Title: Investigation of Groundwater Recharge and Agricultural Runoff Through the Willamette Silt, Oregon

Focus Categories: Groundwater (GW); Nitrate contamination (NC); Management and Planning (M&P)

Principal Investigator: Roy Haggerty, Department of Geosciences, Oregon State University

Congressional District: 5^th

Abstract: The Willamette Silt (WS) is a low-permeability surficial geologic unit comprised of successive dam-burst flood deposits (Glacial Lake Missoula Floods) that underlies 3100 km² (1200 mi²) of arable land in the Willamette Valley, Oregon. The Willamette Silt protects the underlying, regionally important Willamette Aquifer (WA) from agricultural contamination while acting as a semi-confining unit and diffuse recharge source. Our primary study of the hydrogeologic and geochemical properties of the Willamette Silt incorporated extensive field data collection, laboratory analyses, and numerical modeling to characterize the hydraulic and groundwater flow properties and chemical buffering capacity of the Willamette Silt.

At our field site next to the Pudding River near Mt. Angel, Oregon, groundwater flows downward at approximately 6 x 10^-7 m/s. At this rate, water crosses the Willamette Silt to the Aquifer in approximately 25 years. However, after approximately 60 years of fertilizer application at the site, the observed phosphorus and nitrate penetration fronts are located only half way through the Willamette Silt. Phosphorus is a non-conservative solute that sorbs, allowing the Willamette Silt to act as a sink for phosphorus. At our field site, the nitrate penetration front does not pass a geochemical reduction-oxidation (RedOx) boundary, providing evidence that the Willamette Silt, where reduced, prevents nitrate transport via denitrification at this boundary. If this hypothesis is correct, the regional extent of the RedOx boundary and the rate at which the RedOx boundary is moving downward through the Willamette Silt is important for management of water quality in the Willamette Valley. Both of the regional extent and the rate of movement of the RedOx boundary are currently unknown and are targets of future research.

Numerical modeling of a pump test conducted in the Willamette Aquifer shows that the Willamette Silt provides a source of diffuse recharge to the Aquifer under stressing conditions. Recharge to the Aquifer is not focussed at rivers or underneath other surface water bodies. The low hydraulic conductivity of the unit provides a hydraulic buffer to depletion of streams bottoming in the WS under pumping stress generated in the Aquifer.

Results from the study are posted at http://terra.geo.orst.edu/~haggertr/WS/

The Willamette Silt provides two functions critical to Willamette Valley water supply: (1) the low hydraulic conductivity and reducing conditions of the Willamette Silt provide a protective barrier to agricultural contamination of the underlying aquifers; and (2) the Willamette Silt acts as a semi-confining unit to the Willamette Aquifer, and thereby reduces overdraft of streams by pumping wells. Unfortunately, the physical and chemical nature of this barrier is very poorly characterized or understood. In fact no comprehensive hydrogeologic or geochemical study of the Willamette Silt has ever been conducted.

The Willamette Silt is the most extensive geologic unit exposed at the surface in the Willamette Valley of Oregon, underlying the majority of the Central and Southern Willamette Valley's arable land. It covers an area of 3100 km² (1200 mi²), virtually all of which either are currently under agricultural production, or are suitable for agricultural production. Over its entire extent, the Willamette Silt immediately overlies an important regional aquifer, the Willamette Aquifer. The Willamette Silt also lies above a second important regional aquifer, the Columbia River Basalt. In areas covered by the Willamette Silt, these two aquifers produce approximately 200,000 acre-ft per year (250,000,000 m³/yr) of water, which is 60% of the Willamette Valley's groundwater production. All of the streams and rivers in the Willamette lowland except the Willamette River bottom in the Willamette Silt.

The goal of our study was to obtain direct information on recharge and transport of agricultural chemicals across the Willamette Silt. Specifically, we intended (1) to directly measure transport rates across the Willamette Silt; (2) to characterize the hydraulic connectivity between a stream bottoming in the Willamette Silt and the Willamette Aquifer in the presence of pumping stresses; (3) to quantify the nitrate and phosphate concentrations across the Willamette Silt, along with a suite of associated ions and cations. Goals 2 and 3 were successfully reached, while the first goal was not reached.

Our study was conducted at a field site approximately one mile SW of Mt. Angel, Oregon, along the Pudding River. The Pudding River is deeply entrenched within the Willamette Silt at this location. The field where the majority of the work took place has been variously cropped in corn, clover and cereal grains from 1945 to 1982, and then onions, seed cabbage, bush beans and flower seeds to 1996. Since 1997 the field has been used to run a wholesale in-ground nursery operation. Details of the field site and fertilizer applications are given in Iverson (2002, pdf) and on the project web site.

At the field site, three locations with a total of seven boreholes were drilled into the Willamette Silt and upper Willamette Aquifer along a transect across the Pudding River. Core samples were taken in six of the boreholes, including a total of 32 m (105 ft) of continuous core. Samples were frozen on site in dry ice. Piezometers, screened at the bottom 79 cm (2.6 ft), were installed in all boreholes. Two push-point piezometers were also installed below the River at two different depths. Piezometers were developed and instrumented with pressure transducers connected to data loggers. A tipping bucket rain gauge was installed at the site and connected to one of the data loggers. A flow meter was installed in the out-flow for the tile drain. A pumping well screened within the upper Willamette Aquifer that lies within our piezometer transect and that is used for irrigation water supply was also monitored periodically.

Core samples from boreholes were analyzed for pH, phosphorus, ammonia, nitrogen as nitrate, sulfate, and a cation suite (total K, Ca, Mg, Na, Zn, Mn, Cu, and Fe). A further set of shallow test holes, traversing the same line as the piezometers, were dug to obtain more samples which were analyzed for the same chemical suite.

Water samples were taken from the Pudding River every 3.5 days from April, 2001 through February, 2002. Every other sample (i.e., weekly) was analyzed for total Kjeldahl N, total phosphorus, phosphorus as phosphate, nitrogen as nitrate, Cl, and sulfate. In two months (May and July), some samples are missing, due to equipment delivery dates and an equipment failure. While we intended to sample for 52 weeks, we sampled for less once we realized that the direct measurement of recharge across the Willamette Silt would not be possible.

Sediment cores were taken from the Pudding River near the beginning of each month from June, 2001 to November, 2001. We had originally planned to sample from the Hook Rd. bridge, making sampling possible through high water. However, we were unable to apply sufficient force through the sampler over the distance from the bridge to the sediment. Consequently, all samples were taken by wading into the water with soil coring equipment. Samples after early November were therefore not possible due to high water levels. Sediment samples were analyzed at depth intervals of 2 cm for nitrogen as nitrate, sulfate, Cl, and total Kjeldahl nitrogen.

A 3-day pump test was conducted at the irrigation well within our piezometer transect. The irrigation well, all seven of our bored piezometers, and five additional wells located within 1.4 km (0.9 mi) of the pumping well were monitored with pressure transducers and data logging equipment. The pump tests provided an average hydraulic conductivity and specific storage for the Willamette Aquifer. Through modeling we conducted later, the pump test also provided an approximate, vertically-averaged value of hydraulic conductivity for the Willamette Silt and uppermost Willamette Aquifer. Slug tests were also conducted at each of the piezometers, providing near-field values of hydraulic conductivity.

Eight samples of the Willamette Silt were analyzed in our laboratory for various properties, including hydraulic conductivity, grain size, and porosity.

A three-dimensional groundwater flow model of the field site was constructed using MODFLOW to analyze the stream-aquifer connectivity and to more fully analyze the pump test data.

The project has been successful at characterizing the hydraulic properties of the Willamette Silt and advancing our understanding of the hydraulic connection between streams that bottom in the Willamette Silt and the Willamette Aquifer. We have also significantly advanced our knowledge of the potential for agricultural leachate to cross the Willamette Silt and contaminate the Willamette Aquifer. The major findings of this study are outlined below. Other findings are described in Iverson (2002, pdf).

(1) Field observations of nitrate penetration fronts provide evidence that, at our field site near Mt. Angel, Oregon, the Willamette Silt currently prevents transport of nitrate to the Willamette Aquifer. A general trend of decreasing nitrate with depth is observed at two sites near the Pudding River. Further, the point at which nitrate concentrations go to background levels, between 7.5 and 9 m bgs (25 and 30 ft bgs), is coincident with a reduction-oxidation (RedOx) boundary which also corresponds to a sharp rise in pH with depth. This RedOx boundary is easily seen in core sample and is defined by a sharp transition from red-brown silt (oxidized) to blue-gray silt (reduced). We hypothesize that autotrophic denitrification is taking place at this boundary.

It is important to note, however, that it is not yet known the extent to which the Willamette Silt prevents nitrate transport elsewhere in the Willamette Valley. Additional research is needed to map the presence of the RedOx boundary and the thickness of the Willamette Silt below the RedOx boundary over the rest of the Valley. We also do not know whether the RedOx boundary is moving downward. If it is moving downward and if the rate of movement is influenced by fertilizer application, then future nitrate contamination of the Willamette Aquifer by nitrate may be possible.

(2) Pumping from the Willamette Aquifer has a minimal effect on streams that bottom in the Willamette Silt. Numerical analysis, supported by pump-tests, slug-tests, and lab measurements of hydraulic conductivity, show that the Willamette Silt is a source of diffuse recharge (as opposed to focussed recharge underneath surface water bodies). This diffuse recharge accounted for more than 98% of the water volume removed from the Willamette Aquifer during a 3-day pump test. Less than 1% of the volume of water removed from the Aquifer at a pumping well located less than 200 ft from the Pudding River was recharged to the Willamette Silt from the River. The Willamette Silt acts as a reservoir, accepting water during the wet winter months, and diffusely recharging that water to the Willamette Aquifer throughout the year. Consequently, the predominant net effect of pumping from the Willamette Aquifer far from the Willamette River is likely to reduce winter flows in streams.

(3) Laboratory measurements of hydraulic conductivity indicate that at our field site the harmonic mean vertical hydraulic conductivity of the Willamette Silt is approximately 2 x 10^-7 m/s (0.5 gal/day/ft²). Slug tests in shallow to intermediate wells within the Silt and with good completion provide estimates of hydraulic conductivity that are in approximate agreement. However, slug tests from wells completed at the top of the Willamette Aquifer and data from our pump tests suggest that the top of the Willamette Aquifer has lower hydraulic conductivity than the Willamette Silt, with values of approximately 1 x 10^-9 m/s ( 0.002 gal/day/ft²). This low value is probably due to cementation (visible in core and noticeable during drilling) of the gravels at the top of the Willamette Aquifer. Consequently, the hydraulic connection between the Willamette Aquifer and the Willamette Silt is lower than indicated by the hydraulic conductivity of the Silt.

While the project had these important findings and successes, the project was unsuccessful at directly measuring the rate of water movement across the Willamette Silt. This was the intended purpose of measuring concentrations of Cl and other ions in both the Pudding River and the underlying sediments. The concentration profile in the sediment was to be modeled, using the concentrations in the Pudding River as a boundary condition, to obtain a rate of transport down through the sediment. This failed because concentrations within the sediment did not show a coherent pattern related to the concentrations in the water above. We believe that this failure is due to a combination of (1) sediment movement; (2) hydraulic gradients in the sediment other than vertical; and (3) sediment-water interactions that modified the pore-water chemistry. An additional, though not fatal, problem was our inability to sample after November due to high water.

A web site has been created that provides access to all data (water chemistry, soil chemistry, head, physical properties, and hydraulic properties of the Willamette Silt) from the project, graphs of most of the data, well logs, and an electronic copy of Iverson (2002, pdf). The web site address is http://terra.geo.orst.edu/~haggertr/WS/

This grant contributed to the training of four students.

Justin Iverson. Iverson completed his MS thesis, "Investigation of the Hydraulic, Physical, and Chemical Buffering Capacity of Missoula Flood Deposits for Water Quality and Supply in the Willamette Valley of Oregon", in April, 2002. This grant supported his research and his results are the primary results presented in this report.

Louis Arighi. Arighi began his MS, "Protection of the Willamette Aquifer Due to Attenuation of Nitrate by Unweathered Quaternary Sediments", in September, 2001. The first data sets for his thesis are provided by this grant.

Bruce Hammelman. Hammelman is an undergraduate student who has worked on the project from the beginning, collecting samples, building equipment, processing data, and building a web site.

Nghi Huynh. Huynh worked on this project during its inception in 2001. His primary role was an assistant during the building of equipment and the first data collection.

The grant resulted in one thesis and one abstract, and will result in one paper currently in preparation for submission.