinot, O., Brentrup, F., Buckley, C., Foray, S., Hvid, S.K.,
Oenema, J., Richards, K.G., Oenema, O., 2020. Exploring nitrogen indicators of farm performance among
farm types across several European case studies. Agr. Syst. 177.
https://doi.Org/10.1016/i.agsv.2019.102689.
Raun, W.R., Johnson, G.V., 1999. Improving nitrogen use efficiency for cereal production. Agron. J. 91, 357-
363. https://doi.org/10.2134/agronil999.00021962009100030Q01x.
44
-------
Schwede, D.B., Lear, G.G., 2014. A novel hybrid approach for estimating total deposition in the United
States. Atmos. Environ. 92, 207-220. https://doi.Org/10.1016/i.atmosenv.2014.04.008.
Sela, S., van Es, H.M., 2018. Dynamic tools unify fragmented 4Rs into an integrative nitrogen management
approach. J. Soil Water Conserv. 73, 107A-112A. https://doi.Org/10.2489/iswc.73.4.107A.
Sigman, D.M., Casciotti, K.L., Andreani, M., Barford, C., Galanter, M., Bohlke, J.K., 2001. A bacterial method
for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal. Chem. 73, 4145-4153.
Simunek, J., Sejna, M., Saito, H., Sakai, M., van Genuthten, M.T., 2013. The HYDRUS-1D software package
for simulating the one-dimensional movement of water, heat, and multiple solutes in variably-saturated
media. Version 4.17. HYDRUS Software Series 3. Department of Environmental Sciences, University of
California Riverside, Riverside, California, p. 342.
Singh, G., Kaur, G., Williard, K., Schoonover, J., Kang, J., 2018. Monitoring of water and solute transport in
the vadose zone: A review. Vadose Zone J. 17. https://doi.org/10.2136/vzi2016.07.0Q58.
Sullivan, A.B., Rounds, S.A., 2004. Modeling streamflow and water temperature in the North Santiam and
Santiam Rivers, Oregon, 2001-02. U.S. Geological Survey Scientific Investigations Report 2004-5001,
Portland, OR, p. 35.
Sullivan, D.M., Andrews, N., Brewer, L.J., 2020a. Estimating plant-available nitrogen release from cover
crops. Pacific Northwest Extension Publication PNW 636, Corvallis, OR, p. 20.
Sullivan, D.M., Cogger, C.G., Bary, A.I., Bittman, S., Brewer, L.J., 2021. Postharvest soil nitrate testing for
manured grass and silage corn (West of the Cascades). Oregon State University Extension Service
Publication EM 8832, Corvallis, OR, p. 12.
Sullivan, D.M., Peachey, E., Heinrich, A., Brewer, L.J., 2020b. Nutrient and soil health managment for sweet
corn (Western Oregon). Oregon State University Extension Service Publication EM 9272, Corvallis, OR, p. 28.
Thomson, S.J., Armstrong, C.F., 1987. Calibration of the Watermark Model 200 soil moisture sensor. Appl.
Eng. Agric. 3,186-189. https://doi.org/10.13031/2013.26670.
Tonitto, C., David, M.B., Drinkwater, L.E., 2006. Replacing bare fallows with cover crops in fertilizer-
intensive cropping systems: A meta-analysis of crop yield and N dynamics. Agr. Ecosyst. Environ. 112, 58-
12. https://doi.Org/10.1016/i.agee.2005.07.003.
Valkama, E., Lemola, R., Kankanen, H., Turtola, E., 2015. Meta-analysis of the effects of undersown catch
crops on nitrogen leaching loss and grain yields in the Nordic countries. Agr. Ecosyst. Environ. 203, 93-101.
https://doi.Org/10.1016/i.agee.2015.01.023.
van Genuchten, M.T., 1980. A closed-form equation for predicting the hydraulic conductivity of unsaturated
soils. Soil Sci. Soc. Am. J. 44, 892-898. https://doi.org/doi.org/10.2136/sssail980.036159950044000500Q2x.
Van Meter, K.J., Basu, N.B., Veenstra, J.J., Burras, C.L., 2016. The nitrogen legacy: emerging evidence of
nitrogen accumulation in anthropogenic landscapes. Environ. Res. Lett. 11. https://doi.org/10.1088/1748-
9326/11/3/035014.
45
-------
Vanderborght, J., Vereecken, H., 2007. Review of dispersivities for transport modeling in soils. Vadose Zone
J. 6, 29-52. https://doi.org/10.2136/vzi2006.0Q96.
Weitzman, J.N., Brooks, J.R., Compton, J.E., Faulkner, B.R., Mayer, P.M., Peachey, R.E., Rugh, W.D.,
Coulombe, R.A., Hatteberg, B., Hutchins, S.R., 2022. Deep soil nitrogen storage slows nitrate leaching
through the vadose zone. Agr. Ecosyst. Environ. 332, 1-13. https://doi.Org/10.1016/i.agee.2022.107949.
Wesseling, J., Elbers, J., Kabat, P., Van den Broek, B., 1991. SWATRE: Instructions for input, internal note.
Winand Staring Center, Wageningen, Netherlands.
Zhang, X., Davidson, E.A., Mauzerall, D.L., Searchinger, T.D., Dumas, P., Shen, Y., 2015. Managing nitrogen
for sustainable development. Nature 528, 51-59. https://doi.org/10.1038/naturel5743.
46
-------
Appendix A. Soil Water Characteristic Curves
Our in-situ sensors provide abundant data on capillary pressure head (h) and volumetric water content (0)
observed during the study (Figure Al). The Richards equation, which can describe the time-evolution of
during flow of water in the unsaturated soil zone, may be written as
dd_d_
dt dz
de
D(e)--K{e)
Where z is depth (cm) in the soil profile, K(h) is the unsaturated hydraulic conductivity, and D(tf), the soil
water diffusivity is
dh
D(B)=K(e)-
It is apparent that the relationship between and h must be known, but no tractable physical theory for
this exists at the field scale. The scatter in the data seen in Figure Al is due to hysteresis, and the high
sensitivity in the h and response during hundreds of episodes of wetting and drying within countless,
practically random, microscopic pore geometries. It suggests a statistical component in their relationship.
We therefore rely on a semi-empirical approach. Perhaps the most common culminated in work by van
Genuchten (1980), which has been expanded in various ways, but is essentially given by
(0S — 0r*)
0(h) = 0r +¦ s r
[1 + (ah)n]m
where &r and &s are the residual and saturated volumetric water contents, a, n, and m are fitting
parameters. In this model, m = 1 — 1/n. For practical purposes, dr and &s are fitting parameters, since
their true values are not directly measurable in-situ (and they only have physical meaning anyway at the
field scale as representative volume averages, i.e., volume/volume). The parameter a is approximately
equal to the reciprocal of the value of h at the inflection point in the curve, so it is given units [1/cm].
These units are practically universal in the soil physics literature; hence we have deviated from using
English units in this appendix. As noted, the HYDRUS-1D model solves the highly nonlinear governing
equation which is made even more challenging for the numerical solver by the nonlinear relationship of the
van Genuchten (1980) model. Convergence was obtained only by a compromise among the fitting
parameters without hysteresis. This was done largely by trial and error. Ultimately, the scatter in the plots
does indicate one source of error due to simplification of wetting and drying in the model.
47
-------
101 102 103
0.42
JT 0.4
0.38
101
102
0.45
0.44
^ 0.43
0.42
0.41
Figure A-l. Observed water content versus capillary pressure
head. Scatter plot color corresponds to soil texture as the
color mixture, red-green-blue, with red being sand percent,
green as silt percent, and blue as clay percent. The continuous
blue line represents the van Genuchten equation curve used
for the HYDRUS soil layer, and black dots along the curve
represent points used during the HYDRUS simulation. Also
shown are the parameters used for the equation.
I 1 1 M
-•3 ?•
Or = 0.400
f)s = 0.450
a = 0.017
n = 2.0
10 ft
h (cm H20)
48
-------
&EPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
Recycled/Recyclable Printed on paper that contains a minimum of
50% postconsumer fiber content processed chlorine free
------- |