Study Approach

The first phase of the study will integrate and build upon the work by recent studies such as RASA, the DEQ/USGS surface-water study, NAWQA, and the Portland Basin ground-water study. The primary goals of this phase will be to: (1) develop a quantitative, regional understanding of the hydrologic system, including the ground- and surface-water subsystems and their relation; (2) identify needs for further data collection; and (3) identify processes and areas for detailed study in the second phase of the study.

Focus studies during the second phase will concentrate data collection and analysis on specific processes and areas of the basin. Goals of this phase will be to: 1) gain a quantitative understanding of the hydrologic and geologic characteristics of the basin that determine the response of the hydrologic system to changes in water use and allocation, land use, and climate, and 2) develop models of the hydrologic system capable of predicting that response. The size of the Willamette River Basin does not permit detailed study of every area, however, the knowledge of processes gained from the focus studies will be applied to other parts of the basin with similar hydrologic and geologic conditions.

It is anticipated that OWRD staff will participate in the study, assisting with the design and completion of tasks in both phases.

Phase I--Regional Analysis

Data Compilation and Collection

A. Literature search --Early in the study a thorough bibliographic search will be made to locate new publications on the hydrogeology, water quality, and geology of the area. A bibliography of hydrogeologic literature with over 500 citations was compiled as part of the Willamette RASA study (Morgan and Weatherby, 1992). This bibliography will be updated with recent publications and the updated bibliography will be made available to other agencies and the general public on the Internet.

B. Establish spatial data library--A complete inventory of spatial (map) data and information stored in geographic information system (GIS) format will also be compiled for the study area. A great number of data themes have been generated for the Willamette Valley by previous USGS studies, including the RASA and NAWQA studies, by other federal agencies, and by state agencies. Some of these themes include land-use, digital topography, precipitation, soils, geology, and vegetation. The inventory of spatial data will include information on the source of data, scale of the original map, the format of the data, availability, and the agency and contact person. The inventory will be indexed and placed on the Internet where it will be available to the general public and other agencies. From the inventory, project personnel will be able to locate existing spatial data for use in their analyses and avoid any duplication of effort. If data are not readily available, plans will be formulated to generate or acquire the data.

C. Well data--Well information filed by drillers are currently being entered into a statewide computer database by OWRD. This effort has been accelerated in anticipation of the proposed study and the database is scheduled to be completed for the study area by January 1996. The State's Water Rights Information System (WRIS) is also being updated with digitized locations for the point of diversion and place us use for each water right.

Information for over 3,000 wells was entered in the USGS National Water Information System (NWIS) database as part of the Willamette RASA study. Since their entry in NWIS, the State has implemented a consistent statewide well-numbering system. In order to make the State and NWIS databases most useful and consistent, the wells in NWIS will be cross-referenced with the State well-identification numbers; this will make it possible to establish connections between the wells and other data maintained by the State such as water-rights, water-use, and well tests.

Although many wells were inventoried for the Willamette RASA, additional wells will be inventoried and accurately located for this study. First, the distribution of wells located for the RASA study will be reviewed and criteria will be developed to prioritize areas for additional well location work. Wells will be located with the purpose of filling gaps in our knowledge of the subsurface geology, water-levels and flow directions, water chemistry, and water-use. Likely areas include the basalt uplands in the valley, the foothills of the Cascades and Coast Range, the Oregon City uplands, and parts of the valley lowlands. Each located well will be evaluated as a potential geochemical or water-quality sampling site, as an observation well site, and as a candidate for borehole geophysical logging. Locations will be determined in the field using Global Positioning Systems (GPS) for accuracy. Each newly located well will be entered in the NWIS database.

D. Borehole geophysics--Existing borehole geophysical data will be compiled and evaluated for use in interpreting the subsurface geology. The location, type and quality of each log will be determined and entered in the well database as ancillary data. Locational data will also be entered in the GIS database. After reviewing available data, various borehole geophysical methods will be considered for use in collection of new data. Methods will be selected based on their likelihood of providing data for interpretation of the thickness and extent of hydrogeologic units. Borehole geophysical methods to be considered include natural gamma, temperature, electrical, neutron, and flow-meter logging. These data will augment the basic data from the driller's report on the geology and hydrology of the aquifers.

E. Surface geophysics--Surface geophysical data collected by previous studies will be compiled and evaluated for use in refining subsurface geologic mapping. An ongoing aeromagnetic study by the Geologic Division of USGS may provide valuable information on bedrock structure and thickness of sediments in the valley. The USGS Branch of Geophysical Applications and Support will be consulted regarding surface geophysical methods that may be applicable and cost-effective in for the types of geologic materials found in the study area.

F. Geochemical sampling--A program of sampling wells, streams, and springs will be implemented to determine the feasibility of using geochemical indicators to identify ground-water flow paths and residence times, and the nature of ground-water interactions with streams. The first step in planning this program will be to review existing literature on the geochemistry of ground water in the valley. The work by the Willamette NAWQA study will provide the most comprehensive overview of water quality and data availability in the valley. A limited number of sites (approximately 20) will be selected for the initial sampling. Most of these sites will lie on ground-water flow paths inferred from potentiometric surface maps. Two-dimensional ground-water models were constructed along two such flow paths for the RASA study; these flow paths will be targeted for sampling to provide geochemical data that can be used to compare with flowpaths and residence times computed by the models. The concentrations of major ions, stable isotopes of oxygen and hydrogen will be analyzed in most samples and tritium and carbon isotopes will be analyzed in a subset of samples. Work done as part of the Willamette Basin NAWQA suggests that chlorofluorocarbons (CFCs) may also be useful for investigating ground-water/surface-water relations in the basin. If the initial sampling proves that geochemical indicators can help delineate the flow system, additional sampling will be done in areas where more information is needed.

G. Stream gain/loss analysis--Work by previous investigators shows that relations between the ground-water system and the surface water system are both spatially and temporally complex. Gaining and losing reaches of the main stem Willamette River and its major tributaries will be delineated for various times of the year. Data from existing gages will be used along with discharge measurements. The seasonal variation in ground-water flux to and from the surface-water system will be investigated by making multiple discharge measurements on selected river reaches at various times during the year. The basic elements of this task include review of existing data and previous studies, identification of river and stream reaches for additional analysis, determination of the need for and location of additional gaging stations, selection of timing and frequency of discharge measurements, and selection of individual measurement sites. Measurements will be made using an acoustic doppler flow meter for highest accuracy in resolving gains and losses. Dye tracer tests may also be used to identify reaches of streams where water leaves the channel and moves through bed and bank deposits before returning the channel. This phenomenon, known as hyporheic flow, may be an important to interpreting the results of seepage measurements and understanding the nature of ground-water/surface-water relations.

H. Ground-water levels--Networks of wells will be established to monitor fluctuations in ground-water levels in response to recharge, surface water conditions, and ground-water pumping. Monitoring wells will be selected after reviewing historic observation well data and water level maps. Water levels will be measured annually, quarterly, or monthly in most wells, however, as many as 20 wells will be instrumented with continuous recorders to monitor water-level changes on an hourly or daily frequency.

I. Water-use data--The Willamette RASA study estimated water use bysource (ground water or surface water) for industrial, municipal, and agricultural needs in 1990. These estimates will be refined and revised for 1995. Water-rights data from the State WRIS will be used with current information on municipal and industrial water use to update estimates of water use where significant changes have occurred. Agricultural water use will be estimated using crop mapping based on processing of satellite imagery and power consumption records for well pumps in selected areas. Ground-water pumpage will be estimated for each hydrogeologic units by determining the hydrogeologic unit tapped by each well.

Estimated water-use for 1990 and 1995 will be compared to total water-right allocations for those years to determine the fraction of allocated water that is actually used. Using this relation, water use for other years can be estimated to assess historic trends in the amount and location of water use by source.

Data Analysis and Refinement of Concepts

This task involves the analysis and evaluation of both previous work and data collected during phase I to expand and refine the conceptual model of the Willamette Basin hydrologic system. The goal of this task is to provide the foundation for constructing computer models of the hydrologic system. Specific work elements include:

A. Analysis of hydrographic data--Available hydrographic data for wells and streams and climatic data will be analyzed to determine temporal trends (both short and long term) in ground-water levels and stream discharge. In addition, relations between factors such as water-use, land-use, and climate, and the response of the ground- and surface-water subsystems can be determined from the hydrographic data.

Gaging station histories and statistical summaries will be prepared for all long-term stations and their locations stored in the GIS database. Hydrographs will be plotted for selected time periods and analyzed for correlation with climatic and water-use trends. Baseflow recession analysis will be performed on records for unregulated basins to provide data for comparison with the watershed models to be used in the study.

Data for long-term ground water observation wells operated by OWRD will be inventoried and their records entered into the NWIS database. Hydrographs will be plotted for each well and analyzed for long-term trends related to climate and/or water use and to determine ranges of seasonal fluctuation caused by climate and surface-water effects. As data become available from the ground-water level recorders deployed for this study, they will be compared with long-term trends to determine the state of the hydrologic system. Finally, the location of all water-level monitoring wells will be entered in the GIS database.

B. Evaluation of watershed and streamflow routing models for the basin--The models developed during the USGS/DEQ study will be used to calculate estimates of ground-water recharge from infiltration of precipitation and ground-water discharge (baseflow) to streams. Results from the models will be compared with information from other sources in order to evaluate their uncertainty. Recharge estimates made during the RASA study will be compared with the watershed models, and baseflow separation techniques will be used on unregulated streams to independently estimate the ground-water component of streamflow for several watersheds.

C. Refine geologic framework--The hydrogeologic units mapped in the study area during the RASA study will be further refined and expanded to allow this study to address the issues of water-level declines in basalt aquifers and ground-water/surface-water relations. The focus of this task will be to map the thickness and extent of the basalt aquifers in the valley and to differentiate the alluvial gravel deposits adjacent to the Willamette River. Work elements for this task include coding lithologic data from the driller's logs of new field-located wells, plotting geologic sections, interpreting borehole geophysical logs, and incorporating the results of any surface geophysical studies. Differentiating the alluvial gravel deposits will be difficult from drillers logs and other sources of data such as aerial photography, vegetation maps, and topographic maps will be used.

D. Concept testing with cross-sectional ground-water models--Two cross sectional (two dimensional) ground-water flow models were constructed as part of the RASA study. These models, and perhaps one or two additional cross sectional models, will be used to test and refine concepts of how the ground-water system works and estimate the hydraulic characteristics of the system, prior to developing the regional scale three dimensional model. Using cross-sectional models greatly simplifies the data preparation and reduce the computational time for making simulations. Such models can yield important insights as to the relative sensitivity of the system to parameters and boundary conditions. For example, the cross sectional models will be used to assess the distribution of evapotranspiration and ground-water discharge to streams, and to test concepts of subsurface ground-water inflow from the Cascade and Coast Ranges.

Regional Model Development

The goal of this task is to integrate the data and concepts into computer models of the hydrologic system which can be used to quantitatively evaluate the regional hydrology of the Willamette Basin. A detailed modeling strategy will be developed during the Data Analysis and Refinement of Concepts task, however, a preliminary approach to model development is presented below.

A single model of the entire aquifer system would be the most convenient approach, however, the size of the area and the amount of detail required in subdividing the system (both horizontally and vertically), might make a single model cumbersome to develop due to the size of the data files and the computational load it would place on the computer system. If it is necessary to divide the area into two or more sub-areas for modeling purposes, natural hydrogeologic boundaries will be used. For example, if the basin were divided into two areas for modeling, a northern model extending from the Clackamas River (the southern boundary of the model constructed for the Portland Basin study) and a southern model that includes the part of the Willamette Basin south of the Salem Hills might be constructed. For most analyses, only one model would need to be run, however a seamless interface between the two models would be developed so that if basin-wide management scenarios were being developed, the models could be used easily.

The regional model(s) will be calibrated to a steady-state condition based on hydrologic data from the 1996-97. Most of the basin is in a steady-state condition where annual recharge and discharge are in balance and there is no net change in ground-water storage from year to year. Long-term observation well hydrographs will be analyzed to verify this. In some areas, however, particularly where basalt aquifers have been developed, ground-water levels are declining and there is a net loss of storage; in these areas adjustments will be made for the changes in ground-water storage so that the model can be calibrated to steady-state conditions. This type of calibration is called a time-averaged steady-state calibration because stresses on the system (recharge and discharge) and the change in storage are averaged over the calibration period and the model is calibrated to average ground-water levels for the period.

The MODFLOW program and the version of MODFLOW that includes a non-linear regression package for estimating parameters (MODFLOWP) will be used to develop the models. Several options are available for simulating the interaction of streams with MODFLOW and a decision as to which to use will be made as part of developing the modeling strategy. The primary options are the STREAM package (Prudic, 1989), and implementations of the more rigorous BRANCH and DAFLOW surface water models that have been coupled with MODFLOW by USGS researchers.

The following discussion of tasks relates to the development and use of the regional model or models.

A. Construct and calibrate regional models--Model input files describing the horizontal and vertical subdivisions of the system, boundary conditions, hydraulic characteristics, recharge, and ground-water withdrawals will be compiled and formatted from data collected and analyzed in previously described work elements.

Calibration will be done by comparing simulated ground-water levels and ground-water discharge to streams, with water-levels mapped by the RASA study and baseflow estimates from the precipitation-runoff models and measured seepage. Parameter estimation will be done with the MODFLOWP version (Hill, 1992) of the USGS modular finite-difference ground-water flow model (McDonald and Harbaugh, 1988).

B. Model analysis--Following calibration, the regional model will be used to evaluate the adequacy of available data, identify additional data needs, determine annual ground-water fluxes to and from various reaches of the Willamette River and major tributaries, determine directions and rates of ground-water movement, and assess the quantities of ground-water inflow and outflow across boundaries. In addition to providing a quantitative understanding of the flow system, the steady-state regional models can be used to begin to assess the effects of regional water resource management alternatives on various parts of the hydrologic system. The OWRD and other water-management agencies and groups will be consulted to define reasonable management strategies that can be evaluated with the model.

C. Selection of topics and areas for focus studies--As early as possible in phase I, USGS and OWRD staff will select focus study topics and areas. Criteria for selection of areas to conduct focus studies might include: areas of current water availability problems, areas where data is already available from previous or ongoing studies, and areas that are representative of conditions in others parts of the basin (for transfer value).

Statistical Analysis of Well Yield in Low-Permeability Geologic Units

The factors affecting well yield in areas underlain by marine sedimentary rocks and western Cascade volcanic rocks will be analyzed using statistical analysis. In areas of the basin underlain by low-yield aquifers, relations between well yield and factors such as well construction (diameter and depth), geology, and siting will be investigated. Information to be used in the analysis includes existing geologic maps, geomorphic setting, well logs, and aquifer test data. The results will provide potential well constructors with information on the probability of obtaining a given well yield in various geologic and topographic settings.

Distribution of Naturally Occurring Poor-Quality Ground-Water

A large scale reconnaissance of arsenic in ground water will be made by sampling from existing wells. Approximately 300-400 sites in the valley will be sampled during well inventory and water level measurement field work that will be done for other parts of phase I. The specific conductance of water from wells inventoried and measured will also be measured. Statistical relations between the occurrence of arsenic and high conductivity (salinity) and various factors will be explored. Existing water-quality data will also be used and new data will be collected in areas where no data exist. Factors that will be investigated include mineralogy of the rocks along the flowpath between recharge areas and sampling points, flowpath lengths and residence times, and geomorphic settings. The regional ground-water flow model will be used with particle tracking techniques to identify recharge areas, flowpaths, and residence times throughout the basin. Based on statistical relations between water-quality and these factors, water-quality hazard maps could be generated for various contaminants.

Phase II--Focus Studies

The following sections describe the steps that would be common to these focus area studies. At the beginning of phase II, detailed workplans will be developed for each focus study.

Data Collection

Focus Study Model Development