EPA/600/R-16/222 | July 2016 | www.epa.gov/research
United States
Environmental Protection
Agency
Decision Support System for Aquifer
Recharge (AR) and Aquifer Storage and
Recovery (ASR) Planning, Design, and
Evaluation - Principles and Technical Basis
Office of Research and Development
Water Supply and Water Resources Division
-------
EPA 600-R-16-222
July 2016
Decision Support System for Aquifer Recharge
(AR) and Aquifer Storage and Recovery (ASR)
Planning, Design, and Evaluation - Principles and
Technical Basis
Prepared By
Y.Jeffrey Yang1
Chelsea Neil2
and
Jill Neal1, James A. Goodrich3, Michelle Simon1,
Youngshin Jun4,
Daniel K. Burnell5, Robert Cohen5,
Donald Schupp6, and Rhoda Krishnan6
1. U.S.EPA, Office of Research and Development, National Risk Management Research Laboratory, Water Supply and Water
Resources Division, 26WestMartinLutherKingDrive, Cincinnati, Ohio 45268
2. U.S.EPA, Office of Research and Development, National Risk Management Research Laboratory, ORISE Fellowship
Program, 26WestMartinLutherKingDrive, Cincinnati, Ohio 45268
3. U.S.EPA, Office of Research and Development, National Homeland Security Research Center, Water Infrastructure
Protection Division, 26 WestMartinLutherKingDrive, Cincinnati, Ohio 45268
4. Washington University in St. Louis, Department of Energy, Environmental and Chemical Engineering, One Brookings
Drive, St. Louis, Missouri 6313—4899
5. Tetra Tech Inc., 45610 Woodland Road, Suite 400, Sterling, Virginia 20166
6. CB&I Federal Services LLC, 5050 Section Avenue, Cincinnati, OH 45212
Prepared For
U.S. Environmental Protection Agency
Office of Research and Development
National Risk Management Research Laboratory
Water Supply and Water Resources Division
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
July 2016
-------
Disclaimer
The U.S. Environmental Protection Agency (EPA), through the Office of Research and Development
(ORD) National Risk Management Research Laboratory (NRMRL), conducted research and development on
aquifer storage and recovery (ASR), a commonly used practice to store water in the subsurface for later
recovery and beneficial use. Research and development activities were implemented by ORD technical
personnel, contractors, and cooperative organizations. One product of the research is an ASR Decision Support
System (DSS) for ASR planning and site evaluation. This report (EPA 600-R-16-222) discusses the research
results, and describes the principles and technical basis of the DSS.
The report has been peer-reviewed, and administratively reviewed and approved for publication as an
EPA document. It is intended for informational use only. Any opinions contained in the reports are those of the
authors, and should not be construed to represent the position of EPA. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use of a specific product.
1
-------
Abstract
Aquifer recharge (AR) is a technical method being utilized to enhance groundwater resources through
man-made replenishment means, such as infiltration basins and injections wells. Aquifer storage and recovery
(ASR) furthers the AR techniques by withdrawal of stored groundwater at a later time for beneficial use. It is a
viable adaptation technique for water availability problems. Variants of the water storage practices include
recharge through urban green infrastructure and the subsurface injection of reclaimed water, i.e., wastewater,
which has been treated to remove solids and impurities. In addition to a general overview of ASR variations,
this report focuses on the principles and technical basis for an ASR decision support system (DSS), with the
necessary technical references provided.
The DSS consists of three levels of tools and methods for ASR system planning and assessment,
design, and evaluation. Level 1 of the system is focused on ASR feasibility, for which four types of data and
technical information are organized around: 1) ASR regulations and permitting needs, 2) Water demand
projections, 3) Climate change and water availability, and 4) ASR sites and technical information. These
technical resources are integrated to quantify water availability gaps and the feasibility of using ASR to meet
the volume and timing of the water resource shortages. A systemic analysis of water resources was conducted
for sustainable water supplies in Las Vegas, Nevada for illustration purposes. The Level 2 components of the
ASR DSS are intended to support ASR planning and assessment, while the Level 3 components are intended to
assist in the design and evaluation. Quantitative tools in the DSS include analytical and numerical models
capable of examining four key attributes of an ASR system: 1) ASR-Need in water availability, 2) Hydraulic
control and rate of recovery, 3) Contaminant fate and transport, and 4) Geochemical change and arsenic
mobilization. The principles and technical basis in each of these areas are described and illustrative examples
are provided.
li
-------
Acknowledgments
The research described in this report is a part of Project 1.08 "Subsurface Practice" under the EPA's
Safe and Sustainable Water Resources (SSWR) research program. It was implemented as a part of the U.S. EPA
Water Resources Adaptation Program (WRAP), partially at the contractual assistance from CB&I Federal
Services, Inc. through EPA Contract EP-C-04-034 and EP-C-14-012. The research is also supplemental to the
program needs and activities of Climate Impact, Vulnerabilities and Adaptation (CIVA) projects in the U.S.
EPA Air Climate and Energy (ACE) research program.
The project and writing team would like to acknowledge the participation of numerous technical and
administrative staff from the EPA and contracting research organizations. The Office of Water and EPA
Regions are acknowledged for participating in and guiding this research. The individuals from these
organizations include Angela Restivo (Region 6), Linda Bowling and Craig Boomgaard (Region 8), Jill Dean,
Jason Todd, Marilyn Ginsberg, Keara Moore, Matt Colombo, Joseph Tiago, and other EPA UIC colleagues in
the Office of Water. Also acknowledged are support and guidance from ORD management and individuals:
Tom Speth, Sam Hayes, Michelle Simon, Chris Impellitteri, Barbara Butler, and Hale Thurston. The
contributing teams to this report include:
Principal Investigator and Lead Author:
Dr. Y. Jeffrey Yang, P.E., D.WRE, ORD/NRMRL
EPA project research team:
Dr. Chelsea Neil, ORD/NRMRL - ORISE Program
Jill Neal, ORD/NRMRL
Dr. James Goodrich, ORD/NHSRC
Dr. Michelle Simon, P.E., ORD/NRMRL
Contract Research Organizations and Individuals:
TetraTech - GEO
Dr. Daniel Burnell, P.G.
Robert Cohen, P.G.
CB&I Federal Services LLC
Donald Schupp
Rhoda Krishnan
Washington University in St. Louis
Dr. Young-Shin Jun
Peer Reviewers:
ORD
Mark Rodgers, Barbara Butler
EPAOW
Jill Dean, Matt Colombo, Jason Todd, Richard Hall
EPA Regions
Linda Bowling, Kurt Hildebrandt, Craig Boomgaard, Janette E. Hansen
ill
-------
Executive Summary
Aquifer storage and recovery (ASR) is a widely used technical method for the storage of water in a
groundwater aquifer for later withdrawal and beneficial use. ASR application and sustainability are judged by
the rate of recovery, system efficiency, and benign environmental impact. Practices without intentional recovery
is generally called as aquifer recharge (AR). Examples include recharge through urban green infrastructure and
the subsurface injection of excess water such as treated wastewater (i.e., reclaimed water). Over the past few
years, the Safe and Sustainable Water Resources (SSWR) research on ASR practice has focused on developing
an ASR decision support system (DSS) for planning and assessment, design, and evaluation. The research
results are contained in two reports. This report provides an overview of ASR, AR, and other subsurface
activities, with a focus on the ASR technique, its principle, and technical basis. The second report will describe
the DSS software and its applications.
The DSS tools and methods are structured in three levels with the goal to facilitate ASR system design
and related permitting. Level 1 tools and methods are focused on ASR need and feasibility as they are related to
four types of data and technical information: 1) ASR regulations and permitting needs, 2) Water demand
projections, 3) Climate change and water availability, and 4) ASR sites and technical information. These
technical resources are useful to users in assessing water availability gaps and evaluating whether ASR can be
used to address volume shortages or flow imbalances in local water supplies. For illustration, a system-scale
analysis in master-planning was conducted for sustainable water supplies in Las Vegas, Nevada. The results are
presented in this report.
Levels 2 and 3 of the DSS can be used to assist ASR planning and assessment, design, and evaluation at
specific sites. During the planning and assessment, the ASR site characterization and the analysis of water
treatment needs prior to injection are two major elements of investigation. The Level 3 analysis consists of
engineering design and evaluation, for which detailed hydrogeological and geochemical characterization is
conducted, including contaminant mobilization analysis. Together, the analyses are intended to produce
technical data necessary to answer the following questions:
• What is the likely recovery rate of injected water? This planning question is pertinent for water storage
operations that are intended to address temporal or chronic water shortages. Poor recovery rates can also
negatively affect the economics of an ASR project.
• What hydrological changes occur during ASR operation? Vertical hydraulic conductivity and soil-clogging
in the vadose zone are important considerations for ASR operations that utilize spreading basins and soil
infiltration. For ASR wells into saturated zones, aquifer permeability and near-well clogging from
biological growth and inorganic precipitation are key assessment factors.
• Can geochemical reactions between the injected water and native groundwater and/or the geological
formation deteriorate the groundwater quality? During ASR operation, the injected water forms a "bubble"
by displacing the native water closest to the point of introduction and mixing with native water for some
distance away from the injection point. The point at which only native groundwater is present in pore space
defines the edge of the injection bubble. The cycle of injection-withdrawal operations will encourage
geochemical reactions and, in some cases, mobilize contaminants within the bubble.
• Given the analysis, what type of site-specific monitoring program should be used to monitor potential water
quality changes? Water quality can be impacted by both geochemical reactions and hydrological changes.
The water quality impacts must be monitored to ensure that ASR operation is not endangering the
groundwater source.
• Would treating the water prior to injection decrease or eliminate the likelihood of adverse geochemical
interactions at an ASR site? For example, can the injected water be treated to prevent arsenic mobilization
from an aquifer formation where arsenic mobilization may otherwise occur?
iv
-------
Level 1 DSS for ASR Feasibility Analysis
The first level of DSS consists of databases and user inputs for the ASR-Need and feasibility analysis.
One database contains state ASR regulations promulgated under the Safe Drinking Water Act (SDWA) and web
links to the EPA Underground Injection Control (UIC) program's guidelines in permitting and developing ASR
applications. From the DSS database, users can review specific parameter limits of state water quality criteria.
Another database contains the location and purpose of application (e.g., aquifer recharge, potable reuse, etc.) of
existing ASR sites across the contiguous U.S. This database will help user assess whether the local geological
strata at a proposed site are suitable for an ASR operation. Hydrological investigations often aim to determine
formation properties and groundwater flow fields under ambient and ASR operation conditions. Spatial
distribution of faults and fault networks, along with geotechnical instability at or near the ASR site, are also
important considerations. It has been reported that fluid injection in the vicinity of pre-existing faults can trigger
seismic activity in the form of local earthquakes (Ellsworth, 2013). In the DSS Level 1 analysis, the site-
specific geological and hydrogeological data can be analyzed in the context of other ASR operations in the
database. This technical information is organized in a geographic information system (GIS) graphical user
interface (GUI), allowing the browsing of existing ASR example sites for reference in the feasibility analysis.
Location-specific ASR-Need analysis also consists of two other components - climate change and
water availability, and water demand projections. Factors this analysis includes are precipitation projections
using climate change modeling, and projected water demand for socioeconomic development and land use
change scenarios. Alternatively, local master planning documents in economic developments, land use zoning,
and population projection are the preferred basis for water demand projections. Model-generated projections
can be used when socioeconomic data is not available. Suitable models include the U.S. Census Bureau's
projection of population changes, EPA's Integrated Climate and Land-Use Scenarios (ICLUS) projections, and
the Cellular Automata (CA) - Markov land use model projections. The land use projection methods were
systematically reviewed in the National Water Infrastructure Adaptation Assessment, Part II: Characterize
Climatic Change andlmpacts for Water Adaptation Planning and Engineering (U.S. EPA, 2015a).
Projecting climate change for master planning purposes typically involves defining precipitation
intensity and variation over a period of at least 30 years. Data collected from an ensemble of global circulation
models (GCMs) and downscaled regional climate models (RCMs) are commonly used to project a future
climate condition. Due to large uncertainties in the projection for local watershed applications, climate
downscaling results are often used and the model is validated against long-range local precipitation data. The
U.S. EPA (2015a) report describes the datasets and precipitation projections of climate models for local water
resource engineering. The models determined appropriate were included in the ASR DSS.
Estimated water availability gaps and water shortage durations are the essential variables that define
ASR needs and scale requirements. This quantitative analysis, which involves of the determination of water
availability and water demand at a given location, is often a core component of water resource master planning.
Water resource master planning activities within a municipality are commonly based on water resource
inventory and water demand projections alone. The inability of available water resources to provide enough
water to meet current or future water demand is referred to as a water availability gap. Gap analysis in master
planning determines when and how much water should be stored and used to meet water demands. For many
municipalities, the total water management concept can be used in more comprehensive ASR-Need analysis.
This analysis includes quantifying the water demand, water availability gap, water reuse, and economics, along
with consideration of climate and land use changes, and their impacts on local hydrology and water usage.
Equations used in the water budget analysis are provided in Section 3.2.1 and a description of the analysis can
be found in Section 3.3.3. A practical example for the Las Vegas water district is described in Section 3.3.
v
-------
The Las Vegas example
To illustrate the total water management concept, the quantitative water budget gap in an ASR-Need
analysis for the Las Vegas metropolitan area were analyzed. The Las Vegas Valley in southern Nevada faces
chronic water shortages due to both drought conditions related to climate change and a growing water demand
from the increasing population (SWNA, 2009; USGS, 2000). The water level of Lake Mead, the only surface
water source for the region, has dropped over 130 feet from 2000 to 2014 in response to seasonal precipitation
changes, decreasing snowmelt-derived water flow, water diversions from the Colorado River, and chronic
droughts (SWNA, 2015). To reduce the water availability gap in water supplies, the Southern Nevada Water
Authority (SWNA) has operated a large ASR facility since 1985. Operation of this "water bank" relies on water
from Lake Mead. Water withdrawn from the Lake during high flow periods, such as in the spring when
snowmelt occurs, is stored in the aquifer for use during dry seasons.
The total balance between water supply and water demand from 2004 to 2050 was quantified using the
methodology adopted in this DSS. By comparing the time-evolution of major fixed fresh water supplies with
the total demand estimated using population projections, it was determined that the existing sources supplying
the Las Vegas Valley system - Lake Mead and the ASR facility - would provide adequate water supply through
the early 2020s. After 2024, the demand will exceed the total amount of water available, and the system will
need to find alternative sources of fresh water to meet the future demand. Planning how to address a potential
deficit after 2024 can be a challenge. The total water deficit is projected to increase up to nearly 2.46 x 10s m3
per year by 2050, accounting for a 39% increase in water demand from 9.55 x 10s m3 per year in 2024 to 1.33 x
109 m3 per year in 2050. Four options to augment the water supplies were evaluated: 1) Draw water from other
groundwater resources in Clark, Lincoln, and White Pine Counties, 2) Transfer groundwater through a massive
pipeline from the Great Basin aquifer system about 482.8 km (300 miles) north of Las Vegas, 3) Promote water
conservation to reduce the per capita water use by 0.753 m3 per day by 2035, and 4) As described in this
research, increase the volume of water stored through ASR by capturing the reclaimed water returning to Lake
Mead and storing more water in Lake Mead during peak flow conditions. Quantitative analysis indicates that
the ASR and water reclamation option may increase the long-term water supply resilience for the city.
Level 2 DSS for ASR vlannins and assessment
The second phase of planning and assessment (Level 2) involves the site characterization and the
determination of necessary water treatment and/or conditioning prior to injection or infiltration. For these
purposes, less-intensive computational tools are used for the analysis of groundwater flow, contaminant fate and
transport, and groundwater geochemistry. Contaminants of interest include both residual contaminants, such as
endocrine disrupting compounds (EDC) which may not be removed completely by conventional activated
sludge wastewater treatment, and contaminants mobilized through geochemical reactions during ASR
operation.
The DSS provides a step-by-step guide to site characterization. Geological data, hydrological data, and
site information are gathered to answer three major questions regarding site storage capacity, the fate and
transport of residual and mobilized contaminants in injected water, and the mobilization of contaminants from
aquifer formation. Mobilized contaminants can include arsenic, uranium, gross alpha, and gross beta, while
residual contaminants can include EDCs, nitrosamines and other disinfection by-products (DBPs). Nitrosamines
and DBPs are of particular interest because the carcinogenic compounds can form in disinfection of reclaimed
water before injection. It is difficult to remove nitrosamines once they are introduced into groundwater aquifers.
The site assessment can also be used to determine any treatment requirements for water prior to injection
operations. This analysis is focused on two major investigative components: the flow paths of injected water
and the geochemical compatibility of injected water with both native groundwater and the storage aquifer
formation. Based on analysis results, one can assess the treatment needs and determine the injected water
composition that has the least groundwater impact.
Modeling and assessment tools are described in this report. Groundwater flow modeling tools include:
1) Hantush (1967), a 2-D transient groundwater mounding model in an Excel spreadsheet, 2) SuperQ, a 2-D
vi
-------
transient well hydraulics superposition model in an Excel spreadsheet, 3) WhAEM2000 version 3.2.1, a 2-D
analytical and semi-numerical model, and 4) A variant Visual AEM released in February 2009. For water
quality variation, the quantitative models include: 1) 2-D AT123D-AT mode for fate and transport of residual
contaminants in injected water, and 2) Analytical multi-species sequential first-order reaction model over the
AT123D-AT codes.
These models yield approximate solutions, but are quick and appropriate in the planning and
assessment stage. This analysis relies on simplified model simulations for groundwater systems and ASR
operations. Therefore, Level 3 tools are recommended for actual ASR site engineering design and system
evaluation.
Level 3 Tools for ASR Design and Evaluation
Level 3 of the DSS is intended for engineering design and evaluation of ASR systems to achieve the
planning objectives in both recovery rate and water quality impact control. Technical analysis is necessary to
maximize accuracy and precision of the quantitative results for the following three system performance criteria:
• Residence time of injected water in storage. Many states, such as Washington, California, Florida, and
Hawaii, have established minimum residence times that injected water must spend in groundwater storage
to ensure biological safety. The criterion is often specified in an ASR operational permit. The minimum
residence time for biological safety is generally determined based on the inactivation rates of pathogens in
the aquifer. It is also noted that actual residence time varies among ASR projects. Through the DSS'
particle tracking functions, the water residence time can be calculated from simulation results. For aquifer
recharge operations with no water recovery, particle tracking provides valuable information for evaluating
the long-term recharge performance and location-specific groundwater management objectives.
• Flow field simulation and evaluation. The ambient groundwater flow field and its changes during ASR
operation are used to assess hydraulic control of the injected volume. Capture zone analysis is often
accomplished through modeling of groundwater head distributions and particle tracking. In the DSS, three-
dimensional groundwater simulation packages using MODFLOW are recommended.
• Contaminant fate and transport simulation. Water quality simulation can help characterize the residual and
mobilized contaminant distribution in the injected water and the aquifer to assess groundwater impacts.
Customized, site-specific field pilot testing and demonstration, a common engineering practice used for
groundwater remediation programs, can be used as the alternative. The transport simulation relies on site-
specific hydrological parameters and requires greater precision than the assessment used in the planning
phase (Level 2). In such engineering analysis, available tools include semi-analytical models such
asWhAEM2000 or VisualAEM, and numerical packages such as MODFLOW or PHAST.
Evaluations of the groundwater chemistry change and contaminant mobilization are assessed stepwise
in the DSS. First, the simulation package PHAST is used to assess water chemistry change in the injection
"bubble," including the perimeter mixing zones. This investigation may utilize the results from site
hydrogeology and groundwater chemistry investigations in the Level 2 analysis. Second, the quantitative data
on water chemistry changes are used along with hydrogeological information to evaluate the likelihood of
contaminant mobilization and to estimate contaminant concentration distributions under the existing or
proposed ASR design and operational options.
Arsenic is one notable contaminant that can be mobilized in groundwater during ASR. The arsenic
mobilization evaluation was developed from an extensive literature review and experimental studies on arsenic
mobilization. Experimental studies of ASR operation were carried out using reclaimed wastewater and
arsenopyrite, a common arsenic-bearing accessory minerals in aquifer materials. Appendix A contains details of
the review, experimental investigation, and the technical conclusions. In summary, arsenic mobilization
depends on a combination of site hydrology, injected water chemistry, aquifer petrologic composition, and the
resulting groundwater geochemistry at an ASR site. The influential factors are: 1) Water chemistry (e.g., pH,
Eh, ORP) differences between injected water and native groundwater, 2) Presence of natural organic matter
vii
-------
(NOM), S, microbiota, and nutrients in the injected water that may promote biological activities in the
subsurface, and 3) Cyclic operation resulting in groundwater fluctuation and oxygenation in the subsurface.
-------
Table of Contents
DISCAIMER i
ABSTRACT ii
ACKNOWLEDGEMENT iii
EXECUTIVE SUMMARY iv
ABBREVIATIONS / ACRONYMS xiv
1.0 INTRODUCTION 1
2.0 ASR PRACTICE AND APPLICATIONS 2
2.1 ASR Types and General Considerations 2
2.1.1 Storage and recovery operation 2
2.1.2 Water sources and end use 6
2.1.3 Water quality changes in ASR processes 7
2.2 ASR Practice in the U.S 10
2.2.1 ASR Applications in water stressed regions 10
2.2.2 Water quality standards in ASR operations 12
2.3 ASR Decision-Support Framework 13
2.3.1 ASR Decision-Support Framework 13
2.3.2 Technical models in hydrological and geochemical simulations 17
2.3.3 Groundwater chemistry changes and arsenic mobilization 20
3.0 ASR-NEED ANALYSIS IN PLANNING 21
3.1 ASR Assessment and Evaluation 21
3.1.1 Water quantity, quality and water availability 21
3.1.2 Technical feasibility analysis 21
3.2 Water Availability in ASR Feasibility Investigation 22
3.2.1 Water availability analysis 22
3.2.2 The climate change consideration 25
3.2.3 The socioeconomic factor 30
3.3 ASR-Need Analysis in Las Vegas-A Case Study 31
3.3.1 Geophysical settings 31
3.3.2 Future conditions for planning 33
3.3.3 Water budget analysis 36
3.3.4 The role of water conservation and storage in meeting future water demands 38
4.0 ASR FACILITY PLANNING AND ASSESSMENT 41
4.1 Assessment for ASR Planning 41
4.1.1 Infiltration rate and storage capacity 41
4.1.2 Simplified fate and transport analysis 43
4.1.3 Arsenic mobilization assessment 44
4.1.4 Data sources and assessment limitations 45
4.2 Assessment tools in the ASR DSS 46
4.2.1 Hantush (1967) 2-D Transient Mounding Excel Spreadsheet Model 46
4.2.2 SuperQ: 2-D Transient Well Hydraulics Superposition Excel Spreadsheet Model 47
4.2.3 2-D model: WhAEM2000 version 3.2.1 48
IX
-------
4.2.4 2-D model: Visual AEM released in February 2009 49
4.2.5 Fate and transport of residual contaminants in injected water 49
5.0 ASR EVALUATION AND ENGINEERING DESIGN 51
5.1 ASR Hydraulic Properties and Hydrologic Control 51
5.1.1 Particle tracking, capture zone and rate of ASR recovery 51
5.1.2 USGS MODFLOW Transient Numerical Groundwater Flow Model Code and MODPATH Transient Particle
Tracking Code 53
5.1.3 Particle Tracking Example 53
5.2 Fate and Transport of Residual Contaminants in Injected Water 56
5.2.1 MT3DMSfor multi-species transport in groundwater systems 57
5.2.2 SEAWATfor three-dimensional variable-density groundwater flow and transport 58
5.3 Geochemical Compatibility and Water Quality Changes 58
5.3.1 Arsenic mobilization from aquifer materials 58
5.3.2 Geochemical simulation using PHREEQC and PHREEQCI 59
5.3.3 3-D Modeling Tool -PHAST. 60
5.4 Simulation example: arsenic transport in long-term aquifer storage 60
5.4.1 Model Background in PHAST/MODELMuse simulation 60
6.0 CONCLUSION 63
7.0 REFERENCES 64
APPENIX A: LITERATURE REVIEW AND EXPERIMENTAL ANALYSIS OF ARSENIC RE-
MOBILIZATION AT ASR SITES
73
1 Arsenic Occurrence and Natural Sources 74
2 Arsenic remobilization and dissolution mechanisms 77
2.1 Oxidation of arsenic-bearing minerals 77
2.2 Reduction of arsenic-containing ferrihydrite 82
2.3 Impact of microbial activity 83
3 Experimental Investigations 85
4 Investigation Results 86
4.1 Arsenic dissolution rate 86
4.2 Secondary mineral precipitate morphology and mineralogy 88
4.3 Water matrix effects on arsenic remobilization 92
5 Manage Aquifer Conditions for Reduced Arsenic Remobilization 95
5.1 The hydrogeological factor 95
5.2 The chemistry factors and geochemical processes 97
5.3 Pretreatment and monitoring for enhanced reliability 99
6 References 100
x
-------
List of Tables
Table 1. Possible ASR operation types for water storage (and recovery) 5
Table 2. U.S. EPA wasterwater reuse water quality guidelines and the state regulations for five states 14
Table 3. Climate change scenarios for the Southwest U.S 34
Table 4. Land use projection for 2050 36
Table 5. Projected stream flow of Las Vegas Wash in 2050 with returned wastewater under a set of cimate and
land use scenarios 38
Table A-l. Location and conditions for recharge-influenced arsenic mobilization 75
Table A-2. Processes impacting aqueous arsenic mobility 78
Table A-3. Empirically derived rate laws for arsenopyrite oxidation by compounds of interest 79
Table A-4. Microbes impacting Fe/As oxidation and reduction in aquifers 84
Table A-5. Empirically determined activation energies for arsenic mobilization from arsenopyrite 88
XI
-------
List of Figures
Figure 1. Major types of ASR operations for water storage and reclamation 3
Figure 2. Potential scenarios of municipal wastewater reuse involving in ASR and other types of storage
operations 6
Figure 3. Process schematics of flow and water quality changes to consider in ASR design and operations 8
Figure 4. (A) The states with water reuse and ASR rules (Yang et al., 2007 and references therein), and (B) a
map of arsenic-contaminated wells and ASR/AR wells in the 10 EPA regions 11
Figure 5. Outlay of the ASR DDS framework, consisting of programs in three levels: 1) Feasibility analysis; 2)
Planning and assessment; and 3) Design and evaluation 15
Figure 6. Line-up of groundwater and vadose zone simulation programs for ASR decision support system 19
Figure 7. Conceptual schematic showing water distributions among major and secondary water process units
23
Figure 8. Spatial distributions of long-term precipitation changes and population change in the contiguous U.S.
26
Figure 9. Statistics of the rates of precipitation change (RI, %yr_1) in long-range historical monthly
precipitations measured at USHCN climate stations 28
Figure 10. A typical procedure in climate downscaling yielding the RCM dataset 29
Figure 11. Schematic of a typical integrated modeling approach in projecting surface water quality and quantity
changes in a watershed 30
Figure 12. Location of the Las Vegas Wash watershed, Nevada 32
Figure 13. Future population and wastewater projection of Las Vegas Wash watershed 35
Figure 14. Projected 2050 land use/land cover map of Las Vegas Wash watershed 36
Figure 15. HSPF Simulated continuous stream discharge with wastewater projections 37
Figure 16. The projections of total water demand and supply, showing the importance of return flow credit from
the Las Vegas Wash stream flow in the sustainable water supply for the region 40
Figure 17. Example of particle tracking from an injection well (upper) to a recovery well (lower) 52
Figure 18. Hypothetical example of a three-layer sandy aquifer in 3-D groundwater flow modeling in ASR
system design using the ASR DSS models 54
Figure 19. Particle tracking in profile across the injection well. The flow vector at each time step is shown at the
middle in each of the three layers 55
Figure 20. Particle tracking for three design scenarios of two paired injection-recovery wells 56
Figure 21. Computer-simulated well head at the recovery well at a distance of 800 feet in the pair well design
scenario 57
Figure 22. Model-predicted distribution of chloride concentrations (mg/L) 61
Figure 23. Model-predicted distribution of calcium concentrations (mg/L) 62
Figure 24. Model-projected spatial distribution of pH values in groundwater 63
Figure 25. Model projected distribution of arsenic concentrations ((.ig/L) in groundwater 63
Xll
-------
Figure A-l. Structures of arsenic-containing minerals. From Neil et al. (2012) 74
Figure A-2. Eh-pH diagrams from Salzsauler et al (2005). Activity of arsenate=10~3, Fe(II)=10"4 and
SO4=10"2.Adopted from Neil et al. (2012) 80
Figure A-3. Aqueous arsenic concentration in batch reactors at 5, 22, and 35°C 87
Figure A-4. AFM height mode images after 1 day (Al, Bl, CI) and 7 days (A2, B2, C2) in the 10 mM sodium
chloride, 10 mM sodium nitrate, and wastewater systems at room temperature (22°C) and under
aerobic conditions 89
Figure A-5. Tapping mode AFM Images of reacted FeAsS coupons in 10 mM sodium nitrate or 10 mM sodium
chloride 90
Figure A-6. Comparison between secondary mineral precipitation in the aerobic and anaerobic systems for 10
mM sodium nitrate and 10 mM sodium chloride 91
Figure A-7. Optical microscope images and Raman spectra for arsenopyrite coupons reacted in sodium nitrate
(A, B), sodium chloride (C, D), and wastewater (E, F) systems 92
Figure A-8. Evolutions of pH and ORP in batch reactors over the 7-day reaction period 94
Figure A-9. ASR Bubble formation during secondary water injection 96
-------
Abbreviations/Acronyms
AR
Aquifer Recharge
ASR
Aquifer Storage and Recovery
BLM
Bureau of Land Management
BOD5
5-day Biological Oxygen Demand
CA
Cellular Automata
CERCLA
Comprehensive Environmental Response, Compensation, and Liability Act of 1980
CFR
Code of Federal Register
CMIP5
Coupled Model Intercomparison Project Phase 5
DOC
Dissolved Organic Carbon
DOM
Dissolved Organic Matter
DSS
Decision Support System
EDC
Endocrine Disrupting Compounds
ENSO
El Nino Southern Oscillation
ET
Evapotranspiration
GCM
Global Circulation Models
GIS
Geographic Information System
GUI
Graphical User Interface
HSPF
Hydrological Simulation Program in Fortran
ICLUS
Integrated Climate and Land Use Scenarios
IPCC
Intergovernmental Panel on Climate Change
NCAR
National Center for Atmospheric Research
NLCD
National Land Cover Dataset
NOM
Natural Organic Matter
NPDES
National Pollution Discharge Elimination System
NRMRL
EPA National Risk Management Research Laboratory
ORP
Oxidation Reduction Potential
ORD
EPA Office of Research and Development
RCRA
Resource Conservation and Recovery Act
RCM
Regional Climate Model
SDWA
Safe Drinking Water Act
SNWA
Southern Nevada Water Authority
TN
Total Nitrogen
TOC
Total Organic Carbon
TP
Total Phosphorus
TSS
Total Suspended Solid
UIC
Underground Injection Control
USBR
U.S. Bureau of Reclamation
USCCSP
United States Climate Change Science Program
USDW
Underground Source of Drinking Water
USEPA
U.S. Environmental Protection Agency
USGCRP
U.S. Global Change Research Program
USGS
U.S. Geological Survey
USHCN
U.S. Historical Climate Network
xiv
-------
1.0 Introduction
Aquifer storage and recovery (ASR) is the technological means to enhance natural groundwater
recharge through man-made infiltration basins or injection wells for the purpose of recovering the water
at a later date. Without the recovering operation, the practice is often referred as aquifer recharge ( AR).
ASR practices are widely used in the U.S. to improve water availability during droughts and to counter
chronic water shortage. Climate change and continuous socioeconomic development are compounding
factors that make water availability a pressing challenge in many parts of the country (e.g., Barsugli et al.,
2009; Mearns et al., 2009, 2015). One approach to adapt to these challenges is the reuse and storage of
water to make up for water volume deficits. This technical method has two essential components: water
reuse to expand water availability, and water storage to overcome temporal water budget deficits and
imbalances. It is worthy to note, however, that the ASR operation can result in changes in water quality
and water chemistry in the injected water and the native groundwater aquifer. Thus, sustainable and
effective ASR practice requires management and reduction of negative environmental impacts through
proper planning, operation and monitoring.
Currently, ASR with wastewater is being implemented or is under consideration around the world
as a means to combat water deficits. In 2002, the total wastewater reuse reached 6.4 * 106 m3/d (U.S.
EPA, 2004), prompting the U.S. EPA (2001, 2004) to publish guidelines forthis practice. These
developments were summarized in Miller (2006), which articulated the need for an integrated wastewater
reuse program to systematically address technological, regulatory and public perception difficulties. As of
2015, ASR operations involving water and wastewater reuse are used in 27 states (Bloetscher. 2015).
Elsewhere in the Middle East, North Africa and Southern Europe, Angelakis et al. (1999) evaluated water
demand statuses and described the need for wastewater reuse as an unconventional resource. Qadir et al.
(2007) further analyzed water demand and supply imbalance in water scarce countries, mostly in the
Middle East, and related sustainable wastewater reuse to food security. Such analysis and conclusions
about the need for widespread wastewater reuse are echoed in a number of publications (Haruvy, 1998;
Angelakis et al., 1999; Thomas and Durham, 2003; Pasch and Macy, 2005; and Khan et al., 2006).
Water availability is also a particular concern as it relates to the impacts of global climate change.
The United States Climate Change Science Program (USCCSP) (2001) and the Intergovernmental Panel
on Climate Change (IPCC) (2013) provided strong evidence that future climates will be characterized by
increased precipitation extremes, leading to increased rainfall and flooding in some areas and prolonged
droughts in others. These changes will likely induce water availability stress in many parts of the U.S. and
the world. ASR can be implemented to mitigate stress related to this precipitation variability because it
allows for the storage of water during periods of increased rainfall for later use during periods of drought.
Adequate water storage capacity is critical to the sustainability of water resource development,
and is essential to minimizing the impacts of large changes in water availability and demand. In response,
the EPA Office of Research and Development (ORD) has conducted systematic research through the
National Risk Management Research Laboratory (NRMRL) to evaluate ASR technical feasibility,
regulatory implications, and engineering techniques for field applications. The study covers three areas:
1) Need for ASR to mitigate water budget imbalances, 2) Environmental impacts and potential regulatory
implications, and 3) Adaptation techniques and engineering guidelines for sustainable ASR development.
This research has led to the development of a decision support system (DSS) for ASR planning,
design, and evaluation. ASR has been widely used in the U.S. and other parts of the world. However,
1
-------
there are widespread reports of ASR-related water quality concerns and technical challenges in ASR
siting, planning, design, and operation, which makes the need for a DSS apparent. The desire for
sustainable ASR demands for a systematic investigation into ASR viability and engineering requirements.
This report describes the scientific basis of the ASR DSS. The DSS computer program, software
functionality, and application examples will be provided in a separate EPA report and software manual.
The report describes the three levels of analysis which will be used to develop ASR. Level 1 tools are
focused on ASR feasibility as it pertains to regulations and permitting needs, water demand and
availability, and climate change impacts. Level 2 of the DSS can be used to assist ASR planning and
assessment to characterize ASR sites and analyze pre-injection water treatment options. Level 3 analysis
consists of engineering design and evaluation using detailed hydrogeological and geochemical
characterization. This report consists of four main sections to cover these three levels:
1) Section 2.0 describes ASR practice and applications in the U.S., including state regulations and the
EPA UIC programs that manage ASR operations;
2) Section 3.0 discusses the factors controlling ASR practice and sustainability. A focus is placed on
water availability needs in both quantity and quality, climate change, and socioeconomic factors. This
section further introduces the ASR DSS framework in levels of feasibility analysis, planning and
assessment, and design and evaluation (Level 1).
3) Section 4.0 describes the theoretical basis and models for ASR planning and assessment (Level 2)
along with illustrative examples.
4) Section 5.0 describes the theoretical basis and models for ASR design and evaluation (Level 3) along
with illustrative examples.
2.0 ASR Practice and Applications
2.1 ASR Types and General Considerations
2.1.1 Storage and recovery operation
Water storage operations in typical hydrological engineering include reservoirs, above-ground
storage facilities, and underground ASR operations. Comparatively, ASR is often preferred for large-
scale, long-term and economic water storage and recovery. It has been long practiced in the U.S. and
other parts of the world such as the Middle East (See U.S. EPA, 2004; Weeks, 2002; Shelef and Azov,
1996). Groundwater injection wells, spreading basins, and infiltration galleries are major mechanisms for
water injection into suitable aquifers. Two major types of ASR operations exist, differing based on the
type of geological strata into which injection takes place (Figure 1):
• Injection into a confined aquifer. In this case, water from secondary sources, such as treated
wastewater or collected rainwater, is pre-treated and injected into a confined geological unit. The
water can then be recovered from the same well, or designated recovery well(s), and treated for a
specific end use. The water piezometric surface changes in accordance with pressure changes induced
during injection and withdrawal. Components of this ASR type are shown in Figure 1A.
• Injection into an unconfined aquifer. For many applications, water is injected into an unconfined
aquifer. Injection through a spreading basin, infiltration basin (or gallery), or well can result in
2
-------
mounding of the groundwater table under these conditions. These practices are considered AR rather
than ASR if there is no recovery component. Components of this ASR type are shown in Figure lb.
In both cases, injected water occupies an "injection bubble" within the formation. The injected
water fills formation voids in the core of an injection bubble and mixes with native groundwater on the
periphery. Regional groundwater flow outside of the injection bubble can change both direction and
velocity due to ASR injection effects on the hydraulic field.
CONFINED AQUIFER
0
Water Source
©
Pre-treatment
rim-1
0 0 0 0
Recharge Recovery Post treatment End Use
Piezometric level
Low permeability confining layer
Ambient
groundwater
/ Subsurface,
^storage ^ S
Confineil aquifer
UNCONFINED AQUIFER
O 0
Water Source Pre-treatment
JUL
l~lnrO
Infiltration basins
Water table
Ambient
groundwater
Recharge @ * _
^ oT "
% Subsurface storage
@00
Recovery Post treatment End Use
IP
Permeable soil
Unconfined aquifer
ModrfM from Dillon
-------
An ASR system consists of components for water injection into the aquifer and components for
water recovery. Groundwater injection may utilize several engineering means including injection wells,
infiltration basins, and infiltration galleries. The EPA's UIC program regulates AR and ASR
injection wells, but not infiltration basins or galleries (U.S. EPA, undated). Infiltration basins, also
known as spreading basins, are commonly used such as those by the Water Replenishment District of
Southern California.1 The volume of water stored is maximized by inducing infiltration over a large area.
In the case of an infiltration gallery, multiple installations of infiltration media are completed deep in the
vadose zone, commonly arranged as linear features in map view. Green infrastructure infiltration systems
(e.g., rain gardens, grass swales, and permeable pavements) function in much the same way. However,
these systems are not a focus of this report because they cannot be used to inject high enough volumes of
water to meet growing water demands, but for the purpose of groundwater replenishment
The groundwater recovery process is relatively straight forward. Recovery wells can be used for
both confined and unconfined aquifers. Intercept trenches are also commonly used for shallow unconfined
groundwater due to high water yield and economic efficiency. An intercept trench is physically similar to
an infiltration gallery, but is trenched into an unconfined aquifer as opposed to being completed in the
vadose zone. It is also important that the UIC and drinking water programs coordinate to discuss site
specific ASR projects, as there may be special testing and treatment requirements that must be addressed
prior to recovering the water for various beneficial uses.
Table 1 shows the potential injection and withdrawal pairs in ASR operations and their
objectives. For example, injection and withdrawal in a confined aquifer can be accomplished by a single
well in a sequential injection-storage-recovery operation (Figure 1A). Storage and recovery can also be
accomplished by using a pair of wells for injection and withdrawal, respectively. Unconfined aquifers
have seen more combinations in practice. Depending on combinations of injection-withdrawal system
pairs, ASR operations can lead to diverse types of groundwater flow fields and geochemical changes
within the injection bubble, where reactions with both aquifer formation minerals and native groundwater
may occur.
The difference in AR and ASR operation types is related to the end use purpose (See subsequent
Section 2.1.2). They differ in the volume injected, aquifer depth, and injection rate of an operation. For
example, most green infrastructure installations are intended to facilitate low-rate infiltration over a large
area. Green infrastructure uses vegetation, soils, and other practices to augment the natural processes by
promoting water infiltration downward into surficial aquifers, thus reducing overland runoff, enhancing
groundwater recharge, and creating healthier urban environments. Other examples of low rate systems
include decentralized wastewater treatment operations such as individual household septic tanks and drain
fields. By comparison, large water volumes are generated from municipal and industrial wastewater
treatment plants. Natural infiltration in green infrastructure installations are typically inefficient in
handling the large reclaimed water volumes generated by these operations. Therefore, direct injection into
aquifers is the more common solution.
In considering these variations, ASR and AR operations are generalized into four major
categories for analysis in this report:
a) Confined aquifer - Case (a). The injection-withdrawal operation is carried out either by a single well
or a pair of injection-withdrawal wells (Figure 1A). Injected water forms an injection bubble around
the well screen and influences local groundwater flow at the ASR site. The injection bubble core is
nearly pure injected water, while mixing occurs at the bubble perimeter. The water residence time
1 http://www.wrd.org/engineering/groundwater-replenishment-spreading-grounds.php
4
-------
Table 1. Possible ASR operation types for water storage (and recovery).
Injection Zone
Water Source
Vadose Zone
Aquifer
Combination*
Objective
Objective
Objective
Treatment effluent
Municipal wastewater plants
Storm water runoff
Mining water**
Industrial wastewater
Reuse
Reuse
Disposal/recharge
Disposal/recharge
Reuse
Reuse
Green infrastructure
Permeable pavements
Rain garden"
Green roof / Ciston
Runoff swales, etc.
Replenshment
Replenshment
Replenshment
Potential recharge
Replenshment
Replenshment
Decentralized treatment
Community septic tanks
Household dainfield
Replenshment
Replenshment
Reuse
Note: * - Operation involves vadose zone and underlying aquifer.
** - Large water volumes generated from mining operations such as hydraulic fracturing.
#- Rain gardens may or may not have gravel water trenches for deep infiltration.
(average time that injected water spends in storage) and capture zone (region of water extracted by
the withdrawal well) are two basic hydraulic planning parameters.
b) Unconfined aquifer without vadose zone treatment - Case (b). This operation is a variation of Case
(a). Wells are used for injection and withdrawal, while intercept trenches can be used when the
groundwater table is shallow. As for Case (a), the capture zone and residence time are two basic
hydraulic planning parameters.
c) Unconfined aquifer with vadose zone treatment - Case (c). As shown in Figure IB, infiltration water
from a spreading basin or infiltration gallery passes through the oxygen-rich vadose zone, where
several geochemical processes may occur including soil adsorption, ventilation (i.e., the drawing of
air into the space between soil particles), and biodegradation. Water infiltration rate is determined by
native soil in the vadose zone. The water is recovered through the use of a withdrawal well or an
intercept trench. The basic hydraulic parameters in this case include the infiltration rate through the
vadose zone, hydraulic control, and residence time.
d) Unconfined aquifer with no withdrawal operations (i.e., AR) - Case (d). Groundwater recharge comes
from infiltration processes (e.g., spreading basins or green infrastructure, such as a green garden or
gravel infiltration layer connected to a roof rain water collection system). Case (d) is hydraulically
similar to Figure IB, but with no groundwater withdrawal operations. The basic hydraulic parameters
in this case include infiltration rate and receiving aquifer properties (e.g., permeability, porosity, etc.).
5
-------
2.1.2 Water sources an d en d use
AR and ASR processes have been applied to store various water sources, including wastewater
treatment plant effluent for recycling and reuse, storm water runoff, household and commercial
greywater, and reclaimed industrial water such as treated cooling tower blowdown. A previous EPA
report (U.S.EPA, 2004) on water reuse describes the characteristics of the various water sources. Major
types of AR and ASR applications include:
• Groundwater storage and recovery for bene ficial use (ASR). Examples include mitigation of seasonal
or chronic water shortages, crop and landscape irrigation, and industrial usage. This end use variation
is shown schematically in Figure 2, and involves storing water in an aquifer for later recovery.
Understanding the different combinations of water sources and potential reuses can help to identify
important aspects related to ASR sustainability and regulatory programs.
• Groundwater recharge (AR). This category of end use includes replenishment of depleted aquifers,
augmentation of stream flows through natural aquifer-to-stream discharge, development of a
groundwater barrier against salt water intrusion, and long-term storage of excess water such as
drilling fluids in mining operations. However, because AR and ASR wells are considered Class V
injection wells, in order for this drilling fluid to be injected it must first be proven to not
endanger the underground source of drinking water (USDW). Thus, advanced water treatment
of injected water is common (Figure 2), in order to protect groundwater sources at an ASR site. Due
to this important regulation, pre-injection water treatment and water quality requirements are a focus
of this study, and will be discussed in later sections.
Industrial
reuse
Aquifer
-—^storage
Agricultural
-v/reuse
Urban
Storage
Community and small systems
for wastewater reuse
Stream
s^s^ugmentation
Storage,
treatment
Soil treatment
Percolation
Disinfection
Collection,
Grit removal
Biological
filtration
Package
plants
Figure 2 Potential scenarios of municipal wastewater reuse involving in ASR and other types of
storage operations. Note the last scenario for percolation is the form of uncontrolled aquifer
recharge.
Community wastewater
systems for reuse
Headworks
screening
an
Primary
treatment
J
Distribution,
conveyance
Secondary
~ Disinfection
treatment
Tertiary
treatment
Additional
treatment
Additional
treatment
Storage
6
-------
• Groundwater recharge by enhanced infiltration (AR). Unlike centralized injection in the above two
categories, engineered process units are used to increase infiltration for the purpose of replenishing
soil moisture and shallow groundwater. Examples include infiltration basins and infiltration galleries.
Figure 2 shows a variety of end use scenarios for wastewater, a common source of reclaimed water.
For many communities, wastewater and storm water are collected and conveyed to a central location for
treatment before discharge. Some of these centralized wastewater treatment systems are increasingly
engaged in water reuse, including water utilities in California, Texas, and Florida, as well as in other
water-stressed regions. Centralized systems in the U.S. are generally equipped with a secondary
wastewater treatment system (Figure 2), consisting of head works with screening of large debris, a
primary clarifier, activated sludge process and secondary clarifier in secondary treatment, followed by
disinfection before discharge under a National Pollution Discharge Elimination System (NPDES) permit.
These systems are generally capable of achieving treatment standards. EPA regulations on wastewater
treatment effluent commonly include limits on the 5-day biological oxygen demand (BOD5), N, P, and
biological plate counts. Typical standards are 45 mg/L BOD5, 45 mg/L total suspended solids (TSS), and
400 per 100ml fecal coliform in 7-day averages. Many states have stricter discharge standards particularly
on total phosphorus (TP) and total nitrogen (TN). Groundwater standards can also vary by region. For
example, the standards that are complied with in Region 8 are background concentrations of natural
groundwater above maximum contaminant levels, national drinking water standards and water standards
that EPA toxicologists have found to be appropriate for regional conditions.
A tertiary treatment process can be employed in the interest of water reuse safety (Figure 2). The
purpose varies; for example, further removal of nitrogen and phosphorus may be necessary to protect
sensitive environments in stream augmentation. The reclaimed water is then conveyed to the application
site for reuse. Some typical end use types are shown in Figure 2, including direct industrial reuse,
agriculture reuse, urban reuse (e.g., landscape irrigation), and stream augmentation. These operations do
not require long-term, large-volume storage. Large-volume storage involves the use of storage options
such as ASR. This report is only focused on operations involving long-term ASR, uncontrolled aquifer
recharge through infiltration, or Case (d) AR operations.
Small-scale wastewater treatment systems are often used for communities and individual
households (U.S.EPA, 2004). These systems normally have a limited wastewater treatment capacity; for
example, inadequate or no unit process for nitrification and denitrification. As a consequence,
micronutrient management in small-system wastewater effluent is often a concern. The challenges for
decentralized, small-system management have been well documented in states such as Maryland,
Virginia, and other states where septic tank systems are prevalent (Katz et al., 2011). Some additional N
and macronutrient removal from small system effluent occurs during infiltration in systems such as drain
fields. This type of end use, as in Case (d), recharges local aquifers.
2.1.3 Water quality changes in ASR processes
The quality of injected water and native groundwater is subject to change during ASR operation.
Figure 3 schematically shows a conceptual model of such changes during a paired injection and recovery
ASR operation. In the schematic, a constituent of concern has a concentration CWW_GW in the injected
water. Geochemical changes in the vadose zone are accounted for by ACST to become at the
groundwater table for ASR operations involving infiltration (Figure 3).
The injected water is mixed with native groundwater, which has a concentration of C°gw.
Resulting concentrations in the core and perimeter of the injection bubble are subject to compliance with
groundwater standards, {Cqw}- The injected water enters the aquifer formation at the original
concentration CWW_GW when direct injection through a well occurs. Subsequently, the injected water
7
-------
Effluent
Treatment &
%^ww-aw )
Conditioning
1
r *
r i
Infiltration
Spreading basin gallery
Ah
^ww-aw} I I I I
AC.
for}
T
u
l^'rnr aw I
fc"° 1 1cs,d 1
IrgwJ t-awt
> ~
Monitoring
well
Production
well ^
//
Ax (At)
Monitoring
well
In ection
B.
Concentrations
Variable Changes
Standards
Cww-Gw Injected water concentration
^ Change in water concentration due to
ST geochemical interactions in vadose zone
r ^ sf rfi Groundwater
^ standard
Native groundwater
gw concentration
Change in water concentration due to
mixing with native groundwater and
transport to recovery wells
Drinking
{C^} water
standard
Injected water concentration
CST after geochemical changes in
vadose zone
^ Residence time of injected water in the
groundwater
Permit
minimum
concentration
Injected water concentration
Cgw after mixing with native
groundwater and transport to
recovery wells
^ Distance travelled by injected water
during its residence time
Figure 3 (A) Process schematic of flow and water quality changes to consider in ASR design and operation.
The dotted line indicates the location of the water table. The end-use water quality requirements are
first determined to define the requirement for effluent composition after treatment/conditioning, the
soil treatment, ACST, and change in groundwater, ACGW. (B) Definitions of variables included in the
process schematic.
transports and resides in groundwater for a period of time (At). Geochemical changes that occur during
transport to recovery (production) wells are accounted for by ACcm/. The resulting groundwater
concentration, Cgw. should be in compliance with the groundwater standards, and the water quality in the
recovery well is required to meet drinking water standards, {C^}, or a minimal water quality threshold
in accordance with the site permit, {Cmin} (Figure 3). The major geochemical processes leading to water
quality change include:
1) Residual contaminants in reclaimed water for ASR operations. Some recalcitrant contaminants like
endocrine disrupting compounds (EDC) may not be removed during conventional activated sludge
8
-------
treatment processes. As a result, these contaminants can persist in reclaimed water. In many cases, the
contaminants can be removed by implementing appropriate treatment processes before injection. The
impact of these contaminants on groundwater depends on initial concentrations in the injected water
and geochemical processes in the vadose zone and/or in the aquifer formation.
2) Leaching from vadose zone. The literature (e.g., Pitt et al., 1999) indicates that contaminant transport
can be retarded in the vadose zone during infiltration, resulting in contaminant accumulation in the
root zone and/or top soil. When geochemical conditions change, these previously immobilized
contaminants, along with contaminants native to the soil, may become mobilized and transported into
the underlying groundwater.
3) Mobilization of multi-valent transition metals from native aquifer formation minerals. Introduction
of oxygen-rich injected water and additional organic matter may change the groundwater conditions;
for example, the reduction potential (Eh) of groundwater can increase, i.e. become more oxidative.
The change in Eh, and perhaps pH, of groundwater can lead to the oxidation of some native minerals
such as arsenic (As)-bearing pyrite and arsenopyrite. Oxidation of these minerals can lead to the
release of arsenic and other contaminants into the groundwater. The resulting contaminant
concentrations depend on the presence and abundance of native contaminant-bearing minerals,
geochemical condition changes, and reaction rates.
These general geochemical processes have been investigated using laboratory soil column
studies, field investigations, and computer modeling (Bouwer et al., 1981; Yates and Ouyang, 1992;
Westerhoff and Pinney, 2000; Lo et al., 2002; Sen et al., 2005; Scheytt et al., 2006). However,
knowledge gaps exist with respect to the geochemical interactions between injected water and native
groundwater during ASR operation. This limitation can lead to negative public perception and uncertainty
in regulatory oversight (Friedler, 2001; WHO, 2006; and Weber et al., 2006). For example, some studies
suggested that most contaminants are removed from injected water in the upper 1-1.5 m of soil (Westhoff
and Pinney, 2000; Greskowiak et al., 2005). However, toxic organic contaminants, organic matter,
emerging contaminants, inorganic compounds, and pathogens have been observed entering the
groundwater during laboratory and field studies (Bouwer et al., 1981; Manka and Rebhun, 1982; Lucho-
Constantino et al., 2005; Scheytt et al., 2006).
The inconsistent findings potentially result from varying soil and groundwater conditions, organic
carbon content, the contaminant matrix in wastewater effluent, and AR engineering and operations. These
variables and engineering factors can affect contaminant adsorption, biological and abiotic degradation,
transport in unsaturated and saturated soils, and geochemical conditions in the soil and groundwater,
which may promote or suppress re-dissolution of contaminants into the water phase. For this reason, ASR
system planning and evaluation may consider the following factors that can affect water quality changes:
¦ Fate and transport of emerging and recalcitrant contaminants in ASR operations. Scheytt et al.
(2006) described the mobility of pharmaceutical compounds in soil at wastewater reuse sites, and
Toze (2004) listed these contaminants along with endocrine disrupting compounds and pathogens as a
concern in reuse applications.
¦ Long-term changes in the receiving aquifers. Sheng (2005) showed that there had been no substantial
impact to the storage characteristics and groundwater quality at El Paso, Texas ASR sites after 20
years of operation. However, significant impacts have been observed at other locations. DeSimone et
al (1997) determined that alteration of geochemical properties had occurred due to ASR operation in
Cape Cod, MA, where an aquifer of glacial deposits was contaminated with partially treated effluent.
ASR impacts have also been observed in the Central groundwater basin in Los Angeles after 40 years
9
-------
of operation (Schroeder, 2002). Apparently the nature and extent of long-term changes in water
quality and aquifer properties can differ by location and operation.
¦ Hydrogeological impacts on contaminant attenuation. Aquifer properties and groundwater
composition can impact chemical and biological contaminant attenuation at an ASR site (e.g.,
Greskowiak et al., 2005). The impact is a function of aquifer hydrogeology and ASR operation
characteristics, which provides the basis for a thorough and systematic investigation.
The occurrence and magnitude of these negative impacts are functions of a number of variables.
Therefore, site-specific groundwater assessment and ASR evaluation are important. Quantitative analysis
and water quality modeling are useful and often necessary to address potential complications at a given
location. Such analyses normally take place in two steps. Based on site investigation results, hydrological
modeling is conducted, followed by contaminant fate and transport analysis. The ASR DSS framework
outlines these procedures in Section 2.3. This report also describes the hydrological assessment
component. Based on the hydrological and water quality analysis results, one can assess the minimum
injected water composition requirements for different soil and hydrological site conditions. This is
important for engineering control over virus, bacteria, and protozoa viability in soil profiles.
2.2 ASR Practice in the U.S.
2.2.1 ASR Applications in water stressed regions
ASR practice is increasingly used to manage water resources and mitigate water shortages in the
U.S. Long-term precipitation measurements across the contiguous U.S. show uneven precipitation
distribution and changes through the time. The details of these long-term changes and regional
distributions are provided in a recent EPA report (U.S. EPA, 2015a). Based on historical precipitation
measurements, Oregon and Washington received the greatest average monthly precipitation (up to 5.5
in/month), while much of the Great Plains region and California received the least (less than 1.5
in/month) (U.S. EPA, 2015a). Existing ASR projects are mostly located in the water-stressed states,
including Florida, California, and the Southwest (Figure 4). There are also a number of active ASR sites
in New Jersey and South Carolina (Bloetscher et al., 2014). By 2002, at least 27 states had water
reclamation facilities and associated water reuse guidelines, and nine states (Wisconsin, South Carolina,
Texas, Iowa, Missouri, Wyoming, Oregon, North Carolina, and Colorado) had enacted strict regulations
over ASR practices according to the UIC program. In early national assessment, U.S. EPA (2004)
reported that ASR facilities in the U.S. emplaced a total of 6 Ax 106 m3 of reused water into the subsurface,
with Florida and California accounting for 34% and 31% of this total, respectively. The ASR operation is
expected to expand for the increasing water availability gap. For example, in a recent focus report for the
Texas House of Representatives, the House Research Organization expected a large increase of water
storage in ASR practice nearly by three times from 2012 to estimated 152,000 acre-feet per year by
20 702. According to the U.S. EPA's 2012 Guidelines for Water Reuse, the U.S. population uses 121
million m3 of water per day. Currently only 7-8% of this water is reused, while the NRC Water Science &
Technology board reports that one-third of the total water being used could potentially be reused (U.S.
EPA, 2012a).
ASR operation can be adjusted in response to water demand and demographic changes. The U.S.
population has increased since 1900 and the rate of increase has accelerated since the 1970s. This rapid
2 House Research Organization report "Addressing Water Needs Using Aquifer Storage and Recovery". July 12,
2016. http://www.hro.house.state.tx.us/pdf/focus/asr.pdf
10
-------
Legend
State with water
reuse regulations
State with water
reuse guidelines
States with water reuse and
groundwater recharge rules
(as of2007)
B.
O
A'i °f
' Q ' M
Elevated Arsenic
Concentration
» >10-25 ppb
• > 25 - 50 ppb ASR/AR
O '
> > 50 - 75 ppb Well Status
>75-100 ppb 6 Active
1 > 100 ppb 6 Inactive
Figure 4 (A) The states with water reuse and ASR rules (Yang et al., 2007 and references therein), and (B) a map of arsenic-contaminated wells
and ASR/AR wells in the 10 EPA regions of the contiguous United States (created with Ersi® ArcMap™ 10.3)
11
-------
population growth has increased water demand, necessitating the implementation of techniques such as
ASR to supplement natural water supplies. It is worthwhile to note that the largest increases in population
have occurred in 5 water-poor states: Nebraska, Arizona, Utah, Nevada, and Florida with population
growth particularly concentrated in the southern region (See Figure 8B later in Section 3.2.2). It is
important to consider these population growth projections in the larger context of industrial and
agricultural water demands, as well as precipitation variability. For example, in the case of Nebraska, the
population growth rate is expected to level off by 2030 (Y ang et al., 2007). However, like other Great
Plains states, Nebraska has high fresh water usage per capita, largely due to agriculture irrigation. Yet
more than 50% of Nebraska's fresh water is derived from groundwater aquifers that are vulnerable to
overdrafting. The extensive use of groundwater resources can lead to overdrafting where groundwater
pumping exceeds natural recharge rates. Similar observations on the imbalance can be found in many
other areas of the Great Plains region (U.S. EPA, 2015a).
Therefore, unless steps are taken to mitigate stresses to the water supply, population/agriculture-
induced water shortages are likely to occur in the future. ASR using wastewater can be a means of
addressing these shortages by supplementing non-potable agricultural irrigation. The water budget can be
assessed using numerical modeling techniques, such as the neural network model described in Chen et al.
(2003). Many ASR applications in the U.S. are intended for non-potable water and wastewater reuse in
irrigation, industrial, and urban landscape applications (Figure 2) (Asano and Levine, 1996). However,
ASR application for indirect potable water reuse is increasing (Miller, 2006). Reported field-scale indirect
potable examples include the spreading basins in the Southern California Groundwater Replenishment
program, wastewater reuse for aquifer recharge in Las Vegas, Nevada (Ranatunga et al., 2014), and ASR
applications in master planning for water supplies in Manatee County, Florida (Chang et al., 2012).
2.2.2 Water quality standards in ASR operations
There are 1185 aquifer recharge and ASR wells documented in the U.S., which were counted in a
state and U.S. EPA Regional survey as part of a Class V Underground Injection Control Study. The total
number is estimated to be between 1695 and 2000, since not every ASR well is documented due to
variations in the permitting and reporting requirements for Class V wells on a state-by-state basis (U.S.
EPA., undated). As of 1999, approximately 89% of documented wells were located in ten states: FL (488,
including storm drainage and connector wells), ID(48), NV(110), TX(67), SC(55), ID(48), OK(44),
OR(16), WA(12), and CO(9). For these and future ASR sites, regulatory oversight at the U.S. EPA is
administrated by the Underground Injection Control (UIC) program3. ASR wells are regulated as Class V
injection wells in the regulatory framework. The EPA may directly implement a program, or a state may
have primary enforcement authority, or "primacy". ASR system owners and operators need to submit
basic inventory information to the primacy enforcement agency prior to constructing the well. In many
states, the state's regulating department, for primacy states, or the EPA, for states without primacy, will
require a permit for a Class V well. In some states, the operation can be authorized by rule if the owner or
operator submits the inventory information for a fresh water source and demonstrates that the well will
operate in a manner that does not endanger an underground source of drinking water (USDW).
Minimum water quality standards have been promulgated under the Safe Drinking Water Act
(SDWA) to protect groundwater and surface water sources in the United States. Table 2 lists water quality
criteria in ASR, urban reuse, and agricultural reuse for food crops. Florida regulates fecal, TSS and nitrate
concentrations for ASR operations (Table 2). As of 2015, states that require individual ASR permits
include Florida, Idaho, Nevada, Oregon, South Carolina, New Jersey, and Washington. Five states (CA,
3 http://water.epa.gOv/tvpe/groundwater/uic/aauiferrecharge.cfm#uicregulations
12
-------
CO, ID, OK and TX) require permits that satisfy groundwater rules. Other states do not have state-
specific water quality criteria, but generally follow the federal drinking water standards.
Regulations adopted by primacy states for ASR wells vary. As of 2007, nine states (Wisconsin,
South Carolina, Texas, Iowa, Missouri, Wyoming, Oregon, North Carolina, and Colorado) require that
water used for ASR injection be potable water or drinking water treated to national or state Drinking
Water Standards or state ground water standards. Potable water is defined differently in each state but
generally refers to water that is high quality and poses no immediate or long term health risk when
consumed. Some states require that the injected water not exceed the background levels of natural
groundwater, while other primacy states allow additional types of water to be used in ASR, including
treated effluent, untreated surface and ground water, and reclaimed water subject to state recycled water
criteria. Nevada also allows "any" injected water. However, these state-specific ASR regulations do not
supersede the prohibition of movement of contaminated fluid into a U.S. drinking water source.
Specifically, EPA regulations provide that "no owner or operator shall construct, operate, maintain,
convert, plug, abandon, or conduct any other injection activity in a manner that allows the movement of
fluid containing any contaminant into underground sources of drinking water, if the presence of the
contaminant may cause a violation of any primary drinking water regulation under 40 CFR part 142 or
may otherwise adversely affect the health of persons." (40 CFR 144.12).
2.3 ASR Decision-Support Framework
ASR development at a municipality often involves several consecutive planning and engineering
actions. These include ASR-Need analysis, feasibility studies, planning and assessment, and ASR design,
construction, and operation. A technical evaluation may also involve many of these elements for existing
ASR facilities. In this research, an ASR DSS framework was developed to support ASR planning, design,
and evaluation. The DSS can be applied to assist technical investigation and management analysis in three
levels: 1) ASR feasibility analysis; 2) ASR planning and assessment; and 3) ASR design and evaluation.
2.3.1 ASR Decision-Support Framework
These levels of technical and engineering investigations are interrelated and can provide a
structured analysis for ASR development, permitting, and evaluation. The DSS principles, tools and
methods are presented in the remainder of this report along with illustration examples.
Level 1: Feasibility analysis
Level 1 of the DSS framework consists of databases and user inputs. One database contains state
ASR regulations promulgated under the SDWA and weblinks to the EPA UIC program. Users can review
specific parameter limits such as those in Table 2. Another database contains the location and purpose of
application (e.g., aquifer recharge, portable reuse, etc.) for existing ASR sites across the contiguous U.S.
which will be kept current. Technical information on ASR example sites is provided in a GIS GUI for
planning and design reference purposes.
Additional components of feasibility analysis are location-specific ASR site information and
ASR-Need analysis (Figure 5). Factors considered in the ASR-Need analysis include projections of
climate change, water demand for projected socioeconomic development, and land use changes. Local
master planning documents for economic developments, land-use zoning, and population projection, are
preferably used as the basis for the water demand projections. Model-generated projections, such as those
from the U.S. Bureau of Census projections on population or the Cellular Automata (CA) - Markov land
use projections, can be used in the absence of local master planning data. Land use projection methods
were systematically reviewed by the U.S. EPA (2015a), which recommended the CA-Markov modeling
13
-------
End-use
US EPA (2004)*
Arizona
California
Florida
Nevada
Texas
Urban Reuse
Turbidity
Fecal coliform
Biochemical oxygen demand (BOD5)
Total suspended solids (TSS)
2NTU
ND/lOOml
lOmg/L
<2mg/L
2 (5) NTU (Class A)
ND 4/7-D (Class A)
23/100ml, Max (Class A)
200 (800V100 ml ("Class B~)
2 (5) NTU (1 -D avg, max)
2.2/100 ml (7-D mean)
23/100 ml (30-D max)
240/100 ml (max)
ND at 75% samples
25/100 ml (30-D, max)
20 mg/L (yr avg)
5ng/L
NS
2.2/100 ml (30-D avg)
23/100 ml f30-D max")
30 mg/L
Not specified
2 NTU (1-D avg)
5 NTU (max)
14/100 ml
200/100 ml (7-D aw)
800/100 ml (max)
5 mg/L (10 mg/L)
30(20) me/L. Tvnell
Not specified
PH
6-9
00
6-9
Free chlorine
l.Omg/L
1.0ng/L
1.0 mg/L
Agriculture, flood crops
Turbidity
2 (5) NTU (Class A)
0.5-2 NTU (1-D max)
Not specified
3 NTU (1-D avg)
Fecal coliform
<200/100 ml
ND 4/7-D (Class A)
200 (800yi00 ml (Class B)
2.2/100 ml (7-D mean)
23/100 ml (30-day max)
25/100 ml (30-D, 75%)
200/100 ml (avg)
400/100 ml (30-day max)
20/100 (75/100) ml
200/100 ml (7-D avg)
800/100 ml (max)
BOD5
TSS
30 mg/L
30 mg/L
Not specified
Not specified
Not specified
20 mg/L (yr avg)
5ng/L
30 mg/L
Not specified
5 mg/L (10 mg/L)
30 mg/L (Type II)
3 NTU
pH
6-9
00
Free chlorine
1 mg/L
1.0 mg/L
Aquifer
recharge
Turbidity
Site specific and use dependant
Not regulated
Case-by-case
Not specified
Case-by-case
Case-by-case
Fecal coliform
Site specific and use dependant
Not regulated
Case-by-case
200/100 (yr avg)
800/100 (max)
Case-by-case
Case-by-case
Carbonaceous BOD5
Site specific and use dependant
Not regulated
Case-by-case
Not specified
Case-by-case
Case-by-case
TSS
Site specific and use dependant
Not regulated
Case-by-case
10 mg/L
Case-by-case
Case-by-case
pH
Site specific and use dependant
Not regulated
Case-by-case
Case-by-case
Case-by-case
Free chlorine
Site specific and use dependant
Not regulated
Case-by-case
Case-by-case
Case-by-case
Nitrate
Site specific and use dependant
Not regulated
Case-by-case
12 mg/L
Case-by-case
Case-by-case
Note: Data source - US EPA (2004)
* Suggested guidelines for various types of wastewater reuse.
14
-------
Feasibility
analysis
ASR regulations,
permitting needs
Climate change &
water availability
Projected water
demand
ASR sites and
information
Planning and
assessment
Site
characterization
DSS analytical modules
ori groundwater flow
Geochemical
assessment
DSS water quality modules on
groundwater compatibility
Design and
evaluation
Full-scale
groundwater
modeling
¦ Flow field prediction
¦ Residence age (particle tracking)
¦ Fate & transport - Rec. water
¦ Remobilization - As
^ DSS numerical modules on
flow and geochemical analysis
I) As remobilization module, etc.
ASR permitting and
operation
~ Permitting (water
treatment, monitoring)
~ ASR system design
{injectant conditioning,
ASR design, monitoring
well system, etc.)
Figure 5 Layout of the ASR DDS framework, consisting of programs in three levels: 1) Feasibility analysis; 2) Planning and
assessment; and 3) Design and evaluation.
15
-------
method to predict dynamic urban land changes. Climate change analyses typically projects precipitation
changes over a period of -30 years. As described by the U.S. EPA (2015a), future climate considerations
are commonly projected based on climate data, an ensemble of global circulation models (GCMs), or the
RCM downscaling data. These models are validated against historical precipitation data. However,
climate down-scaling results are often used for projection of local watershed applications due to the large
uncertainties that are inherent in the science. The EPA (2015a) describes the datasets and precipitation
projections of climate models for local water resource engineering.
Feasibility analyses are used to determine the magnitude of water availability gaps and durations
of water shortages, and are used to define ASR need and scale requirements. The quantitative analysis is a
core component of water resource master planning, and involves detailed assessment of water availability
and demand. A practical example for the Las Vegas water district is given in Section 3.2.4.
The feasibility analysis portion of the DSS is used to answer key questions about geological strata
suitability for ASR operation. Hydrological investigations are conducted to obtain aquifer transmissivity
or formation permeability, porosity, and geochemical characteristics, and to determine groundwater flow
fields under ambient and ASR operation conditions. Spatial distribution of faults and fraction networks,
along with geotechnical instability at or near the ASR site, are also important considerations. It has been
reported that fluid injection in the vicinity of pre-existing faults can trigger seismic activity in the form of
local earthquakes. In the DSS Level 1 analysis, the site-specific geological and hydrogeological data can
be analyzed in the context of other ASR operations in the database. This may lead to a preliminary
assessment of the ASR feasibility at a given location.
Level 2: ASR planning, and assessment
Site characterization and analysis of water treatment prior to injection are two major components
in engineering investigation at the planning and assessment level (Figure 5). Hydrogeological
characterization is necessary to obtain, at a minimum, the technical parameters necessary to answer the
following questions:
1) What is the likely recovery rate of injected water? This planning question is pertinent for temporary
storage operations that are intended to address temporal or chronic water shortages. Poor recovery
rates can also negatively affect the economics of an ASR project.
2) What hydrological changes occur during ASR operation? Vertical hydraulic conductivity and soil-
clogging in the vadose zone are important considerations for spreading basin and soil infiltration
operations. Aquifer permeability and near-well clogging from biological growth and inorganic
precipitation are parameters assessed in the planning phase for wells that inject into saturated zones.
3) Is there potential for adverse geochemical reactions between the injected water and native
groundwater and/or geological formations? During ASR, the injected water forms a "bubble" by
displacing the native water closest to the point of introduction and mixing with native water for some
distance away from this point. The point at which only native groundwater is present in pore space
defines the edge of the injection bubble. This spatial distribution of injected water defines the nature
and extent of geochemical reactions and contaminant mobility during injection and withdrawal
operations. These geochemical reactions can be simulated for anticipated geochemical conditions
(e.g., Eh, pH, cation and anion species, etc.).
In the DSS's step-by-step site characterization, geological data, hydrological data, and site
information are gathered to answer the three major questions listed above. The assessment can also be
used to define injected water treatment requirements prior to recharge operations. In this analysis,
geochemical compatibility of the injected water with aquifer materials and the flow paths of injected
16
-------
water are two major investigative components. Based on treatment scenario analysis, one can determine
the treatment options and the injected water composition that has the least impact on groundwater.
Level 3: ASR design and evaluation
Level 3 of the DSS framework is designed to aid in engineering design and evaluation of ASR
systems to achieve the planned objectives in recovery rate and water quality impact control (Figure 5).
Engineering analysis and groundwater modeling are necessary to determine the following three system
performance variables:
• Residence time of injected water in storage. Many states, such as Washington, California, Florida,
and Hawaii, have established minimum residence times to ensure biological safety. The required
residence time is specified for water flow from injection to recovery, which varies among ASR
projects. Water residence times are calculated by applying particle tracking functions in model
simulations. For AR operations with no water recovery, such as groundwater replenishment, particle
tracking provides valuable information for evaluating long-term AR performance and location-
specific management objectives.
• Flow field simulation and evaluation. The ambient groundwater flow field and ASR operation criteria
are used to assess hydraulic control of the injected water volume. Capture zone analysis is often
accomplished through modeling of groundwater head distributions and particle tracking. In the DSS,
3-D groundwater simulation packages using MODFLOW are recommended.
• Contaminant fate and transport simulation. Characterization of the distribution of residual
contaminants in injected water and mobilized contaminants in the aquifer is used to assess
groundwater impacts. The alternative is customized, site-specific field pilot testing and
demonstration, a common engineering practice used during CERCLA or RCRA programs. For ASR
design and evaluation, the water quality simulation requires site-specific hydrological parameters and
a higher degree of precision than during the planning and assessment phase (Level 2).
Field pilot tests are an important aspect of ASR planning and will help to determine the hydraulic
parameters used in modeling. However, this DSS does not deal directly with field pilot testing and cannot
recommend one procedure. Instead, each region implementing ASR must develop their own region-
specific procedure. Nevertheless, the results in design and evaluation can lead to a technical basis for the
design of monitoring programs and permitting requirements to meet the ASR objectives and performance
criteria. Therefore, site investigations must produce detailed, accurate, and site-specific hydrogeological
and engineering parameters.
2.3.2 Technical models in hydrological and geochemical simulations
The DSS models for the Level 2 and 3 analyses are listed in Figure 6, with technical descriptions
provided in Appendix A. These hydrological and geochemical simulators can be applied for ASR
planning, design, and evaluation purposes. The software is organized and executed through a custom-
designed GUI interface for model interoperability.
ASR planning models
The models in the first row of Figure 6 are the least data-intensive, but are appropriate for ASR
planning purposes. These models provide fairly accurate two-dimensional analyses of groundwater flow
under ambient or ASR operation conditions. Hantush (1967) and SuperQ are Excel spreadsheet programs
that are simple to use while preserving essential groundwater hydraulics. SuperQ utilizes the Theis and
Hantush-Jacob equations to determine groundwater flow fields, and uses superposition principles to
17
-------
calculate groundwater flow effects from multiple wells with different pumping rates. It can be used for
estimating and assessing the groundwater hydraulic response from ASR injection and recovery wells.
WhAEM2000 (U.S. EPA, 2007) and a variation, Visual AEM (Craig and Matott, 2009), are better
suited for simulating 2-D, steady-state groundwater flow under complex aquifer conditions such as
multilayer aquifers, irregular recharge boundaries, and aquifer heterogeneity. The program can be
augmented to perform 3-D flow simulations (Kraemer, personal communication). Visual AEM is capable
of analytically and numerically simulating contaminant transport in vertically averaged water flow. This
capability is valuable for assessing ASR operations when residual contaminants are of concerns.
The 1-D multi-species semi-analytical model and the semi-analytical AT123D-AT (Burnell et al.,
2012) model in the ASR DSS can be used to assess contaminant fate and transport for ASR planning
purposes. These two programs include simplifying assumptions with respect to groundwater solute
transport. For example, the 1-D multi-species transport model yields steady-state longitudinal dispersion
of parent and daughter contaminants from a continuous point source. The semi-analytical AT123D-AT,
on the other hand, is more computationally complex because of its improved numerical solver and the
new use of Green's function for a finite-depth aquifer, and thus is more versatile. However, by
simplifying the hydrological conditions through model assumptions, outcomes from these models are
limited in their usefulness. For example, while these steady-state solutions can be used for planning
purposes, such as to predict the potential for arsenic contamination, steady-state solutions must be
combined with more complex 3-D groundwater flow models to be used for engineering design.
ASR design and evaluation models
As compared to planning analysis, system design and evaluation analyses require models with a
greater fidelity. Important engineering parameters include injection rate and hydraulic control, as well as
recovery rate and residence time when recovery is involved. These design considerations are described in
Section 4.1.
The DSS utilizes 3-D groundwater flow and geochemical models to analyze the groundwater
system and the 1-D AT123D-AT model for vadose zone analysis. These models are chosen due to the
physical condition of water injection. As shown in Figure 1, ASR water injection can cause groundwater
mounding at the injection location. Water injection or spreading through the vadose zone further
complicates the model simulations because, while vertical water infiltration and contaminant transport
can be simulated using 1-D models, groundwater mounding is superimposed on a regional flow field. As
a result, the 3-D variation in flow and solute transport is difficult to quantify using 1-D or 2-D
groundwater models.
MODFLOW is a widely used and verified modeling package for groundwater simulation. The
modular structure of MODFLOW has provided a robust framework for the integration of additional
simulation capabilities since it was originally introduced in 1984. The current MODFLOW-related family
of programs are capable of simulating 3-D steady or transient flow fields in confined and/or unconfined
aquifers. In addition, the MODPATH module allows for tracking of particle pathways using a semi-
analytical particle-tracking scheme. The results are useful to evaluate the residence time of injected water
and to calculate the rate of recovery. MODFLOW and MODPATH technical details are provided in
Appendix A.
MT3DMS is a multi-species transport model. The numerical model, originally developed by the
U.S. Army of Engineers (Zheng, 2010), considers advection, dispersion, matrix diffusion, and chemical
reactions. Simulation results can provide contaminant concentration distribution at an ASR site, along
with time-series pressure and chemical concentrations at observation wells for a variety of hydrologic
18
-------
Hantush (1967):
2-D Water Table
Mounding Model
The USGS (Carleton,
2010) developed an
Excel spreadsheet
to calculate
transient water
table mounding
beneath a
rectangular
recharge basin
based on the
Hantush (1967)
equation. An Excel-
based spreadsheet
program.
MODFLOW:
3-D Groundwater
Flow Model
The USGS
MODFLOW
(Harbaugh, 2005)
finite-difference
code (with its family
of compatible
programs) is used
to simulate 3-D
groundwater flow,
groundwater/
surface-water
interactions, etc.
ModelMuse is a
Windows GUI for
MODFLOW and
other model codes.
SuperQ:
2-D Well Hydraulics
Superposition
Model
Uses the Theis and
Hantush-Jacob
equations and
superposition to
evaluate transient
effects of multiple
wells, variable rate
pumping, and
simple boundary
conditions on
hydraulic heads in a
uniform 2-D
aquifer. SuperQ is
an Excel-based
spreadsheet
program.
MODPATH:
3-D Particle-
Tracking Pathline
Model
MODPATH is a
USGS (Pollock,
2012) post-
processing program
that computes 3-D
flow paths and
travel times of
groundwater
particles and
retarded solutes
using MODFLOW
output. ModelMuse
is a GUI for
MODPATH.
WhAEM2000:
2-D Groundwater
Flow and Particle
Tracking Model
USEPA (2007)
public domain
Analytical Element
Model (AEM) code
simulates 2-D
steady flow caused
by pumping wells,
hydrologic
boundaries (river,
recharge, and no-
flow conditions),
and inhomogeneity
zones. Easy-to-use
Graphical User
Interface (GUI).
MT3DMS:
3-D Particle-
Tracking Pathline
Model
Links with
MODFLOW to
simulate variable
velocity from
injection/pumping,
hydrodynamic
dispersion, linear
sorption, and first
order sequential
reactions of
multiple pollutants
in groundwater. No
free Windows GUI
is available.
Visual AEM:
Groundwater Flow,
Particle Tracking, &
Transport Model
Visual AEM (Craig
and Mattott, 2009)
is a GUI for single
and multi-layer
AEM modeling of
steady-state
groundwater flow
particle-tracking,
and numerical/
analytical modeling
of vertically-
averaged
contaminant
transport.
Incorporates an
easy-to-use GUI.
SEAWAT:
3-D Particle-
Tracking Pathline
Model
The USGS SEAWAT
code (Langevin et
al., 2007) is a
generic MODFLOW/
MT3DMS-based
computer program
designed to
simulate 3-D
variable-density
groundwater flow
coupled with multi-
species solute and
heat transport. No
free Windows GUI
is available.
AT123D-AT:
1-D Multi-Species
Semi-Analytical
Semi-Analytical
Transport Model
Transport Model
• Simulates
• Simulates
advection,
advection,
hydrodynamic
hydrodynamic
dispersion, linear
dispersion, linear
sorption, and first
sorption, and
order reaction of 1-
sequential
D, 2-D, or 3-D
degradation
dissolved pollutants
reactions of
in groundwater
multiple pollutants
from an
along uniform or
Instantaneous,
variable velocity
continuous, pulse.
flow path for use in
or time-dependent
risk assessment.
source mass flux.
Incorporates an
Incorporates an
easy-to-use GUI.
easy-to-use GUI.
PHAST:
PHREEQC:
3-D Reactive
Geochemical Model
Geochemical
Transport Model
• The USGS PHREEQC
• Using PHREEQC to
(Parkhurstand
handle equilibrium
Appelo, 2013)
and kinetic
geochemical
geochemical
program performs
reactions, the USGS
(1) speciation and
PHAST code
saturation-index
(Parkhurst et al..
calculations, (2)
2010) can simulate
batch reaction and
multicomponent,
1-D transport
reactive solute
calculations with
transport in 3-D
reactions, and (3)
groundwater flow
inverse modeling.
systems.
PHREEQC-I is a
ModelMuse is a
Windows GUI for
Windows GUI for
PHREEQC.
PHAST.
Figure 6 Groundwater and vadose zone simulation programs involved in the ASR decision support system.
19
-------
conditions and well configurations. MT3DMS is used to simulate the fate and transport of residual
miscible contaminants, mostly organic trace constituents, in injected water.
SEAWAT is a coupled version of MODFLOW and MT3DMS designed to simulate 3-D,
variable-density, saturated groundwater flow. Although MT3DMS and SEAWAT are not explicitly
designed to simulate heat transport, temperature can be simulated a chemical species by using a
mathematical analog to Fickian diffusion with appropriate transport coefficients. Thus, the SEAWAT
module allows simulation of various "unusual" ASR applications, including the injection of oil field brine
water after treatment, or water injection into a saline aquifer.
Design and evaluation models for aquifer recharge
On the other end, the AR operations do not include recovery components. As these represent
simplified cases, the models listed above are capable of providing accurate simulation results for design
and evaluation. This approach is applicable for groundwater recharge through green infrastructure
applications, water percolation from septic tank to drain fields, and water injection using wells and
trenches (Figure 2).
2.3.3 Groundwater chemistry changes and arsenic mobilization
An ASR operation introduces water into the aquifer that is geochemically different from native
groundwater, forming an injected water bubble of different water composition. This condition is
schematically illustrated in Figure 1. Displacement of native groundwater at the core and mixing at the
peripheries of the bubble are the processes that lead to potentially detrimental geochemical reactions. One
example is arsenic mobilization from formation minerals into groundwater (Jones and Pichler, 2007; Neil
et al., 2012; 2014; Wallis et al., 2010). ASR-induced mobilization has also been reported for other
metallic constituents (e.g., Pb, Cr, U, Fe, etc.) (Arthur et al., 2002).
PHAST is a 3-D groundwater flow and geochemical package that is used to characterize transient
changes of geochemical parameters (e.g., pH, Eh, etc.) (See Figure 6). PHAST combines the geochemical
program PHREEQC and the 3-D groundwater flow program, HST3D, both from the USGS. PHREEQC is
designed to simulate cation and anion speciation and saturation index, transport of species with reversible
and irreversible reactions, as well as inverse modeling from field measurements. This allows PHAST
application for a diverse set of geochemical conditions in fate and transport analysis.
Arsenic-bearing minerals, such as arsenopyrite (FeAsS), can be present in aquifer formations
under anoxic conditions. However, when conditions become oxic due to the introduction of injected
water, arsenic becomes soluble in groundwater (Jones and Pichler, 2007; Wallis et al., 2010; Neil et al.,
2012). Principle reactions governing arsenic mobility involve arsenopyrite oxidation (an arsenic source)
and iron (hydr)oxide mineral formation (an arsenic sink) (Neil et al., 2014):
FeAsS + 1.5H20 + 2.7502(aq) ^ Fe2+ + H3As03 + S042
Fe2+ + 3H20 Fe(OH)3 + 3H+ + e
PHREEQC is capable of simulating arsenic mobilization in groundwater under various
geochemical conditions (e.g., Appelo et al, 2002; Zheng et al., 2009; Tabelin et al., 2012; Keeting et al.,
2010; Wallis et al., 2010). Arsenic mobilization assessment is achieved using the following steps:
• Define environmental conditions using PHREEQC;
• Identify arsenic mobilization potential by comparing arsenic mineral phase diagrams with existing
mineral phases in the aquifer.
20
-------
• Perform PHAST simulation to determine dissolved arsenic concentrations in groundwater.
3.0 ASR-Need Analysis in Planning
3.1 ASR Assessment and Evaluation
3.1.1 Water quantity, quality and water availability
Water master planning for municipalities is conducted to assure adequate water supply for
domestic, industrial, and environmental needs. ASR operation is necessary in many water-stressed
regions to augment available water resources and meet water demand. Major criteria involved in water
budget analysis include water quantity, water quality, and economics:
• Water quantity. This factor includes both water volume and flow. In many cases, temporary or
seasonal water shortages occur due to an imbalance in the flow requirements for consumptive and
environmental flow demand. This can occur even when an adequate volume of water is available due
to precipitation being underutilized. A common mitigation strategy is to have sufficient storage
capacity (e.g., ASR or reservoirs) to capture all available water and use it to mitigate seasonal
shortages.
• Water quality. Naturally available water resources at a location may be not suitable for intended
consumptive or environmental use. For example, brackish water is not potable unless advanced water
treatment is applied to decrease the salt concentration.
• Economics. Water availability is a function of economics for local managers and water resource
management. Water can be made available from inter-basin or inter-watershed transfers, desalination,
rainwater harvesting, or through storage of different forms (e.g., reservoirs, impoundment, storage
tanks, and ASR). Each option is associated with a set of costs and economic/environmental benefits.
The first step in master planning is often a feasibility analysis of different technical and management
options.
Water master planning involves analysis of water budget and flow components to determine
water availability and water demand for the master planning period. Master planning technical details and
an application example are provided in Section 3.2. A technical feasibility analysis is needed after
determining the need for water storage through ASR operations. ASR operations designated for water
injection must demonstrate that there will be no adverse impacts to native groundwater resources. This
criterion applies to residual contaminants in the injected water, as well as the consequences of any
geochemical reactions between the injected water and native aquifer materials. Other criteria to consider
for water injection are the storage capacity of the formation, and the operational economics in comparison
to other alternatives.
3.1.2 Technical feasibility analysis
ASR feasibility depends on several technical conditions, including formation suitability for
storage and recovery, injected water and native groundwater quality, injection and recovery system
engineering, and economic comparisons among alternatives. The analysis results often serve as the basis
for planning and subsequent detailed engineering design. They can also be part of existing ASR
evaluations.
Not all locations have geological formations that are a potable groundwater resource, and/or are
otherwise suitable for ASR operation. A hydrogeological investigation must be conducted to determine
important attributes such as:
21
-------
• Aquifer storage characteristics. The permeability, porosity, and spatial boundaries of the aquifer will
determine injection/extraction rates and the injection bubble geometry for storage. The location of
boundaries such as aquitards, faults, and surface water bodies influence the long-term pressure
distributions and injection/extraction rates. Aquifer depth is also an important consideration from an
economic perspective.
• Petrology and mineralogy of the aquifer formation. Mineral composition often determines
geochemical compatibility with injected water, and thus the potential for groundwater contamination
at an ASR site. For example, formations composed of sandy sediments tend to have high permeability
and porosity, making them good ASR candidates. However, these geological formations can contain
accessory minerals such as arsenic-bearing pyrite, which contain arsenic or other heavy transition
metals that can leach out from rocks, leading to groundwater contamination. Another potential issue
is precipitation of minerals during geochemical reactions with injected water leading to plugging and
reduction in aquifer permeability.
• Geologic structures and water recoverability. Preferential groundwater flow occurs along faults and
fractures, which have important impacts on hydraulic control and recovery of injected water at a
storage site. Furthermore, water injection under pressure can generate hydraulic fractures in the
geologic formations. This type of secondary permeability has been reported numerous times and is
the main mechanism for hydraulic fracturing engineering.
In addition, ASR operation is only one of many storage options in addressing water deficits.
Other options include surface impoundment, reservoirs, storage tanks, inter-basin or inter-watershed
water transfer, and even water credit trading. Each option has its own advantages and disadvantages. A
compromising optimization technique, for example, was used for option evaluation in water infrastructure
master planning in Manatee County, Florida (Chang et al., 2012). More details can be found in EPA
(2015b).
Investigation focus should be placed on system hydraulic control and water quality when
conducting ASR evaluation and design analyses. Section 3.3 provides an overview of ASR evaluation and
design, and the rest of the report illustrates tools and methods.
3.2 Water Availability in ASR Feasibility Investigation
Water availability and climate change can significantly affect estimates of ASR needs for a given
location. This component of the Level 1 DSS is further discussed here.
3.2.1 Water availability analysis
Current Practice
Water resource master planning is focused on water resource inventory and water demand
projections. A quantitative analysis is used to determine the current water availability gap, and to make
projections for future conditions. Water gap mitigation options are identified that either increase water
availability, such as water reuse and ASR practice, or reduce water demand, such as residential and
agricultural conservation practices (e.g., Wang et al., 2013; Chang et al., 2012). This type of water
resource management is well documented in literature and planning guidelines.
Total water management analysis
ASR need for a municipality can be defined using a more comprehensive approach, which treats
ASR operation as a storage unit process and quantifies the parameters necessary to meet water demand.
ASR is especially needed to address large seasonal and time variations in water demand or supply,
22
-------
ET
Evapotrans-
pi ration
AG
Agriculture
Human
Consumptive Use
Wastewater
Continental
Precipitation
sw
* Surface Water
Ocean Discharge
Habitat &
Ecological Needs
Figure 7 Conceptual schematic showing water distributions among major and secondary water
process units. Also shown as red lines are possible flow vectors for water reclamation and re-
distribution. Reclaimed water reuse options are labeled: (1) Groundwater recharge; (2) ASR for
human reuse; (3) ASR for agricultural reuse; and (4) Surface water in ecological flow
augmentation.
resulting in temporal water shortages. Total water management is necessary in this case to identify which
components of water usage affect downstream environments and human activities.
Figure 7 illustrates a conceptual schematic showing water distributions. In this schematic, the
water input of continental precipitation is partitioned into four primary water unit processes:
evapotranspiration, net surface water storage, net groundwater storage, and net agriculture consumption.
The balance defines water resource abundance and availability. In this evaluation, the unit processes are
balanced in terms of volumes and flow rates. Water flow into each unit must equal the outflow to
secondary processes. These secondary processes include human consumption, habitat and ecological
needs, and oceanic discharge, each of which have specific water quality requirements (Figure 7). These
principles form a basis to evaluate water resource sustainability and to optimize management techniques
to achieve objectives. Water transfer and water quality interactions among the process units are
considered in addition to flow balance during the optimization process. An example can be found in the
quantitative analysis of water resource master planning in Manatee County, Florida (Chang et al., 2012;
Board of County Commissioner, 2008; and references therein).
Water resource imbalance for a municipality or watershed occurs when the flow rates in Figure 7
are mismatched. Under reduced precipitation, Qp I, and increased evapotranspiration, QET t, which can
occur in some regions due to global climate change, the water unit processes can become imbalanced
based on the current state of agriculture, surface water and groundwater, leading to water shortages. This
imbalance can be managed by water redistribution among end water uses. Examples include the reuse of
wastewater in groundwater recharge and ASR for agricultural reuse. Agricultural practices use a
significant amount of fresh water and nutrients. Agricultural irrigation in water-poor regions can account
for volumes of potable water comparable to human consumption.
23
-------
A systemic analysis of the water budget components in Figure 7 can produce a technical basis in
master planning for water reuse and ASR needs. There are variations in the principle sources of reclaimed
water among municipalities. The most predominant volumetric sources are centralized water treatment
facilities, which are listed in the first category of Table 1. The volumetric flow rate injected during ASR
operation (Figure 7) can be written as:
Qrw->gw(T, Y) = Qrw — Qrw->sw(T) — Qrw^dr(T,Y,C,A,ET) — Qgw->ag(T, Y, C,A), (1)
where subscript RW is reclaimed wastewater; GW, SW, and DW are groundwater, surface water, and
drinking water, respectively. Variables T, Y, C, A, ET are time (season), geographic location, crop types,
acreage, and evapotranspiration, respectively. Flow vectors are indicated by paired subscript; for example,
Qrw^sw is the flow rate of reclaimed water to surface water.
The water quality limitations are set by groundwater standards drinking water standards
(Cdw)- or requirements for specific uses such as landscape water use (C?rur! ). Often times, water
recovered from ASR operations is further treated for potable drinking water, urban, or agricultural reuse;
ACwt in the following equations represents the change in concentrations due to these treatments.
C,
RW^SW
C (2)
CrW->DR — Cmin (3)
r' a- at = rstd
^GW T ST — UGW
-RW->GW ~ 1 (-GW + ACST + ACWT — Cmin (4)
-GW + AC ST + ACwt = Cqw
The injection rate during ASR operation depends on the reclaimed water flow rate,
\-Qrw ~ Qrw^sw(P)\, agriculture use of recovered water in ASR storage, [Qcw^AG(.t, Y, C,A, W)], and
the direct usage of reclaimed water in agricultural and urban landscaping, [Qgw^ag (.t, Y, C, A, ET)].
Water quality requirements for these end use scenarios are incorporated as variables into Eqs.2-4.
Operation criteria and secondary treatment systems are tailored during the design phase to ensure that
effluent water quality (CRW^GW) meets the drinking water standards (C^). agricultural irrigation
requirements (Cmin). or groundwater standards (C^). depending on the end use of the effluent. Effluent
treatment (ACWT) accounts for water quality changes that occur during the ASR process (ACST). In
compliance with NPDES permits, effluent quality must comply with effluent concentration limits in
monthly average (C) and maximum (Cmax) values (Eq.2). Effluent water quality (Cmin) is defined in
consideration of crop type, long-term soil salinity and sodicity management, and pathogen dispersion
when used for direct irrigation. Reverse engineering in the quantitative analysis is used to optimize
process assemblies among CWW^GW, ACST, ACGW, and engineering economics.
Applications and data needs
The total water management concept and analytical framework (Figure 7) have been applied to
analyze water resource components, evaluate component interactions, and develop optimal solutions
through ASR operations (Thomas and Durham, 2003; Al-Zubari, 1998; Friedler, 2001). Water
distribution and redistribution through human activities and natural processes are treated as unit
processes. This allows quantitative evaluation of management and engineering options and in depth
characterization of water quantity and quality interrelationships. Wastewater reuse was often overlooked
in conventional resource development in the past, but is now viewed as a viable component in water
supply deficiency mitigation conceptually (Pereira et al., 2002; Thomas and Durham, 2003), in Texas
24
-------
(Sheng, 2005), in China (Chu et al., 2004), in Europe (Angelakis et al.., 1999), and in water-stressed Gulf
countries and the Mediterranean region (Georgopoulou et al., 2001; Friedler, 2001; Al-Zubari, 1998).
Variations exist between the current practice and the comprehensive total water analysis
discussed above. The ASR-Need analysis is designed to determine water supply gaps, in total volume or
flow rate, over the master planning period. Ranatunga et al (2014) reported an example water budget
analysis for the Las Vegas Valley region in Nevada, which is described in Section 3.2.4. In a similar
approach, Chang et al (2012) analyzed the engineering costs and environmental footprints of 21
engineering options including components of ASR operations in water supply master planning in Manatee
County, Florida. This type of water availability analysis involves characterization of the following
technical dimensions:
• Land use and population projections to estimate future water demand;
• Water resource inventory and water availability changes with time;
• Management options to reduce water demand
• Options to increase water availability, including water reuse and ASR operation, water credit trading,
water harvesting, etc.
3.2.2 The climate change consideration
Master water resource planning typically involves temperature and precipitation projection over
a period of 30 years, a time frame over which significant climate change may occur (Yang and Goodrich,
2014). Therefore, it is important to consider potential climate change impacts on water availability during
master planning. The U.S. Southwest, Florida, Southern California, and parts of the Basin and Range
physiographic province, will very likely experience further decreases in precipitation and temperature-
induced increases in evapotranspiration (IPCC, 2001, 2013; U.S. EPA, 2015a). As a result, the water
availability gap and general water stress will likely increase.
Precipitation vroiection to analyze climate change impacts
There are two major precipitation projection approaches available for master planning. One uses
historically observed precipitation data in the analysis. A comprehensive statistical analysis of historical
climatic records was conducted by the EPA (U.S. EPA, 2015a). The analysis utilized >100 years of
precipitation data from 1084 monitoring stations in the contiguous U.S. Precipitation data from each
station was characterized by time-series spectrum analysis to reveal the time dependence of inter-annual,
decadal, and multi-decadal variability. The results lead to the delineation of major hydroclimatic
provinces with unique precipitation variability characteristics (Figure 8).
On a national average, precipitation changes in the historical measurement period are relatively
small. While the average temperature and precipitation in recent decades have increased in the U.S. and
worldwide, precipitation has only increased ~ 6% in the lower 48 states and nearly 2% worldwide since
1901. Climate models project that changes in the national precipitation average through 2100 will be
small (IPCC, 2007, 2013), however, large variability is expected among and within different geographic
regions (Figure 8A). This has practical significance to watershed-scale water resource adaptations.
Changes in precipitation rates with time also were determined using linear regressions. Figure 9
shows the mean and spread of determined precipitation change rates for stations within each
hydroclimatic province. The approximate boundaries of the provinces are shown in Figure 8. There are
many stations with persistent precipitation increase or decrease over the term represented in the historical
data. Areas characterized by large changes in precipitation rate are defined as congregate areas with
25
-------
Los Angeles
(A)
Observed Long-term
Precipitation Changes
0 Long-term precipitation increase
Long-term precipitation decrease
§ 10th percentile increased rate of change
4 10th percentile decreased rate of change
Boundary
(B)
2012-2017 Projected
Population Change
Source: ESRI, 2013
H > 1-5 %
n 1.0-1.5%
0 - 0.9%
<0%
'Vs Boundary
Figure 8 Spatial distributions of population change and long-term precipitation change in the
contiguous U.S. (A). Areas of long-term precipitation decrease (red) and increase (blue) are
delineated based on spatial aggregation of precipitation change rate over a "98 year period.
The six hydroclimatic provinces include: (I) Florida and Southeast coast, (II) Lower Mississippi
— Ohio River valley — New England region, (III) Great Plains and Midwest, (IV) Ranges and
Basins, (V) Western Coast, and (IV) Great Lakes. Detailed information on hydroclimatic
provinces is available in U.S. EPA (2015a); (B). ESRI population data is presented for 2012-2017
on a county scale. Red lines mark the boundaries of hydroclimatic provinces.
26
-------
increases or decreases in the 90th or 10th percentile. Several conclusions can be made in relation to the
future water availability:
• Analysis of long-term precipitation variability has led to the identification of six hydroclimatic
provinces, each possessing a unique set of precipitation variability and trend of change. These
regional distributions reflect combined effects of a synoptic climate state and regional/local climatic
factors.
• Among the regions expected to suffer worsening water availability conditions, portions of the
Appalachian Mountains (P-II) and the Northeast (P-IIIb) have experienced a long-term precipitation
decrease and increased precipitation variances with time (Figure 8A). Such changes will likely
continue under intensified Atlantic atmospheric circulation.
• A region roughly centered about Phoenix, AZ in the Southwest U.S. (southern portions of P-IV and
P-V in Figure 8A) has experienced an overall decrease in precipitation in response to the dynamics of
El Nino Southern Oscillation (ENSO) and other climatic systems. The areal size of this region is
expected to increase in the future, with the boundaries expanding eastward and northward into the
Basin and Range physiographic province.
• Precipitation variance and the frequency of extreme precipitation events (high-intensity and droughts)
have increased markedly since the mid-1900s (Bates et al., 2008). Regions that have experienced
increased variance include the Northeast and the Atlantic coast, Ohio River basin, Lower Mississippi
River basin, and the Southwest U.S. Intense storm events, preferential rain as opposed to snow, and
prolonged drought are major types of climatic change that studies have projected for different parts of
the U.S. (e.g., USCCSP, 2001; Barsugli, 2009; and references therein).
The second approach to precipitation projection involves using the Global Circulation Model
(GCM) and/or Regional Climate Model (RCM). Several down-scaled nation-wide RCM databases are
available including those of the U.S. Bureau of Reclamation (USBR)4 and the NCAR5. Results from these
climate model simulations indicate a small change in precipitation average, but a large variance over
North America for the next 90 years (2010-2100) (IPCC, 2001, 2013). The EPA Office of Water has
produced a national web portal that provides projected precipitation data.6 These models are often
validated against long-term precipitation observation data sets. Additional data from an adaptive, remote-
sensing monitoring scheme can be used to produce projections for water resources planning and
engineering design when the model accuracy is questionable (U.S. EPA, 2015b).
Figure 10 shows a generalized process in a down-scale modeling approach to generate
precipitation projections at a local watershed. The details and technical basis of the process are given in
U.S. EPA (2015a). The basic steps are:
First, it is necessary to evaluate whether the RCM projection is based on bias-corrected GCM in pre-
processing.
Second, RCM downscaling requires a full consideration of the regional and local climate factors
(Figure 10). Convective and orographic precipitation, and local effects of large water bodies can
significantly affect local precipitation variability, especially during high-intensity precipitation events
4 http://qdo-dcp.ucllnl.orq/downscaled cmip projections
5 http://www.narccap.ucar.edu/about/index.html
6 http://water.epa.gov/infrastructure/watersecuritv/climate/scenario.cfm
27
-------
co
«:
CO
«:
CO
«:
I
0
120
ill
m=0.004 (-0.004)
ct=0.139
N=43 (33)
Skewness: 0.59
Kurtosis: 0.41
P-ll
LONE
0
100
m,
m=0.085 (0.077)
ct=0.162
N=315 (251)
Skewness: 1.94
Kurtosis: 11.47
P-llb
LMRB
P-lll
Great Plains.
Midest
0
10
P-VI
Great Lakes
0
40
5 20 ¦
CO
m=0.084 (0.088)
ct=0.181
N=274 (218)
Skewness: -0.30
Kurtosis: 1.33
P-lllb
SE Mixing Zone
m=0.130 (0.126)
ct=0.183
N=28 (22)
Skewness: 0.13
Kurtosis: -0.83
R. (%year )
P-IV
Basin & Ranges
0
40
TO 20 ¦
CO
m=0.093 (0.086)
ct=0.228
N=143 (113)
Skewness: 0.24
Kurtosis: 0.83
P-IVb
Columbia and
Snake river basins
V$7\
P-V
West Coast
Ik
m=0.052(0.049)
<7=0.15
N=99 (79)
Skewness: 0.10
Turtosis: 5.76
R, (% year"1)
m=0.126 (0.121)
ct=0.172
N=47 (37)
Skewness: 0.25
Kurtosis: 0.001
CO
0
100
m=0.085 (0.078)
<7=0.183
N=227 (181)
Skewness: 0.25
Kurtosis: 3.53
1.0
CO
m=0.065 (0.073)
ct=0.125
N=31(23)
Skewness: -0.45
Kurtosis: 1.01
CO
1.0
R, (%yr )
Figure 9 Statistics of the rates of precipitation change (Ri, %yr_1) were calculated from long-
range (~98 years) historical monthly precipitations measured at USHCN climate
stations. The red line indicates the mean percent change and the dotted lines indicate
the standard deviation range. (U.S. EPA, 2015a)
28
-------
(U.S. EPA, 2015a). In some cases, the post-processing in bias correction is conducted to account for
some, if not all, of these location-specific climatic consequences (Figure 10).
Third, and most importantly, the RCM datasets are validated using independent data sets to improve
their reliability. The process of model calibration and validation involves simulated reproduction of
long-duration historical data (e.g., >60 years). This is done to ensure that the RCM simulations have
properly captured the climate variability and local climate factors. Climate variability and local
factors are shown in Table 2-4, and examined in detail in Section 4.0.
Watershed hvdrolosical modeling of climate change impact
In regions such as the
southwestern U.S., impacts from
climate change can lead to less
precipitation, QP, and more
evapotranspiration, QET. This may
result in decreased surface water, Qsw-
and groundwater, QGW. flow (Figure
7), thus reducing water availability at
watershed-scales. Climate change
impacts on stream flow and water
quality have been observed in
watersheds across the U.S. for many
communities relying on surface water
(IPCC, 2007, 2013; Johnson et al.,
2012; Tong et al., 2012; Ranatunga et
al., 2014).
The effect that precipitation
changes has on stream flow is not
straightforward, but is critical to
determine for quantification of water
availability gaps. Changes in
precipitation and land use can affect
both stream flow and surface water
quality for most watersheds. Tong et al
(2012) quantitatively analyzed the
contribution of these two factors to
CMIP5 or GCM
outputs
Bias
correction
RCM experiment -
Physical model
Regional/local
climate actors
RCM experiment -
Post-processing
Model validation on
historical periods
Bias
correction
RCM Datasets
Comparison with long-
term observation dataset
Projections Range (uncertainty)
Figure 10 A typical procedure in climate downscaling
yielding the RCM dataset. Also shown is the validation
process of precipitation projections for water resource
planning.
stream flows in the Little Miami River watershed in southwestern Ohio. U.S. EPA (2015a) published a
generalized procedure for projecting stream flow changes as a result of land use change and climate
change impacts. The procedure is shown in Figure 11.
Future precipitation and temperature at a location are used as the input values in a watershed
hydrologic model using the Hydrological Simulation Program in Fortran (HSPF). A land use model is
separately established in response to projected populations. Several land use projection models can be
used (U.S. EPA, 2015a). One is the CA - Markov model (Figure 11). The CA-MC outputs are
incorporated into the watershed HSPF model to generate stream flow projections. This integrated land use
and climate change modeling has been applied successfully in urban and rural watersheds of diverse size
(Tong et al., 2012; Ranatunga et al., 2014; Chen et al., 2015).
29
-------
Hydrologic model
Population model
Land use model
Future climate inputs
HSPF stream projections
CA-MC land use projection @
Figure 11 Schematic of a typical integrated modeling approach in projecting surface water quality
and quantity changes in a watershed.
Water reclamation and ARS practices for the mitigation of water shortages represents a viable
means of adapting to potential future climate change. Initial investigation results have indicated that water
imbalances will be region-specific. In some regions, water availability stress will be driven by
precipitation changes, while in other regions the predominant drivers will be by population change and
socioeconomic activities.
3.2.3 The socioeconomic factor
Socioeconomic developments affect water demand with time. Larger population size and greater
economic activity can generate greater water demand and lead to greater stress on sustainable water
resource development. As a result, socioeconomic factors are pertinent considerations in ASR planning
and engineering. Population growth thus far has been independent of water availability in the contiguous
U.S. As shown in Figure 8, population growth in the U.S. is concentrated in the Southwest, West, Great
Plains, Texas, and the Atlantic coastal states south of Virginia. In regions where there is increased
population growth in the immediate future (Figure 8B) without a corresponding increase in the historical
precipitation rate (Figure 8A), it may be difficult to meet increased water demands. These regions
include:
• Most of the Basin-and-Range Hydroclimatic Province IV, particularly in Arizona, New Mexico, and
southern Nevada, as well as Nebraska (Region III) and western Texas (Regions Illb);
• The southern California area in the Western Coast Hydroclimatic Province V;
• Florida and southeast Georgia in the Florida and Southeast Coast Province I. The water shortage is
largely due to rapid population growth in the past decades.
30
-------
• Coastal Georgia, coastal and central North Carolina, and regions of Virginia in the Blue Ridge and
Appalachian Mountains.
These regional assessments provide an overview of current water availability statuses and trends
in the future climate. The results can lend support to regional planning and identification of focus areas
for sustainable water resources management. Detailed water resource investigations performed by water
managers typically focus on the watershed or local municipality-scale.
3.3 ASR-Need Analysis in Las Vegas - A Case Study
Total water management including the use of ASR applications was investigated to analyze
potential options in water planning for Las Vegas, Nevada. This case study provides an example of the
type of analysis which can be used in Level 1 of the DSS to determine whether water conservation is
needed (e.g., what are the current and future water gaps) and whether ASR is a feasible means of closing
these gaps. The investigation methodology, results and conclusions are described in Ranatunga et al
(2014), forming the technical basis for this section.
3.3.1 Geophysical settings
The Las Vegas Valley in southern Nevada lies within the Great Basin and Mojave Desert sections
of the Basin and Range physiographic province, bordering the West Spring Mountains to the west and
Ground Gunnery Range to the north (Figure 12). The arid watershed has an area of approximately 4850
km2 and an elevation at the valley floor of -610 m. Most of the storm drains and stream channels within
the valley are dry or low flow due to the arid Mediterranean climate, but some intermittent streams have
become perennial streams due to increased wastewater discharge from urban areas (Piechota and Bastista,
2003).
Las Vegas was a small city during the early 20th Century. Most surface water in the basin was
from summer flash floods, winter rains, and flow from artesian springs (Morris et al., 1997). The runoff
drained to the Las Vegas Wash, a generally barren, gently-sloping, sandy channel that conveyed storm
runoff and wastewater from the Las Vegas Valley to the Las Vegas Bay, an arm of Lake Mead. The Las
Vegas Wash was dry for most of the year, and contained discharge only during brief periods of major
storm runoff. As communities in the Las Vegas Valley grew, the amount of effluent discharges and
surface flow that drained to the Las Vegas Wash and Lake Mead increased. The growing urban area
discharged enough wastewater into the Las Vegas Wash to create a small, but perennial, stream flow by
the 1950s. The ephemeral stream flowing into Lake Mead had been transformed into an active, 40-m-
wide river channel by the end of the 20th Century (Buckingham and Whitney, 2007; USBR, 2009). Today,
the flow of the Las Vegas Wash is composed of treated domestic wastewater effluent, treated industrial
wastewater effluents, dry and wet weather runoff, and groundwater seepage. Industrial and domestic
wastewater effluent discharges account for about 90% of the flow (Cooley et al., 2007). There are three
major municipal wastewater treatment plants located along the Las Vegas Wash, which collect and treat
the municipal wastewater generated in the Las Vegas Valley. There are nine additional permitted
discharges along the Wash that contribute significantly to the flow (Piechota and Bastista, 2003).
The Las Vegas Valley climate is hot and arid. The average annual precipitation is 106 mm,
occurring mostly as high-intensity, short-duration storms in July and August and low intensity rainfall in
the winter months. The average monthly temperatures range from 9°C to 34°C, with an average annual
temperature of 21°C. Average daily relative humidity ranges from 32 to 56 percent in mid-winter and
from 11 to 28 percent in mid-summer. Evapotranspiration is high because of the high summer
temperatures, high solar radiation in cloudless skies, low humidity, and frequently windy summer
31
-------
UkeM««d
Figure 12 Location of the Las Vegas Wash watershed, Nevada. The color-shaded areas are green
for shrub and forest, red for the built area of the Las Vegas metropolitan area, and blue for
water bodies of Lake Mead.
conditions (Stave, 2001, Morris et al., 1997). Soils in the Las Vegas Valley are generally composed of
gravel, windblown sands, and fine grained silts and clays. Valley floor soils typically have a low field
capacity and high penneability. On steep slopes, especially along the Wash, the disturbed soils are
particularly susceptible to erosion (Bureau of Land Management, 2004).
Over the last century, social and economic developments, including legalized gaming, the
construction of Hoover Dam, industrial production for the Second World War, atomic testing, tourism,
and the advent of the modern mega-resort, have steadily increased local populations and associated
demands for water in Las Vegas (SNWA, 2009). Rising population has increased the volume of
wastewater discharged into the Las Vegas Wash, The rapid urbanization and increase in impervious
surfaces cause more storm water runoff to flow directly into the Wash rather than be absorbed by the soil.
The increased water flow in the Wash not only accelerated soil erosion and destabilized the stream
channel, but also significantly degraded wetland areas and contributed excessive sediment to the Las
Vegas Bay (U.S. EPA, 2012b). Climatic change in the form of an overall drying of the region and
increased frequency of extremely high precipitation events has also had a considerable impact on the
region (USGS, 2000; Christensen et al., 2004).
32
-------
Because Las Vegas is a major financial center and one of the fastest growing cities of the region,
it is important to understand the conditions that affect future water resources. With increasing water
demand and decreasing natural water supply, the need to investigate the watershed hydrology and
accurately forecast the future water balance is becoming ever more essential. This confluence of factors
makes the Las Vegas Wash a unique setting for examining the hydrologic impacts of urban growth and
climatic changes.
3.3.2 Future conditions for planning
Future climate change scenarios
The future climate impacts of concern for planning in the Las Vegas Wash watershed include
extreme weather events and a shift in winter precipitation patterns, which influence the winter and spring
discharge (e.g., Barnet and Pierce, 2008; Karl et al., 2009; IPCC, 2007). Alterations in precipitation and
evapotranspiration (ET) rates can affect the amount of annual runoff, groundwater, and soil moisture. The
shifting weather patterns have caused certain areas, such as Las Vegas and other regions in the U.S.
Southwest, to experience less precipitation and changes in precipitation seasonality. Most published
reports (e.g.,, Barnet and Pierce, 2008; Karl et al., 2009; IPCC, 2007) predict that the watershed and
region as a whole will very likely experience a decreasing availability of water resources. There is a large
uncertainty and variations in these future precipitation projections. As a result, this investigation was
based on the ranges of climate change obtained from the 2000 and 2009 annual reports published by the
United States Global Change Research Program (USGCRP). Climate models from the Hadley Centre in
the United Kingdom and the Canadian Centre for Climate Modeling and Analysis were utilized for
analysis. To generate future likely precipitation and temperature scenarios, these modeling results were
further combined with the climate change information from IPCC (2007) and the historical observations
of USGCRP (2000) and Karl et al. (2009).
Based on the scenario with no explicit climate policies put into place to reduce greenhouse gas
emissions, the global average air temperature is projected to rise by 2.4°C to 6.4°C by the end of this
century, and the temperature of the study area will increase by 4°C by the horizon year of 2050 (IPCC,
2007; Karl et al., 2009). In terms of precipitation in 2050, the projection results differ among the global
climate models. The Hadley model projects that there will be a substantial increase in precipitation, with a
percentage increase of 80-100% over California and Nevada and -20% elsewhere in the U.S. The
Canadian model also predicted a large increase of 80-100% over southern California and an increase of
-20% in the Great Lakes and Northern Plains, however the Canadian model also found regions with
precipitation decreases exceeding 20% in the Oklahoma panhandle and the eastern U.S. The IPCC
projection found a dryer condition occurring in the Southwest U.S. and little change or increasing
precipitation in the northern U.S. Because of the incongruent results, both the wet and dry scenarios for
future precipitation were considering during the investigation. The wet scenario (Wet) considered a 20%
increase in precipitation and 4°C increase in temperature, whereas the dry scenario (Dry) considered a
20% decrease in precipitation and 4°C increase in temperature (Table 3). The percentages were chosen to
be within the range for precipitation increases and decreases observed for the Hadley and Canadian
models. According to these climate scenarios, historical temperature and precipitation time series data
were adjusted and inserted to the HSPF hydrological model for simulating the watershed.
Population growth projections
The close relationship between population growth of an area and its watershed hydrology is well
known (e.g., Buckingham and Whitney, 2007). Prior to the rapid population increase and urban
development of the area, the Las Vegas Wash was an intermittent stream that carried storm water to the
Colorado River. Due to the recent urban developments and population growth, the Wash is now carrying
not only storm water runoff, but also urban wastewater treatment effluent. The average daily flow has
33
-------
increased from 0.736 m3/s in 1970 to 6.23 m3/s in 2000 (Piechota and Bastista, 2003). Therefore, in order
to better postulate the future hydrologic conditions in the Wash, it is essential to estimate the population
growth of the study area. This estimate can be used as a surrogate to estimate the amounts of urban
discharge to the Wash and can be used to facilitate a more accurate projection of land use changes.
Table 3 Climate change scenarios for the Southwest U.S.*
Climate Scenario
Changes in Temperature/Precipitation
Wet
Dry
+ 4 °C, + 20% Precipitation
+ 4 °C, - 20% Precipitation
Note: * - Sources; USGCRP (2000) and Karl et al. (2009).
Previous publications (Gabriel and Accinelli, 2007; Harris, 2005; Miranda and Lima, 2010) show
that the growth of a population can often be accurately portrayed by the logistic function. The logistic
population growth model uses an 'S' shaped curve, implying that when the environment has adequate
resources, the population will grow exponentially, and when the population is in proportion to the natural
resources and in alignment with the environmental carrying capacity, it will grow at a much slower rate or
even at a constant rate. However, when resources become limited and the population is larger than the
environmental carrying capacity, the growth rate will decrease, and will ultimately become zero (Gabriel
and Accinelli, 2007). In the Las Vegas Valley, the population growth pattern resembles that of a logistic
curve (Figure 13). Before the mid"20th century, the population growth was slow. From 1970 through 2000,
the average annual population growth in Southern Nevada was 7 percent per year. This is the time period
with an exponential growth of population. By 2000, Southern Nevada's population had increased to
nearly 1.5 million people (SNWA, 2009). By the end of the 20th century, the population growth had
become constant, growing at a rate of 3.4 percent in 2009. According to the model predictions from the
Center for Business and Economic Research at the University of Nevada, by 2015, the growth rate will be
at 2.6 percent, and by 2030, it will level off at around 1 percent. The forecast predicts a growth rate of 0.8
percent in the year 2050.
In this investigation, the logistic function available in the SPSS Statistics software package,
release version 16.0.2, was employed for population analysis. A curve fitting equation was generated and
statistics were estimated based on the urban area population data. The SPSS output for the population
logistic function provided a curve fitting R-squared value of 0.995, indicating a high level of significance.
Using the logistic equation generated, the future population was estimated up to the horizon year of 2050
(Figure 13). According to the model outputs, the population of the valley will be about 3.77 million by the
2050, which is more or less similar to the predictions by the Center for Business and Economic Research
at the University of Nevada, which estimated a population of 3.85 million.
Future land use change scenario
With the city expansion and population increase, the Las Vegas region has undergone significant
changes in land use. Specifically, for the last fifty years, there has been an increase in urban areas and a
decrease in agriculture and pasture lands (Adhikari et al., 2011).
Changes in land use/land cover will affect the hydrologic conditions. Generally, an increase in
impervious surfaces will cause local decreases in natural interception, infiltration, percolation, and soil
moisture storage. Hence, the amount of runoff will increase. Less infiltration during storms and increasing
overflow of effluents will create higher peak flows (SNWA, 2009). It is common that watersheds with
large amounts of impervious cover have significantly reduced groundwater recharge and increased storm
34
-------
23.0
13.0
¦Population
¦Wastewater
16.0
14.0
12.0
10.0
5.0
6.0
4.0
2.0
0
0.0
1970
1990
2010
2020
2030
2050
Figure 13 Future population and wastewater projection for the Las Vegas Wash watershed.
flow and flood frequency. With an increase in impervious cover, surface runoff has greater velocities,
larger volumes, and shorter lag-times between peak rainfall and highest flow concentrations (Brun and
Band, 2000). As the main drainage in a highly urbanized watershed, the Las Vegas Wash also shows
similar characteristics in its hydrograph with sudden peak discharges that last for a short time period. To
better understand the likely hydrologic impacts of urbanization, it is therefore essential to be able to
predict future land use change.
A common approach to derive future land use scenarios is to use a land use model to simulate the
future land use conditions. In this study, the CA land use model in IDRISI, developed for the Las Vegas
Wash watershed by Sun et al. (2013), was used to simulate the land use conditions in the Las Vegas
watershed for the year 2050. This model is an enhanced land use model coupling the CA-Markov model
with a population variant to depict the effects of population growth on land use (Tong et al., 2012). Two
sets of historical land use records were used to determine the pattern and the trend of land use and land
cover changes, and an additional map was used for validation. The 1992 and 2001 land use/land cover
maps from the USGS National Land Cover Dataset (NLCD) were adopted to develop the model, whereas
the 2006 NLCD was used to validate the model. Additionally, the 2050 population estimate from the
logistic equation was incorporated to the CA-Markov land use model to postulate the land use pattern in
2050. The projection results suggest that there would be a great increase in urban area and city expansion
throughout the valley (Table 4). Figure 14 shows the postulated 2050 land use map of the Las Vegas
watershed.
Estimating future wastewater discharge
More than 90% of the total water use of the Las Vegas Valley is population-dependent (SNWA,
2009). With the increase in urban population, more water is needed, and as Morris et al. (1997) and
Buckingham and Whitney (2007) have suggested, more wastewater will be discharged into the Las Vegas
Wash. The effects of the increase in population and tourism on stream discharge can be related to the
increase in the discharge of treated sewage effluents and the volume of urban runoff from the increased
amount of impervious surfaces. The wastewater generation and discharge are closely related to population
in the watershed area (Figure 13). A linear regression between population and wastewater yielded a
35
-------
highly significant Pearson correlation coefficient of 0.97 and a standard error of 0.002 (Ranatunga et al.,
2014). Assuming the correlation remains constant in the future, if the water policy is unchanged, future
wastewater generation can be estimated from the population alone.
Table 4. Land use projection for 2050
Land Cover/Use Type
1992 (Km2)
2050 (Km2)
% change from 1992
Urban
411.47
1875.77
355.87
Agriculture
29.48
2.78
-90.57
Forest
287.13
367.31
27.93
Water
2.54
43.41
1610.20
Range/Grass land
3730.44
2258.34
-39.46
3.3.3 Water budget analysis
Ranatunga et al (2014)
described the modeling results of water
budget analysis for the 2050 hydrologic
conditions in the Las Vegas Wash
watershed. Using a validated
Hydrological Simulation Program in
Fortran (HSPF), the quantitative
analysis considers the effects of future
climate, wastewater discharges
generated from a growing population
(WW), and the projected 2050 land use
patterns (LU). The model was then used
to simulate the impacts of wastewater
discharge, climate change, and land use
change under the following scenarios:
• Wastewater discharge with no
change in climate and land use
(Base + WW)
• Wastewater discharge with a wet
climate and no land use change
(Wet + WW)
• Wastewater discharge with a dry
climate and no land use change
(Dry + WW)
• Wastewater discharge with a wet
climate and land use change (Wet +
WW + LU)
• Wastewater discharge with a dry
climate and land use change (Dry +WW + LU)
Currently, water in the Colorado River is apportioned among the seven Colorado River Basin
states for their total consumption or "net" use. Nevada receives 3.7* 10s m3 of water per year from the
Urban
ForeM
Barren
Water
Shrubs/grass
Figure 14 Projected 2050 land use/land cover map of Las
Vegas Wash watershed.
36
-------
Colorado River for consumptive use (SNWA, 2009). Since Las Vegas is the main population center of the
state, almost all of the apportioned water to Nevada is drawn by the city. A water credit for use in the
same year is claimed when treated reclaimed water is returned to Lake Mead through the Las Vegas
Wash. Each year, through this return flow credit, southern Nevada is able to divert more water than its
consumptive use (SNWA, 2009; Cooley et al., 2007). As about 90% of the total stream discharge of the
Wash is composed of treated wastewater (Piechota and Bastista, 2003; Cooley et al., 2007), the amount of
urban wastewater generated from the population in the watershed is one of the most important factors
determining future flow estimates.
lOO
90
So
60
50
40
30
20
10
Simiated How
Ob&erad Flow
— IT —
Figure 15 HSPF Simulated continuous stream discharge with wastewater projections for the
Base + WW condition
The model simulations indicated that future land use and climate conditions will have a
considerable effect on return water flows. Although climate and land use changes affect only 10% of the
total stream discharge, changes in the peak flow can be observed under different climate and land use
scenarios. As shown in Figure 15, the HSPF simulation reproduced the baseline flow fairly well during
the calibration period, while it did miss several flow peaks. Even if no climate or land use changes occur
in the future, the wastewater discharge to the Wash would still increase in response to the population
increase. This increase is shown in both base flow and peak flows. The projected daily discharge of the
Wash is a 2.5 fold greater in 2050 when compared to the observed flow in 1992. The average annual flow
rate is projected to increase from 6.23 m7s in 1992 to 17.69 m Vs by 2050.
The modeling results that include both climate and land use change show that land use changes
have a significant impact on stream discharge. From Sun et al. (2013), the urban area of the Las Vegas
Valley is projected to increase by 355% from 1992 to 2050 (Table 4). The increase in impervious surfaces
has a significant impact on stream discharge, especially for the peak flow during storm events. This effect
is amplified by climate change effects. Under the Wet climate change scenario (Wet + WW + LU), larger
flow events occur at a higher frequency of 33 times a year (Table 5). Most importantly, the peak flow can
37
-------
reach 373.78 m3/s. This scenario also gives the highest percentage increase of total discharge, 16.59%,
with an average flow rate of 20.6 m3/s (Table 5).
Table 5. Projected high stream flows and average annual flow of the Las Vegas Wash in 2050 with
returned wastewater under a set of climate anc
land use scenarios*
Highest Flow Occurrences
Average Annual Flow
Frequency
Highest value
Flow rate
% change from
Scenario
over 70 m3/s
(m3/s)
(m3/s)
Base + WW
Base + WW
1
70.51
17.66
Wet + WW
4
83.25
18.01
1.95
Dry + WW
0
57.77
17.35
-1.78
Wet + WW + LU
33
373.78
20.60
16.59
Dry + WW + LU
16
106.19
18.42
4.28
Note: * - Modified from Ranatunga et al. (2014). Significant digits are from reference; m3/s -cubic
meter per second. Scenarios modeled include: Wastewater (WW) discharge with no change in
climate or land use (LU) = Base + WW; WW discharge with a wet climate and no LU change = Wet +
WW; WW discharge with a dry climate and no LU change = Dry + WW; Wet + WW with land use
change = Wet + WW + LU; Dry + WW with LU change = Dry +WW + LU.
For the scenario of a Dry climate with incorporated land use changes (Dry + WW + LU), the peak
flow also increases, but to a lesser extent than for the Wet climate scenario (Table 5). On the other hand,
the frequency of peak flow events is higher than for the base scenario (Base + WW) for the Dry scenario
with no land use changes (Dry + WW). Among all the scenarios studied, the projections of this scenario
are critical. Most of the other recent modeling efforts predict hot and dry climate trends and increased
urbanization in this region in the future. According to the model projections, there would be 16 flow
events with an average flow larger than 70 m3/s, and the largest flow event would be 106.19 m3/s (Table
5). Overall, for the Dry + WW + LU scenario, the total average flow would increase by 4.28% compared
to the base scenario. This could be due to larger runoff volumes during storm events as a result of
increasing impervious surfaces.
Generally, the results show a 2.5 fold increase in the average daily discharge into the Wash by the
mid-21st century if the projected population growth continues. Apart from the increase in the average
flows, there will be more extreme flood events caused by climate and land use changes. These kinds of
destructive events can be controlled by applying best management practices (BMPs), such as constructing
detention basins or ponds (Welty, 2009). It is also important to note that these projections are not able to
account for technological advances which can occur in the future, leading to the increased efficiency of
these BMPs. In this watershed, excessive use of detention basins is not encouraged as it may reduce the
amount of runoff that will channel to Lake Mead, and the Las Vegas Valley may not get as many return
flow credits. However, in order to control floods, it is necessary to have some detention basins. Currently,
there are 39 detention basins throughout the watershed and another 30 basins have been planned
(LVWCC, 1999). These detention basins are designed to reduce peak storm water flows by detaining
water and releasing it over a period of less than seven days.
3.3.4 The role of water conservation and storage in meeting future water demands
Currently, the Southern Nevada Water Authority has operated a large ASR facility for consistent
water supplies since 1985. Operation of this ASR system relies on water from Lake Mead. Water
38
-------
withdrawn from the Lake is stored in the aquifer for use in other seasons. However, due to a decade of
chronic drought in the area, water level in Lake Mead has dropped over 130 feet from 2000 to 2014
(SWNA, 2015). The inflow from Colorado River and other tributaries, including the Muddy River and
Lower Virgin River, has shown consistent decline. In addition, these areas heavily depend upon the
snowmelt from snow caps on high altitude mountains. As the climate warms, the peak flow at the
snowmelt-fed streams has shifted to be earlier to the spring season.
Based on the quantitative investigation results, the total water balance between water supply and
demand was further investigated for a period up to the year 2050. Figure 16 shows the time-evolution of
major fixed fresh water supplies and the total demand estimated using population projections. Outcomes
indicate that the Las Vegas Valley system should have an adequate water supply from the existing sources
until the early 2020s. The plot analysis shows the cutoff year as 2024. After this time, the demand will
exceed the total amount of water available, and the valley will need to find other alternative sources of
fresh water. The estimated demand of water by 2024 is 9.55 * 10s m3 per year, and this estimated demand
increases up to 1.33*109m3 peryear by 2050. Based on the 2010 SNWA annual report, the main sources
of water supply for the area are the Colorado River, Las Vegas Valley groundwater supplemented by
existing ASR operations, Virgin/Muddy Rivers Tributary and Coyote Spring Valley Groundwater
Intentionally Created Surplus conservations, and Drop 2 Reservoir System Efficiency Intentionally
Created Surplus. These supplies can provide adequate water until 2024. After that time, the model
predicts that there may not be enough water to meet the demand.
The potential deficit after 2024 can be a challenge to water authorities in the Las Vegas Valley.
The total water deficit is projected to increase up to nearly 2.46* 10s m3 per year by 2050 (Figure 16).
According to the 2009 SNWA water resources plan, there are multiple potential sources that the
authorities have targeted to bring more water to the valley. There are plans to draw water from other
groundwater resources in Clark, Lincoln, and White Pine Counties. SNWA is also planning to build a
massive pipeline system that would take underground water from the Great Basin aquifer system, located
about 482.8 km north of Las Vegas, and pump it to Las Vegas. The plan calls for transferring up to
2.22* 10s m3 of water per year from rural Nevada to the Las Vegas Valley (Progressive Leadership
Alliance of Nevada, 2006). However, these long-distance water transfer options can have detrimental
environmental impacts; for example, water transfer from the Great Basin area could result in declining
groundwater levels and could impact the area's biodiversity. Other potential water supplies include sea
water desalinization, brackish water desalinization, and withdrawal from water banks, such as the
Arizona, California, and Southern Nevada water banks.
Water conservation is the other option to reduce the future gap between water supply and
demand. Through conservation practices, such as education, water pricing, regulations and incentives, and
water-less landscaping, SNWA is planning to reduce the per capita water use by 0.753 m3 per capita per
day by 2035. With conservation practices, SNWA has already reduced consumptive use by roughly
7.9X107 m3 annually between 2002 and 2008 (SNWA, 2009). It is anticipated that this conservation plan
could save 3.4X108 m3 of water annually by 2035.
The projected discharge analysis of Las Vegas Wash in this investigation indicates that the return
flow credit from the Wash is a significant source of water. With an increase in population, the total indoor
and outdoor water use will increase. Studies suggest that a higher outdoor water use may lead to a higher
loss of water from the hydrologic system through evaporation (Qaiser et al., 2011; Stave, 2003).
Therefore wastewater reuse for outdoor applications is less attractive than return flow and discharge back
to Lake Mead for the return flow credit.
39
-------
1.4E+9
1.2E+9
1.0E+9
Return Flow Credit
¦ Drop 2 Reservoir
l Vi rgi n/M uddy/Coyote ICS
¦ Ground Water
Colorado River Allocation
•Total Demand
¦Tata I Supply
b.OE+8
4.0E+S -
Q.QE+O
Figure 16 The projections of total water demand and supply, showing the importance of return
flow credit from the Las Vegas Wash stream flow in the sustainable water supply for the region.
Adopted from Ranatunga et al. (2014).
It is necessary to note the magnitude and composition of the Las Vegas Wash return flow.
Reclaimed wastewater effluent is a large portion of the base flow, while overland runoff and flash flood
water predominate the stream during peak flow events. The peak flow rate is many times greater than the
base flow (See Figure 15 and Table 5). Lake Mead is one primary storage facility for surface runoff and
return flows. Additional storage capacity could help capture the maximum amount of return flows during
the peak flow periods, and can be later used to supplement water during the dry season. ASR expansion
from the current "water bank" operation is therefore a viable option for planning, with the capture and
storage of such large flows at the center of adaptation.
The Las Vegas case study shows how to conduct the climate change and water availability, as
well as water demand in the ASR-Need analysis (See Figure 5). The results demonstrate the potential for
ASR application to prevent future water gaps, which are predicted to occur by the year 2024. This
analysis allows organizations such as the Southern Nevada Water Authority to be proactive when
planning how to meet growing water demands, as well as incorporating the impacts of population growth
and global climate change. This analysis method can be applied to other locations considering whether the
implementation of ASR is a viable means of meeting future water demands. The Level 1 tool is designed
40
-------
to support this analysis by including projections for current water demands, projected population
increases, and projections for temperature and precipitation changes under different representative
concentration pathways for GHG emissions.
4.0 ASR Facility Planning and Assessment
4.1 Assessment for ASR Planning
The initial stage of ASR site planning is often focused on hydrological and environmental
assessment. Major objectives are: 1) Evaluate the hydrological viability of candidate ASR site, including
injection rate, storage capacity, and injection-related geotechnical factors; 2) Assess likely geochemical
compatibility of the storing groundwater aquifer and approximate the likelihood of negative groundwater
impacts; and 3) When recovery is involved, determine the degree of hydraulic control over injected water
and calculate the rate of recovery.
4.1.1 Infiltration rate and storage capacity
The overall water infiltration rate depends on the vertical water infiltration rate in vadose zone
and, when directly injected into an aquifer, is a function of the injection well or infiltration trench design
and aquifer hydraulic conductivity. Related groundwater principles for these determinations have long
been established by Darcy's law for incompressible groundwater. For groundwater head (h) distribution
in an aquifer with a flow sink/source term, G, the generalized Darcy's law is:
^ = aV2h - G (5)
Hydraulic diffiisivity (a) is a ratio of hydraulic conductivity (k) and specific storage (Ss) for
confined aquifer (^/^ ). or a ratio of transmissivity and specific storage (^/$ ) for an unconfined aquifer,
where the transmissivity T(= k ¦ b) is a product of the hydraulic conductivity and aquifer thickness (b)
(See Section 4.1.3 below).
The general Darcy's law has been applied to describe various groundwater systems and hydraulic
devices such as wells, trenches and fractures. The familiar Theis (1935) equation, Jacob equation for
wells in a confined aquifer, and the Boussinesq equation for homogeneous and isotropic aquifers are also
commonly used. For practical engineers, the fundamentals of groundwater hydraulics, their complexity
and their relevance to groundwater system design can be found in the book "Groundwater and Wells" by
Driscoll (1986).
Furthermore, groundwater systems for ASR operation vary between locations. Commonly
encountered types include groundwater flow in homogeneous isotropic, homogeneous anisotropic,
inhomogeneous anisotropic, and leaky confined aquifers, and those counterparts in unconfined aquifers.
In addition, natural groundwater aquifers may vary in lateral extent, contain impermeable layers or lenses,
or contain geological discontinuities such as faults for conduit flows. This degree of complexity in
hydrogeological groundwater systems will require numerical solutions to Eq.5 to determine the
groundwater flow fields in ASR design and evaluation. This can be accomplished using 3-D flow
packages such as MODFLOW from the U.S. Geological Survey. Such numerical modeling tools will be
described in Section 5.0.
ASR planning and assessment commonly relies on fast and less accurate hydrogeological
characterization of the flow systems. The planning focus is often limited to the assessment of hydraulic
properties (e.g., infiltration rate and storage capacity), assessment of chemical fate and transport, and
41
-------
groundwater geochemical compatibility. For simplicity, the governing Eq.5 is solved to yield an
analytical equation for simple calculations. For hydraulics of an ASRwell or trench, the following
equations are widely utilized, which assume an infinite homogenous and isotropic aquifer.
Theiss solution for groundwater pumping or an injection well:
The most widely used Theiss equation specifies that the drawdown from a pumping well or
mounding from injection is positively related to pumping or injection rate (Q), and inversely related to
aquifer transmissivity T(= k ¦ b),
s = — w(—). (6)
4nT \4TtJ v '
In this equation, the well function, W (u), consists of u = r ^Iwhere r is the well radius and
t is the time lapsed after the start of pumping or injection. The well function is often provided in graphic
form in groundwater handbooks or engineering guidelines (e.g., Driscoll, 1986).
The SuperQ Excel spreadsheet model included in this DDS is based on the simple Theiss
equation, and is suitable for rapid assessment in the planning phase. It is noted that the Theiss equation
assumes a negligible regional groundwater flow field compared to the pumping or injection rate. There
are many ASR applications which have multiple wells and are under a significant regional flow field. In
these cases, the groundwater hydraulics of the ASR system is 2-D in nature. Thus, programs using 2-D
flows are necessary in planning assessment. Such programs include the Hantush (1967) model and the
more sophisticated WhAEM2000 (U.S. EPA, 2007) model in the DSS tool box.
Groundwater flow velocity or infiltration rate
Derived from Eq.5, the groundwater flow velocity or infiltration rate (v) is simply given by:
v = nk h2 hl = nki, (7)
x2-x1
where n is the effective aquifer/soil porosity, h2 and are the groundwater heads at locations x2 and x1
along a flow path, respectively, and i is the hydraulic gradient. For vertical infiltration in the vadose
zone, i =1.
The modified Darcy equation in Eq.7 can be used to approximate the groundwater infiltration rate
in the vadose zone. The groundwater flux or infiltration rate from a spreading basin or green
infrastructure (e.g., rain garden) of area A is equal to (nki) A . Eq.7 can also be applied to estimate
injection rate from an infiltration gallery penetrating into the top of an aquifer. In this analysis, the side
wall area of the trench (A) is used and the hydraulic gradient i is averaged from the trench to the edge of
groundwater mounding.
Groundwater storage capacity
Storing aquifers can have a limited storage capacity, particularly if they are shallow, unconfined
aquifers. Many permeable geological formations, such as unconsolidated sands, can have a large variation
of thickness and poor lateral extension, affecting their storage capacity. For example, sand lenses
interbedded with impermeable clay layers can have a limited capacity.
In practice, the overall storage capacity for the target injection formation is estimated during the
planning process, particularly for large ASR operations. The total storage volume is approximately equal
to the total effective porosity in the storage aquifer formation.
42
-------
4.1.2 Simplified fate and transport analysis
One important planning consideration is the hydraulic control of injected water, which can impact
both the injected water recovery rate and the potential for groundwater contamination from residual and
mobilized contaminants. In ASR planning, these considerations can be assessed in three areas.
Hydraulic controls
Hydraulic control over the migration of injected water in an ASR system has several management
implications in ASR planning and assessment. The degree of hydraulic control determines the rate of
recovery and thus ASR efficiency. The hydraulic control will also affect the fate and transport of residual
contaminants in the injected water and the mobilization of contaminants from aquifer materials. Because
of this significance, ASR design and evaluation may demand high confidence when assessing the
hydraulic control and numerical methods, such as MODFLOW, are commonly used. In the planning
phase, limited technical assessment may focus on achieving hydraulic control through a set of
investigative options:
• Hydrogeological assessment. The ASR local and regional hydrogeology data may be collected and
analyzed to assess the suitability of groundwater aquifers for potential ASR operations. For example,
a shallow and self-contained aquifer is often more preferable than a deep potable aquifer under strong
regional flow gradients. The aquifer depth increases ASR capital and operational costs, and a strong
regional flow makes hydraulic control more difficult to achieve.
• Groundwater flow modeling. Groundwater modeling is conducted to quantitatively evaluate hydraulic
control in ASR planning. For this objective, the groundwater system is simplified by applying a 2-D
model instead of a 3-D numerical model. Modeling results are the basis for estimating the rate of
recovery and groundwater flow fields. WhAEM2000 or its visual version are capable groundwater
modeling during the ASR planning phase.
Fate and transport analysis
Contaminant transport during ASR operations is shown in Figure 3 and discussed in Section
2.1.3. The geochemical processes involved in transport include adsorption, ventilation, reaction and
transformation. A generalized governing equation for solute transport in a porous media is given by:
e ^ = 0V2 {D ¦ C) - V(vXiyiZC) - Qs, (8)
where C is the solute concentration in groundwater, 9 is volumetric water content or degrees of soil
saturation, and D is solute dispersion coefficient. On the right side of Eq.8, the first and the second term
describe solute diffusion and advection, respectively, while the third term, Qs. is a lump sum of the solute
reaction, formation, and transformation terms.
Eq.8 is adopted to simulate solute transport in soil and groundwater even for complex
hydrogeological conditions. MT3DMS, a multi-species 3-D transport model included in this DSS, can be
used to simulate the advection, dispersion, and chemical reactions of contaminants in groundwater
systems. These analysis results are suitable for the ASR system's design and evaluation phase.
For planning purposes, less accurate yet computationally-efficient models using analytical or
semi-analytical solutions have been developed and used. Examples include AT123D (Yeh, 1981) and its
updated version, AT123-AT (Burnell et al., 2012). These 1-D or 2-D models were developed by
simplifying Eq.8 under a given set of assumptions. More details on analytical solute transport models are
available in reviews, guides and books (e.g., Dagan, 1987; Anderson and Woessner, 2002). For the ASR
DSS, the AT123D-AT and a related 2-D multi-species reaction model are described in Section 4.2.
Geochemical compatibility analysis
43
-------
Injected water in an ASR operation is geochemically different from the native groundwater,
which can potentially trigger reactions between the injected water and minerals in the hosting aquifer.
Examples include arsenic mobilization and the dissolution of other heavy metals into the aqueous phase.
Mineral precipitation and dissolution may also occur, leading to changes in aquifer hydrological
properties. All of these potential scenarios are assessed during geochemical compatibility analysis.
The geochemical compatibility assessment can take several steps. The EPA UIC program has
guidelines on the ASR geochemical compatibility analysis. The steps in this technical analysis include 1)
Chemical characterization of injected water, native groundwater, and geological formations, 2) Modeling
and analysis of geochemical conditions in ASR operation scenarios, 3) Qualitative and quantitative
assessment of potential geochemical reactions, 4) Development of potential technical solutions through
pre-injection treatment (See Figures 3 and 5) and monitoring program development.
ASR planning may take the first two steps and conduct a qualitative assessment of the
geochemical compatibility. This analysis may hinge on hydrological and geochemical data collected
during ASR DSS Level 1 activities (See Figure 5). Commonly used analysis methods include:
• Geochemical change monitoring and analysis. Groundwater chemistry changes at a given location
during water injection and recovery phases. The changes in major ionic species concentrations are
shown in Piper diagrams, which serve as geochemical evidence to support an operation permit
application at an aquifer recharge site. These types of geochemical analyses are often integrated with
the investigation of potential mineral reactions, providing direct evidence for likely geochemical
changes.
• ASR reference site information. Additionally, geochemical information about other relevant sites can
help the compatibility analysis during the planning phase. Geochemical data from nearby or regional
ASR operations occurring in the same or similar geological formations can help when evaluating the
injected water compatibility and identifying major controlling environmental factors
4.1.3 Arsenic mobilization assessment
The mobilization of arsenic and other heavy metals at ASR facilities is a major concern to the
EPA UIC program due to their detrimental environmental and human health impacts. Natural and
secondary arsenic contamination in groundwater has been reported worldwide and at many ASR sites in
the U.S. (Welch et al., 2000; Neil et al., 2012). This ASR DSS has singled out arsenic as an indicative
contaminant for planning assessment and design/evaluation analysis.
Multivalent arsenic species (As3+, As5+) have different geochemical mobility in groundwater.
Changes in water chemistry (e.g., pH, Eh, TSS) inside and near the injection bubble can lead to the
breakdown of arsenic-bearing minerals and the dissolution of soluble arsenic complexes into the
groundwater. In Appendix A, an extensive technical investigation is described on the specific
geochemical processes and mechanisms of arsenic mobilization from arsenopyrite in reclaimed water and
model wastewaters. These EPA investigation results have been published in Neil et al. (2014; 2012).
In summary, there are several nano- to micro-scale processes controlling arsenic fate and
transport during ASR, and hence their environmental impact. The outcome of these processes will affect
ASR planning, design and operation. Currently there are no established arsenic control guidelines for the
implementation of ASR, in part due to inadequate knowledge of the soil-water interactions and factors
controlling arsenic mobilization (Asano and Cotruvo, 2004). Nonetheless, these new observations on the
geochemical pathways in the As-Fe-S-Cl-N system have implications for the longer term fate and
transport of arsenic in groundwater aquifers. Major geochemical inferences from the EPA investigations
include:
44
-------
• Arsenic mobilization in groundwater is balanced between the oxidative breakdown of host minerals,
such as arsenopyrite, and the precipitation of iron oxides and iron-oxyhydroxides. The latter promotes
the co-precipitation or sorption of soluble arsenic in groundwater.
• Arsenic associated with the stable iron oxide minerals will be trapped as long as the aqueous
environment is favorable for Fe3+ (e.g., oxidative environments). High TOC content in the injected
water can enhance biological activities, creating locally reductive conditions. These reductive
conditions could prevent arsenopyrite oxidative dissolution, but could also potentially lead to the
destabilization of arsenic trapped in iron oxides where present.
• Activation energies for arsenic mobilization in aerobic and anaerobic systems containing sodium
nitrate, sodium chloride, and wastewater samples were experimentally determined. Differences in
activation energies between the systems indicate that the mechanisms controlling arsenopyrite
dissolution and the propensity for arsenic mobilization can vary with dissolved oxygen presence.
These considerations form a basis for developing ASR monitoring programs, modeling arsenic
fate and transport, and determining pretreatment requirements for the injected water. Pretreatment is often
a necessary part of ASR systems (See Figure 1). There is currently a knowledge gap on how water
pretreatment and water withdrawal will affect arsenic mobilization. The EPA investigation on
arsenopyrite-water interactions has revealed the following observations:
1) The difference in water chemistry (pH, Eh, ORP, etc.) between the injected water and native
groundwater can cause arsenic mobilization in groundwater. Pretreatment of the injected water to
reduce this difference can thus minimize adverse reactions.
2) Several geochemical pathways are involved in dissolution and precipitation of arsenic-bearing iron
oxyhydroxides, and thus control arsenic mobility. The processes are facilitated by the presence of
DOM, chloride ions, nitrate, sulphur and oxidants (or ORP) under a given set of pH-Eh conditions.
On the other side, excessive Fe3+ concentrations in groundwater can lead to iron oxyhydroxide
precipitation and enhanced arsenic encapsulation.
3) Biological activities enhanced by DOM can lead to local reductive environmental conditions, which
can prevent arsenopyrite oxidation but may also promote iron oxide and iron oxyhydroxide
dissolution.
4) Injection-withdrawal operation and groundwater cycling can change the environmental conditions in
the ASR formation, thus affecting the arsenic mobility. Predictive modeling of groundwater
hydrology can help with ASR monitoring and with developing injected water pretreatment
requirements.
Therefore, during ASR planning and assessment, geochemical analysis of the injected water and
native groundwater is necessary to estimate the likelihood of arsenic mobilization. Such analysis may
help to assess whether the aquifer is suitable for ASR operation and whether there is a need for above-
ground pretreatment of the injected water. As described in Appendix A, multiple factors, including
aquifer site hydrology, can impact arsenic mobility. Thus, detailed geochemical modeling and
engineering analysis is necessary during the ASR design and evaluation phase. This consideration is
described in Section 5.0.
4.1.4 Data sources and assessment limitations
ASR planning and assessment, as described above, is based on the quantitative analysis of site
hydrology and contaminant fate and transport. For this analysis, the pertinent site-specific data include:
45
-------
• Site hydrology. Published geological and hydrological investigations are available for many parts of
the U.S. Investigations of ASR sites that are in close geographic proximity or have similar
hydrogeologic settings can be a valuable reference for ASR planning and assessment. The ASR DSS
Level 1 provides pertinent weblinks to USGS, EPA, and other data sources. Because of large
variations in ASR site hydrology, site-specific hydrologic data and analyses are preferable. Technical
information and data sources are available from the EPA UIC program7. One of the most important
hydrologic parameters is the soil permeability and aquifer conductivity. Their variations are used to
quantify the hydrological heterogeneity, and are important to ASR planning and assessment.
• Hydrological investigation methods include soil grain size analysis for the vadose zone, and slug tests
and pumping tests for aquifers. Unconfined aquifers are most commonly used for in ASR operations.
Methods the hydrological characterization of unconfined aquifers are well examined and documented
in the literature (e.g., Fetter, 1994; Yang and Gates, 1997; Batu, 1998; and references therein).
• Geochemical characterization of native groundwater and injected water. Chemical compositions of
the native groundwater and injected water are necessary to assess their interactions at an ASR site.
For the areas of analysis described in the preceding section, a set of the geochemical data is necessary
including ORP, pH, Eh, DOM, cation species (particularly Fe3+ and Fe2+), as well as S, P, CI, and N.
These geochemical data can be obtained during conventional sampling of groundwater and reclaimed
water.
• Modeling and quantitative analysis. In the ASR planning phase, quantitative modeling of injected
water infiltration, flow and geochemical reactions in groundwater only has the objective of assessing
the overall groundwater flow, hydraulic control, and adverse geochemical reactions. The results are
limited to determining the feasibility of ASR operations for a given site. Detailed ASR operational
design and evaluation often require more detailed technical and modeling analysis, particularly on site
heterogeneity and geochemical reactions including arsenic mobilization. The technical approach
supporting this analysis is presented in Section 5.0.
4.2 Assessment tools in the ASR DSS
Analytical models (partial differential equations with initial and boundary conditions that
mathematically describe solute transport) can be used to estimate the flux and concentration of dissolved
pollutants in groundwater. These models typically simulate advection, hydrodynamic dispersion, linear
sorption, and first order reactions that affect pollutant fate and transport. A number of solutions have been
presented over the past 50 years for various source terms, initial conditions, and boundary conditions.
When equations have a closed form solution, they are called analytical models. When numerical methods
(e.g. integration) are necessary to calculate the concentrations in the closed form solution, they are called
semi-analytical models.
4.2.1 Hantush (1967) 2-1) Transient Moun ding Excel Spreadsh eet Model8
Analytical equations (partial differential equations with initial and boundary conditions that
mathematically describe groundwater flow) can be used to estimate the magnitude and radius of
groundwater mounding beneath an infiltration basin or dry well, but the accuracy of these results is
limited by simplifying assumptions that are inherent to solving non-linear differential equations. Hantush
(1967) proposed an equation describing the "growth and decay of groundwater mounds in response to
7 http://water.epa.gOv/tvpe/groundwater/uic/aauiferrecharge.cfm#links
8 Modified from Carleton (2010), pp. 22-24
46
-------
uniform percolation." The Hantush equation and similar equations are widely implemented to estimate
water-table mounding beneath septic systems and other similar infiltration structures that can reasonably
be considered steady-state (i.e., infiltration is constant overtime) (Finnemore, 1995). However, few
studies have implemented this equation to the more challenging scenario of transient conditions
(infiltration occurs over a limited duration, then ceases). This scenario is addressed by the numerical
solution, which utilizes Microsoft Excel. The numerical solution is distributed by the USGS and linked to
herein.
Hantush (1967) assumes a water-table aquifer of infinite extent and finite thickness with a
horizontal, impermeable base. The solution also includes the Dupuit assumptions of horizontal flow and a
negligible change of transmissivity with a change in the head. Hantush (1967) solves the general 2-D
groundwater flow equation by assuming boundary conditions that allows the use of a Laplace transform
with respect to time and the Fourier cosine transform with respect to x and then y to derive an integral that
can be solved. The solution that Hantush derived by making these assumptions provides results that
correspond well with similar analytical solutions and with some field measurements.
Finite-difference numerical simulations of groundwater mounding show that vertical anisotropy
can lead to simulated groundwater-mound heights on the order of 15 percent higher than those simulated
using an analytical solution with the assumption that flow is strictly horizontal. In addition, simulations
that include storage in, and delayed yield from, the unsaturated zone result in less groundwater mounding
than that obtained by neglecting the unsaturated zone. Therefore, the height of groundwater mounding is
underestimated by the Hantush equation where vertical anisotropy is present and overestimated where an
unsaturated zone is present.
The Excel spreadsheet developed by the USGS using the Hantush equation calculates the
magnitude of groundwater mounding. The required input values (aquifer thickness, horizontal hydraulic
conductivity, specific yield, basin size, and recharge rate and duration) are straightforward and can be
measured or estimated from published values. This spreadsheet allows users to specify input variables and
generate reasonable, quantified, reproducible estimates of groundwater mounding beneath water
infiltration structures.
4.2.2 SuperQ: 2-D Transient Well Hydraulics Superposition Excel Spreadsheet Model
Based on simplifying assumptions (e.g., uniform aquifer transmissivity and storage coefficient)
and superposition theory, analytical solutions for well hydraulics equations, such as the Theis (1935)
equation, that are implemented in aquifer test analysis software programs (e.g., Aqtesolv. AquiferWin32.
and AquifcrTcstPro) can be used to forward calculate the time-dependent drawdown (and mounding)
caused by extraction and injection from multiple wells with time-varying pumping rates, considering
barrier and recharge boundary effects. These commercial software programs can thus be used to model
aquifer storage and recovery operations and evaluate related design and system performance issues.
Free software developed to calculate hydraulic head changes caused by multiple wells with
variable pumping rates and aquifer boundaries based on the Theis (1935) or other well hydraulics
equations were not found during recent searches. Thus, Tetra Tech developed an Excel spreadsheet
program named SuperQ that can be used to model the 2-D transient effects of aquifer storage and
recovery pumping involving multiple wells, time-varying pumping rates, and boundary conditions
represented using image well theory (Ferris et al., 1962). Two analytical solutions are currently included
in SuperQ: the Theis (1935) solution for transient flow to a fully-penetrating well in a uniform confined
aquifer with no leakage and the Hantush and Jacob (1955) solution for transient flow to a fully
47
-------
penetrating well in a uniform leaky confined aquifer. Drawdown calculations are made using functions
programmed by Hunt (2005).
The SuperQ program utilizes an Excel spreadsheet to document user inputs. The user specifies
the aquifer transmissivity and storage coefficient, aquitard properties (thickness and vertical hydraulic
conductivity), pumping well information (well locations, pumping periods, pumping rates), initial
hydraulic head information, head change calculations, times and locations, and boundary conditions.
Calculated drawdown and head values are reported in table format, which can be exported to contouring
programs such as Surfer, in Excel hydrographs, and as areal drawdown and head plots using Excel's
'clunky' surface graphing utility. An output report is also written. Compiled values of storage coefficients
and transmissivities for different types of geologic media are provided for reference. Different
spreadsheet tabs are used to input program data, report output data, and provide reference hydraulic
property data.
SuperQ has been tested by comparing results for various simulation scenarios with the Aqtesolv
program. Simulation results match between the two programs. Note, however, that SuperQ is undergoing
further development and testing. Not all boundary condition options have been incorporated into SuperQ,
Additional error checking will also be added into the program (for example, to prevent input of pumping
wells on the wrong side of boundaries, which could add erroneous pumping in the pertinent model
domain). An updated version of SuperQ will be provided within the next few months.
4.2.3 2-D model: WhAEM2000 version 3.2.19
WhAEM2000 (U.S.EPA, 2007) is a public domain AEM code that simulates 2-D steady flow
caused by pumping wells, hydrologic boundaries (river, recharge, and no-flow conditions), and
inhomogeneous zones. It incorporates an easy-to-use graphical user interface (GUI). Base maps for a
project can be selected from a graphical index map for the State on an EPA webserver or input from
others sources and file types (including shp, dwg, dxf, jpg, bmp, sid, tiff and other formats). Program
operation and modeling practice is documented in the EPA report "Working with WhAEM2000" using
Vincennes, Indiana as a case study (U.S.EPA, 2007). Frequently asked questions are addressed by
Kraemer (2005).
The AEM method, which is described in detail by Strack (1989) and Haitjema (1995), avoids the
discretization of a groundwater flow domain by grids or element networks. Instead, only the surface water
features in the domain are discretized, broken up in sections (usually a few hundred), and entered into the
model as input data. Each of these stream sections or lake sections is represented by a closed form
analytic solution: the analytic element. The comprehensive solution to a complex, regional groundwater
flow problem is then obtained by the superposition of all analytic elements in the model. Traditionally,
the superposition of analytic functions was considered to be limited to homogeneous aquifers with a
constant transmissivity. However, by formulating the groundwater flow problem in terms of appropriately
chosen discharge potentials, rather than piezometric heads, the analytic element method becomes
applicable to both confined and unconfined flow conditions, as well as to heterogeneous aquifers. The
analytic elements are chosen to best represent certain hydrologic features. For instance, stream sections
and lake boundaries are represented by line sinks, while small lakes or wetlands may be represented by
areal sink distributions. Areal recharge is modeled by areal source distributions (areal sinks with a
negative strength). Streams and lakes that are not fully connected to the aquifer are modeled by line sinks
9 Modified from U.S. EPA WhAEM 2000 website https://www.epa.gov/exposure-assessment-models/whaem2000.
48
-------
or area sinks with a bottom resistance. Discontinuities in aquifer thickness or hydraulic conductivity are
modeled by using line doublets (double layers). Specialized analytic elements may be used for special
features, such as drains, cracks, slurry walls, etc.
WhAEM2000 can be used to simulate steady-state 2-D hydraulic head contours, streamlines, and
particle tracks associated with aquifer storage and recovery operations. Visual AEM is another free AEM
program, while commercial AEM groundwater modeling software are also available.
4.2.4 2-D model: Visual AEM released in February 200910
Visual AEM is freeware authored by James R. Craig and Shawn Matott of the University of
Waterloo. It provides a graphical user interface (GUI) for single and multi-layer analytic element
modeling of (mostly) steady-state groundwater flow and numerical/analytical modeling of vertically-
averaged contaminant transport with the object-oriented codes Bluebird and Cardinal, the multilayer code
TimML, and the public domain AEM code Split.
Visual AEM is designed for simplicity of use, but includes many robust tools and methods for
developing regional scale groundwater flow and transport models or local scale models nested in larger
hydrogeologic domains. Version 1.04 (released in 2009) supports: regional and local scale flow and
transport modeling; simulation of flow in confined and unconfined single-layer or multi-layer aquifers;
capture zone delineation; particle tracking; analytical, finite element, and finite difference simulation of
multi-species solute transport; plume animation and visualization; automated calibration; advanced output
options (ArcView®, ArcMap®, Surfer®); multiple basemaps and DEMs in various formats (.jpg, .gif,
.bmp, .ddf, .grd, .bna, .dxf); pre- and post-processing in Surfer®; grid and mesh generation and editing
tools; and an editable geologic media database.
Visual AEM features include: 2-D single-layer or vertical groundwater flow modeling (steady-
state or with Theis wells); quasi-3-D multi-layer groundwater flow modeling (steady-state); 2-D finite
element and finite difference multispecies contaminant transport modeling; 2-D analytical contaminant
transport modeling; particle tracking; parameter estimation; and transect/cross-sectional analysis. The
program facilitates modeling of 2-D single and multi-layer steady state groundwater flow using a wide
range of hydrogeologic features, including rivers, drains, and lakes, which can be specified based on:
head, resistance, or extraction rate; recharge/leakage zones (circular or polygonal); vertical and horizontal
wells; inhomogeneities (polygonal, circular, or elliptical) in aquifer base, thickness, porosity, hydraulic
conductivity; slurry walls; fractures; and flux-specified and no-flow boundaries. The transport features of
Visual AEM include advection, dispersion, diffusion, sorption, first-order decay of multiple solutes, and
source zones (with specified dissolved and sorbed initial conditions).
4.2.5 Fate and transport of residual contaminants in injected water
4.2.5.1 2-D model: AT123D-AT
Analytical models (partial differential equations with initial and boundary conditions that
mathematically describe solute transport) can be used to estimate the flux and concentration of dissolved
pollutants in groundwater. These models typically simulate advection, hydrodynamic dispersion, linear
sorption, and first order reactions that affect pollutant fate and transport. A number of solutions have been
10 Modified from Visual AEM Homepage
49
-------
presented over the past 50 years for various source terms, initial conditions, and boundary conditions.
When equations have a closed form solution, they are called analytical models. When numerical methods
(e.g. integration) are necessary to calculate concentrations in the closed form solution, they are called
semi-analytical models.
Developed in 1981, the AT123D program consists of semi-analytical solutions for 1-, 2-, or 3-D
dissolved pollutant transport in a homogeneous aquifer subject to a uniform, stationary regional flow field
(Yeh, 1981). AT123D-AT (Burnell et al., 2012) updates this model to include modern programming
methods, correct errors in the pulse source, improve numerical integration solvers, include the new
Green's functions for a finite-depth aquifer, dynamically allocate arrays for large numbers of nodes, and
improve computational efficiency. In recent years, the model has been widely utilized to estimate
dissolved chemical concentrations at receptor wells for use in risk assessments. AT123D-AT is a flexible
semi-analytical model capable of simulating the advective-dispersive transport of pollutants from a wide
variety of source configurations in aquifers bounded by various boundary conditions. The analytical
solutions to the governing equations are based on the use of Green's functions. This approach combines
the product of point and/or integrated line source solutions in each of the three principle directions to
solve advective-dispersive transport from point, line, planar, or rectangular sources. The modeled aquifer
is infinite in the horizontal direction of flow but can be approximated as finite or infinite in the transverse-
horizontal and vertical directions. The source considered may be instantaneous, steady (finite or infinite in
length), or variable over time. The strength of the model lies in the user's ability to choose various source
and aquifer configurations and the time-variant source option. This flexibility makes the model ideally
suited for coupling with unsaturated zone transport models.
A 1-D finite difference model is included in the ASR DSS to simulate unsaturated advection-
dominated transport. The unsaturated zone software simulates 1-D transport to the water table from
infiltration basins in the unsaturated zone. The unsaturated zone model is a 1-D finite difference code that
represents the source as a Dirichlet boundary (specified concentration). The model simulates advection-
dominated transport and sorption in the unsaturated zone. The model is flexible in that it allows the user
to specify initial concentrations in the unsaturated zone based on available field data. The calculated mass
flux at the water table from this unsaturated zone model is the input for the source zone in the AT123D-
AT model.
4.2.5.2 2-D model: Analytical multi-species sequential first-order reaction model
This 1-D, 2-D, or 3-D multi-species model is a useful risk-based screening model to examine not
only the parent compounds but also their potentially more toxic degradation products at downgradient
groundwater receptor locations. This analytical solute transport model assumes a continuous point, line,
or vertical planar source and simulates advection, linear sorption, and first-order sequential reaction. The
model simulates the steady-state plume extent of up to 5 sequentially degrading dissolved chemicals (e.g.
pesticides or chlorinated solvents) in groundwater. The 1-D model uses a point source and includes
longitudinal hydrodynamic dispersion. The 2-D and 3-D models use a line or vertical planar source with
advection-dominated transport and transverse horizontal and/or vertical dispersion.
Given particular assumptions, the exact analytical solutions in the multi-species model are not
subject to significant errors (Burnell et al., 2012, West et al., 2007). This modeling tool includes an
estimate of the time for the plume to reach steady-state and estimates the spatial moments (center of mass,
and plume spread around the center of mass) of the steady-state plume.
A MS Windows-based GUI has been developed for the DSS for easy input of model parameters
and post-processing. The required input values (source concentration, source width and depth,
50
-------
groundwater velocity, longitudinal and transverse dispersion coefficients, retardation factor, and rate
constants) are straightforward and can be measured or estimated from published values. Once the user
enters these input parameters, the model run button rapidly calculates the steady-state plume
concentrations. The 1-D model results of each chemical are presented as concentration vs. distance along
the plume centerline. For 2-D visualization of the model results, the model post-processing allows the
user to overlays plume contours of each chemical over a user-specified base map.
5.0 ASR Evaluation and Engineering Design
ASR design and performance evaluation are essential for managers and engineers to achieve
water resource management objectives. Following the ASR planning phase, engineering design and
evaluation are the next step to determine the hydraulic control and water quality changes in an ASR
system (See Figure 5). Important design and evaluation variables include hydraulic control, residence
time, and water quality constraints.
5.1 ASR Hydraulic Properties and Hydrologic Control
As noted in previous sections, ASR system types vary from place to place. Some are only limited to
aquifer recharge or soil infiltration in the vadose zone. For this application, groundwater vertical
infiltration rate is the most important design variable. On the other hand, injection-withdrawal ASR
operations have the function of making up water shortages or storing excess water. Thus, the objective of
these operations will depend on several engineering factors including the rate of recovery, the steady state
capture zone size, and the minimum groundwater residence time.
5.1.1 Particle tracking, capture zone and rate of ASR recovery
Capture Zone
Capture zone of an ASR system is defined as the region which is hydraulically influenced by the
groundwater withdrawal wells. For complete hydraulic control of the injected water volume, the capture
zone size must be large enough to encompass the injected water and its flow path over time during ASR
operation. Technically, the hydraulic control is evaluated under steady state conditions.
The capture zone size is a function of ASR site hydrology and the hydraulic design of the water
injection-recovery system. Factors affecting the efficiency include aquifer properties (leakiness,
transmissivity, groundwater flow velocity), well design and operation (e.g., capture zone, injection and
withdrawal rates, well conditions), and the duration of water storage. A detailed hydrological
investigation, the basis for proper design and evaluation, often involves modeling the groundwater flow
fields under various ASR operation conditions.
The capture zone theory was first developed in the 1990s for hydrological application to
groundwater pump-and-treat remediation. Mathematical modeling of capture zone size has been outlined
in the literature and several numerical and analytical methods are available. Analytical capture zone
models (Yang et al., 1997; Matott et al., 2008; U.S. EPA, 2007, 2008) can be used to determine the
location and capture size in relatively simple groundwater systems. These 2-D analytical solutions can be
used to approximate the composite capture zone of multiple groundwater pumping wells using the
hydraulic superposition principles of groundwater flow. One of these methods is WhAEM2000 (U. S.
EPA, 2007), a public domain AEM code that simulates 2-D steady flow caused by pumping wells,
hydrologic boundaries (river, recharge, and no-flow conditions), and inhomogeneous zones. The
applicability of this model has been verified in practice.
51
-------
For more complex groundwater flow systems or injection-recovery system configurations, robust
and computationally intensive numerical modeling packages are more appropriate than analytical
solutions. The MODFLOW numerical model developed by USGS is a widely used modeling tool and has
several extensions available. It is briefly described in subsequent sections, while more detailed technical
documentation is available from the USGS.11
Rate of Recovery
The rate of recovery depends on ASR site hydrology and the hydraulic design of the water
injection-recovery system. Factors affecting the efficiency include aquifer properties (leakiness,
transmissivity, groundwater flow velocity), well design and operation (e.g., capture zone, injection and
withdrawal rates, well conditions), and the duration of water storage. In all cases, a detailed hydrological
investigation is needed for proper design and evaluation. This investigation often involves computer
modeling of the groundwater flow fields
under various ASR operation conditions.
One utility of the hydrological module
of the ASR DSS is to calculate the
groundwater capture zone and track particles
in the groundwater (Figure 6). For ASR
operation, the injection-withdrawal induced
local groundwater flow is superimposed upon
a regional flow field. The composite flow
field is then the basis to determine the rate of
recovery and residence time of injected water
in the aquifer undergoing ASR.
Transient Particle Tracking
Transient particle tracking is the other
important quantitative method for ASR design
and evaluation. Based on the solution to the
governing Eq. 1, the particle position {x, y, zj
in the aquifer is computed at each time step.
Post-processing of the simulation data
produces flow pathways at each time interval
for injected water in the injection-recovery
well pairs. An example is shown in Figure 17.
One important objective of transient
particle tracking is to calculate the volume-weighted average residence time and range of injected water
in the storage. The residence time of injected water in an aquifer is one important criterion in ASR design
to ensure biological integrity in recovered water. Many states have adopted a minimum residence time,
including Texas, California, and Florida. Particle tracking results, such as those in Figure 17, help when
developing ASR configuration and operation standards. When necessary, testing well and groundwater
tracer tests are often prescribed. An example is given in Section 5.1.1.2.
The second objective of groundwater particle tracking is to investigate the capture zone under
ASR operational scenarios for sites with complex hydrogeological conditions. Capture zone can be
\
Figure 17 Example of particle tracking from an
injection well (top well in the figure) to a
recovery well (bottom well in figure). Head
contours of groundwater elevations are marked.
The simulation was carried out using MODFLOW.
11 http://water.usgs.gov/ogw/modflow/MODFLOW.html
52
-------
estimated by using the analytical solutions, as discussed earlier in this section. Transient particle tracking
enables more accurate of the groundwater capture by the ASR pumping wells
5.1.2 USGS MODFLOW Transient Numerical Groundwater Flow Model Code and MODPATH
Transient Particle Tracking Code
MODFLOW is the USGS's 3-D finite-difference groundwater model. MODFLOW is considered
an international standard for simulating and predicting groundwater conditions and groundwater-surface
water interactions.
Originally developed and released in 1984 solely as a groundwater-flow simulator, MODFLOW's
modular structure has provided a robust framework for the integration of additional simulation
capabilities, which build on and enhance its original scope. The family of MODFLOW-related programs
now includes programs which can simulate coupled ground- and surface water systems, solute transport,
variable-density flow (including saltwater), aquifer system compaction and land subsidence, parameter
estimation, and groundwater management. Many new capabilities have been added to the original model.
MODFLOW-2005 (v. 1.11.00), the most current release of MODFLOW, is the most stable and well-tested
version of the code.
MODFLOW-2005 simulates steady and transient (non-steady) groundwater flow in aquifer
layers, which can be confined, unconfined, or a combination of confined and unconfined, while aquitards
restrict groundwater flow. This model can be applied to simulate flow to wells, areal recharge,
evapotranspiration, flow to drains, and flow through river beds. The hydraulic conductivity may differ
spatially and anisotropically, and the storage coefficient may be heterogeneous. Specified head and
specified flux boundaries can be simulated as a head dependent flux across the model's outer boundary.
This allows water to be supplied to a boundary block in the modeled area at a rate proportional to the
current head difference between a "source" of water outside the modeled area and the boundary block.
MODPATH is a particle-tracking post-processing model that computes 3-D flow paths using
output from groundwater flow simulations based on MODFLOW, the USGS finite-difference
groundwater flow model. The program uses a semi-analytical particle-tracking scheme that allows an
analytical expression of a particle's flow path to be obtained within each finite-difference grid cell. A
particle's path is computed by tracking the particle from one cell to the next until it reaches a boundary, an
internal sink/source, or satisfies another termination criterion.
Data input to MODPATH consists of a combination of MODFLOW input data files, MODFLOW
head and flow output files, and other input files specific to MODPATH. Output from MODPATH
consists of several output files, including a number of particle coordinate output files intended to serve as
input data for other programs that process, analyze, and display the results in various ways.
5.1.3 Particle Tracking Example
Model Construction
A hypothetical example of flow field prediction and particle tracking analysis is illustrated for ASR
design. In this exercise, the ASR DSS was applied to examine the effects of an injection and extraction
well design configuration assuming no net change in groundwater storage in a hypothetical aquifer
undergoing ASR with typical hydraulic properties. The USGS groundwater flow model, MODFLOW,
was applied to examine transient changes in groundwater levels over time and the effect of the spacing
between an injection well (200 gpm) and extraction well (200 gpm) that operate over time. Based on the
simulated groundwater flow field from MODFLOW, the USGS groundwater particle tracking model,
MODPATH, was then applied to examine changes in 3-D groundwater particle pathways overtime.
53
-------
The model area was one square
miles. The sandy aquifer was assumed to
have a saturated thickness of 100 feet and
was divided into 3 layers of equal
thickness (33.3 ft) consisting of fine- or
coarse-grained sand (Figure 18). All three
layers were assumed to be anisotropic
with a horizontal and vertical conductivity
k I
ratio ( vj , ) of 0.1, and a uniform
K-H
storage coefficient of 0.0001. The kH
values were 100, 30, and 50 feet/day for
layers 1, 2, and 3, respectively, allowing
for differential groundwater flow
velocities in the well field.
3-D groundwater flow from the
injection well toward the extraction well
was simulated. Model grid spacing varied
from 12.5 feet near the injection well to
50 feet at the outer regions of the model
domain. The injection well and extraction well were assumed to operate during alternate months, with
initial recharge of treated water and then extraction of groundwater every other month. The following
ASR well arrangement scenarios were investigated:
Base Scenario: One injection well; no recovery well;
Paired injection-recovery well scenarios: Well spacing at 400, 800 and 1400 feet.
Model Results
The effects of well spacing on the recovery of treated recharge water were evaluated for different
well designs. The groundwater flow model results were presented using both spatial maps and time series
plots of groundwater elevations overtime in both the injection and extraction wells.
The injection of water into the stratified aquifer causes groundwater mounding near the well and
differential groundwater flow in the three layers of the groundwater aquifer, which have different
hydraulic conductivities. The particle trajectory is shown in Figure 19. Regional groundwater flows from
the left to the right of the cross section. Groundwater mounding at the injection well dissipates away from
this point. The differential response of the layers to water injection is apparent in the spacing of flow
vectors at each time step. The unit-time flow vector is the most spaced in the first layer, which has the
highest conductivity of kH = 10 Oft/day, in contrast to the closely spaced flow vectors for the middle
layer, which has the lowest conductivity of kH = 3 Oft/day. This difference indicates that the least
permeable layer to the water injection can cause hydraulic resistance. For the same reason, the injection
well loses its efficiency when boreholes are plugged by biological growth or precipitation.
Figure 20 shows the spatial map of the hydraulic head distribution for the paired well
configurations at three well distances. Also shown is the particle tracking among the design scenarios. For
the comparison, all aquifer properties, well designs, and operations remained the same among the three
scenarios except for spacing of the recovery wells. In these scenarios, there are limited changes in the
groundwater elevation near the ASR injection and extraction wells because the cone of depression from
pumping is limited by recharge from the injection well.
Top elevation = 100 ft
Layer 1, KH = 100 ft/day
> Layer 2, KH = 30 ft/day
Layer 3, KH = 50 ft/day
Bottom elevation = 0 ft
Figure 18 Hypothetical example of a three-layer sandy
aquifer in 3-D groundwater flow modeling for ASR
system design using the ASR DSS models.
54
-------
Cross-Section Along Row 55
kH=100 ft/day
¦ - - - -
-> -» -> -
> -> -> -> -> ->
kH=30 ft/day
- - - - *-
esttSrsC -
>,r attsess
-> -> -> -> -> -
ktf=50 ft/day
Figure 19 Particle tracking in profile across the injection well, with injection occurring across the
entire well depth. Flow vector at each time step is shown at the middle in each of the three layers.
When the injection well and extraction well are closely spaced (<800 feet apart), the results of
particle tracking indicate that there will be full recovery of recharge water particles by the recovery well.
On the other hand, when the injection and extraction wells are far apart (1400 ft), recharge water may not
be recovered by the extraction well. This poor recovery is clearly shown in Figure 20.
During the ASR cycling operations, the aquifer dewatering effects are worth noting for
unconfined aquifers. Groundwater injection can raise the groundwater table, while pumping by recovery
wells dewaters the aquifer. Frequently, an ASR system is operated in sequential injection-storage-
withdrawal modes. This operation leads to fluctuation of the groundwater level in the aquifer.
For the pair of ASR wells at 800 feet spacing (Figure 20), for example, the groundwater level was
calculated using MODFLOW simulation. Figure 21 shows the time series plot of the water level at the
recovery well. For monthly cycling of the injection-recovery operation, the level will fluctuate by
approximately 4 feet with no net change in groundwater storage. The groundwater fluctuation in other
location varies. However, these simulations indicate that the aquifer zone will experience cyclic changes
in redox potential and other environmental conditions. These changes have implications for the
mobilization of arsenic and other redox-sensitive contaminants.
Implications
This simple example is only intended for DSS illustration. Aquifer hydrology and ASR
configurations will be more complicated in the field. Nonetheless, this example illustrates how the DSS
can assist in ASR design and evaluation to determine important parameters, such as the hydraulic control,
rate of recovery, and water table fluctuations.
55
-------
Q=200 gpm
Spacing = 800 feet
in well
Q =200 gpm
Spacing = 400 feet
Injection well
Recovery w/ell
Recoven
s
a
%
Q =200 gpm
Spacing = 1400 feet
Injection well
Recovery wel
Figure 20 Particle tracking for three design scenarios of two paired injection-recovery wells, with
the injection well remaining stationary and the recovery well moving farther away. Physical
model is shown in Figure 18,
5.2 Fate and Transport of Residual Contaminants in Injected Water
In an ASR operation, the residual contaminants in injected water are transported and transformed
as the water flows in the vadose zone and in the saturated groundwater aquifer. These processes are
schematically shown in Figure 3. For ASR design and evaluation, detailed quantitative model simulation
can help evaluate contaminant concentrations and determine pre-treatment requirements.
Water quality impact analysis can be conducted using analytical fate and transport models for
simple hydrologic settings. The models included in the ASR DSS include AT123D, the 2-D multi-species
fate and transport model, WhAEM2000, and Visual AEM. The latter two analytical models also
incorporate finite element and finite difference numerical schemes for 2-D multispecies contaminant
transport and particle tracking. These analytical models are described in Section 4.2.
Two numerical models, MT3DMS and SEAWAT, are described here for complex
hydrogeological settings and ASR configurations. Both models were developed by USGS and are based
56
-------
88
78
l tnu February March April May June July July August September October November Phc etfibpr
Month
Figure 21 Computer-simulated well head at the recovery well in the well distance of 800 feet in
the pair well design scenario. Nearly 10 feet of groundwater fluctuation is predicted for the
monthly injection-recovery ASR operation.
on the fundamental groundwater flow and mass transport equations in Eq.5 and 8, respectively. Recently,
these models have been incorporated into the groundwater modeling platform PHAST.
5.2.1 MT3DMSfor multi-species transport in groundwater systems
MT3DMS is a new version of the Modular 3-D Transport model, where MS denotes the multi-
species structure for accommodating add-on reaction packages. MT3DMS has a comprehensive set of
options and capabilities for simulating advection, dispersion/diffusion, and chemical reactions of
contaminants in groundwater flow systems under general hydrogeological conditions.
MT3DMS is unique in that it includes three major classes of transport solution techniques in a
single code: 1) The standard finite difference method; 2) The particle-tracking-based Eulerian-Lagrangian
methods; and 3) The higher-order finite-volume TVD method. Since no single numerical technique has
been shown to be effective for all transport conditions, the combination of these solution techniques, each
having its own strengths and limitations, is believed to offer the best approach for solving the most wide-
ranging transport problems with desired efficiency and accuracy.
MT3DMS can be used to simulate changes in concentrations of miscible contaminants in
groundwater considering advection, dispersion, diffusion and some basic chemical reactions, with various
types of boundary conditions and external sources or sinks. The chemical reactions included in the model
are equilibrium-controlled or rate-limited linear or non-linear sorption, and first-order irreversible or
reversible kinetic reactions. It should be noted that the basic chemical reaction package included in
MT3DMS is intended for single-species systems. An add-on reaction package such as RT3D or SEAM3D
must be used to model more sophisticated, multi-species reactions.
57
-------
MT3DMS can accommodate very general spatial discretization schemes and transport boundary
conditions, including: 1) Confined, unconfined or variably confined/unconfined aquifer layers; 2) Inclined
model layers and variable cell thickness within the same layer; 3) Specified concentration or mass flux
boundaries; and 4) The solute transport effects of external hydraulic sources and sinks such as wells,
drains, rivers, areal recharge and evapotranspiration.
5.2.2 SEAWATfor three-dimensional variable-density groundwater flow and transport
The SEAWAT program is a coupled version of MODFLOW and MT3DMS designed to simulate
three dimensional, variable-density, saturated groundwater flow. Flexible equations were added to the
program to allow fluid density to be calculated as a function of one or more MT3DMS species. Fluid
density may also be calculated as a function of fluid pressure. The effect of fluid viscosity variations on
groundwater flow was included as an option. Fluid viscosity can be calculated as a function of one or
more MT3DMS species, and the program includes additional functions for representing the dependence
on temperature. Although MT3DMS and SEAWAT are not explicitly designed to simulate heat transport,
temperature can be simulated as one of the species by entering appropriate transport coefficients. For
example, the process of heat conduction is mathematically analogous to Fickian diffusion. Heat
conduction can be represented in SEAWAT by assigning a thermal diffusivity for the temperature species
(instead of a molecular diffusion coefficient for a solute species). Heat exchange with the solid matrix can
be treated in a similar manner by using the mathematically equivalent process of solute sorption. By
combining flexible equations for fluid density and viscosity with multi-species transport, SEAWAT
Version 4 represents variable-density groundwater flow coupled with multi-species solute and heat
transport. SEAWAT Version 4 is based on MODFLOW-2000 and MT3DMS and retains all of the
functionality of SEAWAT-2000.
SEAWAT Version 4 also supports new simulation options for coupling flow and transport, and
for representing constant-head boundaries. In previous versions of SEAWAT, the flow equation was
solved for every transport time step, regardless of whether or not there was a large change in fluid density.
A new option was implemented in SEAWAT Version 4 that allows users to control how often the flow
field is updated. New options were also implemented for representing constant-head boundaries with the
Time-Variant Constant-Head Package. These options allow for increased flexibility when using constant-
head flow boundaries with the zero-dispersive flux solute boundaries implemented by MT3DMS at
constant-head cells.
The report contains revised input instructions for the MT3DMS Dispersion Package, Variable-
Density Flow Package, Viscosity Package, and Constant-Head Package. The report concludes with seven
cases of an example problem designed to highlight many of the new features.
5.3 Geochemical Compatibility and Water Quality Changes
5.3.1 Arsenic mobilization from aquifer materials
One aspect of geochemical compatibility which must be incorporated into ASR models is arsenic
mobilization from aquifer materials. Section 4.1.3 outlined the preliminary considerations for arsenic
mobilization assessment. Appendix A provides a more detailed summary of arsenic mobilization
mechanisms and ASR processes. These assessments together indicate several geochemical conditions
which control arsenic mobilization into groundwater. The redox environment is one of the primary
factors. Redox cycling of iron regulates the fate and transport of many elements of concern due to the
formation of iron oxyhydroxides, which can act as a powerful sorbent for aqueous contaminants. During
58
-------
ASR, the redox potential of groundwater environments can change, leading to the oxidative dissolution of
reduced iron minerals such as arsenopyrite. In reaching a geochemical steady-state condition, arsenic
released into groundwater is counter-balanced by attenuation processes. Attenuation processes include co-
precipitation with or sorption by iron oxyhydroxides and the precipitation of arsenic-containing minerals.
Based on these results, the geochemical pathways leading to arsenic mobilization and attenuation will
depend on the following major geochemical conditions at the ASR site.
• The difference in water chemistry (pH, Eh, ORP, etc.) between injected water and native groundwater
can cause of arsenic mobilization in groundwater. Pretreatment of injected water to reduce the water
chemistry differences can thus minimize adverse reactions.
• Several geochemical pathways are involved of the dissolution and precipitation of arsenic-bearing
iron oxyhydroxides, influencing arsenic mobility. These processes are facilitated by the presence of
DOM, chloride ions, nitrate, sulphur and oxidants (or ORP) under a given pH-Eh condition.
• Biological activities enhanced by DOM can lead to local reductive environmental conditions,
preventing arsenopyrite oxidative dissolution and inhibiting iron oxyhydroxide formation. Although
this effect is not currently quantified for ASR sites, DOM presence and concentration variation in
space can be used as an indicator.
Controlling arsenic mobilization will thus involve the following major steps: 1) Geochemical and
mineralogical analysis of arsenic-bearing minerals in the native groundwater aquifer; 2) Model projection
of groundwater conditions in the injection bubble and mixing zones; and 3) Pre-treatment specification
for injected water. The purpose of these investigations is to reduce arsenic mobilization and promote
arsenic co-precipitation with iron oxides and oxyhydroxides. Alternatively, arsenic mobilization can be
minimized by chemically conditioning injected water to resemble the composition of native groundwater.
The applicability of this treatment is case specific, and largely depends on the engineering and operational
economics.
5.3.2 Geochemical simulation using PHREEQC andPHREEQCI
In the ASR DSS, the geochemical package PHREEQC Version 3 (Parkhurst and Appelo, 2013)
and PHREEQCI, a GUI to PHREEQC, are available from the USGS. PHREEQC implements several
types of aqueous models to perform a wide variety of aqueous geochemical calculations: 1) Speciation
and saturation-index calculations; 2) Batch-reaction and 1-D transport calculations with reversible and
irreversible reactions, which include aqueous, mineral, gas, solid-solution, surface-complexation, and ion-
exchange equilibria, and specified mole transfers of reactants, kinetically-controlled reactions, mixing of
solutions, and pressure and temperature changes; and 3) Inverse modeling, which finds sets of mineral
and gas mole transfers that account for differences in composition between waters within specified
compositional uncertainty limits
A 1-D transport algorithm in PHREEQC can be used to simulate dispersion and diffusion, solute
movement in dual porosity media, and multicomponent diffusion, where species have individual,
temperature-dependent diffusion coefficients, but ion fluxes are modified to maintain charge balance
during transport. The inverse modeling capability of PHREEQC allows identification of reactions that
account for observed water composition along a flowline or with time during an experiment. Sorption and
desorption can be modeled as surface complexation reactions or as (charge neutral) ion exchange
reactions. It has two models for surface complexation: 1) A model based on the Dzombak and Morel
(1990) database for the complexation of heavy metal ions on hydrous ferric oxide (Hfo, referred to as
ferrihydrite), and 2) The CD-MUSIC model, which also allows for multiple binding sites on each surface.
The CD-MUSIC model has more options to fit experiment data and was developed for sorption on
59
-------
goethite. Ion exchange can be modeled using several conventions. Kinetically-controlled reactions can
also be modeled.
PHREEQC is used to evaluate geochemical processes associated with mine drainage, radioactive-
waste isolation, contaminant migration, natural and engineered aquifer remediation, aquifer storage and
recovery, water treatment, natural systems, and laboratory experiments. More technical information and
application examples can be found on the USGS website.
5.3.3 3-D Modeling Tool -PHAST12
The ASR DSS contains a more integrated program PHAST (PHREEQC and HST3D) for
simulation of multicomponent, reactive solute transport in three-dimensional saturated groundwater flow
systems. PHAST can model a wide range of equilibrium and kinetic geochemical reactions. The flow and
transport calculations are based on a modified version of HST3D and geochemical reactions are simulated
using PHREEQC, which is embedded in PHAST. PHAST and Phast4Windows are available at the
USGS.
PHAST is applicable to the study of natural and contaminated groundwater systems at a variety
of scales ranging from laboratory experiments to local and regional field scales. PHAST can be used in
studies of migration of nutrients, inorganic and organic contaminants, and radionuclides; in projects such
as aquifer storage and recovery or engineered remediation; and in investigations of the natural rock/water
interactions in aquifers. PHAST is not appropriate for unsaturated-zone flow, multiphase flow, or density-
dependent flow.
A variety of boundary conditions can be accommodated in PHAST to simulate flow and
transport, including specified- head, flux (specified-flux), and leaky (head-dependent) conditions, as well
as the special cases of rivers, drains, and wells. Chemical reactions in PHAST include: 1) Homogeneous
equilibria using an ion-association or Pitzer specific interaction thermodynamic model; 2) Heterogeneous
equilibria between the aqueous solution and minerals, ion exchange sites, surface complexation sites,
solid solutions, and gases; and 3) Kinetic reactions with rates that are a function of solution composition.
The aqueous model (elements, chemical reactions, and equilibrium constants), minerals, exchangers,
surfaces, gases, kinetic reactants, and rate expressions may be defined or modified by the user.
The PHAST simulator may require large amounts of memory and long Central Processing Unit
(CPU) times. To reduce the long CPU times, a parallel version of PHAST has been developed that runs
on a multiprocessor computer or on a collection of computers that are networked. Only the flow and
transport file is described in detail in the PHAST documentation report. The other two files, the chemistry
data file and the database file, are identical to PHREEQC files, and a detailed description of these files is
in the PHREEQC documentation. ModelMuse and Phast4Windows are Windows GUIs for PHAST.
5.4 Simulation example: arsenic transport in long-term aquifer storage
5.4.1 Model Background in PHASTVMODELMuse simulation
This example reproduces the PHAST example 4 (Parkhurst et al., 2010) for a reactive-transport
model to simulate the water composition evolution over a geologic time frame in the Central Oklahoma
aquifer. It is assumed that the aquifer initially contained brines similar to those found at a similar depth in
the area. Aquifer recharge with a constant influx of freshwater from precipitation is considered to be the
12 Modified from Parkhurst et al. (2010).
60
-------
mechanism for chemical evolution in the aquifer over a geologic time frame. This hypothesis is
examined in the case study. A similar approach can be used for the simulation of arsenic fate and
transport during ASR injection and recovery over the operational time frame.
In the Oklahoma aquifer example, the PHAST/ModelMuse package in the DSS was used to
simulate flow, transport, and reactions at a regional scale (90 km [kilometers] by 48 km) for an aquifer
with both confined and unconfined regions and a complex 3-D flow pattern. The conceptual model for the
calculation assumes that brines initially filled the aquifer. The aquifer contains calcite, dolomite, and
clays with cation exchange capacity, and hydrous ferric oxide surfaces. The initial compositions of the
cation exchangers and surfaces are in equilibrium with the brine, which contains arsenic. Arsenic is
initially sorbed on the hydrous feme oxide surfaces. The aquifer is assumed to be recharged with
rainwater that is concentrated by evaporation and equilibrated with calcite and dolomite in the unsaturated
zone. This water then enters the saturated zone and reacts with calcite and dolomite in the presence of the
cation exchanger and hydrous ferric oxide surfaces. A period of 1,000,000 years of flushing the brine-
filled aquifer with freshwater is simulated.
Model Construction
The model domain is approximately 31 nodes in the X direction (3,000-m node spacing), 17
nodes in the Y direction (3,000-m node spacing), and 9 nodes in the Z direction (50-m node spacing). The
northern and southern boundaries of the model are near rivers that provide satisfactory boundary
conditions. The eastern boundary of the model coincides with the eastern extent of the geologic units of
the aquifer. The extent of freshwater in the aquifer is used to set the western boundary of the model.
The hydraulic conductivity was taken from Parkhurst et al. (1996), but the horizontal hydraulic
conductivity was decreased to attain a maximum head in the aquifer that was consistent with the measured
water table (Parkhurst et al., 1996). The longitudinal dispersivity (2,000 m) and horizontal and vertical
transverse dispersivities (50 m) were set arbitrarily to be less than or equal to the node spacing.
Simulation Results
Figure 22 shows the simulation results for concentration of chloride after 240,000 years, 500,000 years,
740,000 years, and 1,000,000 years. The three-dimensional view of the active grid region is from
10.0
0.000
5.00
Time = 240,000 years
years Time = 500,000 years
- 0.000
5.00
10.0
Time = 740,000 years
Time = 1,000,000 years
Figure 22. Model-predicted distribution of chloride concentrations (mg/L).
61
-------
10.0
0.000
5.00
Time = 240,000 years
years Time = 500,000 years
0.000
5.00
10.0
Time = 740,000 years
Time = 1,000,000 years
Figure 23. Model-predi cted distribution of calcium concentrations (mg/L).
the southwest, which shows a part of the grid region is missing—the inactive grid region—at the western
end of the grid region. Also, the top layer of cells is not shown because the cells are dry. Parts of the
second layer of cells near the rivers are also not shown because they are dry. Figure 22 clearly shows that
the concentration of chloride will significantly decrease overtime.
Figure 23 shows the simulation results for concentration of calcium after 240,000 years, 500,000
years, 740,000 years, and 1,000,000 years. The concentration of calcium also significantly decreases over
time. The yellow areas in the eastern two-thirds of the active grid region in Figure 23 represent areas
where sodium has been removed from the exchanger, and calcium and magnesium are the dominant
cations in solution and on the exchanger. In the western one-third of the active grid region, brines have
been mostly removed, but sodium persists as the dominant cation in solution and on the exchanger.
The low concentration zones of Figure 23 correspond to high pFI zones in Figure 24. Figure 24
shows the simulation results for pFI after 240,000 years, 500,000 years, 740,000 years, and 1,000,000
years. The pH is high in these zones because of the dissolution of calcite and dolomite, which is enhanced
due to the exchange of calcium and magnesium for sodium on the exchanger. Figure 25 shows the
simulation results for concentration of arsenic after 240,000 years, 500,000 years, 740,000 years, and
1,000,000 years. The zones of high pH correspond closely with large arsenic concentrations in Figure 25.
The level of detail in these models makes the Level 3 analysis particularly important for
developing safe and sustainable ASR, since results can be used to predict both what geochemical
reactions can occur and the extent of these reactions. Groundwater sources are a vital resource which
must be protected. While ASR can protect these sources from the detrimental impacts of groundwater
overdrafting, it is equally important to ensure that ASR itself does not deteriorate groundwater quality.
Level 3 analysis will be the strongest tool in developing ASR operations that protect and enhance
groundwater resources.
62
-------
- 9.0
Time = 240,000 years
Time = 740,000 years Time = 1,000,000 years
Figure 24. Model-projected spatial distribution of pH values in groundwater.
M 200
Time = 240,000 years
Time = 500,000 years
Time = 740,000 years Time = 1,000,000 years
Figure 25. Model projected distribution of arsenic concentrations (|ug/L) in groundwater.
6.0 Conclusion
One of the greatest challenges in the coming years will be meeting increased water demands.
Population growth and dynamic climate change have led to widespread gaps in water availability, and the
challenge to our groundwater resources is projected to continue. The development of AR and ASR
technologies is a vital component in addressing these water gaps by closing our cycle of water use and
63
-------
capturing water from extreme precipitation events, as well as protecting natural ecosystems by mitigating
the impacts of groundwater overdrafting, such as the drying of wetlands and land subsidence. AR and
ASR operations must be run in a manner that protects groundwater resources while giving us the
maximum benefit through informed site selection and operational parameters. The Level 1-3 tools
presented in this DSS form a technical basis for AR and ASR best practices.
The three levels of tools cover AR and ASR system development from the earliest stages of ASR-
Need analysis, to planning and assessment, design, and evaluation. A case study for Las Vegas shows
how ASR-Need analysis can be used to determine the potential for ASR to meet future water gaps. This
analysis will be supported by the Level 1 tool, which focuses on ASR feasibility. The Level 1 tool
includes resources to examine water demand, population growth projections, and climate change
scenarios, which can inform projections for future water demand and availability. This tool is also a
valuable resource for information on state-specific ASR regulations and permitting needs, as well as
current information on ASR site locations and technical information, such as records of arsenic-
contamination of groundwater and site mineralogy, which can be used to inform site selection.
The Level 2 tool has additional capabilities to support ASR planning and assessment. This level of
analysis includes simple models for hydrological and environmental assessment. Outcomes from the
Level 2 analysis can be used to further assess potential ASR sites in order to evaluate operational
parameters including injection rate, storage capacity, and injection-related geotechnical factors. This tool
can also be used to determine the geochemical compatibility of the groundwater aquifer formation
minerals with injected water to determine the likelihood of negative groundwater impacts.
Finally, Level 3 tools will be used to evaluate ASR design and performance to achieve water
resource management objectives. Following Level 2 analysis, engineering design and evaluation
assessment incorporated into Level 3 tools must be applied to determine the hydraulic control and provide
a more detailed analysis of water quality changes in during ASR operation, including particle transport
analysis and long term evaluation of the changes in groundwater quality that result both from
geochemical reactions between injected water and aquifer bedrock, and mixing between injected water
and resident groundwater. Important design and evaluation variables which are assessed at this level
include hydraulic control, residence time, and water quality constraints.
Together, the three tools presented in this DSS can be used to develop and support the safe and
sustainable implementation of AR and ASR, ensuring that these technologies can be widely applied to
meet growing water demands and prevent water availability gaps, while protecting our most abundant
source of freshwater for residential, industrial, and agricultural needs.
7.0 References
Adhikari, A. R., K. Acharya, S.A. Shanahan, and X. Zhou, (2011). Removal of nutrients and metals by
constructed and naturally created wetlands in the Las Vegas Valley, Nevada. Environmental
Monitoring and Assessment, 180, 97-113.
Al-Zubari, W.K., (1998). Towards the establishment of a total water cycle management and re-use
program in the GCC countries. Desalination, 120, 3-14.
Anderson, M.P., and W.W. Woessner, (2002). Applied groundwater modeling: Simulation of flow and
advective transport, Volume 4. Academic Press, San Diego, 38lp.
Angelakis, A.N., M.H.F. Marecos Do Monte, L. Bontoux, and T. Asano, (1999). The status of wastewater
reuse practice in the Mediterranean Basin: Need for guidelines. Wat. Res., 33 (10): 2201-2217.
64
-------
Appelo, C.A., M. J. J. Van Der Weiden, C. Tournassat, and L. Charlet, (2002). Surface Complexation of
Ferrous Iron and Carbonate on Ferrihydrite and the Mobilization of Arsenic. Environ. Sci.
Technol., 36 (14), 3096-3103.
Arthur. J. D.. A. A. Dabous. and J.B. Cowart, (2002). Mobilization of arsenic and other trace elements
during aquifer storage and recovery, southw est Florida. In U.S. Geological Survey Artificial
Recharge Workshop Proceedings (pp. 47-50).
Asano, T., and J.A. Cotruvo, (2004). Groundwater recharge with reclaimed municipal wastewater: health
and regulatory considerations. Wat. Res., 38, 1941-1951.
Asano, T., and A. D. Levine, (1996). Wastewater reclamation, recycling and reuse: past, present, and
future. Water science and technology, 33(10-11), 1-14.
Barnett, T.P., and D.W. Pierce, (2008). When will Lake Mead go dry? Water Resources Research, 44, 1-
10.
Barsugli, J., C. Anderson, J. B. Smith, and J. M. Vogel, (2009). Options for improving climate modeling
to assist water utility planning for climate change. Water Utility Climate Alliance, San Francisco,
CA. 144p.
Bates, B.C., Z.W. Kundzewicz, S. Wu and J.P. Palutikof, Eds., (2008). Climate Change and Water.
Technical Paper of the Intergovernmental Panel on Climate Change, IPCC Secretariat, Geneva,
210 pp.
Batu, V., (1998). Aquifer Hydraulics: A Comprehensive Guide to Hydrogeologic Data Analysis,. John
Wiley & Sons, Inc. New York, NY.
Bloetscher. F., C.H. Sham. J.J. Danko III. and S. Ratick, (2014). Lessons Learned from Aquifer Storage
and Recovery (ASR) Systems in the United States. Journal of Water Resource and Protection,
6(17), 1603.
Bloetscher. F. (2015). M63 Aquifer Storage and Recovery (Vol. 63). American Water Works Association.
Board of County Commissioner, (2008). Manatee County Water Supply Facilities Work Plan, Manatee
County Administrative Center Commission Office Commission, Bradenton, Florida
Bouwer, E.J., P.L. McCarty, and J.C. Lance, (1981). Trace organic behavior in soil columns during rapid
infiltration of secondary wastewater. Wat. Res., 15, 151-159.
Brun, S.E., and L.E. Band, (2000). Simulating runoff behavior in an urbanizing watershed. Computers,
Environment and Urban Systems, 24, 5-22.
Buckingham, S. E., and J.W. Whitney, (2007). GIS methodology for quantifying channel change in Las
Vegas, Nevada. JAWWA, 43, 888-898.
Bureau of Land Management (BLM), (2004). Las Vegas valley disposal boundary: Environmental impact
statement. Retrieved from
www.blm.gov/nv/st/en/fo/lvfo/blm_programs/planning/las_vegas_valley_disposal.html
Burnell, D.K., Lester, B.H. and J.W. Mercer, (2012). Improvements and corrections to AT1230 Code,
Ground Water, 50(6), 943-953.
Carleton, G.B., (2010). Simulation of Groundwater Mounding Beneath Hypothetical Stormwater
Infiltration Basins, USGS Scientific Investigations Report 2010-5102, 64 p.
65
-------
Chang, N.-B., C. Qi, and Y.J. Yang, (2012). Optimal expansion of a drinking water infrastructure system
with respect to carbon footprint, cost-effectiveness and water demand. J. Environmental
Management, 110, 194-206.
Chen, J.C., N.B. Chang, and W.K. Shieh, (2003). Assessing wastewater reclamation potential by neural
network model. ling. A pp. Artificial Intel, 16, 149-157.
Chen, H., Tong, S.T.Y., Yang, H., and Yang, Y.J., (2015). Simulating the hydrologic impacts of land-
cover and climate changes in a semi-arid watershed. Hydrological Sciences Journal, DOI:
10.1080/02626667.2014.948445.
Christensen, N. S., A. W. Wood, N. Voisin, D.P. Lettenmaier, and R. N. Palmer, (2004). The effects of
climate change on the hydrology and water resources of the Colorado River basin. Climatic
Change, 62, 337-363.
Chu, J., J. Chen, C. Wang, and P. Fu, (2004). Wastewater reuse potential analysis: Implications for
China's water resources management. Wat. Res., 38, 2746-2756.
Cooley, H., T. Hutchins-Cabibi, M. Cohen, P. H. Gleick, and M. Herger, (2007). Hidden Oasis: water
conservation and efficiency in Las Vegas. Retrieved from the Pacific Institute and Western
Resources Advocates website at http://www.pacinst.org/reports/las_vegas/hidden_oasis.pdf
Craig, J.R., and S. Matott, (2009). Visual AEM User's Manual, downloaded from program Help, 115 p.
Dagan, G., (1987). Theory of solute transport by groundwater. Ann. Rev. FluidMech., 19, 183-215.
DeSimone, L.A., B.L. Howes, P.M. Barlow, (1997). Mass-balance analysis of reactive transport and
cation exchange in a plume of wastewater-contaminated groundwater. J. Hydrol., 203, 228-249.
Driscoll, F.G., (1986). Groundwater and wells. Johnson Screens, 1089p.
Dzombak, D.A., and F.M. Morel, (1990). Surface complexation modeling: Hydrous ferric oxide. Wiley
ISBN 978-471-63731-8. 416p.
Ellsworth, W. L. (2013). Injection-induced earthquakes. Science, 341(6142), 1225942.
Ferris, J.G., D.B. Knowles, R.H. Brown, and R.W. Stallman, (1962). Theory of Aquifer Tests. USGS
Water-Supply Paper 1536 E, 174p.
Fetter, C.W., (1994). AppliedHydrogeology; Third Edition. Prentice-Hall, Inc., Englewood Cliffs, NJ.
Finnemore, E. J., (1995). A program to calculate ground-water mound heights. Ground Water, 33(6),
139-143.
Friedler, E., (2001). Water reuse - an integral part of water resources management: Israel as a case study.
Wat. Policy, 3, 29-39.
Gabriel, B. J., and E. Accinelli, (2007). The Ramsey model with logistic population growth. Economics
Bulletin, 3, 1-8.
Georgopoulou, E., A. Kotronarou, A. Koussis, P.J. Restrepo, A. Gomez-Gotor, and J.J. Rodriguez
Kimenez, (2001). A methodology to investigate brackish groundwater desalination coupled with
aquifer recharge by treated wastewater as an alternative strategy for water supply in
Mediterranean areas. Desalination, 136, 307-315.
Greskowiak, J., H. Prommer, G. Massmann, C.D. Johnston, G. Niitzmann, and A. Pekdeger, (2005). The
impact of variably saturated conditions on hydrogeochemical changes during artificial recharge of
groundwater. App. Geoch.,20, 1409-1426.
66
-------
Haitjema, H.M., (1995). Analytic Element Modeling of Groundwater Flow, Academic Press, New York,
394.
Hantush, M.S., (1967). Growth and decay of groundwater-mounds in response to uniform percolation,
Water Resources Research, 3(1), 227-234.
Hantush, M.S. and C.E. Jacob, (1955). Non-steady radial flow in an infinite leaky aquifer. American
Geophysical Union Trans., 36, 95-100.
Harbaugh, A.W., (2005). MODFLOW-2005, The U.S. Geological Survey Modular Ground-Water Model
- the Ground-Water Flow Process. USGS Techniques and Methods 6-A16, 253 p.
Harris, S., (2005). Diffusive logistic population growth with immigration. Applied Mathematics Letters,
18, 261-265.
Haruvy, N., (1998). Wastewater reuse: Regional and economic considerations. Res., Cons. Recyc., 23: 57-
66.
Hunt, B., (2005). Visual basic programs for spreadsheet analysis. Ground Water, 43(1), 138-141.
Intergovernmental Panel on Climate Change (IPCC), (2001). Climate Change: Synthesis Report. In.
Watson, R.T. and the Core Writing Team, eds., A Contribution of Working Groups I, II, and III to
the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge, United Kingdom, and New York, NY, USA, 398 pp.
IPCC, (2007). Contribution of Working Group II to the fourth assessment report of the Intergovernmental
Panel on Climate Change. Cambridge University Press. Cambridge, UK.
IPCC, (2013). Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the
Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stacker, T.F., D.
Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M.
Midgley (eds.)]. Cambridge University Press, New York, 1535 pp.
Karl, T.R., J.M. Melillo, and T.C. Peterson, (2009). Global Climate Change Impacts in the United States.
New York, NY: Cambridge University Press.
Katz, Brian G., Sandra M. Eberts, and Leon J. Kauffman, (2011). Using CI/Br ratios and other indicators
to assess potential impacts on groundwater quality from septic systems: a review and examples
from principal aquifers in the United States. Journal of Hydrology, 397(3), 151-166.
Keeting, E. H., J. Fessenden, N. Kanjorski, (2010). The impact of CO2 on shallow groundwater
chemistry: observations at a natural analog site and implications for carbon sequestration.
Environmental Earth Sciences, 60(3), 521-536
Khan, S., R. Tariq, C. Yuanlai, and J. Blackwell, (2006). Can irrigation be sustainable? Ag. Wat.
Management, 81, 87-99
Kraemer, S. R., (2005). Frequently Asked Questions about WhAEM2000, USEPA website, Athens, GA.
Johnson, T., J. Butcher, A. Parker, and C. Weaver, (2012). Investigating the sensitivity of U.S.
streamflow and water quality to climate change: U.S. EPA Global Change Research Program's
20 Watersheds Project. J. Water Resour. Plann. Manage., 138(5), 453-464.
Jones, G. W., and T. Pichler, (2007). Relationship between pyrite stability and arsenic mobility during
aquifer storage and recovery in Southwest Central Florida. Environmental Science & Technology,
41(3), 723-730.
67
-------
Langevin, C.D., D.T. Thome Jr., A.M. Dausman, M.C. Sukop, and W. Guo, (2007), SEA WAT Version 4:
A Computer Program for Simulation of Multi-Species Solute and Heat Transport. USGS
Techniques and Methods. Book 6, Chapter A22, 39 p.
Las Vegas Wash Coordination Committee (LVWCC), (1999). Las Vegas Wash comprehensive adaptive
management plan. Retrieved from Las Vegas Wash Project Coordination Team website at
http://www.lvwash.org/html/resources_library_lvwcamp.html
Lo, K.W., Y-C. Jin, and T. Viraraghvan, (2002). Transport of bacteria in heterogeneous media under
leaching conditions. J. Env. Eng. Sci., 1, 383-395.
Lucho-Constantino, C.A., M. Alvarez-Suarez, R.I. Beltran-Hernandex, F. Prieto-Garcia, H.M. Poggi-
Varaldo, (2005). A multivariate analysis of the accumulation and fractionation of major and trace
elements in agricultural soils in Hidalgo state, Mexico irrigated with raw wastewater. Env. Int 'I,
31, 313-323.
Manka, J., and M. Rebhun, (1982). Organic groups and molecular weight distribution in tertiary effluents
and renovated waters. Wat. Res., 16: 399-403.
Matott, L.S., A. J. Rabideau, and J. R. Craig, (2008). Pump-and-treat optimization using analytic element
method flow models. Advances in Water Resources, 29, 760-775
Mearns, L. O., W. J. Gutowski, R. Jones, L.-Y. Leung, S. McGinnis, A. M. B. Nunes, and Y. Qian,
(2009). A regional climate change assessment program for North America. EOS, 90(36), 311-312
Mearns, L.O., D.P. Lettenmaier, and S. McGinnis, (2015). Uses of results of regional climate model
experiments for impacts and adaptation studies: the example ofNARCCAP. Curr. Clim. Change
Rep., 1:1-9.
Milhous, R. (1998) Restoring River Substrate Using Instream Flows: The Gunnison and Trinity Rivers.
Engineering Approaches to Ecosystem Restoration: pp. 925-930.
Miller, G.W., (2006). Integrated concepts in water reuse: managing global water needs. Desalination,
187: 65-75
Miranda, L. C. M., and C. A. S. Lima, (2010). On the logistic modeling and forecasting of evolutionary
processes: Application to human population dynamics. Technological Forecasting and Social
Change, 11, 699-711.
Morris, R. L., D. A. Devitt, A.M. Crites, G. Borden, and L. N. Allen, (1997). Urbanization and water
conservation in Las Vegas valley, Nevada. J. Water Resources Planning and Management, 123,
189-195.
Neil, C. W., Y. J. Yang, and Y.-S. Jun, (2012). Arsenic mobilization and attenuation by mineral-water
interactions: implications for managed aquifer recharge. Journal of Environmental Monitoring
14, (7), 1772-1788.
Neil C., Y.J. Yang, D. Schupp, and Y-S. Jun, (2014). Water chemistry impacts on arsenic mobilization
from arsenopyrite dissolution and secondary mineral precipitation: Implication for managed
aquifer recharge. Environmental Science and Technology, 48(8), 4395-4405.
Parkhurst, D.L., S. Christenson, and G.N. Breit, (1996). Chapter C, Ground-Water-Quality Assessment
of the Central Oklahoma Aquifer, Oklahoma: Results of Investigations. USGS Water-Supply
Paper 2357.
68
-------
Parkhurst, D.L., K.L. Kipp, and S.R. Charlton, (2010), PHAST Version 2—A Program for Simulating
Groundwater Flow, Solute Transport, and Multicomponent Geochemical Reactions. USGS
Techniques and Methods 6-A35, 235 p.
Parkhurst, D.L., and C.A.J. Appelo, (2013). Description of input and examples for PHREEQC version
3—A computer program for speciation, batch-reaction, one-dimensional transport, and inverse
geochemical calculations. U.S. Geological Survey Techniques and Methods, book 6, chap. A43,
497 p., available only at http://pubs.usgs.gov/tm/06/a43.
Pasch, J., and P. Macy, (2005). Building sustainable wastewater reuse in Jordan. Wat. Supply, 5: 17-25.
Pereira, L.S., T. Oweis, A. Zairi, (2002). Irrigation management under water scarcity. Ag. Wat.
Management, 57, 175-206.
Piechota, T., and J. Bastista, (2003). Source water assessment for the Las Vegas valley surface waters.
Carson City, NV. State of Nevada, Bureau of Health Protection Services.
Pitt, R., Clark, S., and Field, R. (1999). Groundwater contamination potential from stormwater infiltration
practices. Urban water, 7(3), 217-236.
Pollock, D.W., (2012). User Guide for MODPATH Version 6—A Particle-tracking Model for
MODFLOW. USGS Techniques and Methods, book 6, chap. A41, 58 p.
Progressive Leadership Alliance of Nevada (PLAN). (2006). Las Vegas valley groundwater development
project: Where does it start? Where will it end? Retrieved from Progressive Leadership Alliance
of Nevada Web site atwww.planevada.org
Qadir, M., B.R. Sharma, A. Bruggeman, R. Choukr-Allah, and F. Karajeh, (2007). Non-conventional
water resources and opportunities for water augmentation to achieve food security in water scare
countries. Ag. Wat. Management, 87, 2-22.
Qaiser, K., S. Ahmad, W. Johnson, and J. Batista, (2011). Evaluating the impact of water conservation on
fate of outdoor water use: a study in an arid region. Journal of Environmental Management, 92,
2061-2068
Ranatunga T., S. T.Y. Tong, Y. Sun, and J.Y. Yang, (2014). A total water management analysis of the
Las Vegas wash watershed, Nevada. Physical Geography, 35(3), 220-244
Scheytt, T.J., P. Mersmann, and T. Heberer, (2006). Mobility of pharmaceuticals carbamazepine,
diclofenac, ibuprofen, and propyphenazone in miscible-displacement experiments. Cont. Hydrol.,
83:53-69.
Sen, T.K., D. Das, K.C. Khilar, and G.K. Suraishkumar, (2005). Bacterial transport in porous media: New
aspects of the mathematical model. Colloids and Surface A: Physicochemical and Engineering
Aspect, 260, 53-62.
Shelef, G., and Y. Azov, (1996). The coming era of intensive wastewater reuse in the Mediterranean
region. Wat. Sci. Tech., 33(10-11), 115-125.
Sheng, Z., (2005). An aquifer storage and recovery system with reclaimed wastewater to preserve native
groundwater resources in El Paso, Texas. J. Env. Management, 75, 367-377.
Schroeder, R.A., (2002). Transport and fate of water-quality indicators after 40 years of artificial recharge
with treated municipal wastewater to the Central Ground-water basin in Los Angeles County. In
G.R. Aiken and E.L. Kuniansky, eds., U.S. Geological Survey Artificial Recharge Workshop
Proceedings, April 2-4, 2002, Sacramento, California. U.S. Geological Survey, Open-file Report
02-89, pp.42-46
69
-------
Stave, K. A., (2001). Dynamics of wetland development and resource management in Las Vegas Wash,
Nevada. JAWWA, 37, 1369-1379.
Stave, K.A., (2003). A system dynamics model to facilitate public understanding of water management
options in Las Vegas, Nevada. Journal of Environmental Management, 67, 303-313.
Southern Nevada Water Authority (SNWA), (2009). Water Resources Plan 09. Southern Nevada Water
Authority, Las Vegas, Nevada. Retrieved from Southern Nevada Water Authority Website at
http: //www. snwa. com/ws/resource_plan .html
Southern Nevada Water Authority (SNWA), (2015). Water Resource Plan 2015. Southern Nevada Water
Authority, Las Vegas, Nevada. Retrieved from Southern Nevada Water Authority Website at
https://www.snwa.com/assets/pdf/wr_plan.pdf
Strack, O.D.L., (1989). Groundwater Mechanics, Prentice Hall.
Sun, Y., S.T.Y. Tong, M. Fang, and Y.J. Yang, (2013). Exploring the effects of population growth on
future land use change in the Las Vegas Wash watershed: An integrated approach of geospatial
modeling and analytics. Environment, Development, and Sustainability, 15, 1495-1515
Tabelin, C.B., T. Igarashi, and R. Takahashi, (2012). Mobilization and speciation of arsenic from
hydrothermally altered rock in laboratory column experiments under ambient conditions. Applied
Geochemistry, 27(1), 326-342
Theis, C.V., (1935). The relation between the lowering of the piezometric surface and the rate and
duration of discharge of a well using groundwater storage. American Geophysical Union Trans.,
16, 519-524.
Thomas, J.-S., and B. Durham, (2003). Integrated water resource management: looking at the whole
picture. Desalination, 156, 21-28.
Tong, S.T.Y., Y. Sun, T. Ranatunga, J. He, Y. J. Yang, (2012). Predicting plausible impacts of sets of
climate and land use change scenarios on water resources. Applied Geography, 32, 477-489
Toze, S., (2004). Reuse of effluent water - benefits and risks. Proc. 4th International Crop Science
Congress, Brisbane, Australia.
U.S. Bureau of Reclamation (USBR), (2009). Colorado River Basin Salinity Control Project-Las Vegas
Wash Unit-Title II. Retrieved from U.S. Department of Interior Web site at
http://www.usbr.gov/proiects/Proiect.isp7proi Name=CRBSCP+-+Las+Vegas+Wash+Unit+-
+Title+II#Group608790
U.S. Climate Change Science Program (USCCSP), (2001). Climate change impacts on the United States:
The potential consequences of climate variability and change. Cambridge University Press, New
York.
U.S. EPA. (2001). Technical program overview: Underground injection control regulations. EPA 816-R-
02-025. Washington DC
U.S. EPA. (2004). Guidelines for water reuse. EPA/625/R-04/108. Washington DC, 443p.
U.S. EPA, (2007). Working with WhAEM2000, Capture Zone Delineation for a City Wellfield in a
Valley Fill Glacial Outwash Aquifer Supporting Wellhead Protection, USEPA/600/R-05/151.
U.S. EPA. (2008). A Systematic Approach for Evaluation of Capture Zones at Pump and Treat Systems.
EPA/600/R-08/003, Washington, DC.
70
-------
U.S. EPA, (2012a). Guidelines for water reuse. Environmental Protection Agency, Municipal Support
Division Office of Wastewater Management Office of Water Washington, DC. Agency for
International Development Washington DC EPAJ625/R-04/108, Cincinnati, OH US EPA/625/R-
04/108.
U.S. EPA, (2012b). Nonpoint Source program success story: Best Management Practices Drastically
Reduce Sediment and Restore Water Quality in Las Vegas Wash.
http://water.epa.gov/polwaste/nps/success319/state_nv.cfm
U.S. EPA, (2015a). National Water Infrastructure Adaptation Assessment, Part II: Characterize Climatic
Change Impacts for Water Adaptation Planning and Engineering. Under Review.
U.S. EPA, (2015b). National Water Infrastructure Adaptation Assessment, Part VI: Climate Change
Adaptation Methods and Case Studies. Under Review.
U.S. EPA, (undated). Class V UIC study Sheet - Aquifer recharge wells and aquifer storage and recovery
wells, http://www.epa.gov/UIC
U.S. Geological Survey (USGS), (2000). Flood of July 8, 1999, in Las Vegas Valley, Southern Nevada.
USGS Factsheet 80, U.S. Department of the Interior.
U.S. Global Change Research Program (USGCRP), (2000). Climate Change Impacts on the United
States: The Potential Consequences of Climate Variability and Change. Retrieved from the
National Assessment Synthesis Team Website at http://downloads.globalchange.gov/nca/nca-
2000-report-overview.pdf
Wallis, I., H. Prommer, C. T. Simmons, V. Post, and P. J., Stuyfzand, (2010). Evaluation of conceptual
and numerical models for arsenic mobilization and attenuation during managed aquifer recharge.
Environmental science & technology 44, (13), 5035-5041.
Wang, X., A. Burguess, and Y.J. Yang, (2013). A scenario-based water conservation planning support
system (SB-WCPSS). Stochastic Environmental Research and Risk Assessment, 27(3), 629-641.
Weber, S., S. Khan, and J. Hollender, (2006). Human risk assessment of organic contaminants in
reclaimed wastewater used for irrigation. Desalination, 187:53-64.
Weeks, E.P., (2002). A historical overview of hydrologic studies of artificial recharge in the U.S.
Geological Survey. In G.R. Aiken and E.L. Kuniansky, eds., U..S' Geological Survey Artificial
Recharge Workshop Proceedings, April 2-4, 2002, Sacramento, California. U.S. Geological
Survey, Open-file Report 02-89, pp.5-12.
Welch, A.H., D.B. Westjohn, D.R. Helsel, and R.B. Wanty, (2000). Arsenic in ground water of the
United States— occurrence and geochemistry. Ground Water, 38(4), 589-604.
Welty, C. (2009). The Urban water budget. In L.A. Baker, ed, The Water Environment of Cities. New
York, NY, Springer, pp. 17-28.
West, M. R., B. H. Kueper, and M. J. Ungs, (2007). On the use and error of approximation in the
Domenico (1987) solution. Ground Water, 45(2), 126-135.
Westerhoff, P., and M. Pinney, (2000). Dissolved organic carbon transformations during laboratory-scale
groundwater recharge using lagoon-treated wastewater. Waste Management, 20, 75-83.
World Health Organization (WHO), (2006). Guidelines for the safe use of wastewater, excreta and
greywater: volume 2, Wastewater use in agriculture. World Health Organization, France.
71
-------
Yang Y. J., and J.A. Goodrich, (2014). Toward quantitative analysis of water-energy-urban-climate nexus
for urban adaptation planning. Current Opinion in Chem Engrng, 5, 22-28.
Yang, Y. J., J.A. Goodrich, S.Y. Li, and R.C. Haught, (2007). AST/R-based water reuse as a part of the
total water solution for water-stressed regions: An overview of engineering practice and
regulatory prospective. Water Environmental and Water Resources Congress 2007, Tampa,
Florida.
Yang, Y. J., R. D. Spencer, and T. M. Gates, (1997). Analytical Solutions for Determination of Non-
Steady- State and Steady-State Capture Zones. Groundwater Monitoring & Remediation, 15,
101-106.
Yang, Y. J., and T. M. Gates, (1997). Well-bore skin effect and slug test data analysis for aquifers of low
permeability. Ground Water, 55(6), 931-937.
Yang, Y. J., D. J. Murray, and J. A. Goodrich, (2007). Water Resources Adaptation to Global Changes:
Risk Management through Sustainable Infrastructure Planning and Management. Water resource
adaptation to GCC, 08 Water Down Under, Adelaide, Australia
Yates, M.V., and Y. Ouyang, (1992). VIRTUS, a model of virus transport in unsaturated soils. App. Env.
Microbiol., 58(5), 1609-1616.
Yeh, G. T., (1981), AT123D: Analytical Transient One-, Two-, and Three-Dimensional Simulation of
Waste Transport in the Aquifer System. ORNL-5602, Oak Ridge National Laboratory, Oak
Ridge, TN.
Zheng, L., J. A. Apps, Y. Zhang, T. Xu, and J. T. Birkholzer, (2009). On mobilization of lead and arsenic
in groundwater in response to CO2 leakage from deep geological storage. Chemical Geology, 268,
281-297.
Zheng, C., (2010). MT3DMSv5.3 Supplemental User's Guide. Technical Report to the U.S. Army
Engineer Research and Development Center, Department of Geological Sciences, University of
Alabama, 51 p
72
-------
Appendix A
Literature Review and Experimental Analysis of Arsenic Re-mobilization at ASR Sites*
Modified from Neil et al. (2012) (Journal of Environmental Monitoring, 14, 7, 1772-1788)
and Neil et al. (2014) {Environmental Science and Technology, 48, 8, 4395-4405)
73
-------
1. Arsenic Occurrence and Natural Sources
Arsenic remobilization is a principal challenge in groundwater and at ASR sites. After
conventionally treated effluent undergoes tertiary treatment, the effluent, i.e., reclaimed wastewater, can
be a common water source for ASR operation (Sheng, 2005; Pavelic et al, 2005; Vanderzalm et al.,
2006). Not only is the use of reclaimed water cost effective, but natural attenuation processes in the
vadose zone and underlying aquifer have been shown to remove residual pathogens from the injected
secondary water (Wilson et al., 1995; Asano and Cotruvo, 2004).
However, recent studies at ASR field sites have shown that reclaimed water recharge can trigger
unfavorable soil-water interactions releasing arsenic, a toxic metalloid, from aquifer materials.
Unfavorable soil-water interactions can release arsenic from aquifer materials resulting in dangerously
high levels of arsenic in groundwater in large areas of the U.S., Australia, Germany, and China (e.g.,
Pavelic et al, 2005; Greskowiak et al., 2006). Table A-l shows locations around the globe where elevated
arsenic has been measured as a result anthropogenic groundwater recharge. For example, Jones and
Pichler (2007) reported that while injection waters to an ASR site in South Central Florida contained 3
ug/L of arsenic, recovered levels ranged from 10-130 fig/L. Despite the observations and intense studies
in multiple aquifer systems, consensus has not been reached on the dominant cause of arsenic
remobilization (Jones and Pichler, 2007; Wallis et al., 2010). This ambiguity in knowledge is further
complicated by many potential sources for arsenic, and by the attenuation processes that occur
concurrently within the aquifer (Wang and Mulligan, 2006; Wallis et al., 2011).
Arsenic exists naturally in aquifer formations in several forms (Figure A-l). Frequently, arsenic is
incorporated into pyrite (FeS2), in quantities as large as 10wt%. Substitution of arsenic into pyrite
crystalline structure occurs under both oxidizing and reducing conditions, and the resulting structure
contains AsS dianion groups (Blanchard et al., 2007). The product, called arsenian pyrite, is less stable
than pyrite and will dissolve more
rapidly in water. Arsenic can also be ^ 8 , \ .
substituted into marcasite (FeSz), i ^ * *
which is arranged in an
orthorhombic stmcture, as compared :^
to the cubic structure of pyrite. This Rl * * *
mineral is less geochemically stable
than pyrite, and compared to
arsenian pyrite, arsenic-substituted
marcasite has a larger arsenic
solubility in water (Reich and
Becker, 2006). In addition, the
substitution product can be either a
stable solid solution, or metastable.
The metastable solid will eventually
form nano-domains of pyrite or
marcasite and arsenopyrite (Reich
and Becker, 2006).
Arsenopyrite (FeAsS),
which contains a 1:1:1 ratio of iron,
sulfur, and arsenic, has a monoclinic
structure similar to that
a) Arsenopyrite
b) Scorodite
c) Orpiment
d) Realgar
gure 1: Structures of arsenic-containing minerals created using Crystal Maker V.2.3.
Figure A-l Structures of arsenic-containing minerals. After Neil
et al. (2012).
74
-------
Table A-l. Location and conditions for recharge-influenced arsenic mobilization.
Site
Recharge Water
Aquifer Type
Site Characteristics
Arsenic Level
Reference
ASR Site
Full-scale ASR trial at
Bolivar, South
Australia
Reclaimed water from the
Bolivar Water Reclamation
Plant
Carbonate Aquifer
Injection flow rate: 7.9-11.9 L/s
Recovery flow rate: 8.7-15.9 L/s
Depth: 100-160 m
^average• 3 m/day
Injected: 0.04 ± 0.03 nM
Ambient: 0.04 ±0.03 nM
Recov.: 0.30 ±0.16 ^M
Vanderzalm et al. (2011)
Southwest-Central
Florida Groundwater
Basin, USA
Surface water
Highly permeable
carbonate rocks, Suwannee
Limestone, Ocala
Limestone
K: 0.98-30 m/day
Pyrite: 276-32,406 mg/kg
As wt% pyrite: 0.01-1.12
Injected and storage zone: 3 ng/L
Recovered: 10-130 ng/L
Wallis et al. (2011)
Jones and Pichler (2007)
San Joaquin Valley,
California, USA
Surplus water from the
Stockton East Water District
Water Treatment Plant
Fluvial sediment of the
Pleistocene Modesto and
Riverbank Formations
Injected flux: 2.5xl06 m3/surface
area
Depth: 60 m
Injected: <5 ng/L
Recovered: 7-10 ng/L
McNab et al. (2009)
Fox River Valley, Green
Bay, Wisconsin, USA
Surface water and
groundwater from another
aquifer
Sandstone and limestone
Transmissivity: 102 m2/day
Recovered: 3-60 ng/L
Bahr et al. (2002)
Brown et al. (2006)
Manatee, Florida, USA
Reclaimed water
Carbonate aquifer
Aquifer x: 0.5 months
Salinity: 2000 mg/L
T: 269C; flow rate: 5.26 m3/min
Storage: 19,000 m3
Injected: ND
Ambient: 8 ng/L
Recovered: 24 ng/L
Overacre et al. (2006)
Ruhr Valley, Western
Germany
Bank infiltration
Sandy sediment, anoxic
Pleistocene aquifer
V: 0.21-0.82 m/day
Maximum of .185 ^M for V = 0.21
m/day
Maximum 0.340 nM for V =
0.82 m/day
Schlieker et al. (2001)
Pumping station
Schuwacht (Hydron-
ZH), Netherlands
Treated and aerated
groundwater.
Coarse, sandy sediments of
the Sterksel formation
Water periodically injected, flow
rate: 30 m3/h for 2 days
Depth: 20-30 m
Injected: 0 ng/L
Recovered: 9-14 ng/L
Appelo et al. (2002)
75
-------
Table A-l continued.
Site
Recharge Water
Aquifer Type
Site Characteristics
Arsenic Level
Reference
Aquifer recharge site
Western Snake River
Plain, Idaho, USA
Crop irrigation with surface
water
Alluvial gravels and sands
Recharge rate: > 50 cm/year
Average 02: 4.8 mg/L
Surface irrigation: 7 [ig/L
Seep water: 38 [ig/L
Busbee et al. (2009)
Hetao basin,
Northwest China
Alluvial fan overflow and
irrigation channels
alluvial-pluvial sand, fluvial-
lacustrine sandy silt, silty
clay and organic matter
rich clay
K=10-20 m/day
Moderate flow (recharge) zone:
30.6 Mg/L
Low flow: 131 [ig/L
Discharge zone: 34 [ig/L
Guo et al. (2008)
Madison River Valley,
Montana, USA
Arsenic-rich river water and
irrigation
Quaternary alluvium and
tertiary volcano-clastic
sediment,
Groundwater flow rate: 0.34 m3/s
Transmissivity: 2490 m2/day
Recharge: 41-74 [ig/L
Oxic zone: 25-50 [ig/L
Reduced zone: 60-160 [ig/L
Nimick (1998)
Note: K is the hydraulic conductivity, V is the linear velocity of the fluid, and t is residence time.
76
-------
of marcasite (Brostigen and Kjekshus, 1970). Arsenopyrite is the most commonly occurring As-bearing
mineral in Earth's crust, and is very stable under natural conditions (Mandal and Suzuki, 2002; Gonzalez
et al., 2006). However, the exposure of these minerals to either oxidizing or acidic aqueous conditions
may result in mineral decomposition which may release arsenic into the environment (Verplanck et al.,
2008). Under oxidizing conditions, the solubility of arsenopyrite is limited by the solubility of scorodite
(FeAs04,2H20), an oxidation product of arsenopyrite. Acidic or reducing surficial conditions can result
in the transformation of arsenopyrite into realgar (a-As4S4) or orpiment (AS2S3). Both of these are
monoclinic arsenic sulfides with very low solubilities, and may provide some mitigation of arsenic
mobility in reducing environments. These three minerals would release arsenic through the following
reactions (Sadiq et al., 2002):
l/4a"As4S4 + 8H20 o HAs042 + S042" + 15H+ + 1 le" LogK= -101.58 (Al)
As2S3 + 20H2O <-> 2HAs042" +3S042" +38H++28e" LogK=-219.21 (A2)
FeAs04'2H20 + H+ o Fe3+ + HAs042 + 2H20 LogK= -11.67 (A3)
The most likely culprits of increased arsenic during ASR are arsenopyrite and arsenian pyrite due
to aqueous chemistry changes (e.g. changes in reduction-oxidation potential, dissolved oxygen levels, and
pH) induced by reclaimed water injection.
Physiochemical and biological process mechanisms at nano- to macro-scales are responsible for
arsenic mobilization from arsenic-bearing minerals and for arsenic sinks present within groundwater
aquifers. Table A-2 describes these reactions pathways. Mechanisms that can promote arsenic
remobilization include the oxidation of arsenic-bearing minerals and reduction of arsenic-containing
ferrihydrite, as well as the impact of groundwater aquifer hydrology and microbial activity on the kinetics
of arsenic release. The mechanisms which can attenuate aqueous arsenic concentrations in natural and
ASR systems include arsenic sorption onto iron-oxyhydroxides and co-occurring arsenic-sulfide
precipitation. Some of the processes operate for both remobilization and attenuation.
2. Arsenic remobilization and dissolution mechanisms
Arsenopyrite and arsenian pyrite are redox sensitive minerals. Changes in the oxidation-reduction
potential can trigger physico-chemical processes that affect total aqueous arsenic concentrations in
groundwater. The composition of reclaimed water can differ from resident groundwater in terms of the
concentrations of salts, metal ions, organic compounds, and dissolved oxygen. Injection of reclaimed
water can, therefore, serve as a trigger which destabilizes arsenopyrite leading to an increase in arsenic
mobility.
2.1. Oxidation of Arsenic-bearing minerals
Naturally, dissolved oxygen is not abundant in deep aquifers. However, it can be present in
shallower aquifers or introduced by the injection of oxygenated reclaimed water through wells and other
recharge facilities. Table A-3 provides empirically derived rate laws for arsenopyrite oxidation. In principle,
molecular oxygen will interact with arsenopyrite through the following mechanism (Walker et al., 2006):
4FeAsS + 6H20 + 1102(aq) 4Fe2+ + 4H3As03 + 4S042" (A4)
Both iron and arsenic could then be further oxidized through the following two reactions:
H3As03 + H20 HAs042 + 4H+ + 2e (A5)
Fe2+ + 3H20 Fem(OH)3 + 3H+ + e" (A6)
77
-------
Table A-2: Processes impacting aqueous arsenic mobility*.
Description
Proposed Mechanism
References
Arsenic Mobilization Processes
Arsenic-containing pyrite
dissolution
Oxidation by 02: FeAsS + 1.5H20 + 2.7502(aq)-<-» Fe2+ + H3As03 + S042"
Oxidation by Fe3+: FeAsS(s) + 7H20 + HFe3+ 12Fe2+ + H3As03(aq) +S042" + 11H+
Walker et al. (2006)
Yu et al. (2007)
Reduction of arsenic-
bearing ferrihydrite
Reduction by bio-organisms in the presence of acetate**:
8FeOOH + CH3COO"+ 15H2C03 -> 8Fe2++ 17HC03" + 12H20
McArthur et al. (2001)
Labile arsenic release
Competitive desorption from iron oxides or clay minerals due to anionic ligand
(silicate, phosphate, carbonate, etc.) concentration influx
Violante and Pigma (2002)
Waychunas et al. (2008)
Jain and Loeppert, (2000)
Arsenic Attenuation Processes
Sorption onto ferrihydrite
Fe2+ + 3H20 Fe(OH)3 + 3H+ + e"
Salzsauler et al.(2005)
Precipitation of arsenic-
containing minerals
Precipitation of scorodite: Fe3+ + HAs042" + 2H20 <-» FeAs04-2H20 + H+
Precipitation of arsenic-sulfides
Salzsauler et al. (2005)
Note: * - Adopted from Neil et al. (2012).
** - Used as a model organic compound.
78
-------
Table A-3: Empirically derived rate laws for arsenopyrite oxidation by compounds of interest*.
Conditions
Rate Law
Reference
T = 255C
Ro2(aq) (mol m"2s_1) = 10"7-41±°-47 x a°276±011 x aH°+12±° 01;
ai=activity i
Asta et al. (2010)
T= 25-C, pH = 2, 1= 0.01 M
RFe3+ (mol mineral rrf2 s 1)= -10-500(MFe3+)106±011
McKibben et al.
(2008)
T = 25 5C, pH = 2-4, 1 = 0.01 M
Ro2(aq) (mol mineral m"2 s"1) = -10-611(Mo2(aq))a33±018(MH+)a27±009
McKibben et al.
(2008)
T = 255C, pH = 2
RFe3+ (mol s-1)— -1.45xlO"3(A)(mFe3+)a9S
A=Arsenopyrite surface area (m2); mpmol i kg"1
Rimstidt et al.
(1993)
pH = 1.8-6.4
-2211+57
Ro2(aq) (mol m~2S_1) = 10 T (Mo2(aq))a45±0 05
Yu et al. (2007)
T = 25?C, pH = 6.3-6.7, DO = 0.3-17 mg L1
Ron(aq) (mol m~2s_1) = io~10,14±0,03
Walker et al. (2006)
T= 40?C, pH = 1.6
12% solids and microorganisms
Rbio(kg m~3d_1) = 2.1
Miller and Hansford
(1992)
T = 255C
pH = 1.1
Einitiai— 615 mV
r , zF dE
r (moi l~2 s _i)= le;;:fR\dt
(1 +[Fe2+])([Fe3+] 1 6)
[Fe2+]
z=moles of electrons transferred; F=Faraday constant;
E=reduction-oxidation potential
Ruitenberg et al.
(1999)
Alkaline solution
265.4 < p02< 1053 kPa
—15.1 kj rii-
f = ke RT [OH-]023
dt L J
Koslides and
Ciminelli (1992)
T= 120-1805C
p02= 2-20 atm
0.5 N H2S04
dN . . "8672
—(mol min cm ) = 49.527e t p02
S=total surface area of FeAsS particles
Papangelakis and
Demopoulos(1990)
Note: * - Adopted from Neil et al. (2012), T = temperature, I = Ionic strength, DO=Dissolved oxygen.
79
-------
Salzsauler et
al. (2005) studied a
system containing
these two reactions at
the arsenopyrite
residue stockpile in
Snow Lake, Manitoba
Canada, which formed
as a result of mining
arsenopyrite-associated
gold. Arsenopyrite was
found to be stable
under high pH and low
Eh conditions; however
after oxidation, Fe2+
and As3+ was released
and secondary mineral
precipitation occurred
as predicted in the Eh-pH diagram shown in Figure A-2. Site measurements indicated that while the Eh
was dictated by the Fe2+ /Fe3+ couple, the equilibrium would also be affected by pH, because lower pH
values increased the ratio of As3+ to As As5+. Groundwater pH values in the analysis here generally
ranged between 7 and 9, and Eh values between 0 and 0.2 V. Under these conditions, the formation of
goethite and amorphous Fe(OH)3 were kinetically favorable. Both of these minerals have the capability of
sorbing aqueous arsenic. A decreased Eh under the same pH conditions would result in the formation of
AsS2-, while lowering the pH under the same Eh conditions would form H3ASO3. Neither of these
conditions would mitigate arsenic mobility. As shown in Figure A-2, scorodite also forms in near-neutral
pore solutions (pH ~ 7. Eh ~ 0.1 V) thorough the following reaction:
Fe3+ + HAs042 + 2H20 FeAs04 2H20 + H+ (A7)
The formation of scorodite can immobilize the arsenic in groundwater. However, this reaction
would also lower the pH of the system, decreasing the supply of As5+ due to both consumption during
reaction and less favorable reduction-oxidation conditions. Consequently, the saturation index with
respect to scorodite would decrease, reaching a geochemical equilibrium, because As5+ is needed for
continuous precipitation.
In addition to the influx of oxidants and reductants such as 02, Fe3+, and total organic carbon
(TOC) during ASR, the oxidation states of metal may also be impacted by microbial activity. Nesbitt et
al. (1995) studied the oxidation of arsenopyrite by air-saturated distilled water and found that microbial
processes caused changes to the mineral surface and structure. The surface iron layer in arsenopyrite
transformed into Fe3+-oxyhydroxides, and this transformation provides a potential source of arsenic
attenuation through arsenic sorption onto the mineral surfaces. They also reported diffusion of iron and
arsenic atoms from within the bulk mineral to its surface, as opposed to electron transfer. At the surface,
As(-I) was oxidized, forming As5+ and As3+ oxyhydroxides. These oxidized arsenic species dissolve
readily into solution as aqueous arsenate (AsC>43) or arsenite (AsC>33). Although these surface altering
processes also occurred on arsenopyrite oxidized in air, the rate of oxidation was significantly enhanced
by the presence of water; In Nesbitt's study, following 25 hours of reaction in the air, only 22% of the
surface iron had transformed to Fe3+-oxyhydroxides, compared to 64% on the sample reacted in the
presence of water.
C.
111
Scoicdite
r
k,a;.0.
AS (OH), \
o.
Goethrte
o
u 0
-------
McKibben et al. (2008) investigated the differences between oxidation of arsenopyrite by
dissolved oxygen, Fe3+, and N03", and also examined temperature effects on the oxidation process.
Temperature experiments revealed non-Arrhenius behavior, including the promotion of an inhibitory side
reaction that prevented FeAsS dissolution. This side reaction was hypothesized to be the precipitation of
arsenic-sulfide minerals at higher temperatures. According to Rimstidt et al. (1993), the activation energy
for the oxidation of arsenopyrite becomes negative at temperatures greater than 25°C, implying that
oxidation occurs under ambient conditions. Comparison between the oxidation rates by Fe3+ and O2
showed that Fe3+ oxidized arsenopyrite at a rate at least one order of magnitude faster than by dissolved
oxygen. The molal specific rates were determined to be:
RFe3+ (moles mineral m"2 s"1)= -10"5 00(MFe3+)106±011 (A8)
Ro2(aq) (moles mineral m"2 s"1)= -10-611(Mo2(aq))0-33±018(MH+)0m009 (A9)
Results from the NO3" study did not show dissolution by either oxidative or proton-promoted
dissolution when under anaerobic conditions.
These results are consistent with other studies examining the oxidation of arsenopyrite by Fe(III)
(Yu et al. 2004, 2007). Not only is Fe(III) found often in reclaimed water, but it can be a product of
arsenopyrite or arsenian pyrite dissolution from solids in the subsurface (e.g. Eq.3). Oxidation by Fe(III)
occurs according to the following pathway (Yu et al., 2004):
FeAsS(s) + 7H20 + 1 lFe3+ «>12Fe2+ + H3As03(aq) +S042_ + 11H+ (A10)
Although this reaction will deplete aqueous Fe3+, it can be recreated through the reoxidation of Fe2+ by
dissolved oxygen or by iron-oxidizing bacteria in the environment. In particular, at lower pHs, the
oxidation of FeAsS by aqueous Fe3+would dominate over the oxidation by dissolved oxygen. For
example, atpH 1.8, the activation energy ofFeAsS oxidation by Fe3+is 16 kJ/mol, as compared to 43
kJ/mol for dissolved oxygen, which would allow for fast oxidation by Fe3+ (Yu et al., 2004).
Despite the greater potential impact of Fe3+on arsenopyrite oxidation, the effect of oxygen is
expected to prevail at ASR sites because the occurrence of very acidic (i.e. pH 1.8) groundwater is rare
and oxygen is ubiquitous in injected secondary water. It should be observed, however, that not all studies
agreed on the influence of dissolved oxygen on arsenopyrite dissolution. Walker et al. (2006) found that
changes in dissolved oxygen levels between 0.3 and 17 mg/L did not impact dissolution over a pH range
of 6.3-6.7. As a result, they concluded that the rate-determining step was not the electron donation to the
oxidant, dissolved oxygen, but rather the removal of electrons from As"1 or S"1 and electron transfer to the
oxygen atom in water. Modeling this process is further complicated because the oxidation of arsenopyrite
does not necessarily release arsenic and ferrous iron congruently. While iron will be released at a
stoichiometric ratio during FeAsS dissolution, As and S display lower dissolution rates (McKibben et al.,
2008). This occurs because the Fe can readily leave the lattice structure while As and S are retained on the
surface due to differences in the oxidation rates of the individual elements. By contrast, both Buckley and
Walker (1988) and McGuire et al. (2001) found elemental sulfur remaining on the mineral surface, while
both Fe and As were oxidized and leached out; McGuire et al. hypothesized that this sulfur was the
product of a complex oxidation pathway.
Another mechanism for arsenic mobilization is sulfide-arsenide exchange in the presence of an
oxidant (Zhu et al., 2008). This mechanism likely proceeds through the binding of aqueous sulfide to iron
on the surface of arsenopyrite or arsenian pyrite, destabilizing arsenic within the mineral lattice. As"1 is
consequently more vulnerable to oxidation processes, ultimately forming pyrite and an aqueous arsenite
or arsenate species. For aquifers that contain H2S, this reaction may proceed through the following
mechanism (Heinrich and Eadington, 1986):
4FeAsS + 4H2S + 502(g) + 2H20 4FeS2 + 4H3As03 (Al 1)
81
-------
This reaction can continue until most of the reactive surface arsenic atoms are released, and proceed
further through the oxidation of surface pyrite or fracturing of the surface, which would expose new
arsenic-rich sites (Zhu et al., 2008). In the reaction, the source of aqueous sulfide is often a product of
sulfate-reducing bacteria through the following mechanism (Jegensen, 1983):
2CH3CHOHCOO + S042 <-> 2CH3COO + 2HC03 + H2S(aq) (A 12)
These bacteria are most likely to flourish under anaerobic conditions, which are not conducive to
sulfide-arsenide exchange. Therefore, in order for this process to take place there must first be an
environment that favors H2S generation. The oxidant would then be introduced to the system, potentially
through the injection of oxygenated secondary water during ASR. This could induce the oxic or hypoxic
conditions prerequisite to Reaction (11) taking place. Under conditions where sulfide and an oxidant are
present in groundwater, this reaction is more thermodynamically favorable than oxidation by molecular
oxygen (Zhu et al., 2008). It can occur in the absence of oxygen, when another oxidant such as Fe3+,
Mn4+, or nitrate (i.e. electron acceptors) is present.
2.2 Reduction of arsenic-containing ferrihydrite
While the aforementioned mechanisms point to the solubilization of arsenic as a result of
oxidation, reducing conditions can also mobilize arsenic from aquifer materials. This process generally
occurs through the reduction of arsenic-bearing ferrihydrite [Fe(OH)3] (Zhu et al., 2008; Zheng et al.,
2004; McArthur et al., 2001). Ferrihydrite strongly sorbs arsenic, and desorption is less likely to occur
until the mineral degrades through dissolution. There exist other processes that can affect the sorption of
arsenic to ferrihydrite. Both ferrous iron and carbonate may form surface complexes on ferrihydrite
resulting in arsenic displacement (Appelo et al., 2002). Furthermore, phosphate and silicate will compete
with arsenate (As5+) for sorption sites on the mineral surface (Violante and Pigna, 2002; Jain and
Loeppert, 2000; Waychunas et al., 2007; Dixit and Hering, 2003).
Arsenic exists in many aquifer formations associated with ferrihydrite, either through sorption or
co-precipitation. Reduction processes can result from interactions between As-bearing ferrihydrite and
ammonium, sulfide, and organic matter. Frequently, these reductants are products of microbial processes.
For example, ferrihydrite reduction by organic matter (e.g. acetate) occurs through the following reaction
(McArthur et al., 2001):
8FeOOH + CH3COO + 15H2C03 -~ 8Fe2+ + 17HC03 + 12H20 (A13)
where acetate is a product of microbial metabolism of organic matter. McArthur et al. (2001) speculated
on the impact of peat beds within aquifers as a source for this organic matter. Their study showed that
there was a correlation between areas with high groundwater arsenic levels and the location of these peat
beds. Simeoni et al. (2003) studied the impact of fulvic acid on As5+ sorption to ferrihydrite. It was found
that fulvic acid can both inhibit sorption and displace sorbed arsenic from the surface; the fulvic acid
could reduce ferrihydrite at a rate of 30 nM/h. In addition, humic and fulvic acids can affect the rate of
microbial ferrihydrite reduction. Wolf et al. (2009) found that both humic acid and fulvic acid accelerated
the reduction of ferrihydrite through electron shuttling in the presence of microorganisms. The reducing
capacity of humic acid was 1.8 eq/mol, compared to 0.6 eq/mol for fulvic acid. In addition, citric acid can
prevent As5+ sorption onto ferrihydrite and both citric and fulvic acids can prevent As3+ sorption onto
ferrihydrite (Grafe et al., 2002).
Reduction is not limited to ferric iron within the mineral matrix; reduction of arsenic on the
mineral surfaces can also occur. The efficacy of ferrihydrite as an arsenic sorbent stems from its ability to
form surface complexes with aqueous As3+and As5+. The impact of arsenic speciation on this binding has
been studied extensively (Goldberg and Johnston, 2001; Cheng et al., 2009; Tufano et al., 2008; and
Sveijensky and Fukushi, 2006). Goldberg and Johnston (2001) found that As5+ forms inner-sphere
82
-------
complexes with amorphous Fe oxides, while As3+forms both inner- and outer- sphere complexes. Inner-
sphere complexation refers to the specific binding of the arsenic ion to the mineral surface, while outer-
sphere complexation refers to the electrostatic interactions between the ion and surface (Cheng et al.,
2009). As5+ forms a bidentate binuclear complex, meaning that each arsenic anion chemically forms two
bonds with the surface metal ions. This type of surface complexation is stronger and less likely to
dissociate than the monodentate binuclear complex which forms through the singular bond of As3+ anions
to ferrihydrite. This inner-sphere As3+ complex will dominate under high surface coverage, whereas lower
coverage will result in binuclear outer-sphere complexes (Cheng et al., 2009). As a result of the weaker
binding of As3+to ferrihydrite compared to As5+, Peters and Burkert (2008) and Smedley and Kinnburgh
(2002) showed the reduction of As5+on the ferrihydrite surface will result in a certain degree of
remobilization. Increases in pH can also lead to desorption of As5+.
Tufano et al. (2008) compared the reduction of arsenic and iron in a system containing As5+
sorbed on ferrihydrite. This was accomplished by exposing arsenic-loaded ferrihydrite to mutant
microbial strains which reduced either As or Fe, as well as Shewanella sp. ANA-3, which reduced both.
They found that the largest proportion of arsenic was released due to the As reduction, and the smallest
proportion was released due to Fe reduction. This was explained by an equilibration between dissolution
and reprecipitation of ferrihydrite which can occur during iron reduction. Because ferrihydrite is less
thermodynamically stable, some magnetite (Fe304) in addition to ferrihydrite will precipitate following
reductive dissolution and can further remove arsenic from the system. In addition, intermediate phases
such as green rust [Fe4+2Fe2+3(0H)i2S04 • 3H2O] can form during iron oxyhydroxide mineral precipitation
and will preferentially sorb aqueous arsenic (Jonsson and Sherman, 2008; Randall et al., 2001).
Interestingly, this process was not observed for arsenic sorbed to a more stable mineral, such as goethite,
or under arsenic reducing conditions. The net result is the increased arsenic mobilities in these systems.
These findings are consistent with those of Fendorf et al. (2010) in a study of arsenic variation in
groundwater in South and Southeast Asia. This study speculated that the primary source of arsenic
mobilization in the case described was the weathering of coal and arsenic-containing sulfides. Initially,
arsenic will be transferred to iron oxides; however, arsenic can be freed through microbial reduction of
Fe3+ to Fe2+ or As5+to As3+by organisms described in the following section, and through competition for
surface sites by other ligands such as phosphate and carbonate.
2.3 Impact of Microbial Activity
Often, the reduction-oxidation potential of the system is dictated by microbial activity within the
aquifer. Table A-4 contains a summary of native bacteria which have been studied due to their ability to
metabolize the minerals of concern. The impact of microbial activity on the kinetics of arsenic release is
critical to consider because they can catalyze both the oxidative dissolution of arsenic-bearing pyrite and
the reduction of ferrihydrite (O'Day et al., 2004; Islam et al, 2004; Corkhill and Vaughan, 2009). There
must be a supply of organic carbon for microbial activity to be significant. Therefore, the rate of arsenic
release will be faster in the presence of reactive organic carbon and for mineral-bound arsenic, as
compared to the presence of recalcitrant organic carbon and for arsenic bound to Fe oxides. Under
anaerobic conditions within the aquifer, there could be native iron-oxidizing or -reducing bacteria, as well
as sulfur-oxidizing or -reducing bacteria.
McGuire et al. (2001) studied changes in mineral surface speciation during microbial-mediated
dissolution in order to study sulfide mineral dissolution in natural systems. For this experiment, a sulfur-
oxidizing (Thiobacillus caldus), iron-oxidizing (Ferroplasma acidarmanus), and a mix of sulfur and iron-
oxidizing microorganisms (T. caldus, F. acidarmanus, and Leptospirillum ferrooxidans) were utilized.
Bacteria were cultured with a mixture of crushed and polished samples of arsenopyrite, marcasite, and
pyrite. Cell growth and structural changes on the single crystal surface were observed using scanning
83
-------
Table A-4: Microbes Impacting Fe/As Oxidation and Reduction in Aquifers.
Bacteria Name
Description
Reference
Desulfotomaculum
auripigmentum
Bacteria capable of reducing As(V) to As(lll) and
precipitating orpiment or amorphous As2S3
O'Dayetal. (2004)
Thiobacillus caldus
Sulfur-oxidizing bacteria
McGuire et al. (2001)
Ferroplasma acidarmanus
Iron-oxidizing bacteria
McGuire et al. (2001)
Leptospirillum
ferrooxidans
Iron-oxidizing bacteria
McGuire et al. (2001)
Pseudomonas species
Used in bioremediation, can be pathogenic
Islam et al. (2004)
Clostridium species
Anaerobic bacteria, grow best on carbohydrates
Islam et al. (2004)
Nitrosolobus species
Ammonia-oxidizing bacteria
Islam et al. (2004)
Thiobacillus ferrooxidans
Iron-oxidizing bacteria that thrives in acidic
environments
Jones and Pichler
(2007)
Pseudomonas
arsenitoxidans
Bacteria capable of oxidizing arsenopyrite
llyaletdinov and
Abdrashitova (1981)
electron microscopy (SEM) imaging. For samples reacted with the iron-oxidizing or mixtures of bacteria,
SEM imaging revealed discrete dissolution pits on the pyrite surface, while marcasite and arsenopyrite
developed rough surfaces and linear dissolution pits. Elemental sulfur deposits were also observed on the
arsenopyrite surface during both abiotic control experiments and those utilizing iron-oxidizing bacteria.
For samples reacted with the sulfur-oxidizing bacteria, dissolution pits were observed on all samples,
similar to those seen in the abiotic control, however no sulfur deposits were observed. This is likely due
to the bacteria oxidizing sulfur at the surfaces. Because iron concentrations increased linearly over the
course of the reaction, they concluded that little or no iron minerals were being formed in the
experimental system. Therefore, changes in the iron concentration were used to quantify dissolution rates.
Analysis showed that pyrite dissolved at nearly one-sixth the rate than arsenopyrite and half the rate of
marcasite. It was also observed that arsenopyrite dissolution was enhanced by the presence of both the
iron-oxidizing bacteria, F. acidarmanus, and the mixture of F. acidarmanus, T. caldus, and L.
ferrooxidans, while marcasite dissolution was only enhanced by the mixture. These enhanced rates
resulted from the regeneration of ferric iron by the bacteria.
Arsenopyrite surface morphology changes by acidic, oxidative dissolution by Thiobacillus
ferrooxidans in the presence of a number of salts, including phosphate, were also studied by Jones et al.
(2003). Within a week's time, a layer of Fe+3PC>4 had formed on the surface as a result of iron oxidation.
This layer was not observed during abiotic control experiments. For the system containing T.
ferrooxidans, it was also observed that this Fe+3PC>4 layer did not prevent oxidation and dissolution of the
arsenopyrite below, despite the surface coating. Because the bacteria could not have reached the
arsenopyrite below the overlayer in the experimental systems, they concluded that bacteria may not need
to be attached to the mineral surfaces to promote arsenopyrite dissolution.
Another study conducted by Islam et al. (2004) looked at the mobilization of arsenic from
sediment samples by a mixture of native bacteria, which included Fe+3-reducing and As+5 reducing
bacteria. Results showed that the reduction of As+5 and Fe+3 by this bacteria mixture, following
stimulation by acetate, was not coupled, and that Fe+3 was reduced first, possibly due to its higher redox
potential. In addition, the native As+5 reducing bacteria were not obligate As+5 reducers, meaning that they
can utilize other species as electron acceptors, including Fe+3, before they utilize As+5. When a culture of
84
-------
anaerobic Fe+3-reducing bacteria was used to observe arsenic mobilization, it was also found that arsenic
was freed from sediment samples and a higher concentration of As+3 was measured as compared to As+5,
indicating that the bacteria may have reduced As in addition to Fe.
Some microbes in the aquifer may also be capable of oxidizing arsenite or reducing arsenate.
These include dissimilatory arsenate-respiring prokaryotes (DARPs), heterotrophic arsenite oxidizers
(HOAs), and chemoautotrophic arsenite oxidizers (CAOs) (Oremland and Stolz, 2003). Some strains are
capable of oxidizing arsenite and respiring arsenate (Handley et al., 2009). The microbial arsenic cycle
begins with the oxidation of As3+ within the aquifer to As5+. This can be a result of arsenite oxidizers, or it
can be triggered by human activity such as digging wells and depleting the water table, which provides
both oxidants, such as oxygen and nitrogen, and additional biomass. Evidence of this microbial cycle has
been found in groundwater aquifers in Bangladesh (Harvey et al., 2002). Furthermore, reclaimed water
itself can contain bacterial contaminants. Therefore, it is crucial to examine the impact of both potential
microbes in the injected reclaimed water and native bacteria on mineral dissolution.
3. Experimental Investigations
Although many studies exist on groundwater-arsenopyrite interactions and the subsequent fate
and transport of arsenic in groundwater, no study to date has fully addressed the unique scenario ASR
using reclaimed wastewater. This is in part due to the complicated nature of the interactions, as reclaimed
water not only has many constituents, but also its composition will not be constant during a single ASR
operation or between different ASR sites. Therefore, this study characterized the potential interactions
between prevailing reclaimed water components and arsenic-bearing minerals.
The experimental approaches and procedures were provided in Neil et al. (2014). Major
experiments include:
• Aqueous phase batch reactor testing of dissolution rates. In this set of experiments, the arsenopyrite
dissolution rate were determined under different experimental conditions. Zero-order reaction kinetics
were confirmed by the linear concentration evolution of arsenic in the reactor (trend lines in Figure
A-3). Each batch reactor contained 250 mL of the reaction solution and 0.05 g of acid-washed FeAsS
powder. Reactors were continuously stirred, and temperature was controlled at 5, 22, or 35 ± 1°C
using a hot water or ice bath. 2 mL samples were removed at 0, 0.5, 1, 2, 3, 4, 5, and 6 hours and
filtered immediately using a 0.2 ^m polytetrafluoroethylene (PTFE) membrane syringe filter and
capped to prevent evaporative losses. This time frame was chosen to minimize the effect of secondary
mineral precipitation on aqueous arsenic levels. Finally, samples were acidified to 2% v/v acid and
arsenic concentrations were measured using ICP-MS. At least three experimental replicates were run
to confirm arsenic remobilization trends.
• Characterization of Secondary Mineral Precipitate Morphology and Mineralogy. Changes on the
arsenopyrite mineral surface were examined using polished arsenopyrite thin sections, called
"coupons." Tapping mode AFM (AFM, Veeco Inc.) was used to characterize secondary mineral
precipitates on arsenopyrite coupons by measuring changes in the height, amplitude and phase over
the 7 day reaction period. AFM tapping mode probes were 125 |im long with phosphorus (n) doped
silicon tips (nominal tip radius of 10 nm, MPP-11100-10, Bruker probes). In addition, Raman
spectroscopy was conducted using an inVia Raman Microscope (Renishaw, UK) on reacted
arsenopyrite in order to identify secondary mineral precipitates. Raman measurements were carried
out with a 514 nm laser and a grating of 1800 lines/mm. A 20x objective and decreased power were
used to limit the energy density of the laser, preventing artificial phase transformation of secondary
mineral precipitates (Modesto Lopez et al., 2009).
85
-------
4. Investigation Results
4.1. Arsenic Dissolution Rate
Evolution of Aqieous Arsenic Cmcentration
Figure A-3 shows the arsenic concentration changes with time in aqueous phase for 10 mM
sodium nitrate, 10 mM sodium chloride, and wastewater under aerobic (Al, Bl, and CI) and anaerobic
(A2, B2, and C2) conditions. Among aerobic systems, the highest arsenic mobility was observed in the 10
mM sodium chloride system. Arsenic concentrations were similar between the wastewater and sodium
nitrate systems. The only difference between the two model wastewater systems was the presence of
nitrate versus chloride anions, neither of which are expected to interact significantly with arsenopyrite in
the presence of dissolved oxygen according to the literature (McKibben et al., 2008). In addition, neither
nitrate nor chloride compete with arsenate for Fe(III) adsorption sites (Rau et al., 2003; Youngran et al.,
2007; and Guo and Chen, 2005). Therefore, differences in the arsenic mobility are not anticipated to
result from changes in the oxidative dissolution of arsenopyrite or sorption of arsenic, but, more likely,
from effects on secondary mineral formation and phase transformation, which further impact arsenic
attenuation.
For the anaerobic system, the highest arsenic mobility was observed in the sodium nitrate system
(up to 0.28 mM), while very low concentrations were observed in the 10 mM sodium chloride and
wastewater systems (up to 0.12 mM and 0.08 mM, respectively). For all systems, arsenic mobility was
lower under anaerobic conditions (15%, 78%, and 76% reductions for nitrate, chloride, and wastewater
systems, respectively, compared to aerobic conditions), indicating the role of dissolved oxygen in the
oxidative release of arsenic from arsenopyrite through reaction (1). The decreased percent reduction in the
anaerobic 10 mM sodium nitrate system compared to wastewater and sodium chloride may be due to the
oxidation of arsenopyrite by nitrate anions in the absence of dissolved oxygen, which is a new
observation. However, more work is needed to confirm this reaction pathway.
Activaion Energy Calculations
For all aqueous systems, the activation energies for arsenic remobilization were calculated using
the Arrhenius' equation. Because zero-order reaction kinetics were observed in the early stages of
dissolution, the slope of the concentration evolution at each temperature (e.g., trend lines in Figure A-3)
was assumed to be equal to the rate constant, k, of the reaction. A larger rate constant would therefore
correlate with higher arsenic concentrations at the end of the 6-hour reaction period. The rate constant, k,
is related to the temperature and activation energy in accordance with the Arrhenius' equation:
k = Ae~E^RT (A 14)
Taking the natural logarithm of this equation gives a linear relationship between the rate constant
and temperature, T:
ln(fc) = ~~~ + ln04) (A 15)
The rate constant k for each reaction condition was determined by calculating the slope of the
best fit trend line for the concentration evolution at each temperature. The natural log of k was plotted
86
-------
2
¦S-
£
m
t
c
¦5
T ime (hours)
2 3 4
T ime (hours)
2
£
£
m
t
E
<
0.6-
1«,
2
0.5-
¦
¦ ¦
0.4-
b
m
*,
0.3-
m
5
0.2-
s
E
0.1-
<
0.0-
0.6 h
5
=•
0.5-
m
0.4-
¦
V
0.3-
w
0.2-
¦
£
0.1-
<
;¦ -
2
=-
&
u
£
E
¦5
2 3 4
T ime (ho its)
2 3 4
T ime (hoirs)
T ime (hours)
2 3 4
T ime (hoirs)
Figure A-3 Aqueous arsenic concentration in batch reactors at 5, 22, and 35°C for (Al) pH 7,
10 mM sodium nitrate, aerobic, (A2) pH 7, 10 mM sodium nitrate, anaerobic, (Bl)
pH, 7 10 mM sodium chloride, aerobic, (B2) pH 7, 10 mM sodium chloride,
anaerobic, (CI) pH 7 wastewater, aerobic, (C2) pH 7 wastewater, anaerobic.
Standard deviations between replicate trials are indicated by error bars.
against the inverse of the temperature and the slope of this line was equal to the negative activation
energy, Ea, divided by the gas constant, R. For the aerobic systems, the calculated activation energies for
87
-------
Table A-5.Empirically determined activation energies for arsenic mobilization from arsenopyrite.
Aqueous Media
Temperatures
(°C)
Activation Energies (kJ/mol)
Aerobic Anaerobic
10 mM Sodium Nitrate
10 mM Sodium Chloride
Wastewater
5, 22, and 35
5, 22, and 35
5, 22, and 35
40.8 ±3.5 31.2 ±3.2
36.9 ±2.3 28.4 ±3.6
43.6 ±5.0 44.1 ±6.3
Note: All reactions were carried out at pH 7.0 ± 0.2. The solid-to-liquid ratio was 250 mL: 0.05
g FeAsS powder. See Neil et al. (2014).
arsenic remobilization were 40.8 ± 3.5, 36.9 ± 2.3, and 43.6 ± 5.0 kJ/mol for 10 mM sodium nitrate, 10
mM sodium chloride, and wastewater, respectively. For the anaerobic systems, the calculated activations
energies for arsenic remobilization were 31.2 ± 3.2, 28.4 ± 3.6, and 44.1 ± 6.3 kJ/mol, for 10 mM sodium
nitrate, 10 mM sodium chloride, and secondary effluent samples from the wastewater treatment plant,
respectively (Table A-5). The activation energies for iron release were not calculated because aqueous
iron levels were below the detection limit during the 6-hour reaction period. This may result from the
reprecipitation of aqueous iron as iron (III) (hydr)oxides.
The literature provides activation energies for a number of minerals related to this system,
including the oxidation of arsenopyrite by dissolved oxygen (57 kJ/mol at pH 5.9) (Yu et al., 2007), and
the reductive dissolution of ferrihydrite (40.7 kJ/mol) (Erbs et al., 2010), hematite (88 kJ/mol) and
goethite (94 kJ/mol) (Sidhu et al., 1981). The range of observed activation energies indicates that the
most likely processes occurring are the oxidation of arsenopyrite by DO or the reduction of ferrihydrite,
because all measured activation energies ranged between 30 and 50 kJ/mol. Decreases in the activation
energy from the aerobic to anaerobic system for sodium nitrate and sodium chloride may indicate a
switching of the dominant mechanism for arsenic release from oxidation of arsenopyrite by DO to the
reduction of ferrihydrite, which may take place at low reduction-oxidation potentials characteristic of
anoxic environments (Pedersen et al., 2006). To confirm these hypotheses, in situ X-ray absorption
spectroscopy (XAS) can be utilized to observe time-resolved change in iron and arsenic oxidation states
and geometry for the different aqueous systems. This necessary testing was not conducted in the research.
For the wastewater system, the activation energy did not change between the aerobic and
anaerobic systems. Despite the lower activation energy for 10 mM sodium nitrate and 10 mM sodium
chloride in anaerobic systems, the mobility of arsenic in these systems was 3.5 times higher for nitrate
and 1.5 times higher for chloride, indicating that other factors, such as the availability of reactants,
prevented remobilization. To investigate these observed trends and to determine secondary mineral
effects on aqueous arsenic mobilization, the differences in secondary mineral formation and phase
transformation were studied between sodium nitrate, sodium chloride, and wastewater systems.
4.2. Secondary Mineral Precipitate Morphology and Mineralogy
Secondary Mineral Morphology and Coverage
Differences in secondary mineral precipitation between the three aqueous systems yield further
insight into the observed trends in arsenic remobilization. Arsenic remobilization depends on the balance
between the dissolution of existing As-bearing minerals and the precipitation of arsenic from the water.
Thus the mineral phase and morphology of secondary precipitant yields a basis to assess the degree and
completeness of arsenic remobilization.
88
-------
A. 10 mM Sodium Nitrate B 10 mM Sort nun Chloride C. Secondary effluent
Figure A-4 AFM height mode images after 1 day (Al, Bl, CI) and 7 days (A2, B2, C2) in the 10 mM
sodium chloride, 10 mM sodium nitrate, and wastewater systems at room temperature
(22°C) and under aerobic conditions. The scan size for these images is 3 microns and the
height scale is 100 nm.
Figure A-4 shows the AFM height mode images after 1 day and 7 days in the 10 mM sodium
chloride, 10 mM sodium nitrate, and wastewater systems at room temperature (22°C) and under aerobic
conditions. Images at additional time points are provided in Figure A-5. For all time points, multiple
images were taken over the entire sample surface to confmn observations. The images in Figure A-4
showed very distinct differences in precipitate morphology between the three systems. For the 10 mM
sodium nitrate system, after 1 day there was a significant amount of small precipitates covering the entire
surface (Figure A-4 Al). After 7 days, these precipitates grew in quantity and size, and at the end of the
reaction period there was a variety of both larger and small particles, indicating continued nucleation and
growth for the entire period (Figure A-4 A2). For the 10 mM sodium chloride system (Figure A-4 B),
particles after 1 day were larger in size and sparse on the surface. After 7 days, these particles appeared to
aggregate to form a coating on the surface. Unlike the sodium nitrate system, there was not much
evidence of continued nucleation because the size and morphology of precipitates was very different
between days 1 and 7. For the wastewater system (Figure A-4 C), there was little precipitation after 1 day
and both the size and morphology of precipitates did not change significantly over 7 days.
Under anaerobic conditions, there was no observed precipitation on the coupons for all three
systems even after 7 days (Figure A-6).
89
-------
Reaction 10 mM sodium nitrate 10 mM sodium chloride Wastewater
Time
6 hours
1 day
7 days
Figure A-5 Tapping mode AFM Images of reacted FeAsS coupons in 10 mM sodium nitrate or 10 mM
sodium chloride. All systems were at pH 7.0 ± 0.2, room temperature, and equilibrated with
atmospheric oxygen. Images are 20 x 20 jam and the height scale is 100 nm.
90
-------
10 mM sodium nitrate 10 mM sodium chloride Wastewater
System
Anaerobic
system
7 days
Aerobic
system
7 days
Figure A-6 Comparison between secondary mineral precipitation in the aerobic and anaerobic systems
for 10 mM sodium nitrate and 10 mM sodium chloride. All systems were at pH 7.0 ± 0.2
and room temperature. Images are 20 x 20 jam and the height scale is 100 nm.
Secondary Mineral Phase Identification
Identification of secondary mineral phases in aerobic systems was accomplished using Raman
spectroscopy. The characteristic spectra for different iron oxide minerals were determined by measuring
standard samples on the Raman instrument (Figure A-7). For the anaerobic system, there was no
precipitation in the AFM image. Neither was detectable by Raman spectroscopy.
Early in the reaction period (<1 day), there was no detectable secondary mineral precipitation on
the surface for any system. In the sodium nitrate system, the characteristic peaks of maghemite (y-Fci-O,).
an iron(III) oxide polymorph, become detectable after 1 day of reaction (Figure A-7 B). By 7 days, the
entire coupon surface in the sodium nitrate system is coated in maghemite (Figure A-7 A). For the sodium
chloride system, no precipitation was detected after 1 day. After 7 days, however, the surface was covered
in a non-homogeneous coating of hematite (a-Fe2C>3) and maghemite (Figure A-7 D). The visual
difference between these two mineral phases is apparent on the arsenopyrite surface (Figure A-7 C). For
the wastewater system, there was no detectable precipitation over the 7-day reaction period. 6-line
ferrihydrite, magnetite, and goethite standards were also considered, but the spectra did not match the
reacted samples.
91
-------
4.3 Water Matrix Effects on
Arsenic Remobilization
Nitrate and Chloride Effects on
Secondary Mineral Phase
Transformation
Hematite is the most
therm odynami cally stable iron
oxide polymorph and is the
final form resulting from the
transformation of less stable
iron(III) (hydr)oxides (Jang et
al., 2007). The occurrence of
hematite in the sodium chloride
system and not the sodium
nitrate system after 7-days
reaction time was confirmed by
multiple replicates. The faster
transformation of iron(III)
(hydr)oxides in the presence of
sodium chloride compared to
sodium nitrate is a new
interesting observation, and can
greatly impact arsenic
remobilization from
arsenopyrite.
Previous research
conducted into the effects of
chloride and nitrate on
heterogeneous and
homogeneous iron(III)
(hydr)oxide nucleation and
growth provides insight into
this phenomenon (Hu et al.,
2012). Using time-resolved
small angle x-ray scattering
(SAXS) and grazing-incidence
SAXS, Hu et al. (2012)
observed that in the presence of
chloride ions, Ostwald ripening
was the dominant process
controlling heterogeneous
precipitation, whereas
continuous nucleation, growth,
and aggregation occurred in the
nitrate system. Ostwald
ripening describes the growth
mechanism wherein smaller
precipitates dissolve and
B
£
a
Uftrsac ed 1 day -
Magheaite S tandarti
1
500
1000 1500
Wavelength (cm )
2000
Spot 2
D
MsEtea It* st c Itai&te sto
'I
A lb
500
1000
1500
2000
Wavelength (an L)
—| Lfifasc fr£ 1 day " £ys|
A
i yyfiMki UniiihiMi
.1 'IPAjWip 1 ^ n\
—I—
500
1000 1500
Wavelength (cm1)
2000
Figure A-7 Optical microscope images and Raman spectra for
arsenopyrite coupons reacted in sodium nitrate (A, B),
sodium chloride (C, D); and wastewater (E, F) systems.
Optical microscope images for the 7-day sodium nitrate
system (A) shows a uniform coating of maghemite, as
indicated by the characteristic Raman peaks (B). For the
sodium chloride system, after 7 days, the surface was
covered in a non-homogeneous coating (C) of hematite
(a-Fe203) and maghemite (D). No precipitation was
observed in the wastewater system (E, F).
92
-------
redeposit on the surface of larger precipitates, resulting in an increase in particle size while the total
number of particles decreases. Ostwald ripening is a spontaneous process; the formation of larger
particles with smaller surface-to-volume ratios is more stable and energetically favorable. These reported
differences in the iron(III) (hydr)oxide growth mechanisms are observable in AFM images of
arsenopyrite coupons after 1 and 7 days reaction time(Figure A-4). In the sodium nitrate system, small
particles are always visible on the surface in addition to larger aggregates, indicating continued
nucleation, growth, and aggregation. In the sodium chloride system, larger particles are visible after just 1
day. The difference in morphology is even more obvious after 7 days, with the formation of a uniform
iron(III) (hydr)oxide coating. There is a lack of smaller precipitates in both the 1- and 7-day samples,
indicating that primary particles may have gone through Ostwald ripening processes.
The prevalence of Ostwald ripening as a growth mechanism can also explain the faster phase
transformation observed in the sodium chloride system. Ostwald ripening usually occurs in the late stages
of first-order phase transformations (Slezov, 2009). During Ostwald ripening, metastable Fe3+(hydr)oxide
nanoparticles are dissolving and recrystallizing on the surface of larger particles to minimize surface free
energies. The larger particles formed through Ostwald ripening will tend to be more thermodynamically
stable than their nanoscale precursors, resulting in the phase transformation of less stable Fe3+ (hydr)oxide
polymorphs such as ferrihydrite into more stable forms, such as maghemite and, eventually, hematite
(Dubinina and Lakshtanov, 1997; Liu and Zeng, 2005).
This phenomenon will have secondary effects on arsenic motilities in the sodium nitrate and
sodium chloride systems. Increased iron(III) (hydr)oxide nucleation in the sodium nitrate system leads to
a large number of smaller particles. The high cumulative surface area of these precipitates can lead to
more available surface sites for the sorption of aqueous arsenic anions, resulting in lower arsenic
concentrations. This mechanism is consistent with observations of arsenic remobilization from
arsenopyrite in the sodium nitrate system as compared to sodium chloride. With increased reaction time,
iron(III) (hydr)oxide undergoes aging processes to form maghemite in the sodium nitrate system and a
mixture of maghemite and hematite in the sodium chloride system. Hematite, due to its increased
crystallinity, has less sorption capacity for arsenic than maghemite (Park et al., 2009). However, it is
important to note that the transformation of iron(III) (hydr)oxides into more stable iron(III) oxide
polymorphs can lead to the irreversible sorption of associated arsenic anions. Therefore, although these
systems will have less capacity for arsenic sorption, the arsenic attenuated by the iron(III) (hydr)oxides in
early stages will become strongly bound within the iron(III) oxide matrix (Peterson and Burkert, 2008).
This inferred trapping mechanism can be beneficial for the long term fate and transport of arsenic in oxic
or hypoxic groundwater systems where ferric iron minerals are stable.
Inhibited Secondary Mineral Precipitation in the Wastewater System
Interestingly, this research observed no precipitation in the system containing wastewater in
comparison to both the sodium nitrate and sodium chloride systems. Currently, there are no studies which
have reported on this apparent inhibition of Fe3+(hydr)oxide precipitation. Nonetheless, studies which
model arsenic remobilization during ASR operations have assumed the formation of ferrihydrite as an
attenuation mechanism during arsenic transport in ASR (Willis et al., 2010).
This possibility was further examined by monitoring the reduction-oxidation potential (ORP) and
pH over the 7-day reaction period for the wastewater, sodium nitrate, and sodium chloride aqueous
solutions. ORP is a measure of the tendency of the solution to gain or lose electrons. A positive redox
potential indicates oxidizing conditions, meaning that the aqueous solution is more likely to gain electrons
from arsenopyrite, thereby becoming reduced while arsenopyrite is oxidized. Evolution trends in pH and
ORP can be found in Figure A-8.
93
-------
Time (days)
Nitrate # Chloride A Wastewater
-100
Time (days)
Nitrate + Chloride A Wastewater
Figure A-8 Evolutions of pH and ORP in batch reactors over the 7-day reaction period. The pH value
was not adjusted over this time. All reactors were at room temperature (22°C) and open
to the atmosphere (P02 = 0.21 atm).
For the 10 mM sodium nitrate and sodium chloride systems, similar evolutions were observed for
pH and ORP measurements. pH decreased steadily over the 7-day period. This is likely due to the
continuous oxidative dissolution of arsenopyrite through reaction Eq. 1, which produces arsenous acid.
For the wastewater system, the pH fluctuated between 7.0 and 8.4 over the reaction period. It is likely that
the wastewater effluent from the Cincinnati Milk Creek treatment plant contains a multitude of buffering
agents, such as bicarbonate ions, in addition to the metal ions. Dewettinck et al (2001) examined the
buffering role of the wastewater constituents. These buffering agents may prevent decreases in pH. At a
lower pH, increased arsenic remobilization would occur due to proton-promoted dissolution. Higher iron
concentrations would result in higher saturation indices with regards to iron(III) (hydr)oxide precipitates.
However, the higher pH in the wastewater should also contribute to higher saturation indices due to the
increased hydroxide ion concentration (reaction Eq.2). Because water chemistry effects on iron(III)
(hydr)oxide saturation indices are contradictory, additional factors may be contributing to the inhibited
precipitation for wastewater.
The ORP values provide further insight into precipitation trends. The ORP increased over the 7-
day period and was generally positive for the sodium nitrate and sodium chloride systems. In contrast, the
ORP in the wastewater system fluctuated but always remained negative over the reaction period. The
formation of iron(III) (hydr)oxides is contingent on the oxidation of Fe2+, released through reaction eq. 1,
to Fe3+. The negative redox potential in the wastewater system indicates that the condition is a reducing
environment for arsenopyrite. This could prevent the oxidation of Fe2+and precipitation of iron(III)
(hydr)oxides, a process consistent with experimental observations.
The lower ORP conditions in the wastewater system are prevalent in reclaimed wastewater. In
secondary wastewater treatment, low ORP conditions are necessary to facilitate biological denitrification
and phosphorus removal processes. These redox reactions are fed by the addition of dissolved organic
carbon (DOC) serving as the electron donor. Although much of the DOC present in wastewater is
removed prior to effluent discharge and reuse, the DOC levels can still be elevated when compared to
groundwater concentrations. In the experimental investigation reported here, the wastewater samples had
a non-purgeable TOC concentration of 12.42 mg/L, while concentrations in the two model systems were
negligible. This factor may be the root of observed differences in precipitation, as the presence of DOC
would prevent the oxidation of Fe2+. However, uncertainty exists making it necessary to quantify the
94
-------
effects of organic TOC on both arsenic remobilization from arsenopyrite and its correlation with
heterogeneous and homogeneous iron(III) (hydr)oxide nucleation and growth.
5. Manage Aquifer Conditions for Reduced Arsenic Remobilization
5.1. The hydrogeological factor
In the preceding sections, the literature review and experimental investigations strongly suggest
that multiple environmental factors can facilitate arsenic remobilization into groundwater, and at the same
time, precipitate dissolved arsenic from the groundwater. Chemical and biologically enhanced arsenic
remobilization in aquifers can occur only under specific groundwater conditions, high or low oxidation-
reduction potential, dissolved oxygen presence, and low pH, as discussed in preceding sections (e.g.
Table 2). The extent to which these environmental conditions evolve within a timeframe and location,
depends on the geochemical compatibility of injected water and the hosting aquifer, abiotic and biotic
processes that change the pH and redox conditions, and the water injection process. The resulting
environmental changes also evolve in response to the local and regional groundwater flow pattern. These
factors are generally location-specific and application-dependent, as potential groundwater changes can
enable arsenic remobilization through the arsenic remobilization and reduction reactions in Table A-2.
Geochemical and hydrological flow simulations have been reported for a number of ASR sites
(Stamos et al., 2001; Pavelic et al., 2006; Grovea and Wood, 1979; Lawrence and Upchurch, 1982;
Powelson et al., 1993; Sharif et al., 2008). Literature review and analysis points to the linkage of
groundwater flow regimes and the potential changes in groundwater, soil matrix, and the injected water,
to the changes in groundwater chemistry, and consequently arsenic remobilization. One principal change
occurring during ASR is that the injection-induced radial groundwater flow generates groundwater
zoning, stratification and hence changes in pH-Eh conditions. Other major factors are the water injection
and withdrawal cycle, producing groundwater table fluctuations that result in interchange of the air-filled
vadose zone and upper groundwater layer.
At an ASR site, injection-withdraw or recharging operations lead to local changes in natural
groundwater flow, and consequently the mixing and reactions of oxygenated injection water with native
groundwater. Localized elevation of the water table, known as groundwater mounding, has been observed
since 1961 as a consequence of lateral permeability restrictions to the dissipation of hydraulic head at
injection facilities (Bouwer, 2002; Todd, 1961). The occurrence of groundwater stagnancy in a "bubble"
and "bottle brush" of reclaimed water (Figure A-9), which form due to a lack of mixing between the
injected water and groundwater, varies significantly as a result of aquifer heterogeneity, preferential flow
pathways, leakage and buoyancy, and soil chemical makeups (Lowry and Anderson, 2006; Vacher et al.,
2006; Clark et al., 2004). These subsurface aquifer properties contribute to variations in the water
injection rate and areal extent, and can thus influence many environmental factors controlling arsenic
mobility on local and micro-scales. This explains the large spatial and temporal changes often found in
groundwater arsenic concentrations (Hoque et al., 2009;). In a macro-scale, however, the "bubble" and
"bottle brush" concept in the ASR process (Figure A-9) prescribes two major types of macro-scale
physical boundaries and geochemical domains. In these areas, the injected water interacts with native
groundwater and aquifer formation (Vacher et al., 2006). Figure A-9 generalizes the ASR process and
resulting mixing-replacement phenomenon - limited mixing of injected water and ambient groundwater at
the perimeter and displacement of the native groundwater forming a hydraulic and geochemical bubble.
This generalization agrees with the geochemical and hydrological studies showing that zoned
flow fields marked by ages and fractions of injected water prevail in the injection "bubble" as it spreads
toward undisturbed aquifers (Clark et al., 2004; Ma and Spalding, 1996; Brown and Misut, 2010). On the
other hand, the geochemical conditions can vary significantly at an ASR site depending upon the degree
of aquifer anisotropicity, thermal and density difference between injected water and native groundwater,
95
-------
and the convective flow in the frontier of
the "bubble" during injection and recovery
phases. These factors affect both the ASR
recovery rate, and the conditions of soil-
water reactions (Pavelic et al., 2006; Lowry
and Anderson, 2006; Ward et al., 2008,
2007; Langevin, 2008; Minsley et al.,
2011).
The macro-scale "bubble" or
"bottle brush" formation coupled with
micro-scale soil-water interactions sets up
the basis to investigate the soil-water
interactions in Table A-2 for safer and
more sustainable ASR planning, design and
operations. It must be noted that the spatial
zoning of environmental conditions is not
static, but subject to vertical groundwater fluctuations. ASR operations can change local groundwater
flows. The injection-pumping cycles affect hydraulic communication with surface water bodies.
Groundwater fluctuation introduce both oxygen and labile organic matter to shallower aquifer zones
(generally <150 m), leading to the increased mobilization of arsenic in these zones as compared to deeper
aquifers (Ahmed et al., 2004). The widespread pumping from these deeper aquifer strata for irrigation or
other uses are found to have introduced vertical communication of frequently arsenic-contaminated
shallow groundwater and have consequently eradicated deep aquifers as a arsenic-free water resource
(Burgess et al., 2010).
Therefore, the vertical distribution of arsenic requires attention for investigation at ASR sites. The
trend can be site specific and the causes vary. Yu et al. (2003) reported decreased arsenic concentrations
with increased depth in aquifers in Bangladesh, Harvey et al. (2006) studied the geochemical profile of
the Ganges Delta in South Asia. They found a maximum in arsenic concentration at 30 m depth, and
hypothesized that is was correlated to the zone where older water mixes with younger, recharge water.
Kinniburgh et al. (2003) and McArthur et al. (2004) observed similar bell-shaped arsenic depth profiles in
their sites in Bangladesh and West Bengal, respectively. Decreased arsenic concentrations at very shallow
depths (<15 m) may be due to arsenic adsorption or co-precipitation with insoluble ferrosoferric
hydroxides (Kim et al., 2002).
Seasonal fluctuations in groundwater levels may also impact arsenic mobility even during non-
pumping conditions and thus add complexity in characterization and mitigation. Such an impact in the
Ganges Delta region was studied by Harvey et al. (2006). They observed that for the months of June
through November, a uniformly elevated water table caused by heavy rainfall warranted a small
groundwater gradient. After the raining season, groundwater discharge into the river and groundwater
pumping during the spring for irrigation led to increased groundwater flow and a lower groundwater
table. This cycle of groundwater recharge and discharge could provide a pathway for arsenic mobilizers,
such as organic carbon or oxidants, to enter the aquifer. Seasonal temperature variation can also impact
the reduction-oxidation potential within the groundwater (Greskowiak et al., 2006). Warmer temperatures
during the summer lead to increased microbial activity, subsequently decreasing dissolved oxygen and
nitrate levels and leading to a reductive potential, while during the winter, decreased microbial activity
Injection
1. Injected secondary
water "bubble"
2. Butter zone
3. Native around water
Figure A-9. ASR Bubble formation during secondary
water injection
96
-------
leads to increased dissolved oxygen, and a more oxidizing potential. Harvey et al. (2006) attributed these
hydrological changes as the cause for observed variation of groundwater arsenic concentrations.
Additionally, the wetting and drying history of an aquifer formation can significantly affect its
infiltration rate and thus the time of water residence in the vadose zone and upper groundwater layer. This
occurs because frequent wetting and drying cycles lead to soil hydrophobicity. Arye et al.(2011) observed
this process at an ASR site in Tel-Aviv, Israel, where reclaimed water was recharged through the use of
an infiltration basin, increasing the number of wetting and drying cycles. Due to the sorption of
hydrophobic substances in the reclaimed water, the soil became hydrophobic leading to the retardation of
reclaimed water in the vadose zone and top soil. Hydrophobic soil also tends to sequestering organic
carbon, which may be beneficial in mitigating arsenic mobilization (Spaccini et al., 2002).
5.2 The chemistry factors and geochemical processes
The experimental studies, along with the literature reviewed, indicate several geochemical
conditions in control of arsenic mobility and remobilization into groundwater. Redox condition is one of
the primary factors. The redox cycling of iron in the Earth's subsurface regulates the fate and transport of
many elements of concern. Anthropogenic processes such as ASR can have a drastic effect on the redox
potential of groundwater environments, triggering the oxidative dissolution of reduced iron minerals
including arsenopyrite. In reaching a geochemical steady-state condition, arsenic released from the
remobilization into groundwater is counter-balanced in attenuation processes. These include the
precipitation of iron-oxyhydroxides including ferrihydrite with strong arsenic sorption capacity, and the
precipitation of arsenic-containing minerals.
Formatiai of Iron-oxvhvd'oxides
As discussed in preceding sections, the breakdown of iron oxyhydroxides leads to arsenic
remobilization and the formation of these iron oxyhydroxide minerals also contribute to arsenic
attenuation in aquifers. The dissolution of arsenopyrite produces an abundance of aqueous iron, resulting
in supersaturation with respect to a number of different iron oxyhydroxides or oxide minerals. The
association of arsenic with iron oxyhydroxides or oxide minerals is well documented (McGuire et al.,
2001; Dixit and Hering, 2003; Nickson et al., 1998; Bowell, 1994; Kneebone et al., 2002; Raven et al.,
1998; Richmond et al., 2004; Cances et al., 2005; and references therein). Cances et al. (2005) found that
of the total arsenic present in arsenic-contaminated soil samples, less than 10% was readily mobilized, 10-
37% was sorbed, and more than 65% was associated with iron oxyhydroxides. This arsenic occurred
primarily as As+5, with very small proportions (-7%) of As+3, as would be expected due to the chemical
and structural affinity of these minerals to sorb As+5 preferentially over As+3.
Dixit and Hering (2003) examined how changes in arsenic speciation impact their sorption
behavior onto amorphous ferrihydrite and goethite. For arsenate, it was found that the maximum sorption
density was much higher for ferrihydrite compared to the other minerals, which may be related to the
amorphous structure of ferrihydrite having a higher surface area. However, the capability of ferrihydrite
to sorb arsenic is extremely pH and oxidation state dependent. For example, at pH=10, goethite sorbed
more arsenate than ferrihydrite. For arsenite, goethite sorbed more than ferrihydrite at both high and low
concentrations. However, these results seem to conflict with previous studies which have reported
decreased adsorption capacities for aged ferrihydrite, which is expected to contain some goethite, up to
PH 9 . 129 Additional experiments showed if the water contains high levels of phosphates, As(III) will be
97
-------
preferentially sorbed rather than As(V), because the phosphates will compete for As(V) sorption sites
(Dixit et al., 2003).
Ferrihydrite is only thermodynamically stable at nanoscales (Navrotsky et al., 2008). Overtime, it
will undergo phase changes and can ultimately become one of a number of iron oxide polymorphs.
Dissolution and reprecipitation of ferrihydrite in aqueous systems result in the formation of goethite, and
eventually hematite, the most thermodynamically favorable bulk mineral. This characteristic results in the
unique capability of ferrihydrite to irreversibly immobilize contaminants which are adsorbed during phase
transformation. However, once ferrihydrite transforms into goethite or hematite, which both have
crystalline structures, the adsorptive capability of the iron oxides for foreign ions decreases under
circumneutral pH conditions due to the decrease in reactive surface adsorption sites (Dixit and Hering,
2003).
Arsenie-sulfide precipitation
In addition to arsenic association with iron oxyhydroxides, aqueous arsenic ions may become
more permanently incorporated into aquifer minerals through the precipitation of arsenic sulfides,
including orpiment and realgar (Cances et al., 2005, 2008; O'Day, 2006). This reaction will only occur in
arsenic-impacted aquifer under reductive conditions, as oxidative conditions would lead to the formation
of sulfate as opposed to sulfide. Acidic conditions within the aquifer, leading to a pH<4, can also trigger
the transformation of arsenopyrite to realgar or orpiment (Craw et al., 2003). Furthermore, arseniosiderite
(CaFe3(As04)303 '3H20) can be formed as an oxidation product of realgar and orpiment, as well as
arsenopyrite; this mineral is soluble and will not provide mitigation comparable to its precursor. Arsenic-
containing minerals formed under reducing conditions where arsenic concentration was within |o,M level,
exceeding the solubility of secondary mineral phases (O'Day et al., 2004). Arsenic precipitated from
solution as realgar (AS4S4) when sulfur concentrations were low compared to iron or orpiment (AS2S3)
when sulfur concentrations were high compared to iron. However, there was no evidence of
coprecipitation of arsenic with iron-sulfide minerals. When arsenic concentrations were low, the solubility
of these minerals was not exceeded and the only arsenic sequestration occurred through weak adsorption
processes. Under reducing conditions, this adsorption occurred on FeS or pyrite, while under slightly
reducing conditions this would occur on a variety of Fe(II, III) oxides or hydroxides. This process was not
a stable means of sequestering aqueous arsenic as compared to the precipitation of arsenic-containing
minerals. The importance of H2S concentration is also noted for its role in the Fe-As-S geochemical
interactions. An abundance of H2S would increase FeS precipitation, thus creating competition for the
sulfur needed for AS4S4 or AS2S3 formation.
Kirk et al. (2004) proposed that arsenic sulfide precipitated resulting in decreased arsenic
concentrations following sulfate reduction in unconsolidated glacial aquifers. On the other hand, Zhu et
al. (2008) introduced similar conditions to a sedimentary rock aquifer containing pyritic black shale and
found that sulfate reduction enhanced arsenic mobilization, possibly through sulfide-arsenide exchange.
Therefore, aquifer mineralogy in ASR sites is a key factor which can dictate sulfide mineral formation.
Notably, ferric ions released from these minerals will form iron(III) (hydr)oxide minerals, attenuating
mobilized arsenic. Groundwater and secondary injected water chemistry can greatly impact the
mechanism and overall potential for concurrent arsenic sorption or co-precipitation. This investigation
showed that the presence of high concentrations of chloride ions will inhibit the continued nucleation of
iron(III) (hydr)oxides. In addition, the promotion of Ostwald ripening could lead to the faster phase
transformation of iron(III) (hydr)oxides to maghemite and subsequently, after 7 days, to hematite. As a
result, the arsenic mobility is higher in systems which contain sodium chloride rather than sodium nitrate.
Lastly, it was determined that the presence of wastewater inhibits iron(III) (hydr)oxide precipitation due
to a decreased ORP for this system.
98
-------
5.3 Pretreatment and monitoring for enhanced reliability
The nano- to micro-scale processes controlling arsenic fate and transport have implications for
ASR planning, design and operation, in order to reduce ASR environmental impact. There are no
established arsenic control guidelines for the implementation of ASR design, in part due to inadequate
knowledge of the soil-water interactions and controlling factors (Asano and Cotruvo, 2004). These
observations on geochemical pathways in the As-Fe-S-Cl-N system have implications for the longer term
fate and transport of arsenic in groundwater aquifers and should be considered when managing arsenic
contamination at ASR sites. Major geochemical inferences include:
• Arsenic mobilization in groundwater is balanced between the oxidative breakdown of host minerals
such as arsenopyrite and the precipitation of iron oxides and iron oxyhydroxide. The latter promotes
co-precipitation or sorption of soluble arsenic in groundwater.
• Arsenic associated with stable iron oxide minerals will be trapped as long as the aqueous environment
is favorable for Fe(III) (e.g., oxidative environments). High TOC content in injected water can
enhance biological activities creating local reductive conditions, potentially leading to the
destabilization of arsenic tapped in iron oxides or preventing arsenopyrite oxidation.
• Activation energies for arsenic mobilization in aerobic and anaerobic systems containing sodium
nitrate, sodium chloride, and wastewater samples were experimentally determined. Differences in
activation energies between the systems indicate that the mechanisms controlling arsenopyrite
dissolution and the propensity for arsenic mobilization can vary with dissolved oxygen presence.
These considerations are the basis for developing ASR monitoring programs, modeling arsenic
fate and transport, and developing pretreatment requirements for injected water. Pretreatment is often a
necessary part of ASR systems (See Figure 1 in the main text). There is currently limited knowledge on
how water pretreatment and water withdrawal affect arsenic mobilization. This investigation on
arsenopyrite-water interactions has revealed the observations below:
• The difference in water chemistry (pH, Eh, ORP, etc.) between the injected water and native
groundwater is the cause of arsenic remobilization in groundwater. Thus, pretreatment of injected
water can minimize adverse geochemical reactions by reducing this difference.
• Several geochemical pathways are involved in the dissolution and precipitation of arsenic-bearing
iron oxyhydroxides, and thus the arsenic mobility. The processes are facilitated by the presence of
DOM, chloride ions, nitrate, sulphur and oxidants (or ORP) under a given pH-Eh condition. On the
other hand, high Fe3+ concentrations in groundwater can lead to iron oxyhydroxide precipitation and
enhanced arsenic encapsulation.
• Biological activities enhanced by DOM can lead to local environmental conditions that promote
reductive iron oxide or iron oxyhydroxide dissolution.
• Injection-withdrawal operation and groundwater cycling can change the environmental conditions in
the ASR formation, thus affecting the arsenic mobility. Predictive modeling of groundwater
hydrology during ASR operation can help when monitoring ASR and developing injected water
pretreatment requirements.
99
-------
6. References
Ahmed, K. M., P. Bhattacharya, M. A. Hasan, S. H. Akhter, S. M. M. Alam, M. A. H. Bhuyian, M. B.
Imam, A. A. Khan and O. Sracek, (2004). Ahmed, K. M., P. Bhattacharya, M. A. Hasan, S. H.
Akhter, S. M. M. Alam, M. A. H. Bhuyian, M. B. Imam, A. A. Khan and O. Sracek, (2004).
Applied Geochemistry, 2004, 19, 181-200. Applied Geochemistry, 2004, 19, 181-200.
Appelo, C. A. J., M. J. J. Van Der Weiden, C. Tournassat and L. Charlet, (2002). Surface complexation of
ferrous iron and carbonate on ferrihydrite and the mobilization of arsenic. 'Environmental Science
& Technology, 2002, 36, 3096-3103.
Arye, G., J. Tarchitzky and Y. Chen, (2011). Treated Wastewater Effects on Water Repellency and Soil
Hydraulic Properties of Soil Aquifer Treatment Infiltration Basins. Journal of Hydrology, 397,
136-145.
Asano, T., and Cotruvo, J. A., (2004). Groundwater recharge with reclaimed municipal wastewater:
health and regulatory considerations. Water Research, 2004, 38, (8), 1941-1951.
Asta, M. P., J. Cama, C. Ayora, P. Acero, and G. de Giudici, (2010). Arsenopyrite dissolution rates in O2-
bearing solutions. Chemical Geology, 273, 272-285.
Bahr, J., T. Grundl, G. Harrington, J. Krohelski, and M. Werner, (2002). A review of aquifer storage
recovery techniques. Wisconsin Department of Natural Resources, 40p.
Blanchard, M., M. Alfredsson, J. Brodholt, K. Wright and C. Catlow, (2007). Arsenic incorporation into
FeS2 pyrite and its influence on dissolution: A DFT study. Geochimica et Cosmochimica Acta,
71, 624-630.
Bowell, R. J., (1994). Arsenic sorption by Iron oxy- hydroxides and oxides. Applied Geochemistry, 9,
279-286.
Bouwer, H., (2002). Artificial recharge of groundwater: hydrogeology and engineering. Hydrogeology
Journal, 10, 121-142
Brostigen, G., and A. Kjekshus, (1970). Compounds with the Marcasite Type Crystal Structure. V. The
Crystal Structures of FeS2, FeTe2, and CoTe2. A eta Chem Scandinavica, 24, 1925-1940
Brown, C. J., K. Hatfield, M. Newman, (2006). Lessons learned from a review of 50 ASR projects from
the United States, England, Australia, India, and Africa increasing freshwater supplies. In:
Proceedings of the 2006 UCOWR/NIWR Annual Conference, Santa Fe, New Mexico, USA.
Brown, C. J., and P. E. Misut, (2010). Aquifer geochemistry at potential aquifer storage and recovery sites
in coastal plain aquifers in the New York City area, USA. Applied Geochemistry, 25, 1431-1452.
Buckley, A. N., and G. W. Walker, (1988). The surface composition of arsenopyrite exposed to oxidizing
environments. Applied Surface Science, 35, 227-240.
Burgess, W. G., M.A. Hoque, H.A. Michael, C.I. Voss, G.N. Breit, and K.M. Ahmed, (2010).
Vulnerability of deep groundwater in the Bengal Aquifer System to contamination by arsenic.
Nature Geoscience, 3, 83-87.
Busbee, M.W., B.D. Kocar, S.G. Benner, (2009). Irrigation produces elevated arsenic in the underlying
groundwater of a semi-arid basin in Southwestern Idaho. Applied Geochemistry 24, 843-859
Cances, B., F. Juillot, G. Morin , V. Laperche , L. Alvarez, O. Proux, J-L Hazemann , G. E. Brown Jr. ,
and G. Calas, (2005). XAS evidence of As(V) association with iron oxyhydroxides in a
100
-------
contaminated soil at a former arsenical pesticide processing plant.. Environmental Science &
Technology, 39, 7
Cances, B., F. Juillot, G. Morin, V. Laperche, D. Polya, D. J. Vaughan, J. L. Hazemann, O. Proux, G. E.
Brown, Jr. and G. Calas, (2008). Changes in arsenic speciation through a contaminated soil
profile: A XAS based study. Environmental Science & Technology, 397, 178-189.
Cheng, H.; Hu, Y.; Luo, J.; Xu, B.; Zhao, J., (2009). Geochemical processes controlling fate and transport
of arsenic in acid mine drainage (AMD) and natural systems. Journal of Hazardous Materials,
165(1-3), 13-26.
Clark, J. F., G. B. Hudson, M. L. Davisson, G. Woodside and R. Herndon, (2004). Geochemical imaging
of flow near an artificial recharge facility, Orange County, California. Ground Water, 42, 167-
174.
Corkhill, C. L., and D. J. Vaughan, (2009). Arsenopyrite oxidation - A review. Applied Geochemistry,
24, 2342-2361.
Craw, D., D. Falconer, and J.H. Youngson, (2003). Environmental arsenopyrite stability and dissolution:
theory, experiment, and field observations. Chemical Geology, 199, 71-82.
Dewettinck, T.; Van Hege, K.; Verstraete, W., Development of a rapid pH-based biosensor to monitor
and control the hygienic quality of reclaimed domestic wastewater. Applied Microbiology and
Biotechnology, 56, (5-6), 809-815.
Dixit, S., and J. G. Hering, (2003). Comparison of Arsenic(V) and Arsenic(III) sorption onto iron oxide
minerals: Implications for arsenic mobility. Environmental Science & Technology, 37(18), 4182-
4189.
Dubinina, E. O.; Lakshtanov, L. Z., (1997). A kinetic model of isotopic exchange in dissolution-
precipitation processes. Geochimica et Cosmochimica Acta, 61, (11), 2265-2273.
Erbs, J. J.; Berquo, T. S.; Reinsch, B. C.; Lowry, G. V.; Baneijee, S. K.; Penn, R. L., (2010). Reductive
dissolution of arsenic-bearing ferrihydrite. Geochimica et Cosmochimica Acta, 74(12), 3382-
3395.
Fendorf, S, H. A. Michael and A. van Geen, (2010). Spatial and temporal variations of groundwater
arsenic in South and Southeast Asia. Science, 328, 1123-1127.
Goldberg, S., and C. T. Johnston, (2001). Mechanisms of Arsenic Adsorption on Amorphous Oxides
Evaluated Using Macroscopic Measurements, Vibrational Spectroscopy, and Surface
Complexation Modeling. Journal of Colloid andlnterface Science 234. 204-216.
Gonzalez, Z, A., M. Krachler, A. K. Cheburkin, and W. Shotyk, (2006). Spatial Distribution of Natural
Enrichments of Arsenic, Selenium, and Uranium in a Minerotrophic Peatland, Gola di Lago,
Canton Ticino, Switzerland. Environmental Science & Technology, 2006, 40, 6568-6574.
Greskowiak, J., H. Prommer, G. Massmann and G. Niitzmann, (2006). Modeling seasonal redox
dynamics and the corresponding fate of the pharmaceutical residue phenazone during artificial
recharge of groundwater. Enviroimentd Science & Technology, 40, 6615-6621.
Grafe, M., M. J. Eick, P. R. Grossl, and A. M. Saunders, (2002). Adsorption of arsenate and arsenite on
ferrihydrite in the presence and absence of dissolved organic carbon. Journal of Environmental
Quality, 2002, 31, 1115-1123.
Grovea, D. B., and W. W. Wood, (1979). Prediction and Field Verification of Subsurface-Water Quality
Changes During Artificial Recharge, Lubbock, Texas. Ground Water, 17, 250-257.
101
-------
Guo, X., and F. Chen, (2005). Removal of arsenic by bead cellulose loaded with iron oxyhydroxide from
groundwater. Environmental Science & Technology, 39 (17), 6808-6818.
Guo, H., S. Yang, X. Tang, Y. Li and Z. Shen, (2008). Groundwater geochemistry and its implications for
arsenic mobilization in shallow aquifers of the Hetao Basin, Inner Mongolia. Science of The Total
Environment, 393, 131-144.
Handley, K. M., M. Hery, and J. R. Lloyd, (2009). Redox cycling of arsenic by the hydrothermal marine
bacterium Marinobacter santoriniensis. Environmental Microbiology, 11, 1601-1611.
Harvey, C. F., C. H. Swartz, A. B. M. Badruzzaman, N. Keon-Blute, W. Yu, M. A. Ali, J. Jay, R. Beckie,
V. Niedan, D., P. M. O. Brabander, K. N. Ashfaque, S. Islam, H. F. Hemond, and M. F. Ahmed,
(2002). Arsenic mobility and groundwater extraction in Bangladesh. Science, 298, 1602-1606.
Harvey, C. F., K. N. Ashfaque, W. Yu, A. B. M. Badruzzaman, M. A. Ali, P. M. Oates, H. A. Michael, R.
B. Neumann, R. Beckie, S. Islam and M. F. Ahmed, (2006). Groundwater dynamics and arsenic
contamination in Bangladesh. Chemical Geology, 228, 112-136.
Heinrich, C. A., and P. J., Eadington, (1986). Thermodynamic predictions of the hydrothermal chemistry
of arsenic, and their significance for the paragenetic sequence of some cassiterite-arsenopyrite-
base metal sulfide deposits. Economic Geology, 81, 511-529.
Hoque, M, A. Khan, M. Shamsudduha, M. Hossain, T. Islam and S. Chowdhury, (2009). Near surface
lithology and spatial variation of arsenic in the shallow groundwater: southeastern Bangladesh.
Environmental Geology, 2009, 56, 1687-1695.
Hu, Y.; Lee, B.; Bell, C.; Jun, Y.-S., (2012). Environmentally Abundant Anions Influence the Nucleation,
Growth, Ostwald Ripening, and Aggregation of Hydrous Fe(III) Oxides. Langmuir, 28(20), 7737-
7746.
Ilyaletdinov, A., and SA Abdrashitova, (1981). Autotrophc oxidation of arsenic by Pseudomonas
arsenitoxidants. Mikrobiologiya, 50, 197-205.
Islam, F. S., A. G. Gault, C. Boothman, D. A. Polya, J. M. Charnock; D. Chatteijee, and J. R. Lloyd,
(2004). Role of metal-reducing bacteria in arsenic release from Bengal delta sediments. Nature,
430, 68-71.
Jain, A., and R. H. Loeppert, (2000). Effect of competing anions on the adsorption of arsenate and
arsenite by ferrihydrite. J. Environ. Qual., 29, 1422-1430.
Jang, J.-H.; Dempsey, B. A.; Burgos, W. D., (2007). A model-based evaluation of sorptive reactivities of
hydrous ferric oxide and hematite for U (VI). Environmental Science & Technology, 41(12),
4305-4310.
Jergensen, B. B., (1983). The microbial sulphur cycle. In W.E. Krumbein ed., Microbial Geochemistry,
Blackwell Scientific, Oxford, pp. 91-124.
Jones, G. W., and T. Pichler, (2007). The Relationship between Pyrite Stability and Arsenic Mobility
During Aquifer Storage and Recovery in Southwest Central Florida. Environmental Science &
Technology, 41(3), 723-730.
Jones, R. A., S. F. Koval, and H. W. Nesbitt, (2003). Surface alteration of arsenophyrite (FeAsS) by
Thiobacillus ferrooxidans. Geochimica et Cosmochimica Acta, 2003, 67, 955-965.
Jonsson, J., and D. M. Sherman, (2008). Sorption of As(III) and As(V) to siderite, greenrust (fougerite)
and magnetite: Implications for arsenic release in anoxic groundwaters. Chemical Geology, 255,
173-181
102
-------
Kneebone, P. E., P. A. O'Day, N. Jones and J. G. Hering, (2002). Deposition and fate of arsenic in iron-
and arsenic-enriched reservoir sediments. Environmental Science & Technology, 36, 381-386.
Kim, M.-J., J. Nriagu and S. Haack, (2002). Arsenic species and chemistry in groundwater of southeast
Michigan. Environmental Pollution, 120, 379-390.
Kinniburgh, D., P. Smedley, J. Davies, C. Milne, I. Gaus, J. Trafford, S. Burden, S. Ihtishamul Huq, N.
Ahmad, K. Ahmed, A. H. Welch and K. G. Stollenwerk, (2003). The scale and causes of the
ground-water arsenic problem in Bangladesh. In: Welch A, Stollenwerk KG, edts. Arsenic in
Ground Water: Geochemistry and Occurrence. New York: Springer; pp. 211-257.
Kirk, M. F., T. R. Holm, J. Park, Q. Jin, R. A. Sanford, B. W. Fouke and C. M. Bethke, (2004). Bacterial
sulfate reduction limits natural arsenic contamination in groundwater. Geology, 32, 953-956.
Koslides, T., and V. S. T. Ciminelli, (1992). Pressure oxidation of arsenopyrite and pyrite in alkaline
solutions. Hydrometallurgy, 30, 87-106.
Langevin, C. D., (2008). Modeling axisymmetric flow and transport. Ground Water, 46, 579-590.
Lawrence, F. W., and S. B. Upchurch, (1982). Identification of recharge areas using geochemical factor
analysis. Ground Water, 1982, 20, 680-687.
Liu, B.; Zeng, H. C., (2005). Symmetric and asymmetric Ostwald ripening in the fabrication of
homogeneous core-shell semiconductors. Small, 1, (5), 566-571.
Lowry, C. S., and M. P. Anderson, (2006). An Assessment of Aquifer Storage Recovery Using ground
Water Flow Models. Ground Water, 44, 661-667.
Ma, L., and R. F. Spalding, (1996). Stable isotope characterization of the impacts of artificial ground
water recharge. JA WRA, 32, 1273-1282.
Mandal, B. K., and K. T. Suzuki, (2002). Arsenic round the world: a review. Talanta, 58, 201-235.
Modesto Lopez, L. B.; Pasteris, J. D.; Biswas, P., Sensitivity of micro-Raman spectrum to crystallite size
of electrospray-deposited and post-annealed films of iron-oxide nanoparticle suspensions. Applied
Spectroscopy, 63(6), 627-635.
McArthur, J. M., P. Ravenscroft, S. Safiulla, and M. F. Thirlwall, (2001). Arsenic in groundwater: testing
pollution mechanisms for aquifers in Bangladesh. Water Resources Research, 37, 109-117.
McArthur, J. M., D. M. Baneijee, K. A. Hudson-Edwards, R. Mishra, R. Purohit, P. Ravenscroft, A.
Cronin, R. J. Howarth, A. Chatteijee, T. Talukder, D. Lowry, S. Houghton and D. K. Chadha,
(2004). Natural organic matter in sedimentary basins and its relation to arsenic in anoxic ground
water: the example of West Bengal and its worldwide implications. Applied Geochemistry, 19,
1255-1293.
McGuire, M. M., K. J. Edwards, J. F. Banfield, and R. J. Hamers, (2001). Kinetics, surface chemistry, and
structural evolution of microbially mediated sulfide mineral dissolution. Geochimica et
Cosmochimica Acta, 65, 1243-1258.
McKibben, M.; Tallant, B.; Del Angel, J., (2008). Kinetics of inorganic arsenopyrite oxidation in acidic
aqueous solutions. Applied Geochemistry, 23, (2), 121-135.
McNab Jr, W. W., M. J. Singleton, J. E. Moran and B. K. Esser, (2009). Ion exchange and trace element
surface complexation reactions associated with applied recharge of low-TDS water in the San
Joaquin Valley, California. Applied Geochemistry, 24, 129-137.
103
-------
Miller, D. M., and G. S. Hansford, (1992). Batch biooxidation of a gold-bearing pyrite-arsenopyrite
concentrate. Minerals Engineering, 5, 613-629.
Minsley, B.J., J.B. Ajo-Franklin, A. Mukhopadhyay, and F.D. Morgan, (2011). Hydrogeophysical
methods for analyzing aquifer storage and recovery systems. Ground Water, 49(2), 250-269.
Modesto Lopez, L. B.; Pasteris, J. D.; Biswas, P., Sensitivity of micro-Raman spectrum to crystallite size
of electrospray-deposited and post-annealed films of iron-oxide nanoparticle suspensions. Applied
Spectroscopy, 63(6), 627-635.
Navrotsky, A., L. Mazeina and J. Majzlan, (2008). Size-driven structural and thermodynamic complexity
in iron oxides. Science, 319, 1635-1638.
Neil C., Yang YJ, Schupp D, Jun Y-S, (2014). Water chemistry impacts on arsenic mobilization from
arsenopyrite dissolution and secondary mineral precipitation: Implication for managed aquifer
recharge. Environmental Science & Technology, 48(8), 4395-4405.
Neil, C.W., Y.J. Yang, and Y-S. Jun, (2012). A critical review of the potential impact of managed aquifer
recharge on mineral-water interactions responsible for arsenic mobilization in groundwater. J.
Environ. Monit., 14: 1772-1788.
Nesbitt, H. W., I. J. Muir and A. R. Prarr, (1995). Oxidation of arsenopyrite by air and air-saturated,
distilled water, and implications for mechanism of oxidation. Geochimica et Cosmochimica Acta,
59, 1773-1786.
Nickson, R., J. McArthur, W. Burgess, K. M. Ahmed, P. Ravenscroft and M. Rahmann, (1998). Arsenic
poisoning of Bangladesh groundwater. Nature, 395, 338-338.
Nimick, D. A., (1998). Arsenic Hydrogeochemistry in an Irrigated River Valley—A Reevaluation.
Ground Water, 36, 743-753.
O'Day, P.A., D. Vlassopoulos, R. Root, and N. Rivera, (2004). The influence of sulfur and iron on
dissolved arsenic concentrations in the shallow subsurface under changing redox conditions.
PNAS, 101, 13703-13708.
O'Day, P. A., (2006). Chemistry and mineralogy of arsenic. Elements, 2, 77-83.
Oremland, R. S., and John F. Stolz, (2003). The ecology of arsenic. Science, 300, 939-944.
Overacre, R., T. Clinton, D. Pyne, S. Snyder and P. Dillon, (2006). Reclaimed water aquifer storage and
recovery: Potential changes in water quality. Proceedings of the Water Environment Federation,
2006, 1339-1360.
Papangelakis, V. G., and G. P. Demopoulos, (1990). Acid pressure oxidation of arsenopyrite: Part II,
reaction kinetics. Canadian Metallurgical Quarterly, 29, 13-20.
Park, H.; Myung, N. V.; Jung, H.; Choi, H., (2009). As (V) remediation using electrochemically
synthesized maghemite nanoparticle s. Journal of Nanoparticle Research, 11(8), 1981-1989.
Pavelic, P.; Nicholson, B. C.; Dillon, P. J.; Barry, K. E., (2005). Fate of disinfection by-products in
groundwater during aquifer storage and recovery with reclaimed water. Journal of contaminant
hydrology, 77(4), 351-373.
Pavelic, P., P. J. Dillon, and B. C. Nicholson, (2006). Comparative evaluation of the fate of disinfection
byproducts at eight aquifer storage and recovery sites. Enviroimentd Science & Technology, 40,
501-508.
104
-------
Pavelic, P., P. J. Dillon and C. T. Simmons, (2006). Multiscale characterization of a heterogeneous
aquifer using an ASR operation. Ground Water, 44, 155-164.
Pedersen, H. D.; Postma, D.; Jakobsen, R., (2006). Release of arsenic associated with the reduction and
transformation of iron oxides. Geochimica et Cosmochimica A:ta, 70(16), 4116-4129.
Peters, S. C., and L. Burkert, (2008). The occurrence and geochemistry of arsenic in groundwaters of the
Newark basin of Pennsylvania. Applied Geochemistry, 23, 85-98.
Postma, D.; Larsen, F.; Minh Hue, N. T.; Due, M. T.; Viet, P. H.; Nhan, P. Q.; Jessen, S. r., Arsenic in
groundwater of the Red River floodplain, Vietnam: controlling geochemical processes and
reactive transport modeling. Geochimica et Cosmochimica Acta, 71, (21), 5054-5071.
Powelson, D. K., C. P. Gerba and M. T. Yahya, (1993) Virus transport and removal in wastewater during
aquifer recharge. Water Research, 27, 583-590.
Randall, S. R., D. M. Sherman and K. V. Ragnarsdottir, (2001). Sorption of As(V) on green rust
(Fe4(II)Fe2(III)(0H)i2S043H0) and lepidocrocite (y-FeOOH): surface complexation by EXAFS
spectroscopy. Geochimica et Cosmochimica Acta, 65, 1015-1023.
Rau, I.; Gonzalo, A.; Valiente, M., (2003). Arsenic(V) adsorption by immobilized iron mediation.
Modeling of the adsorption process and influence of interfering anions. Reactive and Functional
Polymers, 54(1-3), 85-94.
Raven, K. P., A. Jain and R. H. Loeppert, (1998). Arsenite and arsenate adsorption on ferridydrite:
Kinetics, equilibrium, and adsorption envelopes. Environmental Science & Technology, 32, 344-
349.
Reich, M., and U. Becker, (2006). First-principles calculations of the thermodynamic mixing properties of
arsenic incorporation into pyrite and marcasite. Chemical Geology, 225, 278-290
Richmond, W. R., M. Loan, J. Morton and G. M. Parkinson, (2004). Arsenic Removal from Aqueous
Solution via Ferrihydrite Crystallization Control. Environmental Science & Technology, 38,
2368-2372.
Rimstidt, J. D., J. A. Chermak, and M. P. Gagen, (1993). Rates of reaction of Galena, Sphalerite,
Chalcopyrite, and Arsenopyrite with Fe(III) in acidic solutions. In C. Alpers et al. edts.,
Environmental Geochemistry of Sulfide Oxidation, ACS Symposium Series, 550, 2-13.
Ruitenberg, R., G. S. Hansford, M. A. Reuter, and A. W. Breed, (1999). The ferric leaching kinetics of
arsenopyrite. Hydrometallurgy, 52, 37-53.
Sadiq, M., A. Locke, G. Spiers and D. Pearson, (2002). Geochemical Behavior of Arsenic in Kelly Lake,
Ontario. Water, Air, & Soil Pollution, 141, 299-312.
Salzsauler, K. A., N. V. Sidenko, and B. L. Sherriff, (2005). Arsenic mobility in alternation products of
sulfide-rich, arsenopyrite-bearing mine wastes, Snow Lake, Manitoba, Canada. Applied
Geochemistry, 20, 2303-2314.
Schlieker, M., J. Sch€uring, J. Hencke and H. D. Schulz, (2001). The influence of redox processes on
trace element mobility in a sandy aquifer-an experimental approach. J. Geochem. Explor., 73,
167-179.
Sharif, M. U., R. K. Davis, K. F. Steele, B. Kim, T. M. Kresse and J. A. Fazio, (2008). Inverse
geochemical modeling of groundwater evolution with emphasis on arsenic in the Mississippi
River valley alluvial aquifer, Arkansas (USA). Journal of Hydrology, 350, 41-55.
105
-------
Sheng, Z., (2005). An aquifer storage and recovery system with reclaimed wastewater to preserve native
groundwater resources in El Paso, Texas. Journal of environmental management, 75(4), 367-377.
Sidhu, P.; Gilkes, R.; Cornell, R.; Posner, A.; Quirk, J., (1981). Dissolution of iron oxides and
oxyhydroxides in hydrochloric and perchloric acids. Clays and Clay Minerals, 29(6), 269-276.
Simeoni, M. A., B. D. Batts and C. McRae, (2003). Effect of groundwater fulvic acid on the adsorption of
arsenate by ferrihydrite and gibbsite. Applied Geochemistry, 18, 1507-1515.
Slezov, V. V., (2009).Theory of the late stages of nucleation-growth processes: Ostwald ripening. In:
Kinetics of First-Order Phase Transitions, Wiley-VCH Verlag GmbH & Co. KGaA: pp 93-169.
Smedley, P., and D.G. Kinnburgh, (2002). A review of the source, behaviour and distribution of arsenic
in natural waters. Applied Geochemistry, 17, 517-568
Spaccini, R., A. Piccolo, P. Conte, G. Haberhauer and M. H. Gerzabek, (2002). Increased soil organic
carbon sequestration through hydrophobic protection by humic substances. Soil Biology and
Biochemistry, 34, 1839-1851.
Stamos, C., P. Martin, T. Nishikawa, and B. F. Fox, (2001). Simulation of ground-water flow in the
Mojave River basin, California: U.S. Geological Survey Water-Resources Investigations Report
01-4002, 129 p.
Stollenwerk, K. G., (2003). Geochemical processes controlling transport of arsenic in groundwater: A
review of adsorption. In, A.H. Welch and K.G. Stollenwerk, edts., Arsenic in Ground Water:
Geochemistry and Occurrence, pp.67-100. Springer, US.
Sveijensky, D. A., and K. Fukushi, (2006). A predictive model (ETLM) for As(III) adsorption and surface
speciation on oxides consistent with spectroscopic data. Geochimica et Cosmochimica Acta, 70,
3778-3802.
Todd, D. T., (1961). The distribution of ground water beneath an artificial recharge area. International
Association of Scientific Hydrology, 57, 254-262
Tufano, K., J., C. Reyes, C. W. Saltikov, and S. Fendorf, (2008). Reductive processes controlling arsenic
retention: Revealing the relative importance of iron and arsenic reduction. Environmental Science
& Technology, 42, 8283-8289.
Vacher, H. L., W. C. Hutchings and D. A. Budd, (2006). Metaphors and models: The ASR bubble in the
Floridan aquifer. Ground Water, 44, 144-154.
Vanderzalm, J., Le Gal La Salle, C.; Dillon, P., (2006). Fate of organic matter during aquifer storage and
recovery (ASR) of reclaimed water in a carbonate aquifer. Applied Geochemistry, 21, (7), 1204-
1215.
Vanderzalm, J. L., P. J. Dillon, K. E. Barry, K. Miotlinski, J. K. Kirby and C. Le Gal La Salle, (2011).
Arsenic mobility and impact on recovered water quality during aquifer storage and recovery
using reclaimed water in a carbonate aquifer. Applied Geochemistry, 26, 1946-1955.
Verplanck, P. L., S. H. Mueller, R. J. Goldfarb, D. K. Nordstrom and E. K. Youcha, (2008), Geochemical
controls of elevated arsenic concentrations in groundwater, Ester Dome, Fairbanks district,
Alaska. Chemical Geology, 255(1-2), 160-172.
Violante, A., and M. Pigna, (2002). Competitive sorption of arsenate and phosphate on different clay
minerals and soils. Soil Sci. Soc. Am. J., 2002, 66, 1788-1796.
106
-------
Walker, F. P., M. E. Schreiber and J. D. Rimstidt, (2006). Kinetics of arsenopyrite oxidative dissolution
by oxygen. Geochimica et Cosmochimica Acta, 70, 1668-1676.
Wallis, I., H. Prommer, C. T. Simmons, V. Post, and P. J. Stuyfzand, (2010). Evaluation of conceptual
and numerical models for arsenic mobilization and attenuation during managed aquifer recharge.
Enviroimentd Science & Technology, 44, 5035-5041.
Wallis, I., H. Prommer, T. Pichler, V. Post, S. B. Norton, M. D. Annable, and C. T. Simmons, (2011).
Process-based reactive transport model to quantify arsenic mobility during aquifer storage and
recovery of potable water. Enviroimentd Science & Technology, 45, 6924-6931.
Wang, S., and Mulligan, C. N., (2006). Natural attenuation processes for remediation of arsenic
contaminated soils and groundwater. Journal of Hazardous Materials, 138, (3), 459-470.
Ward, J. D., C. T. Simmons and P. J. Dillon, (2008). Variable-density modelling of multiple-cycle aquifer
storage and recovery (ASR): importance of anisotropy and layered heterogeneity in brackish
aquifers. Journal of Hydrology, 356, 93-105.
Ward, J.D., C.T. Simmons, and P.J. Dillon, (2007). A theoretical analysis of mixed convection in aquifer
storage and recovery: how important are density effects? Journal of Hydrology, 343 (3-4), 169—
186
Waychunas, G. A., Y.-S. Jun, P. J. Eng, S. K. Ghose, T. P. Trainor, O. B. Mark and B. K. Douglas,
(2008). Anion sorption topology on hematite: comparison of arsenate and silicate. In,
Developments in Earth and Environmental Sciences, 7, 31-65.
Wilson, L.; Amy, G.; Gerba, C.; Gordon, H.; Johnson, B.; Miller, J., (1995). Water quality changes
during soil aquifer treatment of tertiary effluent. Water Environment Research, 67(3), 371-376.
Wolf, M., A. Kappler, J. Jiang, and R. U. Meckenstock, (2009). Effects of humic substances and quinones
at low concentrations on ferrihydrite reduction by Geobacter metallireducens. Environmental
Science & Technology, 43, 5679-5685.
Youngran, J.; FAN, M.; Van Leeuwen, J.; Belczyk, J. F., Effect of competing solutes on arsenic (V)
adsorption using iron and aluminum oxides. Journal of Environmental Sciences, 19, (8), 910-919.
Yu, W. H., C. M. Harvey, and C. F. Harvey, (2003), Arsenic in groundwater in Bangladesh: A
geostatistical and epidemiological framework for evaluating health effects and potential remedies,
Water Resour. Res., 39(6), 1146, doi: 10.1029/2002WR001327
Yu, Y.; Zhu, Y.; Gao, Z.; Gammons, C. H.; Li, D., (2007). Rates of Arsenopyrite Oxidation by Oxygen
and Fe (III) atpH 1.8-12.6 and 15-45 C. Environmental Science & Technology, 41, (18), 6460-
6464.
Yu, Y.; Zhu, Y., Williams-Jones, A.E., Gao, Z., Li, D., (2004). A kinetic study of the oxidation of
arsenopyrite in acidic solutions: implications for the environment. Applied Geochemistry, 19(3),
435-444
Zheng, Y., M. Stute, A. van Geen, I. Gavrieli, R. Dhar, H. J. Simpson, P. Schlosser and K. M. Ahmed,
(2004). Redox control of arsenic mobilization in Bangladesh groundwater. Applied Geochemistry,
19, 201-214.
Zhu, W., L. Y. Young, N. Yee, M. Series, E. D. Rhine and J. R. Reinfelder, (2008). Sulfide-driven
arsenic mobilization from arsenopyrite and black shale pyrite. Geochimica et Cosmochimica
Acta, 72, 5243-5250.
107
------- |