chemical spill: perceptions, water quality, and health

Subscriber access provided by Purdue University Libraries

Environmental Science & Technology is published by the American Chemical Society.1155 Sixteenth Street N.W., Washington, DC 20036Published by American Chemical Society. Copyright © American Chemical Society.However, no copyright claim is made to original U.S. Government works, or worksproduced by employees of any Commonwealth realm Crown government in the courseof their duties.

Article

Residential Tap Water Contamination Following the Freedom IndustriesChemical Spill: Perceptions, Water Quality, and Health Impacts

Andrew James Whelton, LaKia McMillan, Matthew Connell, Keven M Kelley,Jeffrey P. Gill, Kevin D White, Rahul Gupta, Rajarshi Dey, and Caroline Novy

Environ. Sci. Technol., Just Accepted Manuscript • Publication Date (Web): 16 Dec 2014

Downloaded from http://pubs.acs.org on December 16, 2014

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are postedonline prior to technical editing, formatting for publication and author proofing. The American ChemicalSociety provides “Just Accepted” as a free service to the research community to expedite thedissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscriptsappear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have beenfully peer reviewed, but should not be considered the official version of record. They are accessible to allreaders and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offeredto authors. Therefore, the “Just Accepted” Web site may not include all articles that will be publishedin the journal. After a manuscript is technically edited and formatted, it will be removed from the “JustAccepted” Web site and published as an ASAP article. Note that technical editing may introduce minorchanges to the manuscript text and/or graphics which could affect content, and all legal disclaimersand ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errorsor consequences arising from the use of information contained in these “Just Accepted” manuscripts.

1

Residential Tap Water Contamination Following 1

the Freedom Industries Chemical Spill: 2

Perceptions, Water Quality, and Health Impacts 3

Andrew J. WheltonΦ*, LaKia McMillan†, Matt Connell†, Keven M. Kelley†, Jeff P. 4

Gill†, Kevin D. White†, Rahul Gupta‡, Rajarshi Dey∆, Caroline Novy† 5

Φ Division of Environmental and Ecological Engineering and Lyles School of Civil 6

Engineering, Purdue University, West Lafayette, IN USA 7

† Department of Civil Engineering, University of South Alabama, Mobile, AL USA 8

‡ Kanawha Charleston Health Department, Charleston, WV USA 9

∆ Department of Mathematics and Statistics, University of South Alabama, Mobile, AL 10

USA 11

*Corresponding author: Assistant Professor, Division of Environmental and Ecological 12

Engineering and Lyles School of Civil Engineering, Purdue University, 550 Stadium Mall 13

Drive, West Lafayette, IN USA 47907-2051; T: (765) 494-2166; F: (765) 494-0395; E: 14

[email protected]; [email protected] 15

Page 1 of 41

ACS Paragon Plus Environment

Environmental Science & Technology

2

KEYWORDS: drinking water; 4-Methylcyclohexanemethanol; spill; contamination; 16

odor; plumbing; flushing 17

Abstract Word Count: 200 18

Text Word Count: 6,455 19

Page 2 of 41



3

ABSTRACT 20

During January 2014, an industrial solvent contaminated West Virginia’s Elk River and 21

15% of the state population’s tap water. A rapid in-home survey and water testing was 22

conducted two weeks following the spill to understand resident perceptions, tap water 23

chemical levels, and premise plumbing flushing effectiveness. Water odors were detected 24

in all 10 homes sampled before and after premise plumbing flushing. Survey and medical 25

data indicated flushing caused adverse health impacts. Bench-scale experiments and 26

physiochemical property predictions showed flushing promoted chemical volatilization, 27

and contaminants did not appreciably sorb into crosslinked polyethylene (PEX) pipe. 28

Flushing reduced tap water 4-methylcyclohexanemethanol (4-MCHM) concentrations 29

within some but not all homes. 4-MCHM was detected at unflushed (<10 to 420 µg/L) 30

and flushed plumbing systems (<10 to 96 µg/L) and sometimes concentrations differed 31

among faucets within each home. All waters contained less 4-MCHM than the 1,000 32

µg/L Centers for Disease Control drinking water limit but one home exceeded the 120 33

µg/L drinking water limit established by independent toxicologists. Nearly all households 34

refused to resume water use activities after flushing because of water safety concerns. 35

Science based flushing protocols should be developed to expedite recovery, minimize 36

health impacts, and reduce concentrations in homes when future events occur. 37

Page 3 of 41



4

INTRODUCTION 38

Early on January 9, 2014 a Freedom Industries, Inc. chemical storage tank was found leaking. 39

An investigation revealed more than 10,000 gallons of an industrial coal processing liquid had 40

been released into West Virginia’s Elk River. Freedom Industries, Inc. initially reported that 41

“Crude MCHM” was spilled, but twelve days later the company also disclosed that an additional 42

product called “Stripped PPH” was also present in the spilled liquid (Table 1; Figure SI-1).1 The 43

Elk River was the regional water company’s sole drinking water source used to supply the state 44

capitol including 300,000 people, 15% of the State’s population. 45

Contaminated river water traveled downstream and entered West Virginia American Water’s 46

(WVAW) Kanawha Valley 50 million gallon per day (MGD) drinking water treatment plant. 2 In 47

the days leading up to the spill, water demand was approximately 43 MGD, cold weather (-5°C) 48

was attributed to water main breaks, and residents were allowing faucets to drip to prevent their 49

plumbing pipes from freezing. WVAW estimated that it had less than three hours of tap water in 50

reserve. WVAW predicted that if the raw water intake was shutdown water for fire-fighting as 51

well as basic hygiene and sanitation purposes would not be available and at least 45 days would 52

be needed to restore service to large sections of the distribution system. At 4:00 pm, WVAW 53

detected contaminated drinking water entering the distribution system and observed a licorice 54

odor.3 A Do Not Use order was issued at 5:50 pm for the entire service area. Little to no 55

toxicological data and physiochemical properties were available for many of the solvent’s 56

ingredients.4 The Governor declared a State of Emergency for the nine counties affected and at 57

12:46 am January 10 President Obama declared the incident a Federal disaster.5 58

59

Page 4 of 41



5

Table 1. Chemicals Suspected to be in the Spilled Tank Liquid According to Declarations 60

by Eastman Chemical Company and Freedom Industries, Inc. 61

Product Reported Ingredient Estimated Composition

of the Spilled Liquid

Crude

MCHM

4-Methylcyclohexanemethanol (MCHM) 68% to 89%

4-(Methoxymethyl)cyclohexanemethanol 4% to 22%

Water 4% to 10%

Methyl 4-methylcyclohexanecarboxylate 5%

Dimethyl 1,4-cyclohexanedicarboxylate 1%

Methanol 1%

1,4-Cyclohexanedimethanol 1% to 2%

Stripped

PPH*

Propylene glycol phenyl ether (PPH) Amount unclear

Dipropylene glycol phenyl ether (DiPPH) Amount unclear

*Thirteen days after the spill, the Centers for Disease Control and Prevention (CDC) reported 62

that the spilled product contained “88.5% Crude MCHM, 7.3% Stripped PPH and 4.2% water” 63

though the CDC’s calculation methodology was not disclosed. Stripped PPH was blended into 64

Eastman Chemical Company’s Crude MCHM by Poca Blending Company in Nitro, WV. The 65

Stripped PPH product contained both PPH and DiPPH compounds. 66

67

During the next 10 days, the Do Not Use order remained in effect for much of the area. Tap 68

water remained relatively stagnant in premise plumbing systems as only toilet flushing and fire-69

fighting were permitted. The greatest measured concentration of 4-MCHM, the main ingredient 70

of the spilled liquid, entering or leaving the water treatment facility was 3,350 µg/L (Figure 1). 71

The greatest concentration detected during rapid water distribution system sampling by WVAW 72

and the State was 3,773 µg/L. It remains unclear if 3,773 µg/L was the greatest concentration 73

that exited the water plant because testing did not begin until January 10, when contaminated 74

water had already entered the distribution system.3 Tap water samples were also analyzed for 75

propylene glycol phenyl ether (PPH) and dipropylene glycol phenyl ether (DiPPH), but these 76

Page 5 of 41



6

compounds were found in only two water samples collected and at concentrations of 11 µg/L and 77

10 µg/L, respectively. An analytical method for detecting and quantifying the spilled liquid’s 78

ingredients in water did not exist when the incident occurred. This method was developed after 79

WVAW was notified of the spill. 80

Page 6 of 41



7

(a) 81

(b) 82

Figure 1. 4-MCHM Monitoring Results for the (a) Kanawha Valley Water Treatment 83

Facility from January 10 to January 15, 2014 and (b) Water Distribution System from 84

January 10 to March 6, 2014. 4-MCHM has both trans- and cis-isomers9 and 4-MCHM 85

concentrations reported by WVAW, the State, and in the present study were reported as the 86

combined trans- and cis- isomer concentration. Distribution system samples were collected at 87

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

9-Jan 10-Jan 11-Jan 12-Jan 13-Jan 14-Jan 15-Jan 16-Jan 17-Jan

4-M

CH

M C

on

cen

tra

tio

n, µ

g/L

Date

Raw

Treated

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

9-Jan 19-Jan 29-Jan 8-Feb 18-Feb 28-Feb 10-Mar

4-M

CH

M C

on

cen

trati

on

, µ

g/L

Date

(i)

(ii)

(i)

(ii)

Page 7 of 41



8

various locations to include hydrants, storage tanks, booster stations, public buildings including 88

schools, hospitals, and private businesses. The dotted horizontal lines represent the (i) CDC’s 4-89

MCHM screening level (1,000 µg/L), (ii) WVTAP’s 4-MCHM screening level (120 µg/L). Only 90

results where the value was greater than the MDL are shown (i.e., more than 1,100 “non-detect” 91

results are not shown). Also not shown are March 2014 testing results where water leaving the 92

water treatment plant contained 0.42 to 0.60 µg/L 4-MCHM. Testing was also conducted in June 93

2014 after the water treatment plant’s activated carbon filters were replaced. 4-MCHM was not 94

found above a MDL of 0.38 µg/L exiting the water treatment plant or in the distribution system. 95

96

Flushing was conducted by WVAW to remove contaminated water from its 2,200 miles of 97

water mains, 107 storage tanks, and 120 booster stations within 124 pressure zones. Water 98

samples were collected at various locations to monitor flushing effectiveness (Figure 1). 99

WVAW’s initial flushing objective was to reduce 4-MCHM concentrations below the US 100

Centers for Disease Control and Prevention (CDC) health based screening level of 1,000 µg/L.6 101

Subsequent response objectives included reducing 4-MCHM concentrations below 50 µg/L and 102

then below a 10 µg/L, a screening level established by a State of West Virginia. The 10 µg/L 103

concentration was the lowest MDL at the time. During the following months, WVAW would 104

flush its distribution system to achieve 4-MCHM concentrations less than 2 µg/L once a lower 105

MDL was developed. In March 2014, another research team hired by the State called WVTAP, 106

or West Virginia Testing Assessment Project, issued a health based 4-MCHM screening level of 107

120 µg/L using the same toxicological data CDC reviewed, but with different assumptions 108

(Table 2).7 109

Four days after the spill, WVAW had flushed parts of its distribution system and began 110

advising residents in those areas to flush their premise plumbing systems using a step-wise 111

protocol.8 The procedure had been reviewed by government public health officials before 112

release. Residents were told that after flushing, tap water would be ‘appropriate to use’ by health 113

officials but may still have an odor. No in-home tap water or air quality testing was conducted by 114

Page 8 of 41



9

WVAW or responding local, county, State, or Federal organizations before, during, or 115

immediately after plumbing system flushing activities. 116

Table 2. Comparison of Drinking Water Screening Levels Established by the US Centers 117

for Disease Control and Prevention (CDC) and West Virginia Testing Assessment Project 118

(WVTAP) 119

Contaminant Name &

Assumption by Health Officials

CDC

(January 2014)

WVTAP

(March 2014)

4-MCHM, µg/L 1,000 120

PPH, µg/L 1,200 850

DiPPH, µg/L 1,200 250

Exposure Duration 14 days 28 days

Most sensitive population 1 year old child Formula fed infant

Exposure routes Ingestion only Ingestion, inhalation, dermal

WVTAP included toxicologists from academic and public health organizations located in Israel, 120

the United Kingdom, and United States. In January 2014, the State of West Virginia applied a 121

100-fold safety factor to the CDC’s 4-MCHM screening level because of concern that limited 122

toxicology data existed. The State of West Virginia’s screening level for 4-MCHM was 10 µg/L. 123

124

The goal of this rapid response study was to understand resident responses and tap water 125

quality within unflushed and flushed residential plumbing systems. Specific objectives were to 126

rapidly: (1) determine resident behaviors and perceptions following the spill, (2) characterize 127

plumbing system characteristics and chemical levels in homes, and (3) determine the ability of 128

the flushing procedure to reduce chemical levels within homes. To complete this work, the 129

research team visited the affected area from January 18 to 22, 2014. The visit was conducted 130

when residents were being authorized by WVAW to flush contaminated water from their 131

premise plumbing systems. To interpret results of this field effort, syndromic surveillance 132

Page 9 of 41



10

records as well as public health survey results obtained by government agencies were also 133

reviewed. A detailed timeline of events can be found in the Supporting Information. 134

MATERIALS AND METHODS 135

Survey Instrument and Participating Households. More than 80 households were identified 136

by the West Virginia Clean Water Hub and People Concerned About Chemical Safety and 137

wanted to participate in the research project. These nonprofit organizations were helping 138

distribute water to affected households and assisted the authors to make contact with these 139

households. Due to time, financial, and logistical limitations, the authors could only include a 140

fraction of those households expressing interest in this study. Sixteen households participated in 141

the research. 142

All participating households were provided tap water by the affected water distribution system 143

and were located in Kanawha, Putnam, and Boone Counties; specifically, in Cross Lanes, Elk 144

View, Nellis, South Charleston, and Charleston. A map showing the locations of homes where 145

tap water was sampled can be found in Figure SI-2. One representative of each household was 146

interviewed by the authors and completed a 10 question survey. The survey was designed to 147

evaluate resident behaviors and perceptions in households directly affected by the contamination 148

incident (Table SI-1). 149

Sampling Activity and Analysis. Two faucets were sampled at each home. Kitchen faucets 150

were chosen in every home based on their high use frequency, while the second location (usually 151

a bathroom faucet or outside spigot) was chosen to represent low frequency use. Amber glass 152

bottles with polytetrafluoroethylene (PTFE)-lined caps pre-cleaned with HNO3 and/or shipped 153

direct from the manufacturer were used. Glass bottles were used for water sample collection 154

Page 10 of 41



11

where 4-MCHM, total organic carbon (TOC), UV254 absorbance, alkalinity, chloride, fluoride, 155

hardness, nitrate, orthophosphate, and phosphorus analyses were desired. Glass bottles did not 156

contain preservatives. Plastic bottles were used for metal sample collection and contained 16N 157

HNO3 preservative. Water samples were stored on ice at 4°C and refrigerated until analysis and 158

waters were analyzed within 96 hours of collection. 159

Any point-of-use water filtration devices were removed from the faucets before water 160

collection. Next, about 100 mL of first draw tap water was analyzed for pH and temperature 161

(Thermo Scientific Orion 5 Star™ portable pH meter), free and total chlorine (HACH® Pocket 162

Colorimeter™ II with N,N-diethyl-p-phenyldiamine reagents), and turbidity (HACH® 163

turbidimeters). After these initial measurements, tap water odor was evaluated. Field blanks were 164

obtained by filling containers with laboratory-purchased deionized water. Next, a total of four 165

water containers were filled per tap for (a) 4-MCHM, chloride, fluoride, hardness, nitrate, 166

orthophosphate, phosphorus [1 L], (b) TOC [0.125 L], (c) metals [0.250 L], (d) alkalinity, color, 167

ultraviolet absorbance at 254 nm (UV254) [0.250 L]. All containers were headspace free. After 168

premise plumbing flushing, a second series of water samples were collected from flushed taps. 169

Water samples were shipped to three different laboratories. A more detailed water analysis 170

methods description can be found in the Supporting Information. Particulate material released 171

from a Home 8 bathroom faucet was collected and analyzed by X-ray photoelectron 172

spectroscopy (XPS) at the University of New Mexico. 173

Premise Plumbing Flushing Procedure. Six homes that had unflushed pipes were visited. 174

These homes underwent premise plumbing flushing in accordance with the protocol issued by 175

WVAW8 with slight modification. The protocol entailed flushing all hot water taps for 15 176

minutes, all cold water taps for five minutes, and finally flushing all other appliances for five 177

Page 11 of 41



12

minutes. The document stated that any lingering licorice odor detected would not be harmful. 178

The details for the flushing protocol design (i.e., estimated flowrate, water volume removed) 179

were not found. Flushing protocol modifications were made after members of the research team 180

experienced chemical exposure symptoms such as eye-burning and dizziness. Modifications 181

included turning off hot water heaters and allowing them to cool before flushing, flushing one 182

room in the house at a time, opening doors and windows, using fans to ventilate rooms while 183

flushing, and following the flushing protocol more than once to improve the chance of 184

contaminant removal. At the time of this rapid response there was no data pertaining to the 185

interaction of the spilled liquid’s ingredients or its breakdown products with premise plumbing 186

materials (i.e., metal and plastic pipes, gaskets, hot water heaters). The authors requested that 187

children and immunocompromised persons leave the house during flushing activities. Tap water 188

sampling was only conducted before and after the first flushing procedure. 189

Contaminant Interaction with PEX Pipe. Bench-scale sorption experiments were conducted 190

to understand the interaction of 4-MCHM [TCI America, 99.0%] and cyclohexanemethanol 191

(CHM) [Acros Organics, 99.0%] with crosslinked polyethylene (PEX) type A and type B pipes. 192

These two compounds were present in the spilled liquid. The presence of both 4-MCHM and 193

CHM was confirmed by WVTAP investigators who characterized the liquid remaining in the 194

Freedom Industries, Inc. tank. CHM however was not listed on any safety data sheets provided to 195

WVAW, State, or Federal response agencies (Table 1). 196

PEX type A and type B pipes were examined because they were both present in some of the 197

homes visited. PEX-A pipe is also more susceptible than PEX-B pipe to contaminant permeation 198

because of its low bulk density.10 PEX pipes purchased from a local building supply store were 199

cut into 1.1 to 1.3 gram dog-bone shaped specimens (2.5 to 2.6 mm thick), then were immersed 200

Page 12 of 41



13

in pure solvent at room temperature. During several weeks, specimens were periodically 201

weighed.10 PEX interaction with toluene and cyclohexane [Fisher Scientific, 99.8% and 99.0% 202

purity, respectively) was also examined; neither compound was found in the spilled liquid but 203

were used as controls. 204

Syndromic Surveillance. Data on illness frequency from 10 sentinel multi-provider and 205

multi-location medical practices was compiled and analyzed in the present study. Information 206

pertained to 224 patients from the impacted area that sought medical attention with self-reported 207

symptom onset from January 9 to February 10. The list of symptoms included multisystem 208

symptoms (respiratory, digestive, integumentary [skin], neurological); respiratory: cough, sore 209

throat; digestive: nausea, vomiting, diarrhea; skin: rash, skin irritation; neurological: Headache; 210

and “other symptoms” for symptoms that had not been defined. Providers did not report names, 211

addresses or other identifying information on the patients beyond gender and age. Patients were 212

asked whether or not they were in a flushed or unflushed building when tap water exposure 213

occurred, and only those with clear associated exposure were included. 214

Physiochemical Properties and Statistical Analysis. Contaminant physiochemical 215

characteristics were estimated using SPARC11 chemical modeling software (Danielsville, GA 216

USA), and water solubility was also estimated using COSMOS-RS12 (Amsterdam, The 217

Netherlands). These programs allowed for the authors to estimate the fate and transport of 218

contaminants at different temperatures. Water temperatures chosen were representative of the 219

Elk River (5°C) and within plumbing systems (21°C, 60°C). 220

Minitab 14 Student (Minitab, Inc. State College, PA)13 was used to perform two-tailed student 221

t-tests and linear regression statistical analysis was also carried-out. Any water quality result less 222

Page 13 of 41



14

than the method MDL was assigned a value of zero. An alpha value of 0.05 was selected as the 223

significance level for all data interpretation. 224

Logistic regression modeling was applied to syndromic surveillance data. Response variables 225

were binary (i.e., patients answered whether they had a symptom of nausea or not) and were 226

coded as 1 and 0, respectively. In a logistic setting, the odds of an event happening were modeled 227

where Y was a binary random variable. Then, the odds of Y being 1 was given by the ratio of 228

probabilities of Y being 1 and Y being 0. The equation was: Odds (Y=1) = P(Y=1)/P(Y=0). Note 229

that the odds could be any number more than 0; an odd of 1 implied a fair chance. The following 230

model was applied: 0

1

[ 1]K

e i i

i

Log Odds Y Xβ β ε=

= = + +∑ . In the above model, ε implied the 231

unexplained model error; 1 2, ,..., KX X X were covariates (factors) and 0 1, ,..., Kβ β β were 232

unknown coefficients estimated using Minitab. The above model provides an effective 233

interpretation for coefficients involved with each factor in terms of the odds. If the factor was 234

also binary eβ was interpreted as the ratio of the odds for X=1 and X=0. 235

RESULTS AND DISCUSSION 236

In-Home Survey Results and Comparison to Public Health Studies 237

Household Demographics and Premise Plumbing Characteristics. Single story, multi-story, 238

and manufactured homes were visited. Households had similar demographics to the 2012 West 239

Virginia census (Table SI-2).14 Sixteen persons, 32 to 68 years old, representing 16 households 240

were interviewed. A range of 1 to 5 persons lived in each household. Children under 18 years of 241

age and/or adults with medical concerns lived in 12 of the households. Children lived in half of 242

the households. Five residences had pets with at least one cat or dog. 243

Page 14 of 41



15

Ten of the 16 home plumbing systems entirely or partly contained copper pipe. Pipe materials 244

found that were not copper were: Chlorinated polyvinylchloride (cPVC) [five homes] > 245

crosslinked polyethylene (PEX) [three homes], galvanized iron [two homes], and poly(1-butene) 246

(PB) [single home]. Renovated homes typically contained some cPVC, PEX, or PB plastic 247

drinking water pipe. 248

The visit was conducted when WVAW was advising residents to flush contaminated water 249

from their premise plumbing systems. Between January 18 and 22, seven households reported 250

they had not flushed their plumbing systems. Another seven households reported having already 251

flushed their plumbing systems. A single household was considered partially flushed because the 252

second story was not flushed, while remaining faucets had been flushed. 253

Resident Behaviors and Perceptions. Nearly all households surveyed (14 of 16) reported 254

detecting an unusual tap water odor during the first two weeks of the incident, two reported an 255

unusual taste, and six reported an unusual tap water color. The most commonly reported odor 256

descriptors were licorice and sweet. These descriptors agreed with those found by other 257

researchers who characterized odor threshold and recognition concentrations of the contaminated 258

water with sensory panels.9 Most households (10 of 16) reported detecting an unusual tap water 259

odor on January 9, 2014. Two households indicated that they detected a licorice odor before 260

January 9. This is an interesting finding because the US Chemical Safety and Hazard 261

Investigation Board investigation found that at least one other chemical storage tank was leaking 262

before January 9.15 Other tap water odor descriptors mentioned by the respondents included the 263

terms acetone, chemical, metallic, organic chemistry lab, and rotten. Households reported 264

noticing odors with different intensities between January 9 and the date this survey was 265

conducted. The greatest odor intensity ratings occurred during January 9 to 13, and odor intensity 266

Page 15 of 41



16

levels generally decreased with time. These observations agree with findings by the WVTAP 267

investigators who also found residents noticed that tap water odor intensity decreased with 268

time.16 269

Resident Health Impacts. Contaminated water exposure impacted resident health. Almost half 270

of the households in the present study (7 of 16) reported that the water caused at least one person 271

in their home to become ill. In contrast, only a third of households surveyed by the Kanawha 272

Charleston Health Department (KCHD)17 and one fifth of households surveyed by the CDC18 273

experienced health issues they thought were attributed to the spill. 274

An examination of syndromic surveillance records revealed that several exposure routes were 275

significant. Results of a KCHD telephone survey17 and CDC18 in-home survey conducted after 276

this rapid response study support this finding. Syndromic surveillance records show that 277

exposure routes such as drinking, washing or bathing were significant. Patients who drank 278

contaminated water were more likely to report nausea (p<0.001), vomiting (p<0.001), diarrhea 279

(p<0.001), and sore throat (p=0.002) symptoms. Bathing or hand washing with contaminated 280

water resulted in skin irritation (p<0.001) and rash (p=0.002), symptoms, which was intuitively 281

expected. Interestingly, female patients were more susceptible to vomiting (p=0.023) and sore 282

throat (p=0.025) symptoms than male patients. The CDC18 found that symptoms could be 283

grouped into three exposure categories based on their household survey: (1) Bathing, showering, 284

or skin contact [52.6%], (2) Eating, drinking, or swallowing [43.9%], and (3) Breathing mist or 285

vapor [14.6%]. Table 3 compares the symptoms reported by emergency departments and 286

physicians19,20 in-home CDC survey18, a KCHD telephone survey17, and the WVTAP 287

researchers.16 Persons reported experiencing symptoms at home, work, and food facilities 288

(Supporting Information). 289

Page 16 of 41



17

Table 3. Comparison of the Present Work to Public Health Impact Studies Conducted by Local, State, and Federal, and 290

Research Organizations 291

Symptom

Organization Conducting Study and Date Information was Publicly Released

Household

Interview

Survey

(This Study)

Jan. 2014

Syndromic

Surveillance

Record Review

(This Study)

Jan. 2014

WVTAP

Household

Interview

Survey16

Feb. 2014

CDC / BPH

Emergency

Department

Record Review19

Mar. 2014

BPH

Physician

Record

Review20

Mar. 2014

KCHD

Household

Telephone

Survey17

Apr. 2014

CDC

Household

Interview

Survey18

Jul. 2014

Dermatologic

Skin irritation - 40.3 - - - 63.2§ 53.9

Rash 12.5 47.6 40 28.5 21.6 § 43.6

Itching - - 10 19.8 60.0 - -

Eye irritation 12.5 25.3 10 14.6 (pain) 13.3 26.4 5.1

Gastrointestinal

Nausea 31.3 21.0 30 37.9 - 26.42∆ 12.8

Vomiting 0.0 13.7 10 28.2 8.3 ∆ 5.1

Abdominal Pain 6.3 - - 24.4 8.3 27.0Φ 5.1

Diarrhea 6.3 16.3 0 24.4 5.0 Φ 12.8

Respiratory

Unspecified - - - - - 17.0 -

Sore Throat - 9.4 - 14.9 8.3 - 10.3

Cough - 6.9 - 12.7 15.0 - 15.4

Orientation

Dizziness 18.8 - 40 - - 25.2ε 7.7

Headache 12.5 13.7 30 21.9 11.7 ε 10.3

Other 12.5 - 80 - - 14.1 23.1

Page 17 of 41



18

Numbers in columns total to greater than 100% because multiple symptoms were reported by each surveyed person/household. Dashes 292

“-“ indicate that the data set did not classify symptoms in that specific category; Kanawha-Charleston Health Department (KCHD) 293

syndromic surveillance data represent 224 patients from 10 physicians; The household survey as part of WVTAP15 represents 10 294

households in eight of the nine counties affected; The Centers for Disease Control and Prevention (CDC) and West Virginia 295

Department of Health and Human Resources (DHHR) Bureau of Public Health (BPH) emergency department data represent 356 296

patients from 10 emergency departments; The West Virginia BPH physician record review represents 60 persons; The KCHD 297

randomized telephone survey represents 499 persons and the title of the effort was Community Assessment Population Survey; The 298

KCHD telephone survey included categories where multiple symptoms were listed. Symptoms that were used in combined categories 299

are denoted with symbols; The CDC’s Community Assessment for Public Health Emergency Response (CASPER) household survey 300

data represent 171 households; The present study household survey data represents 16 households. For some of the reports there are 301

significant differences between when the report was dated complete and when it was released to the public. Studies are presented 302

chronologically as the data became publicly available.303

Page 18 of 41



19

Of the seven households that reported health impacts in the present study, only two reported 304

their symptoms to a medical professional. Similarly, the KCHD telephone survey17 found few 305

households, one in five, that reported symptoms sought medical attention. The CDC18 however 306

found almost half of the households that reported symptoms sought medical care. All three data 307

sets show that responders who only monitor physician and emergency department records in 308

incidents such as this will significantly underestimate the population affected. 309

Interestingly, two distinct syndromic surveillance symptom peaks were found that 310

corresponded with the January 9 Do Not Use order and plumbing system flushing activity 311

initiated on January 13 (Figure 2). These two peaks were not statistically different. The West 312

Virginia Poison Control Center21 also noticed a surge in call volume as each pressure zone began 313

flushing. Callers reported nausea, reddened skin, and rash symptoms21 and call volume decreased 314

during the next two weeks (Table SI-3). Syndromic surveillance data showed patients whose 315

homes were flushed were likely to report experiencing a sore throat (p=0.000). Another notable 316

finding is that eye irritation was 2.28 times as likely to be reported if the patient became ill 317

during the first symptom peak (Table SI-4). 318

Estimated vapor pressure and Henry’s Law Constants of the spilled contaminants show 319

compound volatility increased as water temperature increased resulting in greater exposure 320

(Tables SI-5 to SI-7). Laboratory testing also showed that a greater mass of 4-MCHM volatilized 321

from water into air under hot water conditions than at room temperature (Figure SI-3). Based on 322

the evidence examined, plumbing system flushing negatively impacted human health. 323

Water Use Activity. Before January 9, all households visited in the present study used tap water 324

for hygiene activities and nearly all used it for drinking purposes. Approximately two weeks 325

after the Do Not Use Order was lifted, few of the visited households chose to resume their pre-326

Page 19 of 41



20

spill water use activities: Drinking (1 of 12), showering (4 of 15), clothes washing (3 of 15), 327

brushing teeth (1 of 15), cooking (1 of 15), water for animals (2 of 12). Households did not 328

resume their pre-spill water use activities because they remained unconvinced that the water was 329

safe to use based on (1) licorice odor observations after flushing, (2) self-reported symptoms, and 330

(3) reports from friends and media organizations that tap water was causing illness. In fact, only 331

a little more than a third of the households believed the tap water was safe after the Do Not Use 332

order was lifted according to the CDC.18 With time, more residents resumed their pre-spill water 333

use activities. Seven weeks after the Do Not Use order was issued roughly 80% of the 334

households were using the tap water to bathe but less than 12% for cooking or drinking.17 335

Households that participated in the present study were using alternate water sources for 336

drinking, cooking, hygiene activities and their pets. Survey results from both the KCHD17 and 337

CDC18 support this finding. The majority of households visited in the present study relied solely 338

on bottled water for all activities; including bathing during the first two weeks of the event. Two 339

of the 16 households utilized rainwater catchment systems and the KCHD survey17 similarly 340

found few households (less than 10%) used rainwater as an alternate water source. One of 16 341

households purchased an outdoor camping shower for use in lieu of using the indoor shower, and 342

others boiled rainwater for bathing. One household was bathing children in a plastic storage tub 343

with bottled water. Representatives of one of the 16 households traveled 60 miles outside of the 344

affected area to wash clothes and bathe. The KCHD survey17 also found few households (20%) 345

had persons who traveled out of the affected area to meet their water needs. A few households in 346

the present study utilized PUR® water filters in attempt to remove the contaminants, and 347

rainwater and bottled water were the most commonly used water sources for feeding pets.348

Page 20 of 41



21

(a) 349

(b) 350

Figure 2. (a) Visits to Emergency Departments Reported by the CDC and BPH (b) Syndromic Surveillance Cases of Clinically 351

Defined Chemical Exposure from January 9 through February 9, 2014. Records represent 10 emergency medical departments 352

with 356 patients and 10 physician offices with 224 patients, respectively.353

0

10

20

30

40

50

Nu

mb

er o

f P

ati

ents

Symptom Onset Date

Plumbing system flushing begins

0

10

20

30

Nu

mb

er o

f P

ati

ents

Symptom Onset Date

Plumbing system flushing begins

Page 21 of 41



22

Surprisingly, a significant number of households surveyed by the KCHD17 (23%) and the 354

CDC18 (37.4%) reported using the tap water during the Do Not Use order. Of those persons who 355

used tap water during the Do Not Use order, the following activities were most popular [KCHD, 356

CDC]: Bathing or showering [78%.0, 80.1%], hand washing [55.2%, 45.9%], clothes washing 357

[44.0%, 37.7%], dish washing [42.2%, 32.2%], and feeding pets [28.6%, 19.2%]. KCHD17 358

further reported residents used water during the Do Not Use order for teeth brushing (40.5%), 359

drinking (37.1%), cooking (29.3%), and watering plants (23.3%), while CDC18 also reported 360

persons ate or drank food prepared with water (26.6%) and drank the water (26.6%). Results of 361

the present study showed only one household of 16 did not learn about the Do Not Use order on 362

January 9. According to both the KCHD17 and CDC18 however, approximately 20% of the 363

households affected did not learn about the Do Not Use order until after January 9. Household 364

water use during the Do Not Use order likely resulted in some of the illnesses reported (Table 3). 365

Organic Contaminant Levels in the Unflushed and Flushed Homes 366

Odor and 4-MCHM Levels. Of the 10 homes where tap water was analyzed, six homes were 367

unflushed upon arrival and four homes had undergone the flushing procedure before the 368

investigators arrived. Table SI-8 describes the plumbing system characteristics and tap water 369

sampling locations. Tap water odors were detected by the authors in all homes at all taps. The 370

authors described these odors as sweet chemical, strong, sweet licorice/chemical, candy-like, and 371

earthy. Most of these descriptors are in agreement with odor analysis conducted by others9 on the 372

liquid removed from the Freedom Industries, Inc. tank. 373

Despite a tap water odor being detected in all six unflushed homes, 4-MCHM was only 374

detected above the 10 µg/L MDL in four of those residences. Detection of an odor when 4-375

MCHM was not present found at a concentration greater than the MDL was likely due to the 376

Page 22 of 41



23

extraordinarily low odor threshold concentration of contaminants in the contaminated water. 377

WVTAP researchers reported that the odor threshold concentration of the liquid removed from 378

the Freedom Industries, Inc. tank was less than 0.15 µg/L.9 Results implied that the human 379

olfactory system was capable of detecting contaminants when advanced analytical methods 380

could not. 381

Chemical modeling software enabled physiochemical property predictions for many of the 382

spilled liquid’s ingredients. SPARC results indicated that the maximum 4-MCHM solubility 383

from 5°C and 60°C ranged from 1,340 mg/L to 14,900 mg/L (Tables SI-5 to SI-7). However, 384

results from COSMOS-RS chemical modeling software implied that the range was nearly 2,600 385

mg/L at 5°C to 2,800 mg/L at 60°C (Figure SI-4). While the exact reason for this prediction 386

discrepancy is unclear, the 4-MCHM concentration found during in-home testing was much less 387

than its maximum estimated water solubility. 388

The maximum 4-MCHM concentration found in the present study was 420 µg/L. This value 389

was much less than the 3,773 µg/L maximum concentration found by WVAW and the State 390

during water distribution system testing, and 3,120 µg/L concentration found at the water 391

treatment plant (Figure 1). Because no in-home testing was carried-out by responders before 392

flushing, a direct comparison between the maximum 4-MCHM concentrations found here to 393

other affected homes cannot be carried-out. Concentrations found in the present study were 394

within the range found in the distribution system, but are not representative of all homes 395

affected. 396

No trend for 4-MCHM concentrations across or within homes was found (Figure 3). Some 397

differences between faucets were very large (∆=120 µg/L in Home 4) while other differences 398

were very small (∆=2 µg/L in Home 3). These observations could be due to a number of 399

Page 23 of 41



24

phenomena: (1) faucet water use frequency, (2) unstudied plumbing system material interactions, 400

and (3) biodegradation. Bench-scale experiments conducted by the authors revealed that 4-401

MCHM had a low affinity for PEX plumbing pipe (Table SI-9). Chemical oxidation experiments 402

conducted by WVTAP investigators using 4-MCHM and free chlorine showed free chlorine did 403

not affect 4-MCHM concentration.22 404

405

406

Figure 3. Tap Water 4-MCHM Concentration for Unflushed Homes at Different In-Home 407

Locations. The dashed line represents the method detection limit (MDL) of 10 µg/L; Homes 5 408

and 6 did not contain 4-MCHM in concentrations above the MDL; Home 1’s bathroom faucet 409

concentration was below the MDL; Distances shown in parenthesis reflect the straight-line 410

distance from each household to the WVAW treatment plant. Single water samples were 411

analyzed from each tap. 412

413

Odors were detected before and after flushing. Prior to flushing, odors were described as 414

“strong” and after flushing odors were described as “faint” or “very faint”. While flushing Home 415

0

100

200

300

400

500

Kitchen Bath Bath Spigot Kitchen Bath Kitchen Spigot

Home 1 (3 mi) Home 2 (15 mi) Home 3 (13 mi) Home 4 (13 mi)

4-M

CH

M C

on

cen

tra

tion

, µ

g/L

Location

Page 24 of 41



25

1’s plumbing system, investigators experienced strong odors in poorly ventilated bathrooms and 416

a kitchen that did not have functional windows or overhead vent fans. One person experienced 417

eye irritation and another person experienced dizziness. Volatilized chemicals likely caused these 418

symptoms. The strongest odors were most frequently associated with hot water and chemical 419

volatilization likely occurred more rapidly from hot water. It should be noted that adverse health 420

effects occurred well below the CDC’s the 4-MCHM screening level of 1,000 µg/L. WVAW, 421

State, and Federal responders did not advise residents about potential chemical volatilization, 422

inhalation, or dermal exposure concerns. 423

The ability of the plumbing system flushing method to reduce 4-MCHM tap water 424

concentrations was evaluated at four unflushed households (Figure 4). Flushing reduced 4-425

MCHM concentrations in Homes 1 and 3 by 86% and 79% respectively, while the 4-MCHM 426

concentration in Home 2 was relatively unchanged. The observed 4-MCHM reduction in Homes 427

1 and 3 can likely be attributed to less contaminated tap water entering the home from the 428

recently flushed WVAW water distribution system. Home 2 however was located on a cul-de-sac 429

and its result implies equally contaminated water was present in the plumbing system after 430

flushing. The 4-MCHM concentration in Home 5’s tap water was not found above the 10 µg/L 431

MDL either before or after flushing. Future work should be carried-out to understand chemical 432

fate and design premise plumbing flushing protocols that reduce organic contaminant 433

concentrations at building taps. 434

Homes that had been flushed before the research team arrived also contained tap water with 435

characteristic odors, but 4-MCHM was only found greater than 10 µg/L in one home. The 4-436

MCHM concentration found at this home’s bathroom tap was 12 µg/L while the kitchen tap 4-437

Page 25 of 41



26

MCHM concentration was 26 µg/L. Variation between 4-MCHM concentrations at different taps 438

could be attributed to similar reasons discussed previously. 439

440

Figure 4. Tap Water 4-MCHM Concentration Before and After the Premise Plumbing 441

System Flush. Results from a single kitchen or bathroom tap at each home is shown; Dark blue 442

bars represent pre-flush concentration, light grey bars represent post-flush; The text in 443

parentheses describes the straight-line distance to the WVAW treatment plant and number of 444

days after the Do Not Use Order was issued before the plumbing system was flushed; The 445

dashed line represents the MDL of 10 µg/L. Post-flush sample for Home 1 was below the MDL; 446

No 4-MCHM was found above the MDL in Home 5 before or after flushing; Single water 447

samples were analyzed from each tap; Water pre- and post-flush was not analyzed from all 10 448

homes visited. 449

450

Surrogate Tap Water Contamination Indicators in Flushed and Unflushed Homes. To 451

indirectly measure the removal of contaminated water from plumbing systems surrogate water 452

0

20

40

60

80

100

120

Home 1 (3 mi, 9days)




4-M

CH

M C

on

cen

tra

tio

n, µ

g/L

Location

Home 1(3 mi., 9 days)




Page 26 of 41



27

quality indicators [TOC, UV254 absorbance, and specific UV absorbance (SUVA)] were 453

evaluated. At the time this study was conducted, the exact composition including some major 454

and minor ingredients of the spilled chemical mixture and breakdown products formed remained 455

unknown. Previous researchers have recommended that TOC concentration should be applied to 456

detect water distribution system contamination.23 It was expected that if aromatic compounds 457

with double bonds were present, differences between their concentrations would be indicated by 458

differences in the amount of ultraviolet light at 254 nm wavelength the tap water absorbed.24,25 459

Tap water TOC values varied across flushed (0.72 to 2.62 mg/L) and unflushed households 460

(0.74 to 2.02 mg/L). While some tap water TOC values exceeded WVAW’s reported TOC 461

concentration exiting their water treatment plant (1.0 and 1.2 mg/L)26, the difference between the 462

flushed and unflushed home TOC ranges was not significant (p=0.658). Theoretical TOC 463

calculations showed that 4-MCHM would have contributed less than 0.5 mg/L of organic carbon 464

to the water at concentrations found during this in-home study (Table SI-10), within the range of 465

the observed variability within and across homes. Responders should conduct this calculation in 466

response to future contamination incidents to determine if TOC is a valid tap water parameter to 467

monitor. 468

No relationship was found between 4-MCHM concentration and UV254 or SUVA for flushed 469

or unflushed systems. No UV254 absorbance was detected for laboratory prepared aqueous 470

solutions of either 30 mg/L 4-MCHM or 30 mg/L Crude MCHM. It is logical that no correlation 471

was found between 4-MCHM concentrations in flushed and unflushed homes with UV254 472

absorbance (p=0.635, 0.537) or with SUVA levels (p=0.376, 0.598). Results implied that minor 473

ingredients of the spilled solvent were not present at in-home tap water or in sufficient quantity 474

to influence surrogate water quality indicator results. The WVTAP researchers did not find any 475

Page 27 of 41



28

breakdown products or alterations in drinking water odor when the spilled mixture was diluted in 476

drinking water and then exposed to free chlorine or potassium permanganate.22 477

Inorganic Contaminants in Unflushed and Flushed Homes. Water pH, chlorine 478

concentrations, water temperature and turbidity results are shown in Table SI-11. Water pH and 479

chlorine levels detected were within ranges reported on WVAW’s 2013 drinking water consumer 480

confidence report.26 Plumbing system flushing had no overall impact on water pH across the 481

homes; Tap water in flushed homes was generally above pH 7, while all unflushed homes had 482

pH levels less than 7 (p=0.010). Free and total chlorine concentrations did not differ between 483

flushed (1.0 + 0.6 mg/L and 2.5 + 0.7 mg/L) and unflushed homes (0.8 + 0.6 mg/L and 1.7 + 1.2 484

mg/L) (p=0.884, 0.859). During 2013, WVAW’s finished water free chlorine concentration 485

ranged from 0.8 to 2.7 mg/L.26 Chlorine concentrations within homes at different faucets 486

sometimes differed by orders of magnitude (2.65 mg/L vs. 0.05 mg/L); Tap water stagnation in 487

pipes likely contributed to chlorine decay.27 488

Alkalinity, hardness, chloride, fluoride, nitrate, and phosphorus concentrations were similar 489

across the 10 homes studied (Table SI-12). However, in unflushed homes copper and lead 490

concentrations were found above EPA health limits and aluminum, iron, and manganese were 491

detected above EPA recommended aesthetic limits. Because tap water metal concentrations were 492

lower after plumbing systems were flushed, elevated metal concentrations were likely due to 493

water distribution and plumbing system corrosion potentially caused by prolonged water 494

stagnation. Table SI-13 describes unregulated metals found in the tap waters. 495

Several homeowners complained of observing “colored” water or particles exiting their faucets 496

during flushing. Testing revealed that copper and iron likely caused color (Supporting 497

Information). Physical material captured exiting a faucet contained a variety of metals typical of 498

Page 28 of 41



29

water treatment coagulants as well as water distribution system and plumbing component 499

corrosion (Figure SI-5). 500

Page 29 of 41



30

Table 4. Tap Water Metal Concentrations Found in Unflushed and Flushed Homes 501

Parameter1

EPA

Limit

Unflushed Homes Flushed Homes

Min Max Above

Limit? Min Max

Above

limit?

Health Standards, Maximum Contaminant Level (MCL)

As 0.010 < 0.00058 0.00066 No < 0.00058 < 0.00058 No

Ba < 2 0.02 0.05 No 0.02 0.03 No

Be 0.004 < 0.000082 < 0.000082 No < 0.000082 0.000094 No

Cd < 0.005 0.00008 0.00020 No 0.000073 0.000703 No

Cr < 0.1 0.0003 0.0009 No 0.0003 0.0011 No

Cu < 1.3 0.006 1.700 Yes 0.006 0.030 No

Pb 0-0.015 0.0001 0.0200 Yes 0.0002 0.0050 No

Aesthetic Standards, Secondary Maximum Contaminant Level (SMCL)

Al 0.05-0.2 0.01 1.00 Yes 0.001 0.15 No

Fe < 0.3 0.006 1.900 Yes 0.010 0.280 No

Mn < 0.05 0.0005 0.06 Yes 0.0002 0.0200 No

Zn < 5 0.19 0.86 No 0.20 0.32 No

All values shown are reported in mg/L; MCL = Maximum contaminant level; Health standards are primary MCLs while aesthetic 502

standards are secondary MCLs.503

Page 30 of 41



31

LIMITATIONS AND IMPLICATIONS 504

This study has several limitations. Although, the results presented remain the only in-home tap 505

water testing data available that describe odor and chemical quality before and after plumbing 506

system flushing. Absence of WVAW as well as responding local, county, State, or Federal 507

organizations to document chemical levels in-homes has resulted in this knowledge-gap. The 508

results presented were obtained during a rapid response while contaminated water was being 509

purged from area plumbing systems. Because of time and funding limitations, a large-scale 510

sampling plan was not feasible. The results presented provide a unique contribution to the 511

literature. 512

This study was designed to understand resident responses and tap water quality within a set of 513

unflushed and flushed residential plumbing systems. There were more than 93,000 utility 514

customers where approximately 83,000 were residential and 5,000 were businesses spanning 515

nine counties. While the households visited were not randomly selected, their comparison to 516

government agency water distribution system testing, in-home surveys, and public health studies 517

enables their interpretation. The data from the present work are unique. 518

By not conducting in-home testing immediately following the contamination incident, 519

responding organizations failed to document chemical exposure differences within and across 520

homes. This rapid response study found that different 4-MCHM concentrations were present at 521

different taps within and across homes visited. In contrast, subsequent in-home testing conducted 522

by WVTAP researchers revealed that 4-MCHM concentrations did not differ between faucets 523

one month after the spill.28 While the maximum 4-MCHM concentration found in the present 524

study was 420 µg/L, water testing of the utility’s distribution system revealed a significantly 525

Page 31 of 41



32

greater concentration, 3,773 ug/L. It remains unknown if the 3,773 ug/L concentration reached 526

resident taps or was the highest 4-MCHM concentration residents experienced. 527

The absence of 4-MCHM concentration data at the water treatment plant, within the 528

distribution system on January 9, and inside homes during the response inhibits a more complete 529

understanding of chemical exposures experienced by the community. Responding organizations 530

should not only develop analytical test methods during the early hours of the response, but, in 531

parallel begin collecting water samples so that they can be analyzed once the analytical methods 532

are developed. This approach could enable a retrospective examination of community exposures. 533

All water utilities and government agencies should determine which organization is 534

responsible for rapid in-home testing during a crisis. In the US, water utilities generally argue 535

their responsibility ends at the water meter and some public health officials counter that they do 536

not understand plumbing system materials and engineering. This issue must be addressed. Any 537

rapid response sampling plan should be representative of the affected area and include sampling 538

sites and environmental conditions where residents were or are being exposed, in their homes. 539

Understanding chemical concentration differences within and between homes is important. 540

Contaminants of concern could breakdown and contribute different chemical exposures to the 541

residents during water use and flushing. Residual contaminant sources in plumbing systems 542

could also pose continual exposure risks to residents. Because 4-MCHM was not found to react 543

with free chlorine and had limited solubility in PEX plumbing pipe, concentration differences 544

observed in the present study are likely due to water usage, other plumbing material interactions, 545

abiotic, and biotic processes. 546

The premise plumbing flushing procedure reduced 4-MCHM concentrations within some, but 547

not all homes visited and caused persons to experience adverse health impacts. The finding that 548

Page 32 of 41



33

the 4-MCHM concentration before and after flushing one of the plumbing systems was relatively 549

unchanged can be attributed to equally contaminated water being drawn into the home during 550

flushing. Because 4-MCHM concentrations found in the present study were below the CDC’s 551

screening level, exposure to this water should not have caused adverse health impacts. However, 552

results of this study and those of others reviewed here show that illnesses were caused due to 553

premise plumbing flushing. This consequence is likely due to the CDC’s 4-MCHM screening 554

level being inadequate for the water exposure conditions. Illnesses were caused because 555

individuals were exposed to chemicals that had volatilized from the tap water into air and 556

volatilization was promoted at higher water temperatures. Poor indoor air exchange conditions 557

also contributed to the exposures. 558

It should be noted that the CDC did not establish an inhalation screening level and 559

toxicological data is lacking for many of the spilled liquid’s ingredients. Interestingly, nine 560

months after the incident, the EPA announced a health based 30-day air screening level of 0.010 561

ppmv35 for 4-MCHM based on much of the same data the CDC used for its screening level 562

calculation. It remains unknown if adverse health effects would occur at this concentration or if 563

this concentration was exceeded inside affected homes during flushing. To date, no inhalation 564

toxicological studies have been conducted regarding the spilled liquid; Oral and dermal toxicity 565

studies are all that exist. 566

Numerous studies exist that describe plumbing system contamination and flushing approaches. 567

Most of the incidents pertain to removing tap water contaminated with inorganic contaminants 568

and there are too many to list. In these incidents, flowrate, volume of water flushed, and water 569

chemistry were found important.29,30 Less available are flushing case studies for organic 570

contaminant incidents. There has been some discussion of these events, but little data is 571

Page 33 of 41



34

available.31-34 None were found that estimated or monitored inhalation risks due to flushing. In 572

some contamination incidents, affected plastic pipes and other plumbing system components 573

including hot water heaters were replaced because of inadequate contaminant removal or 574

decontamination was not deemed possible.32 Also found was that some testing was carried-out 575

on certain plastics not used for drinking water piping systems. Thus, some of the limited bench-576

scale flushing data may not apply to real-world events. 577

Methods are needed for predicting indoor air concentrations when flushing contaminated tap 578

water into buildings. Any plumbing system flushing procedure should, at the minimum, 579

consider: (1) Contaminants present and their maximum quantities expected, (2) Contaminant 580

physiochemical properties (i.e., water solubility, Henry’s Law Constant, vapor pressure, pKa), 581

(3) Contaminant fate as influenced by water temperature, abiotic, and biotic processes, and (4) 582

Residual sources in the premise plumbing and water distribution systems (i.e., biofilm, corrosion 583

scales, plastic materials, unflushed contaminated water). During future incident responses where 584

indoor air contamination is possible, responders should test flushing protocols under worst-case 585

conditions before premise plumbing flushing is recommended. Rapid testing could determine the 586

procedure’s contaminant removal effectiveness and help identify unanticipated indoor 587

environmental quality and public health issues. Once premise plumbing flushing is 588

recommended, responders should also monitor signs of illness in the community and conduct in-589

home surveys in parallel to detect any unanticipated issues. If illnesses occur, the flushing 590

guidance should be modified. Retrospectively, emergency department physician records and 591

surveys of residents demonstrated that resident health was adversely affected by flushing. 592

To better prepare for an incident, water utilities and government agencies should document and 593

describe the fixed (i.e., tanks, pipelines) and transient (i.e., roadway, barge, etc.) contaminant 594

Page 34 of 41



35

threats to their drinking water sources. A list of high use and high volume contaminants in or 595

routinely passing through the watershed may help prepare for spill response. Additional 596

information that should be considered includes the industrial product’s ingredients, ingredient 597

toxicity, analytical methods and sampling equipment needed for a response, as well as fate and 598

reactivity data for the environment and water infrastructure. When a spill occurs, responders 599

must quickly obtain and characterize a sample of the spilled product as some, but not all, product 600

ingredients are listed on safety data sheets. This was important as WVTAP researchers detected 601

several other contaminants in the spilled liquid not present on the safety data sheet and much of 602

the initial information reported by Freedom Industries, Inc. including the spilled product’s 603

purpose, toxicity, volume spilled, and composition reported was inaccurate. 604

More than 11 months after the chemical spill, several investigations are ongoing and the 605

community is still recovering. The US Department of Justice and Chemical Safety and Hazard 606

Investigation Board have ongoing investigations. The State of West Virginia has expended tens 607

of millions dollars in its response. WVAW has spent more than $12 million, is facing 608

approximately 54 lawsuits, and considering the installation of source water monitoring 609

equipment. Most seriously though, many of the 300,000 people in the area suffered adverse 610

health effects due to contact with contaminated tap water and in part due to being ordered to 611

flush their plumbing systems. This incident demonstrated that a sound scientific approach for 612

responding to and recovering from large-scale tap water contamination incidents is lacking and 613

very much needed. 614

ACKNOWLEDGMENT 615

Thanks are extended to the 16 households who participated in this study along with Rob 616

Goodwin and Maya Nye whom helped us identify households visited. We also acknowledge Drs. 617

Page 35 of 41



36

Kateryna Artyushkova and Jose M. Cerrato at the Univ. New Mexico for their XPS analysis. Dr. 618

Kevin West (Univ. South Alabama) is thanked for providing COSMOS-RS water solubility 619

estimation data. The authors appreciate Coleman Miller, Fredrick Avera, and Mahmoud 620

Alkahout (Univ. South Alabama) who conducted the sorption studies. Drs. David Ladner 621

(Clemson Univ.), Simoni Triantafyllidou (EPA), Kevin Morley (AWWA), Maryam Salehi and 622

Chad Javert (Purdue Univ.) also provided feedback on manuscripts contents. WVAW staff, State 623

of West Virginia officials, and WVTAP researchers are also thanked for their insights. The 624

authors greatly appreciate feedback provided by the anonymous reviewers. 625

SUPPORTING INFORMATION 626

Additional tables as well as figures can be found in the supporting information section. This 627

information is available free of charge via the Internet at http://pubs.acs.org. 628

CORRESPONDING AUTHOR 629

Assistant Professor, Environmental and Ecological Engineering Division, Lyles School of Civil 630

Engineering, Purdue University, 550 Stadium Mall Dr., West Lafayette, IN USA 47907; T: (765) 631

494-2166; F: (765) 494-0395; E: [email protected]; [email protected] 632

AUTHOR CONTRIBUTIONS 633

The manuscript was written through contributions of all authors. All authors have given approval 634

to the final version of the manuscript. †,‡,∆,Φ These authors contributed equally. 635

FUNDING SOURCES 636

Funding for this work was provided by the US National Science Foundation award CBET # 637

1424627. 638

Page 36 of 41



37

NOTES 639

Results and opinions presented in this manuscript only represent the opinions of the authors and 640

not NSF or any other person. 641

ABBREVIATIONS 642

CDC, Centers for Disease Control and Prevention; cPVC, chlorinated polyvinylchloride; DiPPH, 643

dipropylene glycol phenyl ether; EPA, Environmental Protection Agency; KCHD, Kanawha-644

Charleston Health Department; 4-MCHM, 4-methylcyclohexanemethanol; GC/MS, gas 645

chromatography-mass spectrometry; LLE, liquid liquid extraction; MCL, primary maximum 646

contaminant level; MDL, Method detection limit; PB, poly(1-butene) (PB); PEX, crosslinked 647

polyethylene; PPH, propylene glycol phenyl ether; SMCL, secondary maximum contaminant 648

level; SPME, Solid phase microextraction; SUVA, Specific ultraviolet absorbance; TERA, 649

Toxicological Excellence in Risk Assessment; TOC, Total organic carbon; WVAW, West 650

Virginia American Water; WV BPH, West Virginia Bureau of Public Health; WV DHHR, West 651

Virginia Department of Health and Human Resources; WVTAP, West Virginia Testing 652

Assessment Project. 653

LITERATURE CITED 654

(1) Rosen, J.; Whelton, A.J.; McGuire, M.J.; Clancy, J.L.; Bartrand, T.; Eaton, A.; Patterson, J.; 655

Dourson, M.; Nance, P.; Adams, C. WV TAP FINAL REPORT. West Virginia Testing 656

Assessment Project: Charleston, WV USA, July, 2014. Accessible at 657

http://www.wvtapprogram.com. 658

(2) McGuire, M. J.; Rosen, J.; Whelton, A. J.; Suffet, I. H. (2014). An Unwanted Licorice Odor 659

in a West Virginia Water Supply. Journal of American Water Works Association. 72-82. 660

DOI: http://dx.doi.org/10.5942/jawwa.2014.106.0091 661

(3) McIntyre, J.L. Direct Testimony of Jeffrey L. McIntyre. Public Service Commisson of West 662

Virginia, Charleston; CASE NO. 14-0872-W-GI; General Investigation Pursuant to W.Va. 663

CODE 24-2-7 into the Actions of WVAWC in Reacting to the January 9, 2014 Chemical 664

Spill. Submitted by Jackson Kelly, PLLC. 46pp. July 2, 2014. 665

(4) Adams, C.; Whelton, A.J.; Rosen, J. Literature Review: Health Effects for Chemicals in 666

2014 West Virginia Chemical Release: Crude MCHM Compounds, PPH and DiPPH. West 667

Page 37 of 41



38

Virginia Testing Assessment Project: Charleston, WV USA, March, 2014. Accessible at 668


(5) Governor Tomblin Declares State of Emergency for Nine Counties. State of West Virginia 670

Governor Earl Ray Tomblin’s Office, Charleston, WV USA, January 9, 2014. Accessible at 671

http://www.governor.wv.gov/media/pressreleases/2014/Pages/governor-tomblin-declares-672

state-of-emergency-in-9-counties.aspx. 673

(6) Summary Reports of Short-term Screening Level Calculation and Analysis of Available 674

Animal Studies for MCHM, PPH, and DiPPH. US Centers for Disease Control and 675

Prevention (CDC). Atlanta, GA USA, January 20, 2014. Accessible at 676

http://www.bt.cdc.gov/chemical/MCHM/westvirginia2014/pdf/MCHM-Summary-677

Report.pdf.and http://www.bt.cdc.gov/chemical/MCHM/westvirginia2014/pdf/DiPPH-PPH-678

calculation.pdf. 679

(7) Toxicological Excellence in Risk Assessment (TERA). Report of Expert Panel Review of 680

Screening Levels for Exposure to Chemicals from the January 2014 Elk River Spill. West 681

Virginia Testing Assessment Project: Charleston, WV USA, March 2014. Accessible at 682


(8) West Virginia American Water (WVAW). How to Flush your Plumbing System and How to 684

Flush Plumbing Faucets and Appliances: Charleston, WV USA, January 2014. Accessible 685

at http://www.amwater.com/files/WV%20-%20How%20to%20flush.pdf. 686

(9) McGuire, M.J.; Suffett, I.H.; Rosen, J. 2014. Consumer panel estimates of odor thresholds 687

for crude 4-methylcyclohexanemethanol. Journal of the American Water Works 688

Association. 106 (10), E445-E458. DOI: http://dx.doi.org/10.5942/jawwa.2014.106.0129. 689

(10) Whelton, A.J.; Dietrich, A.M.; Gallagher D.L. 2009. Contaminant diffusion, solubility, and 690

material property differences between HDPE and PEX potable water pipes. Journal of 691

Environmental Engineering. 136 (2), 227-237. DOI: 692

http://dx.doi.org/10.1061/(ASCE)EE.1943-7870.0000147. 693

(11) SPARC. ARChem, Inc. Danielsville, GA USA. http://www.archemcalc.com/sparc.html. 694

(12) COSMOS-RS. Scientific Computing & Modelling NV: Amsterdam, NDL. 695

https://www.scm.com/COSMO-RS/. 696

(13) Minitab 14 Student. Minitab, Inc. State College, PA USA. http://www.minitab.com/en-us/. 697

(14) State & County QuickFacts January 6, 2014. US Department of Commerce (DOC), Census 698

Bureau: Washington, D.C. USA. Accessible at 699

http://quickfacts.census.gov/qfd/states/54000.html. 700

(15) US Chemical Safety and Hazard Investigation Board. Freedom Industries Investigation 701

Update, Public Meeting. Charleston, WV USA, January 9, 2014. Accessed at 702

http://www.csb.gov/assets/1/19/Public_Presentation_07_16_14_revised.pdf 703

(16) Whelton, A.J.; Rosen, J.S.; Clancy, J.L.; Clancy, T.P.; Ergul, A. 704

The Crude MCHM Chemical Spill 10-705

Home Study: Resident Behaviors, Perceptions, and Residence Characteristics. West 706

Virginia Testing Assessment Project: Charleston, WV USA, May 2014. Accessible at 707


(17) Community Assessment Population Survey (CAPS). Gupta, R.; Latif, D. Kanawha-709

Charleston Health Department: Charleston, WV USA, April 22, 2014. Accessible at 710

http://www.kchdwv.org/KCHD/media/KCHD-Media/PDF%20Files/2014-05-12-KCHD-711

UC-presentation-updated.pdf. 712

Page 38 of 41



39

(18) Disaster Response and Recovery Needs of Communities Affected by the Elk River Chemical 713

Spill, West Virginia. US CDC, National Center for Environmental Health, Division of 714

Environmental Hazards and Health Effects, Health Studies Branch: Atlanta, GA USA, April 715

2014. Accessible at 716

http://www.dhhr.wv.gov/News/2014/Documents/WVCASPERReport.pdf. 717

(19) Elk River Chemical Spill Health Effects Findings of Emergency Department Record Review. 718

US Centers for Disease Control and Prevention (CDC) Agency for Toxic Substances 719

Disease Registry (ATSDR) and West Virginia Bureau of Public Health (BPH): Charleston, 720

WV USA, April 2014. Accessible at 721

http://www.wvdhhr.org/Elk%20River%20Chemical%20Spill%20Health%20Effects%20-722

%20Findings%20of%20Emergency%20Department%20Record%20Review.pdf. 723

(20) Elk River Chemical Spill Medical Data. West Virginia Bureau of Public Health (BPH): 724

Charleston, WV USA, June 2014. Accessible at http://www.dhhr.wv.gov/News/chemical-725

spill/Documents/PRFindings.pdf. 726

(21) WV Poison Center Response Roles and Responsibilities. Presentation to the National 727

Association of City and County Health Officials Webinar, Response and Recovery During 728

an Environmental Disaster: Learning from the Elk River Chemical Spill. Reported by: 729

Scharman EJ. April 22, 2014. West Virginia Bureau of Public Health: Charleston, WV 730

USA. 731

(22) McGuire, M.J. Oxidation Studies with Crude 4-methylcyclohexanemethanol in Water. West 732

Virginia Testing Assessment Project: Charleston, WV USA, April 2014. Accessible at 733


(23) Pilot-Scale Tests and Systems Evaluation for the Containment, Treatment, and 735

Decontamination of Selected Materials from T&E Building Pipe Loop Equipment, 736

EPA/600/R‐08/016. Prepared by: Shaw Environmental, Inc. for US EPA: Cincinnati, OH 737

USA, 2008. 738

(24) Edzwald, J. (1993). Coagulation in Drinking Water Treatment: Particles, Organics, and 739

Coagulants. Water Science and Technology. 27, 21-35. 740

(25) McKnight, D. M.; Boyer, E. W.; Westerhoff, P. K.; Doran, P. T.; Kulbe, T.; Andersen, D. T. 741

(2001). Spectrofluorometric characterization of dissolved organic matter for indication of 742

precursor organic material and aromaticity. American Society of Limnology and 743

Oceanography, 38-48. 744

(26) WVAW. 2013 Consumer Confidence Report. Charleston, WV USA. 745

(27) Nguyen, C. K. Interactions Between Copper and Chlorine Disinfectants: Chlorine Decay, 746

Chloramine Decay and Copper Pitting. Thesis. Department of Civil and Environmental 747

Engineering, Virginia Tech, Blacksburg, VA, USA, August 2005. 748

(28) Whelton, A.J.; Rosen, J.S.; Clancy, J.L.; Clancy, T.P.; Ergul, A. 749

The Crude MCHM Chemical Spill 10‐Home Study: Tap Water Chemical Analysis. West 750

Virginia Testing Assessment Project: Charleston, WV USA, May 2014. Accessible at 751


(29) Edwards, M.; Parks, J.; Griffin, A.; Raetz, M.; Martin, A.; Scardina, P.; Elfland, C. Lead 753

and Copper Corrosion Control in New Construction, Project #4164. Water Research 754

Foundation: Denver, CO USA, March 2011. 755

(30) Clark, B.; Sheldon Masters, and Marc Edwards. 2014. Profile Sampling to Characterize 756

Particulate Lead Risks in Potable Water. Environmental Science and Technology, 48 (12), 757

6836–6843. DOI: http://pubs.acs.org/doi/abs/10.1021/es501342j 758

Page 39 of 41



40

(31) Welter, G.; Rest, G.; LeChevallier, M.; Spangler, S.; Cotruvo, J.; Moser, R. Standard 759

Operating Procedures for Decontamination of Distribution Systems. Water Research 760

Foundation: Denver, CO, USA, 2006. 761

(32) Removing Biological and Chemical Contamination from a Building’s Plumbing System: 762

Method Development and Testing, EPA 600/R-12/032.US EPA: Washington, DC, USA, 763

May 2012. 764

(33) Chemical Contaminant Persistence and Decontamination in Drinking Water Pipes: Results 765

using the EPA Standardized Persistence and Decontamination Experimental Design 766

Protocol., EPA/600/R-12/514.USEPA: Washington, DC, USA, 2012. 767

(34) Szabo, J.; Minamyer, S. 2014. Decontamination of chemical agents from drinking water 768

infrastructure: a literature review and summary. Environment International. 72, 119-23. 769

DOI: 10.1016/j.envint.2014.01.025. 770

(35) Derivation of an Extrapolated Short-term Inhalation Screening Level for 4-771

Methylcyclohexanemethanol (MCHM – CAS# 34885-03-5). US EPA, Office of Research 772

and Development (ORD), National Center for Environmental Assessment (NCEA), 773

Superfund Health Risk Technical Support Center (STSC): Washington, D.C. USA, July 3, 774

2014.775

Page 40 of 41



41

TOC ART 1

2

Page 41 of 41



Supporting Information Submitted to Environmental Science & Technology

SI‐1 This file will be posted along with the manuscript file.

SUPPORTING INFORMATION 1

2

Residential Tap Water Contamination Following the Freedom Industries Chemical Spill: 3

Perceptions, Water Quality, and Health Impacts 4

Andrew J. WheltonΦ*, LaKia McMillan†, Matt Connell†, Keven M. Kelley†, Jeff P. Gill†, Kevin 5

D. White†, Rahul Gupta‡, Rajarshi DeyΔ, Caroline Novy† 6

7

Φ Division of Environmental and Ecological Engineering and Lyles School of Civil Engineering, 8

Purdue University, West Lafayette, IN USA 9

† Department of Civil Engineering, University of South Alabama, Mobile, AL USA 10

‡ Kanawha Charleston Health Department, Charleston, WV USA 11

Δ Department of Mathematics and Statistics, University of South Alabama, Mobile, AL USA 12

Thirty-six pages 13

14

CORRESPONDING AUTHOR 15

Assistant Professor, Division of Environmental and Ecological Engineering and Lyles School of 16

Civil Engineering, Purdue University, 550 Stadium Mall Drive, West Lafayette, IN USA 47907-17

2051; T: (765) 494-2166; F: (765) 494-0395; E: [email protected]; [email protected] 18



SUPPORTING INFORMATION: TABLE OF CONTENTS 19 20

SI-1.0 INTRODUCTION 21 22

Figure SI-1. Chemical Structures and Common Acronyms of Several Known Ingredients 23

of the Liquid Spilled into the Elk River 24

25

SI-2.0 MATERIALS AND METHODS 26 27

Figure SI-2. Map Depicting the Location of the 10 Unflushed () and Flushed ( ) Homes 28

Where Tap Water was Sampled in Comparison to the Freedom Industries Spill Site () 29

and West Virginia American Water Intake on the Elk River (faucet image). 30

31

Table SI-1. Survey Applied for Each Household 32

33

SI-2.1 Water Quality Analysis Methods Applied to Tap Water Collected in Homes 34

35

SI-2.2 Headspace Solid Phase Microextraction (HS-SPME) GC/MS Analysis for the 36

Crude MCHM Temperature Experiment 37

38

SI-3.0 RESULTS AND DISCUSSION 39 40

SI-3.1 Syndromic Surveillance Patient Demographics and Additional Observations 41

42

Table SI-2. Demographics Comparison of Survey Respondents to 2012 West Virginia 43

Census Data 44

45

Table SI-3. West Virginia Poison Center Calls Related to the Drinking Water Chemical 46

Contamination Incident 47

48

Table SI-4. Symptoms Found Statistically Significantly Related to Each Exposure 49

Location and Exposure Route 50

51


53

Table SI-5. Estimated Log Kow and Water Solubility Values of Spilled Solvent 54

Ingredients at 5 °C, 21 °C, and 60 °C using SPARC Chemical Modeling Software 55

56

Table SI-6. Estimated Vapor Pressures and Henry’s Law Constants of Spilled Solvent 57


59

Table SI-7. Estimated Diffusion Coefficients in Air and Water of Spilled Solvent 60


62



Figure SI-3. Relative Abundance of Each Isomer for a Crude MCHM Solution based on 63

GC/MS Chromatograms Obtained by Applying a Headspace-Solid Phase Microextraction 64

(HS-SPME) Method 65

66

Figure SI-4. Estimated 4-MCHM Water Solubility using COSMO-RS Chemical 67

Modeling Software, 0 °C to 100 °C 68

69

Table SI-8. Tap Water Sampling Location and Type of Plumbing Pipe in Homes Visited 70

71

Table SI-9. Toluene, Cyclohexane, 4-Methylcyclohexanemethanol (MCHM), and 72

Cyclohexanemethanol (CHM) Solubility in PEX Plumbing Pipes at Room Temperature 73

74

Table SI-10. Theoretical and Measured Total Organic Carbon (TOC) Concentration of an 75

Aqueous 4-MCHM Solution and Measured TOC Concentration of a Crude MCHM 76

Solution 77

78

Table SI-11. Initial Tap Water Quality in Six Unflushed Homes 79

80

Table SI-12. Tap Water Alkalinity, Chloride, Fluoride, Hardness, Nitrate, and 81

Phosphorus Concentrations in Unflushed and Flushed Homes 82

83

Table SI-13. Tap Water Unregulated Metal Concentrations in Unflushed and Flushed 84

Homes 85

86

SI-2.4 Water and Material Characterization Results of Investigating Colored Tap Water 87

and Particle Complaints 88

89

Figure SI-5. Solid, Blue-Colored Material Collected when it Exited a Bathroom Tub 90

Faucet 91

92

SI-4 EVENT TIMELINE 93



SI-1.0 INTRODUCTION 94

95 4-Methylcyclohexanemethanol (MCHM) Cyclohexanemethanol (CHM) (**) 96 CAS: 34885-03-5 CAS: 100-49-2 (**) 97

98 Methyl 4-methylcyclohexanecarboxylate 1, 4-Cyclohexanedimethanol (CHDM) 99 CAS: 51181-40-9 CAS: 105-08-8 100

OH

OH3C 101 Dimethyl 1,4-cyclohexanedicarboxylate 4-(Methoxymethyl)cyclohexanemethanol 102 CAS: 94-60-0 CAS: 98955-27-2 103

104 2-Methoxyethoxybenzene (MEB) (**) 105 CAS: 41532-81-4 106

107 Methanol Water 108 CAS: 67-56-1 CAS: 7732-18-5 109

110 1-Phenoxy-2-propanol (PPH) Methyl-2-phenoxyethoxy propanol (DiPPH) 111 CAS: 770-35-4 CAS: 517309-94-0 112

Figure SI-1. Chemical Structures and Acronyms of Spilled Liquid Ingredients. Percentages 113 calculated based on the Crude MCHM safety data sheet and Freedom Industries, Inc. disclosures shown 114 in Table 1. CHM and 2-methoxyethoxybenzene (2-MEB) were found by McGuire et al25 after GC/MS 115



analysis of the product removed from the subject tanks by the West Virginia National Guard. CHM and 2-116 MEB were not listed on either the Crude MCHM or Stripped PPH product safety data sheets. 117 SI-2.0 MATERIALS AND METHODS 118 119

120 121

Figure SI-2. Map Depicting the Location of the 10 Unflushed () and Flushed ( ) Homes 122

Where Tap Water was Sampled in Comparison to the Freedom Industries Spill Site () 123 and West Virginia American Water Intake on the Elk River (faucet image). In Elk View, 124

West Virginia (upper right), three homes were sampled next door to one another (1 unflushed 125

and 2 flushed). The square symbol in the upper right covers the two other symbols. 126

127 128

Table SI-1. Survey Applied for Each Household 129

130 1. When did you find out about the drinking water being contaminated? 131

2. Where did you hear about the incident first? 132

a. TV b. Newspaper c. Radio d. Word of mouth e. Other: 133

3. Who/what organization do you feel is most responsible for the problems this incident? 134

4. What type of pipe is installed in your –DRINKING WATER– plumbing system? 135

a. Copper b. PEX c. cPVC d. PVC e. Other: 136

5. When was your plumbing system installed or last renovated? 137

6. Have you flushed out your entire house, if so when? Date/ Time 138

Aesthetic 139 7. When did you first notice the water odor and describe the types? Has the odor(s) 140

changed? 141



Rate the strength of the water odor from 1-5 (1 no odor, 2 slight, 3 moderate, 4 strong, 5 142

unbearable) 143

DAY:_________________ 1 2 3 4 5 144

DAY:_________________ 1 2 3 4 5 145

DAY:_________________ 1 2 3 4 5 146

DAY:_________________ 1 2 3 4 5 147

8. When did you first notice any coloration in your water? Has the color changed? 148

Rate the intensity of the color from 1-5 (1 clear, 2 slight, 3 moderate, 4 dark, 5 very dark) 149

DAY:_________________ 1 2 3 4 5 150

DAY:_________________ 1 2 3 4 5 151

DAY:_________________ 1 2 3 4 5 152

DAY:_________________ 1 2 3 4 5 153

9. When did you first notice any taste of your water? Has the taste changed? 154

Rate the strength of the taste from 1-5 (1 no taste, 2 slight, 3 moderate, 4 strong, 5 155

unbearable) 156

DAY:_________________ 1 2 3 4 5 157

DAY:_________________ 1 2 3 4 5 158

DAY:_________________ 1 2 3 4 5 159

DAY:_________________ 1 2 3 4 5 160

Health 161 10. Do you have any children, 70 year old people, others who have immunocompromised: 162

11. Describe your level of contact with the water before the incident? After the incident? 163

a. Drinking: 164

b. Showering/bathing: 165

c. Washing clothes: 166

d. Brushing teeth: 167

e. Cooking: 168

f. Animals get water: 169

g. Make baby formula: 170

h. 171

12. Have you felt differently after contacting the water? Yes/No 172

(1 No affect; 2 slightly different, 3 moderately differently; 4 very different, 5 severely 173

different) 174

a. Nausea: 1 2 3 4 5 175

b. Vomiting: 1 2 3 4 5 176

c. Diarrhea: 1 2 3 4 5 177

d. Dizziness: 1 2 3 4 5 178

e. Rash: 1 2 3 4 5 179

f. Numbness: 1 2 3 4 5 180

g. Memory loss: 1 2 3 4 5 181

h. Other: 1 2 3 4 5 182

13. Have you talked with your/a medical doctor since the event occurred? Yes/No 183

14. Contact information 184

Name: Age: Gender: M / F 185

Address: Phone: Email: 186

Number of people living in residence 187



188

SI-2.1 Water Quality Analysis Methods Applied to Tap Water Collected in Homes 189

The University of South Alabama laboratory (Mobile, AL) characterized UV254 absorbance 190

using a HACH® DR 5000 UV-Vis spectrophotometer (EPA 2009) and alkalinity concentration 191

by titration. Water sample color was also characterized by filtering 40 mL through a 0.25 µm 192

glass filtration / cellulose (GF/C) filter (Whatman, diam. 4.7 cm) and permeate analysis with the 193

HACH® DR/2000 direct reading spectrophotometer. 194

ALS Environmental Laboratory, Inc. (Charleston, WV) analyzed various tap water parameters 195

with the following minimum detection limits (MDL): Chloride (0.0071 mg/L), fluoride (0.040 196

mg/L), nitrate (0.0050 mg/L), and orthophosphate (0.0050 mg/L). Chloride, fluoride, nitrate, and 197

orthophosphate levels were characterized in accordance with EPA methods 300 (EPA 2003) and 198

365.1 (EPA 2003). 199

Tap water 4-MCHM levels were determined following EPA SW-846 methods. 4-MCHM 200

(99.0% purity) was obtained from TCI America, Inc. Water samples (approximately 1,000 mL) 201

were extracted using method 3510C with methylene chloride under an acidic pH. The extract 202

was concentrated on a steam bath using a Kuderna Danish apparatus, reduced to a final 1.0 mL 203

volume using nitrogen evaporation, and analyzed using method 8270C with an Agilent 204

5890/5973 gas chromatograph/mass spectrometer (GC/MS) system. Prior to analysis internal 205

standards were added to each sample. Before sample analysis, the GC/MS was tuned to meet the 206

method decafluorotriphenylphosphine (DFTPP) relative mass abundance criteria and calibrated 207

using a six calibration standards. 4-MCHM was calibrated from 5 µg/mL to 500 µg/mL. The 208

MDL was 10 µg/L. 209



Inductively coupled plasma mass spectrometry (ICP-MS) was also applied for metals analysis 210

and hardness determination (Al, Sb, As, Ba, Be, B, Cd, Ca, Cr, Co, Cu, Fe, Pb, Mg, Mn, Mb, Ni, 211

K, Se, Ag, Na, St, Tha, Sn, Ti, and Zn) (EPA 1994). MDLs ranged from 0.000036 mg/L (Sb) to 212

0.018 mg/L (Na). A full metal characterization was conducted because at the time of this 213

sampling, little was known about the ingredients, degradation products, and impurities of the 214

spilled solvent. 215

Dauphin Island Sea Laboratory (Dauphin Island, AL) conducted TOC analysis using a 216

Shimadzu TOC-L CPH/CPN (EPA 1999). Calibration standards of 0, 2, 4, and 5 mg/L were used 217

and a coefficient of determination of 0.9989 was obtained. The TOC MDL was 0.10 mg/L. 218

Method 415.3, Determination of Total Organic Carbon and Specific UV Absorbance at 254 nm 219

in Source Water and Drinking Water, Revision 1.2. US Environmental Protection Agency 220

(EPA): Cincinnati, OH USA, 2009. 221

Method 300.1, Determination of Inorganic Anions In Drinking Water By Ion Chromatography, 222

Revision 1.0. US EPA: Cincinnati, OH USA, 1993. 223

Method 365.1, Determination of Phosphorus by Semi-Automated Colorimetry, Revision 2. US 224

EPA: Cincinnati, OH USA, 1993. 225

Method 200.8, Determination Of Trace Elements In Waters And Wastes By Inductively Coupled 226 Plasma - Mass Spectrometry, Revision 5.4. US EPA: Cincinnati, OH USA, 1994. 227

Method 415.1, Organic Carbon, Total (Combustion Or Oxidation). US EPA: Cincinnati, OH 228

USA, 1999. 229

230

SI-2.2 Headspace Solid Phase Microextraction (HS-SPME) GC/MS Analysis for the Crude 231

MCHM Temperature Experiment 232

The influence of water temperature on 4-MCHM volatilization was investigated by applying 233

headspace / solid phase microextraction (HS-SPME) coupled with GC/MS. The HS-SPME-234

GC/MS method applied during this experiment was adapted from procedures described by Sigma 235

Aldrich (2014) which was reported to quantify analytical grade 4-MCHM in water. Crude 236

MCHM obtained from Eastman Chemical Company, Inc. was used in the HS-SPME Experiment 237



and a 3 mg/L Crude MCHM solution was prepared (by weight). The helium carrier gas flowed at 238

a rate of 0.8 mL per minute. Samples extracted by HS-SPME were prepared by adding 12 mL of 239

dilute Crude MCHM solution to 20 mL amber vials resulting in 8 mL of headspace. The Sigma 240

Aldrich method required the use of 10 mL vials. Vials were then agitated at 55˚C and 250 rpm 241

changing direction every 10 seconds for 30 minutes. A Gerstel MPS autosampler was used to 242

inject a 100 µm polydimethylsiloxane (PDMS) fiber into the sample for the last five minutes of 243

agitation for contaminant adsorption. The PDMS fiber underwent two minutes of desorption. An 244

Agilent 7890A GC and Agilent 5975C mass spectrometer detector were used in splitless mode. 245

The column was Agilent HP-5 30 m x 320 µm x 0.25 µm. The GC oven program started at 50˚C 246

for one minute and increased temperature at a rate of 15˚C per minute until it reached 220˚C. 247

This temperature was maintained for five minutes resulting in a total run time of 17 minutes. The 248

MS was operated in scan mode from m/z 50-100. 4-MCHM isomers were confirmed using a 4-249

MCHM analytical standard and confirmation of retention time at several different 250

concentrations. 251

G006292 SUPELCO, GC Analysis of Methylcyclohexane Methanol (MCHM) and Propylene 252

Glycol Phenyl Ether (PPH) in Water on VOCOL® after SPME using 100 μm PDMS Fiber. 253

Sigma Aldrich, Inc.: St. Louis, MO USA, March 2014. Accessible at: 254

http://www.sigmaaldrich.com/catalog/product/supelco/g006292?lang=en&region=US 255



SI-3 RESULTS AND DISCUSSION 256 257

Table SI-2. Demographics Comparison of Survey Respondents to 2012 West Virginia 258

Census Data 259

260

Demographic Survey Respondents

from this Study US Census Data

for West Virginia Persons per Household 2.93 2.43 Persons under 65 years 81.3% 83.2% Persons 65 years and over 18.7% 16.8% Percent Female 75.0% 50.7% The sampled population for this table included 16 households. 261

262

Table SI-3. West Virginia Poison Center Calls Related to the Drinking Water Chemical 263

Contamination Incident 264 265

Type of Inquiry Received

First 10 Days Before the Do Not

Use Order was Lifted for the Entire

Area1

Following the Do Not Use Order being Lifted for All Affected Areas

First Week2 Second Week3

Human Exposure 1,939 95 84 Animal Exposure 100 0 2 General Information 384 101 80

TOTALS 2,423 196 166 1. Jan. 9 (5:57 pm) to Jan. 19 (12:30 am); 2. Jan. 19 (12:31 am) to Jan. 26; 3. Jan. 27 to Feb. 1. 266 267



Table SI-4. Symptoms Found Statistically Significantly Related to Each Exposure Location 268

and Exposure Route 269

Factors (p-value and Odds Ratio) Symptoms

Drinking (0.000, OR=3.48), Work (0.012, OR=4.52), Sex (0.052, OR=0.48) Nausea

Drinking (0.000, OR=6.35), Sex (0.023, OR=0.33) Vomiting

Drinking (0.000, OR=6.26), Food facility (0.008, OR=26.46), Work (0.013,

OR=7.76), Home(0.051, OR=8.03) Diarrhea

Onsetc1/peak 1(0.017, OR=2.28) Eye Irritation

Wash/Bath (0.000, OR=5.49), Onsetc2/peak2 (0.001, OR=0.35), Drinking

(0.074, OR=0.53) Skin Irritation

Wash/Bath (0.003, OR=2.9), Drinking (0.053, OR=0.54) Rash

Work (0.003, OR=4.62), Food Facility (0.015, OR=5.85), Sex (0.096,

OR=0.48) Headache

“None” Cough

Drinking (0.002, OR=5.20), Flushed (0.00, OR=0.20), Sex (0.025,

OR=0.17) Sore Throat

Type I error of 0.05 was applied to all statistical tests; OR= Odds ratio; Peak1= January 9; Peak 2= 270

January 13; Sex= 1 for female, 0 for male; Bold factors indicate statistically significant result; Odds ratios 271

(OR) represent the probability of a patient reporting a symptom if the patient became ill during the first 272

symptom reporting peak. For example, eye irritation is 2.28 times as likely to occur if the patient became 273

ill during the first symptom peak. 274



Of the 224 persons who sought medical attention from one of the 10 sentinel medical 276

providers, 61% were female and 35% were male. The mean patient age was 40.1 years with a 277

minimum age of 2 years and maximum of 87 years. 278

Relationships between symptoms and exposure location (home, work, food facility) were 279

found. However, because the extreme majority of patients were exposed to the water in their 280

home environment, the influence of “home exposure” for many of the symptoms could not be 281

detected. Although, patients who were exposed to the chemicals in their work environment were 282

more likely to report nausea (p=0.012), diarrhea (p=0.013) and headache (p=0.003), patients 283

who got exposed to the chemicals in a food facility were more likely to report diarrhea (p=0.008) 284

and headache (p=0.015). 285


Table SI-5. Estimated Log Kow and Water Solubility Values of Spilled Solvent Ingredients at 5 °C, 21 °C, and 60 °C using 286

SPARC Chemical Modeling Software 287

288

Constituent Log Kow Water Solubility, mg/L

5°C 21°C 60°C 5°C 21°C 60°C 4-Methylcyclohexanemethanol (MCHM) 2.77 2.43 1.71 1,340 2,900 14,900 Methyl 4-methylcyclohexanecarboxylate 2.81 2.54 2.51 508 924 964 Dimethyl 1,4-cyclohexanedicarboxylate 1.65 1.58 2.02 8,210 9,550 3,330 Methanol -0.58 -0.58 -0.59 10,000 10,000 10,000 1,4-Cyclohexanedimethanol (CHDM) 0.78 0.62 0.17 15,200 28,100 176,000 Cyclohexanemethanol (CHM) 2.27 1.98 1.38 3,630 7,040 27,600 2-Methoxyethoxybenzene (MEB) 2.70 2.44 2.39 1,370 2,510 2,790 1-Phenoxy-2-propanol (PPH) 2.49 2.19 1.88 2,020 4,620 11,300 NOTES: SPARC could not estimate physiochemical properties of two compounds: 4-(methoxymethyl)cyclohexanemethanol or methyl-2-289 phenoxyethoxy propanol (DiPPH); 2-MEB and CHM were not reported on any product safety data sheet (SDS), but were found in the solvent 290 remaining in the Freedom Industries, Inc. tank number 396 by the WVTAP independent researchers who applied GC/MS analysis. 291

292


Table SI-6. Estimated Vapor Pressures and Henry’s Law Constants of Spilled Solvent Ingredients at 5 °C, 21 °C, and 60 °C 293

using SPARC Chemical Modeling Software 294

Constituent Vapor Pressure, mmHg Constant, atm-m3/mol

5°C 21°C 60°C 5°C 21°C 60°C 4-Methylcyclohexanemethanol (MCHM) 0.013 0.073 1.50 1.68 x 10-6 4.22 x 10-6 1.69 x 105 Methyl 4-methylcyclohexanecarboxylate 0.11 0.44 4.46 4.28 x 10-5 9.71 x 10-5 9.51 x 10-4 Dimethyl 1,4-cyclohexanedicarboxylate 0.002 0.014 0.22 7.71 x 10-8 3.89 x 107 1.71 x 10-5 Methanol 48.7 118.9 633.5 2.69 x 10-6 6.34 x 10-6 3.51 x 10-5 1,4-Cyclohexanedimethanol (CHDM) 1.67 x 10-6 1.95 x 10-5 0.0017 2.07 x 10-11 1.31 x 10-10 2.61 x 10-9 Cyclohexanemethanol (CHM) 0.028 0.14 2.58 1.14 x 10-6 3.01 x 10-6 1.41 x 10-5 2-Methoxyethoxybenzene (MEB) 0.023 0.11 1.29 3.17 x 10-6 8.75 x 10-6 9.26 x 10-5 1-Phenoxy-2-propanol (PPH) 9.81 x 10-4 0.01 0.20 9.74 x 10-8 *** 3.56 x 10-6 NOTES: *** SPARC did not output a value for this condition; SPARC could not estimate physiochemical properties of 295

two compounds: 4-(methoxymethyl)cyclohexanemethanol or methyl-2-phenoxyethoxy propanol (DiPPH); 2-MEB and 296

CHM were not reported on any product safety data sheet (SDS), but were found in the solvent remaining in the Freedom 297

Industries, Inc. tank number 396 by the WVTAP independent researchers who applied GC/MS analysis. 298


Table SI-7. Estimated Diffusion Coefficients in Air and Water of Spilled Solvent Ingredients at 5 °C, 21 °C, and 60 °C using 299

SPARC Chemical Modeling Software 300

301

Constituent Air, cm2/s Water, cm2/s

5°C 21°C 60°C 5°C 21°C 60°C 4-Methylcyclohexanemethanol (MCHM) 0.052 0.058 0.073 4.8 x 10-6 6.77 x 10-6 1.76 x 10-5 Methyl 4-methylcyclohexanecarboxylate 0.048 0.053 0.0671 4.4 x 10-6 6.21 x 10-6 1.62 x 10-5 Dimethyl 1,4-cyclohexanedicarboxylate 0.042 0.047 0.059 4.12 x 10-6 5.81 x 10-6 1.52 x 10-5 Methanol 0.138 0.153 0.191 1.01 x 10-5 1.42 x 10-5 3.67 x 10-5 1,4-Cyclohexanedimethanol (CHDM) 0.048 0.053 0.0667 4.9 x 10-6 6.91 x 10-6 1.81 x 10-5 Cyclohexanemethanol (CHM) 0.058 0.064 0.0807 5.22 x 10-6 7.36 x 10-6 1.92 x 105 2-Methoxyethoxybenzene (MEB) 0.050 0.055 0.0697 4.61 x 10-6 6.51 x 10-6 1.71 x 10-5 1-Phenoxy-2-propanol (PPH) 0.0507 0.0563 0.0708 4.75 x 10-6 6.7 x 10-6 1.76 x 10-5 NOTES: SPARC could not estimate physiochemical properties of two compounds: 4-(methoxymethyl)cyclohexanemethanol or methyl-2-302 phenoxyethoxy propanol (DiPPH); 2-MEB and CHM were not reported on any product safety data sheet (SDS), but were in the solvent remaining 303 in the Freedom Industries, Inc. tank number 396 by the WVTAP independent researchers who applied GC/MS analysis. 304


305

Figure SI-3. Relative Abundance of Each Isomer for a Crude MCHM Solution based on a Headspace-Solid Phase 306

Microextraction (HS-SPME) GC/MS Method. Crude MCHM concentration (30 mg/L) was determined by weight; Each bar 307

represents the mean and standard deviation of three replicates; Peak height was 31% greater at 60°C than 23°C for both the trans- and 308

cis-isomers; The method could be optimized to obtain greater peak heights.309

0

2,000

4,000

6,000

8,000

10,000

12,000

Trans-isomer Cis-isomer Total isomers

Pea

k H

eigh

t -

cou

nt

Isomers and Total

23C 60C


Table SI-8. Tap Water Sampling Location and Type of Plumbing Pipe in Homes Visited 310

6 Unflushed Households 4 Flushed Households

Location Pipe Type Location Pipe Type

Home 1 Kitchen Cu, Fe

Home 7 Kitchen Fe

Bathroom Cu, Fe Bathroom Fe

Home 2 Bathroom Cu, cPVC

Home 8 Kitchen Cu

Garage Spigot Cu Bathroom Cu

Home 3 Kitchen Cu

Home 9 Kitchen cPVC

Bathroom Cu Bathroom cPVC

Home 4 Kitchen cPVC

Home 10 Kitchen PEX

Outside Spigot cPVC Outside Spigot PEX

Home 5 Kitchen PEX, PB

Outside Spigot PEX, PB

Home 6 Kitchen Cu, PEX

Bathroom Cu, PEX Pipe types: Copper (Cu), iron (Fe), chlorinated polyvinylchloride (cPVC), crosslinked polyethylene (PEX), poly(1-311

butene) (PB) 312

313

Figure SI-4. Estimated 4-MCHM Water Solubility using COSMO-RS Chemical Modeling 314

Software, 0 °C to 100 °C315


Table SI-9. Toluene, Cyclohexane, 4-Methylcyclohexanemethanol (MCHM), and Cyclohexanemethanol (CHM) Solubility in 316

PEX Plumbing Pipes at Room Temperature 317

318

Type of Pipe and Compound Total Mass

Increase, Percent Time to Reach

Half-Saturation, Days Total Test

Duration, Days Solubility,

g/cm3 x 100 PEX-A Type Pipe (medium density resin)

Toluene 9.14 + 0.02 0.48 12.9 8.58 + 0.00 Cyclohexane 8.79 + 0.08 3.13 18.1 8.25 + 0.00 4-Methylcyclohexanemethanol (MCHM) 0.71 + 0.12 < 0.07 47.3 0.66 + 0.11

PEX-B Type Pipe (high density resin) Toluene 8.20 + 0.90 0.62 11.8 7.85 + 0.09 Cyclohexanemethanol (CHM) 0.40 + 0.03 < 0.04 46.6 0.38 + 0.03

Mean and standard deviation values are shown for three replicate specimens (2.5 to 2.6 mm thick) per contaminant; Toluene and cyclohexane were 319

examined as controls. Neither compound was present in the spilled liquid. Toluene results were similar to those reported for PEX pipes by others 320

(Whelton et al. 2009. Contaminant diffusion, solubility, and material property differences between HDPE and PEX potable water pipes. Journal of 321

Environmental Engineering. 136 (2), 227-237). The increase in sample mass indicated diffusion occurred, but 4-MCHM and CHM diffusion 322

coefficients could not be calculated due to the rapid update of contaminant during the short exposure period. Both 4-MCHM and CHM were 323

present in the spilled liquid. The manufacturer reported bulk density of the PEX-A pipe was 0.935 g/cm3. A bulk density value was not reported 324

for PEX-B pipe, but it was estimated to be most nearly 0.957 g/cm3 based on Whelton et al (2009). 325


Table SI-10. Theoretical and Measured Total Organic Carbon (TOC) Concentration of an Aqueous 4-MCHM 326

Solution and Measured TOC Concentration of a Crude MCHM Solution 327

328

Mass of 4-MCHM analytical standard added to 1 L of

Water, mg

Theoretical TOC Concentration for the 4-MCHM Solution, mg/L

Measured TOC Concentration for the 4-MCHM Solution, mg/L

Measured TOC Concentration for the Crude

MCHM Solution, mg/L 0 0 BDL BDL

0.5 0.375 0.58 + 0.02 0.90 + 0.09 1.5 1.125 1.06 + 0.02 0.93 + 0.50 3.1 2.325 1.83 + 0.03 2.18 + 0.05 30 22.5 17.5 + 0.51 15.8 + 0.40

The MDL for TOC analysis for tap water samples applied during this study was 0.10 mg/L; BDL = Below detection limit; The molar mass of 4-329

MCHM is 128 g/mol and the ratio of the total mass of carbon to the 4-MCHM’s molar mass is 0.75; Three replicate measurements were used for 330

preparing the 4-MCHM solutions; The Crude MCHM analyzed in this experiment had an experimentally measured density of 0.917 + 0.010 g/cm3 331

which was not statistically different than the 0.908 + 0.004 g/cm3 manufacturer reported value (Steinbugler M. Crude MCHM Biochemical 332

Oxygen Demand Determination. Eastman Kodak Company: September 30, 1997, Rochester, NY USA).333


Table SI-11. Initial Tap Water Quality in Six Unflushed Homes 334

335

Location and (No. Samples)

Water Quality Characteristic

Temp, °C pH Free Cl2,

mg/L Total Cl2,

mg/L Turbidity,

NTU

Cold Water Taps

Mean (11) 14.4 + 4.6 6.9 + 0.4 0.8 + 0.6 1.7 + 1.2 7.3 + 13.3

Range (11) 7.4 to 19.8 6.2 to 7.3 0 to 1.4 0 to 3.1 0 to 43.0

Hot Water Tap Home #6

Single Tap 20.0 7.3 0.6 1.4 0.2

Mean and standard deviation values shown. Two cold water samples were collected per household, except 336 for Home 6 where one water sample was collected from a hot water line. The hot water line was sampled 337 because the cold water line had been turned off due to a pipe break. The water sample collected from the 338 hot water line was at room temperature. 339 340

341

Table SI-12. Tap Water Alkalinity, Chloride, Fluoride, Hardness, Nitrate, and Phosphorus 342

Concentrations in Unflushed and Flushed Homes 343

344

Parameter1 EPA Limit

Unflushed Homes Flushed Homes

Min Max Above Limit?

Min Max Above limit?

General Alkalinity None 13.75 22.50 - 15.00 18.53 - Hardness None 39 51 - 39 46 - Chloride < 250 10 18 No 10 11 No Fluoride < 4 0.56 0.65 No 0.52 0.58 No Nitrate-N < 1 0.37 0.55 No 0.39 0.41 No OrthoP None 0.17 0.28 No 0.16 0.33 No

All values shown are reported in mg/L; Fluoride and nitrate have health based EPA maximum 345

contaminant levels; The other contaminants have EPA recommended aesthetic limits. 346


Table SI-13. Tap Water Unregulated Metal Concentrations in Unflushed and Flushed 347

Homes 348

349

Element Unflushed Homes Flushed Homes

Maximum Flushed Level Compared to

Unflushed Home Level Min Max Min Max

Sb < 0.000036 0.000230 < 0.000036 0.000150 Lower 34% B 0.0074 0.0510 0.0090 0.0180 Lower 64% Ca 9.0 12.0 9.2 10.0 Lower 16% Co < 0.000041 0.000130 0.000011 0.000240 Higher 84% Mg 4.0 5.5 4.0 4.9 Lower 10% Mo < 0.000088 0.00059 < 0.000088 0.00046 Lower 22% K 1.2 1.4 1.2 1.6 Higher 14% Se < 0.00064 0.00096 < 0.00064 0.00097 Higher 1% Ag < 0.000042 0.000085 < 0.000042 0.000180 Higher 111% Na 8.0 12.0 7.8 8.4 Lower 30% Ni 0.0001 0.0200 0.0004 0.0050 Lower 75% St 0.052 0.071 0.054 0.060 Lower 15% Tha < 0.000062 0.000180 < 0.000062 0.000820 Higher 380% Sn < 0.000054 0.000350 < 0.000054 0.000340 Lower 2% Ti < 0.000086 0.001500 0.000410 0.000940 Lower 37%

All results shown as mg/L 350


SI-2.4 Water and Material Characterization Results of Investigating Colored Tap Water 351

and Particle Complaints 352

Several homeowners complained of observing “colored” water or particles exiting their faucets 353

during flushing. Copper was detected when blue colored water was reported at Home 1 (0.19 to 354

1.7 mg Cu/L) and Home 8 (0.12 to 0.29 mg Cu/L). Both homes contained copper piping. All 355

samples collected by the sampling team appeared to have no coloration. After glass fiber 356

filtration, it was determined that colored tap water existed in three unflushed households: Home 357

1 (Cu, Fe, 8.00 color units), Home 2 (Cu, cPVC, 2.0 color units), and Home 4 (cPVC, 11.5 color 358

units). All observed tap water color values were below the EPA SMCL for color which specifies 359

“most people find color objectionable over 15 color units” (EPA 2014). The tap water was highly 360

susceptible to causing corrosion (Sarver and Edwards 2012) as it had moderate levels of free 361

chlorine, low alkalinity, and had its residence time increased due to the Do Not Use order. 362

XPS analysis of the particulate material released from a bathroom faucet in Home 8 contained 363

less than 1% silica, and primarily consisted of oxygen (58%), aluminum (18%), carbon (17%), 364

zinc (1.4%), phosphorus (1.4%), fluorine (1.0%), chromium (as Cr+6) (0.8%), calcium (0.2%), 365

and copper (0.1%). The source of these elements could be due to the WVAW’s use of aluminum 366

sulfate as a water treatment coagulant (McGuire 2014), orthophosphate as a corrosion inhibitor, 367

as well as water piping and faucet component corrosion. The source of the particulate material 368

could be due to a corrosion deposit released from the home’s copper plumbing system or water 369

distribution system. 370

McGuire, M.J. Oxidation Studies with Crude 4-methylcyclohexanemethanol in Water. West 371

Virginia Testing Assessment Project: Charleston, WV USA, April 2014. Accessible at 372



Secondary Drinking Water Regulations: Guidance for Nuisance Chemicals. US EPA: 374

Washington, DC USA. Accessed November 20 2014 at 375

http://www.water.epa.gov/drink/contaminants/secondarystandards.cfm. 376

Sarver E, Edwards M.A. 2012. Inhibition of Copper Pitting Corrosion in Aggressive Potable 377

Waters. International Journal of Corrosion. 2012 (2012). 16 pp. 378

379

380 381

Figure SI-5. Solid, Blue-Colored Material Collected when it Exited from a Bathroom Tub 382

Faucet 383


SI-4.0 EVENT TIMELINE 384

385 This timeline was constructed based on firsthand experiences of the authors, discussions with 386

government, utility officials, as well as media representatives who reported on the incident. Brief 387

descriptions of events have been provided. 388

389

BEFORE JANUARY 9, 2014 390 During April 2010 a resident complains to the West Virginia Department of Environmental 391

Protection (DEP) of an odor that smells like licorice near the Freedom Industries, Inc. 392

site. 393

Weeks before January 9, 2014 a few residents near Barlow Drive in Charleston, West 394

Virginia and the Elk River smelled licorice odor. 395

396

JANUARY 9, 2014 397 7:30 am – A resident living across the street from the Freedom Industries, Inc. facility calls 398

the DEP and states noticing “something in the air at the 77-79 split each morning when 399

he comes into work. He says it is coating his wife’s throat.” Person had smelled the odor 400

for a few weeks but it was particularly strong and objectionable on January 9. 401

10:30 am – Freedom Industries, Inc. discovers a leak at the bottom of storage tank number 402

396 labeled “CRUDE MCHM.” 403

10:48 am – Local media organizations report detecting “licorice” odors. 404

11:15 am – DEP inspectors arrive at Freedom Industries, Inc. and meet a man who identified 405

himself as the company President. 406

11:25 am – Freedom Industries, Inc. President discloses that there is a problem with a tank 407

leaking. DEP inspectors observe a pool of liquid and “upwelling fountain-like liquid 408

flow.” 409

12:05 pm – Freedom Industries, Inc. official reports the spill to DEP spill hotline, explained 410

that the product is 4-methylcyclohexanemethanol (4-MCHM). 411

12:00 to 12:30 pm – West Virginia American Water (WVAW) is notified of the spill by 412

local emergency personnel. WVAW increases potassium permanganate and free 413

chlorine oxidant doses and begins feeding powdered activated carbon at their 414

treatment plant. 415 2:13 pm – Kanawha County Homeland Security issues a directive to residents to go indoors 416

if they showed signs of “nausea, etc.” 417

3:00 pm – Kanawaha County Health Department (KCHD), the health department serving 418

250,000 people, receives complaints of a strong “licorice” odor. 419

4:00 pm – WVAW indicates that their water treatment plant was overwhelmed by the 420

contaminated water which has now entered the water distribution system. 421 5:23 pm –West Virginia Department of Military Affairs and Public Safety notifies the public 422

of an upcoming “Do Not Use” water order for WVAM customers via Twitter 423

5:36 pm – West Virginia Governor Earl Ray Tomblin issues a warning to citizens: 424

“EMERGENCY: Do NOT use tap water for drinking, cooking, washing or bathing 425

in Boone, Lincoln, Kanawha, Jackson, Putnam counties.” 426 5:45 pm – WVAW issues the Do Not Use tap water order, permitting only toilet flushing and 427

firefighting. 428


5:54 pm – The Governor declares a State of Emergency for “Boone, Cabell, Clay, Jackson, 429

Kanawha, Lincoln, Logan, Putnam, and Roane counties. Residents served by Lincoln 430

Public Service District (PSD), Queen Shoals PSD, Reamer PSD, City of Culloden PSD, 431

and City of Hurricane PSD are also affected. West Virginians in the affected service 432

areas are urged NOT to use tap water for drinking, cooking, washing or bathing. Right 433

now, our priorities are our hospitals, nursing homes, and schools. ” 434

6:00 pm – Local health departments order the closing of all impacted permitted facilities 435

including food establishments and schools across nine counties. 436

7:00 pm – News report of residents seeking bottled water local grocery stores, gas stations 437

and pharmacies. A few media organizations report that some stores have experienced 438

thefts and fights among shoppers. 439

Afternoon / Evening – The West Virginia Department of Health and Human Resources 440

(DHHR) requests assistance from the US Centers for Disease Control and Prevention 441

(CDC) Agency for Toxic Substances and Disease Registry (ATSDR) regarding the 442

drinking water contamination incident. The CDC develops a short-term screening level 443

for the drinking water affected by the chemical spill, specifically for 4-MCHM of 444

1,000 µg/L. Screening levels are not established for any other contaminants at this 445 time. This calculation was reviewed and approved by a federal interagency workgroup 446

composed of representatives from the National Institute of Environmental Health 447

Sciences, National Toxicology Program, National Institutes of Health, U.S 448

Environmental Protection Agency (EPA). 449

Afternoon / Evening – The West Virginia DHHR, National Guard, WVAW, and the 450

chemical manufacturer begin developing analytical methods required for characterization 451

of river and tap waters. A liquid liquid extraction approach with methylene chlorine and 452

gas chromatography - flame ionization detector (GC-FID) method is developed for 4-453

MCHM. Once developed, the method is sent to Federal partners for review. The initial 454

method detection limit (MDL) is 50 µg/L, and then is lowered to 10 µg/L. 455

Evening – A West Virginia Interagency working group that includes West Virginia 456

DHHR, National Guard, DEP, DHSEM, and the Governor’s office decide that due 457

to the limited toxicological data available for 4-MCHM, an additional 100x safety 458

factor should be applied to the CDC screening level. The West Virginia 4-MCHM 459

screening level is determined to be 10 µg/L, the lowest MDL at the time. 460 461

JANUARY 10 462 12:46 am – President Barack Obama declares the incident a federal disaster. US 463

Department of Homeland Security's Federal Emergency Management Agency (FEMA) 464

announces federal emergency aid is now available “to supplement state and local 465

response efforts authorized under Title V of the Stafford Act, to save lives and to protect 466

property and public health and safety, and to lessen or avert the threat of a catastrophe in 467

the counties of Boone, Clay, Jackson, Kanawha, Lincoln, Logan, Putnam, and Roane.” 468

10:00 am – Media organizations report that some water distribution sites are being 469

overwhelmed. 470

10:30 am – DEP officials order Freedom Industries, Inc. to halt operations, remove 471

remaining chemicals from its above-ground tanks, and estimate between 2,000 to 5,000 472

gallons of “Crude MCHM” leaked from tank number 396. 473


12:30 pm – The West Virginia DHHR, National Guard, and WVAW begins collecting 474

water entering and exiting the WVAW treatment plant. Levels show 3,350 µg/L 475

entering and 1,560 µg/L exiting the plant. No 4-MCHM data are available prior to 476

12:30 pm. Water however was transiting through the water treatment facility 477

during this time as toilet flushing and fire-fighting activities were permitted for the 478

300,000 service population. 479

4:00 pm – The drinking water screening level of 1,000 µg/L for 4-MCHM developed by 480

CDC is publicly announced by Major General James A. Hoyer, the Adjutant 481 General of the West Virginia National Guard. The methodology applied and exposure 482

duration (1-day, 3-day, 14-day, 28-day) for the CDC screening level is not made public at 483

this time. 484

5:30 pm – Freedom Industries, Inc. President apologizes for leak and drinks bottled water 485

during press conference. The President states “Look guys, it has been an extremely long 486

day, I'm having trouble talking at the moment. I would appreciate it if we could wrap this 487

thing up.” 488

The CDC declines to establish a 4-MCHM air screening level citing the lack of 489

supporting toxicological information. The effort was revisited in February 2014 by 490

the CDC and again, the agency declined to establish an air screening level. Scientists 491

from the CDC and EPA continue to discuss the air screening level in the months 492

following the spill. In October, however, the EPA establishes an air screening level 493

for 4-MCHM. 494 495

JANUARY 11 496

The 4-MCHM concentration exiting the WVAW treatment facility decreases below 497

1,000 µg/L. 498

The West Virginia DHHR reports that 32 people have sought medical treatment at area 499

hospitals and four people have been hospitalized with symptoms including nausea 500

and vomiting. 501

The West Virginia DHHR also reports that the WV Poison Center was overwhelmed 502 with calls. Questions pertain to health impacts and queries about keeping goats, chicken 503

and other farm animals safe from exposure. Later in July, during a National Association 504

of City and County Health Officials (NACCHO) webinar, the Poison Center reports that 505

the number of persons who called the center and did not get through is unknown because 506

their telephone capacity was reached. 507

U.S. Chemical Safety and Hazard Investigation Board (CSB) investigation begins at the 508

request of West Virginia Senator Jay Rockefeller. 509

Freedom Industries, Inc. announces that the quantity of MCHM leaked was 7,500 gallons. 510

KCHD begins collecting syndromic surveillance data of its 10 sentinel providers. 511 KCHD begins to conditionally reopen permitted food facilities based on their demonstration 512

of provision of potable water for use. 513

514

JANUARY 13 515 9:02 am – Since January 11, the 4-MCHM level exiting the WVAW treatment facility did not 516

exceed 0.920 mg/L. 517

12:00 pm – WVAW announces first “zones” in downtown Charleston are cleared to 518 flush their plumbing systems and residential customers will be given a 1,000 gallon 519


credit. A plumbing system flushing guidance document reviewed by government health 520

officials was released to the public by WVAW. The document states persons should 521

discard any in-home filters, flush hot and cold water from their plumbing system using a 522

prescribed approach, and “any lingering smell, which is expected, is not a health issue.” 523

Hospitals located in Charleston are permitted to flush their plumbing systems first followed 524

by buildings located in other pressure zones. One instance occurred over the next several 525

days where residents within a pressure zone began flushing which resulted in a boil water 526

order due to low pressure. On January 18, plumbing system flushing had been 527

recommended for all pressure zones. 528

Residents begin flushing contaminated tap water from their plumbing systems into 529

their home spaces. Residents discharge contaminated tap water into the sewer 530

system, their septic tanks, and on ground surfaces. 531 532

JANUARY 14 533 The Charleston Sanitary Board operations manager states that the wastewater treatment 534

facility can handle the excess water generated due to plumbing system flushing and 535

estimates that the treatment facility can oxidize 4-MCHM. Treated wastewater treatment 536

plant effluent is discharged to the Kanawha River. No licorice odors are reported at the 537

facility. DEP conducts effluent monitoring and 4-MCHM is not detected above the MDL. 538

539

JANUARY 15 540 10:00 am – KCHD reports 101 people visited area emergency rooms between 7 pm January 541

13 and 7 am January 15 with symptoms claimed to be related to using tap water. 542

12:00 pm – The last water sample in the WVAW treatment plant where time 4-MCHM was 543

detected with an MDL of 10 µg/L. 544

5:00 pm – The Do Not Use order is lifted for approximately half of the affected customers. 545

7:00 pm – The DEP issues five notice of violations (NOV) to Freedom Industries, Inc. 546

because the location where the remaining unspilled solvent was moved to in Nitro, West 547

Virginia does not contain a secondary contaminant wall. 548

8:30 PM – CDC Director, Dr. Thomas Frieden advises West Virginia DHHR through a 549

letter, "Due to limited availability of data, and out of abundance of caution, you 550

may wish to consider an alternative drinking water source for pregnant women 551

until the chemical is at non-detectable levels in the water distribution system." 552 9:00 pm – Media calls Dr. Thomas Freiden at home for clarification. Dr. Freiden asks media 553

to contact the CDC press office. 554

555

JANUARY 17 556

8:00 am – Unfunded University of South Alabama professors Andrew Whelton, Kevin 557

White, and science and engineering students LaKia McMillan, Keven Kelley, Matt 558

Connell, and Jeff Gill arrive from Alabama to provide residents assistance. The 559

group teams up with two nonprofit organizations, the West Virginia Clean Water 560 Hub and People Concerned About Chemical Safety. The team visits several affected 561

homes for tap water testing, helps conduct plumbing system flushing, and distributes 562

bottled water to the community. 563

Freedom Industries, Inc. files for Chapter 11 bankruptcy protection. According to the filing, 564

the company estimates $10 million in assets and $10 million in liabilities. Freedom 565


Industries, Inc. hypothesizes that a local water line break adjacent to the Charleston 566

facility may have caused or contributed to the ground beneath a storage tank at the 567

Charleston facility to freeze in the extraordinary frigid temperatures in the days 568

immediately preceding the incident. 569

Bulk emergency water distribution trailers located at a distribution site in Elk View, West 570

Virginia were found to be distributing contaminated 4-MCHM tap water to the 571

population. Officials halt distribution at that site and direct persons who collected that 572

water to discard it. 573

The CDC first begins posting their 4-MCHM screening level calculation and 574

justification on their website. 575 576

JANUARY 18 577 University of South Alabama researchers experience firsthand that the plumbing system 578

flushing protocol causes acute health impacts (i.e., eye burning and dizziness). The team 579

halts flushing, modifies the approach in the field, then resumes flushing homes. 580

University of South Alabama researchers team up with independent filmmaker and Ohio 581

State University student Krista Bryson to produce a YouTube.com video on how 582

residents can flush plumbing systems and limit their in-home chemical exposure. The 583

video receives more than 5,000 views. 584

12:50 pm – WVAW’s Do Not Use order ends for entire affected area. Plumbing system 585

flushing is authorized for all remaining areas. The MDL for 4-MCHM at the time is 10 586

µg/L. 587

University of South Alabama researchers attend the Governor’s press conference with a West 588

Virginia Clean Water Hub representative and report to the Governor’s staff that the 589

plumbing system flushing protocol is causing negative health impacts. These researchers 590

also approach media representatives to notify them of the observed in-home conditions. 591

The Offices of the West Virginia Governor, Bureau of Public Health, and National Guard 592

host a conference call to addressing community concerns. When asked about why so 593

many people are going to the hospital, the Bureau of Public Health responds that it 594

could be attributed to flu season. When asked why residents may experience nausea 595

when they flush their plumbing system, the Bureau of Public Health implies that it 596

could be due to sediment or copper exposure. 597 598

JANUARY 19 599 Governor Tomblin’s office meets with lobbyists to discuss a new West Virginia Source 600

Water Protection Act and commits to strengthening regulation. 601

602

JANUARY 20 603 WVAW cites the CDC’s health-based drinking water screening level and announces that 4-604

MCHM being present below 1,000 µg/L is an aesthetic issue, not a health-based issue. 605

Governor Tomblin states "I'm not going to say absolutely, 100 percent that everything is safe. 606

But what I can say is if you do not feel comfortable, don't use it." 607

12:00 pm – Governor Tomblin announces the West Virginia Source Water Protection Act. 608

609

610


JANUARY 21 611

9:50 am – Freedom Industries, Inc. publicly reveals for the first time that the leaking 612

tank number 396 also contained a second industrial product called “Stripped PPH” 613

not just “Crude MCHM.” This information was not previously disclosed to State or 614 water utility officials. According to statements by the DEP, Freedom Industries, Inc. 615

reportedly knew about this additional product in the spilled tank on the first day of the 616

incident. The safety data sheet for Stripped PPH indicated that this product contained a 617

mixture of propylene glycol phenyl ethers including polypropylene glycol phenyl ether 618

(PPH) and and dipolypropylene glycol phenyl ether (DiPPH). 619

The West Virginia DHHR, National Guard, and WVAW begins developing analytical 620

method for detecting and quantifying PPH and DiPPH in water. 621

The West Virginia DHHR requests assistance from CDC regarding the drinking water 622

contamination incident. The CDC develops a short-term screening level for PPH and 623

DiPPH at 1,200 and 1,200 µg/L, respectively. This calculation was reviewed and 624

approved by the federal interagency workgroup similar to the CDC’s 4-MCHM level. 625

The West Virginia DHHR and National Guard begins testing archived water treatment 626

facility water samples for PPH and DiPPH chemicals. Only two water samples of 627

hundreds showed presence of PPH and DiPPH at 10 and 11 µg/L, levels far below CDC 628

screening levels. 629

630

JANUARY 22 631 The DEP announces a revised estimate of 6,251 gallons of solvent leaked from tank number 632

396 tank, not 5,000 gallons as Freedom Industries, Inc. previously indicated. 633

634

JANUARY 24 635 Freedom Industries, Inc. is ordered by Governor Tomblin to physically remove all 17 above 636

ground storage tanks at their Elk River site and this action must begin on or before March 637

15. 638

639

JANUARY 27 640 5:00 pm – Kanawha County bottled water distribution ends and is attributed to Federal 641

resources being exhausted. Approximately 11,484,000 bottles of water had been 642

distributed in addition to bulk water distribution. The estimate does not include 643

distributions by private entities such as businesses and nonprofit organizations. 644

Freedom Industries, Inc. updates the estimated quantity of solvent that leaked from the tank 645

to 10,142 gallons. 646

647

JANUARY 28 648 U.S. Federal Bureau of Investigation (FBI) Hazardous Materials Response Unit and U.S. 649

Attorney visit Freedom Industries, Inc. spill site to inspect the tanks. The FBI is under the 650

U.S. Department of Justice. 651

652

JANUARY 29 653 Marshall University professor Scott Simonton testifies to a joint House Transportation and 654

Infrastructure Committee of State lawmakers that through his tap water testing he found 655

4-MCHM as high as 191 to 1,900 µg/L in affected buildings between January 13 and 18. 656


He also testifies that a low level of formaldehyde was in the contaminated tap water on 657

Jan. 13. The professor postulates that 4-MCHM can degrade into formaldehyde and states 658

"any exposure, no matter how slight, is going to increase cancer risks." The West 659

Virginia Bureau of Public Health and WVAW states that the formaldehyde assertion is 660

unfounded. 661

662

JANUARY 30 663 Governor Tomblin requests that WVAW continue making bottled water available. WVAW 664

already had purchased bottled water supplies and bulk water distribution equipment in 665

advance of the request. 666

Governor Tomblin reports that 17.5 million bottles of water have been distributed to date. 667

The US National Science Foundation awards five RAPID response research grants in 668 support of scientific need in West Virginia. The grants go to Dr. Andrew Whelton at 669

University of South Alabama; Drs. Jennifer Weidhass and Lian-Shin Lin at West 670

Virginia University; Drs. Andrea Dietrich, Daniel Gallagher, Paolo Scardina, Virginia 671

Tech; Dr. Beverly Moss with student Krista Bryson, The Ohio State University; Dr. 672

William Alexander and Nathan DeYonker, University of Memphis. 673

674

JANUARY 31 675 Contractors at the Freedom Industries, Inc. site cause an underground pipe to leak solvent 676

onto the ground. A trench onsite catches the newly-spilled solvent, preventing it from 677

entering the Elk River. A licorice odor was detected in the surrounding air. 678

679

FEBRUARY 1 680 Licorice odor complaints are reported at five schools located in Kanawha, Putnam, and 681

Lincoln counties. Follow up tap water testing reveals 4-MCHM present in tap water at 682

levels well below the CDC’s health based screening level for 4-MCHM. 683

684

FEBRUARY 3 685 WVAW states that no 4-MCHM is coming off their 16 filters. Filters contained sand, gravel, 686

anthracite, and a top layer of activated carbon. The 4-MCHM MDL ranges from 2 to 6 687

µg/L depending on the laboratory being used. 688

689

FEBRUARY 4 690 A Federal grand jury begins a criminal investigation into Freedom Industries, Inc. and 691

WVAW. The office of U.S. Attorney Goodwin issues subpoenas. 692

West Virginia legislator Natalie Tennant announces some residents are melting snow to give 693

their children baths because of tap water safety concerns. 694

Eric Olson, Senior Strategic Director of the National Resources Defense Council (NRDC), 695

testifies to the U.S. Senate in Washington, D.C. and reiterates concerns about the lack of 696

in-home testing and the plastic pipe permeation potential that may have taken place 697

inside plumbing systems. 698

699

700


FEBRUARY 5 701

An elementary school and high school, both in Kanawha County, are dismissed because 702

of licorice odor complaints. There is a media report of one teacher fainting and 703

students self-reporting dizziness and eye burning symptoms. 704

The CDC visits Charleston, West Virginia. At a press conference, the acting director of 705

CDC’s National Center for Environmental Health and ATSDR states “we’re not 706

really talking about whether water is safe, we’re talking about is the water 707

‘appropriate for use’ given the information we know about MCHM.” CDC says 708

pregnant women can use the water. 709 710

FEBRUARY 6 711 Fourteen Kanawha County schools report licorice odor complaints. Building tap water was 712

previously classified as “non-detect” before being allowed to reopen. 713

714

FEBRUARY 7 715 The Office of West Virginia Governor Tomblin calls on Professor Andrew Whelton’s team 716

to provide scientific assistance in the ongoing response. Whelton invites Jeffrey Rosen, 717

President of Corona Environmental Consulting to fly-in and also meet with the 718

Governor’s staff. 719

720

FEBRUARY 8 721 Professor Whelton and Mr. Rosen meet with a West Virginia multi-agency task force in 722

Charleston, West Virginia to prioritize the major unanswered scientific questions. 723

Whelton and Rosen engage science and engineering experts Dr. Michael J. McGuire of 724

Michael J. McGuire, Inc., Dr. Andy Eaton of Eurofins Eaton Analytical Laboratory, Dr. 725

Jennifer Clancy and Tim Clancy of Corona Environmental Consulting, Dr. Craig Adams 726

of Utah State University, Dr. Michael Dourson and Jacqueline Patterson of Toxicological 727

Excellence in Risk Assessment (TERA) to expand the depth and breadth of the 728

investigative team. 729

730

FEBRUARY 10 731 The West Virginia Poison Center releases a factsheet for the Elk River spill describing 732

potential health effects, the CDC screening levels, and a definition of “safe” drinking 733

water. 734

The West Virginia Bureau of Public Health testifies at joint West Virginia House 735

Transportation and Infrastructure Committee hearing of State lawmakers. The 736

Bureau of Public Health explained "Everybody has a different definition of safe. I 737

believe the water, based on the standards we have, is usable." Also mentioned is that 738

the 4-MCHM has a low “partition coefficient” and therefore is unlikely to “stick” to 739

pipes. 740 741

FEBRUARY 11 742

Governor Tomblin announces emergency funding of the West Virginia Testing 743

Assessment Project (WVTAP), an independent team of science and engineering 744 experts to provide assistance. The team announces a four part effort to: (1) Convene an 745

international panel of experts to examine the West Virginia safety factor applied to their 746


10 µg/L MCHM drinking water screening level; (2) Determine the drinking water odor 747

threshold for MCHM, (3) Conduct a focused residential drinking water sampling field 748

study to be used to support the design of a larger more comprehensive program for the 749

nine counties affected. 750

751

FEBRUARY 12 752 Complaints of an unusual “carrot-like” odor in the bottled water are reported to KCHD. A 753

KCHD investigation determines that the suspect bottled water was expired and originated 754

from FEMA. 755

756

FEBRUARY 21 757 The CDC refers to the drinking water as “safe”. 758

759

FEBRUARY 28 760 5:30 pm – Governor Tomblin ends the state of emergency. 761

Kanawha County schools switch to tap water for drinking and cooking. 762

763

MARCH 7 764 Freedom Industries, Inc. submits their tank decommissioning plan to DEP. The report is 765

authored by Civil Engineering Consultants, Inc. 766

767

MARCH 12 768 1:00 pm – Professor Whelton and graduate student LaKia McMillan present early findings of 769

plumbing system flushing effectiveness and resident perceptions at a University of New 770

Orleans seminar. Results show flushing was not effective at reducing chemical levels at 771

some of the homes they visited. Also found was that chemical levels before flushing 772

varied based on tap. Whelton and McMillan question whether or not chemicals diffused 773

into plastic pipes during the low water use period and leached out into drinking water 774

after flushing. 775

Residents report licorice odor complaints near a local landfill to the media. One resident 776

formally complains to DEP. DEP sends an inspector to the site and determines the odors 777

were not at objectionable levels. 778

9:00 pm – Local Mayor of Hurricane, West Virginia issues statement that the nearby 779

landfill has been accepting tanker trucks of wastewater from the incident that 780

affected WVAW. His investigation is in response to complaints from citizens about 781

licorice odors. Approximately 36,000 gallons is planned for disposal. Reports are 782

that wastewater is mixed with sawdust before disposal. The DEP approved the 783

permit modification. 4,000 gallons was disposed of on March 7, 12,000 gallons on 784

March 10 and 11, and 8,000 gallons on March 12. 785

786

MARCH 16 787 Freedom Industries, Inc. is supposed to have begun tank removal, but discovery of additional 788

materials that unexpectedly contained asbestos delays tank removal. 789

790

791


MARCH 17 792

WVTAP issues a report that summarizes the literature regarding health effect studies 793

of 4-MCHM, PPH, and DiPPH. 794

WVTAP issues a report that describes results from an expert drinking water odor 795 panel. The WVTAP report states that the panel examined drinking water that contained 796

the Crude MCHM/Stripped PPH mixture and the odor threshold concentration was less 797

than 0.18 µg/L concentration. Testing was conducted on the mixture and not the pure 4-798

MCHM analytical standard because the University of South Alabama and WVTAP 799

research teams detected a sharp licorice odor at in-home taps and less irritating “dull” 800

licorice odor in laboratory prepared drinking water samples that contained pure 4-MCHM 801

analytical standard. 802

803

MARCH 18 and 19 804 After two water main breaks in a nearby utility distribution system, WVAW customers in the 805

Town of Ona, West Virginia complain of licorice odors in their tap water. 806

WVAW investigates and finds 4-MCHM concentrations at 3 µg/L in Ona’s drinking water. 807

The MDL at the time was 2 µg/L. 808

809

MARCH 21 and 22 810 Persistent licorice odors remain in residents’ homes 811

4:00 pm – WVTAP water sampling results show that 4-MCHM is desorbing from 812

activated carbon located within WVAW treatment plant. 4-MCHM was detected in 813

customer homes, but not in Elk River water entering the WVAW facility. Levels 814

found by WVTAP were near the MDL of 0.38 µg/L. WVAM and the National 815

Guard helped enable this water sampling. 816 6:00 pm – WVAW begins their own water sampling in the Elk River and inside their 817

treatment facility for 4-MCHM using the lowest MDL for 4-MCHM of 0.38 µg/L. 818

819

MARCH 25 820 12:18 pm – WVAW announces their follow-up treatment facility water sampling results 821

show low levels of 4-MCHM coming off filters into the treated water ranging from 822

“nondetect” to 0.60 µg/L. WVAW reiterates that the company plans to begin changing 823

out 500 tons of granular activated carbon in the plant’s 16 filters beginning the following 824

week. 825

826

MARCH 26 827 Professor Andrea Dietrich and her Virginia Tech research team announce that the odor 828

threshold concentration of the pure 4-MCHM analytical standard in drinking water is 7 829

µg/L. 830

831

MARCH 27 832 WVAW shutsdown their raw water intake for two hours due to a report of white colored 833

“foam” on Elk River. Water quality testing reveals no changes to source water quality 834

and no characteristics outside of typical water quality parameters. 835

836

837


MARCH 28 838 9:30 am – WVTAP holds public meeting at West Virginia State University to review project 839

status include (1) the Crude MCHM Odor Threshold Study, (2) The Ten Home Study: 840

Resident Interviews and Tap Water Analysis, (3) Examination of Tentatively Identified 841

Compounds and their Implications, (4) Expanded Home Tap Water Monitoring Plan: 842

Preliminary Results, and (5) Expert Health Effects Panel. 843

844

MARCH 29 845 As part of WVTAP, toxicology and drinking water experts from the U.S., Israel, and The 846

United Kingdom fly into Charleston, West Virginia to begin deliberations regarding 847

screening levels for tap water use. 848

849

MARCH 31 850

WVTAP issues a report describing the results of a consumer drinking water odor panel. 851 WVTAP reports that the odor threshold concentration for the spilled solvent in drinking 852

water is less than 0.55 µg/L based on a consumer sensory panel. The level is much lower 853

than previous 4-MCHM MDLs used to analyze water samples by the State and WVAW. 854

855

APRIL 1 856

9:30 am – WVTAP holds a public meeting at West Virginia State University and 857

releases preliminary results of their Expert Health Effects Panel. The WVTAP 858

Expert Health Effects Panel recommends 28 day screening levels for 4-MCHM, 859

PPH, and DiPPH. The recommended values are lower than levels recommended by 860

the CDC. The WVTAP panel assumptions differed than those of the CDC. The 861

panel considered inhalation and dermal exposure to contaminated water in their 862

assessment, selected a bottle-fed infant as their most susceptible population, and 863

considered a 28-day exposure period. 864 WVAW begins replacing the activated carbon in their filters. 865

866

APRIL 3-8 867 The KCHD conducts a randomized telephone study to assess the physical, psychosocial, 868

communication and economic impact of the spill on households. 869

870

APRIL 8 871

The engineering contractor Civil and Environmental Consultants, Inc. (CEC) hired by 872 Freedom Industries, Inc. provides the company a site remediation plan report. This 873

report contains surface and groundwater sampling results for water samples collected 874

onsite January 31 and February 11. 4-MCHM levels were found as high as 76,000 to 875

190,000 µg/L in surface water captured onsite. Groundwater levels were found as high as 876

14 µg/L for 4-MCHM and 22 µg/L for PPH. No testing for DiPPH was carried-out. A 877

sheen was also observed at the banks of the Elk River after a heavy rain. The consultants 878

recommend Freedom Industries, Inc. consider either (1) Natural attenuation, collection, 879

or treatment or (2) Pump and treat as remediation approaches. 880

881

882


APRIL 11 883 The DEP reports that Freedom Industries, Inc. is trucking wastewater and contaminated 884

rainwater to Ohio and North Carolina for deep well-injection and landfill disposal. 885

886

APRIL 22 887

During a NACCHO webinar, Dr. Rahul Gupta of KCHD estimates that 90,000 to 888

108,000 residents experienced a negative health impact due to contaminated water 889

exposure. 890 891

APRIL 24 892 The US Chemical Safety and Hazard Investigation Board and contractors remove portions of 893

the Freedom Industries, Inc. tank and conduct inspections. 894

895

MAY 1 896 It is publicized that, at the request of DEP, EPA is working to come up with a more 897

comprehensive way of monitoring air quality for MCHM. The method is considered for 898

use during the dismantling of Freedom’s chemical storage tanks and the cleanup of the 899

Elk River site where the leak occurred. 900

901

MAY 5 902

WVTAP issues a report on chemical water quality and resident survey results for their 903 Ten Home Study. Tap water testing was conducted in mid-February, weeks following 904

the in-home plumbing system flushing. WVTAP did not find any PPH or DiPPH in the 905

10 homes, but found 4-MCHM in all 10 homes located in eight of nine counties. 4-906

MCHM levels were as high as 6.1 µg/L where an MDL applied was 3.8 µg/L. No PPH or 907

DiPPH was detected in these homes using an MDL of 0.38 µg/L. 908

909

MAY 6 910 10:30 am – WVAW announces that the activated carbon filter change-out process is halfway 911

complete. WVAW had been using a 4-MCHM of 2 µg/L since the change-out began to 912

determine if 4-MCHM was present in the filter effluent. 913

914

MAY 12 915

Results of the KCHD randomized telephone study of 499 participants are released. 916 Seven weeks after the Do Not Use order was issued roughly 80% of the respondents were 917

using the tap water to bathe but less than 12% were using it for cooking or drinking. 918

More than 20% of respondents used the WVAW tap water during the Do Not Use order. 919

Those persons used tap water for bathing/ showering, hand washing, clothes washing, 920

dish washing, teeth brushing, drinking, cooking, gave it to pets, and watering plants. 921

Approximately a third of households surveyed felt that a household member had an 922

illness related to the spill. 923

924

MAY 16 925

WVTAP issues two reports, a document describing its Expert Health Effects Panel 926

conclusions and justification and a document describing chemical oxidation 927 experiments conducted on the spilled solvent. WVTAP examined solvent integrity 928


when dilute solutions were exposed to potassium permanganate and free available 929

chlorine, two disinfectants used by WVAW. The report is referred to as the “oxidation 930

studies” report. No clear effect was found, implying that no breakdown products were 931

generated. 932

933

MAY 21 934

WVTAP releases a report describing numerous tentatively identified compounds found 935 while analyzing tap water collected from the ten homes sampled. Results show that 936

some of the contaminants detected in the tap water were artifacts produced during sample 937

extraction and analysis, not breakdown products from the spilled solvent. 938

MAY 26 939 During the week, the Charleston Sanitary Board wastewater treatment facility permitted the 940

discharge of 4-MCHM contaminated stormwater into its facility. Stormwater originated 941

from the Freedom Industries, Inc. site. No 4-MCHM was found in the wastewater facility 942

effluent above an MDL of 2.7 µg/L . 943

944

MAY 30 945

WVTAP releases a residential in-home tap water sampling plan report created using 946 data from the all completed WVTAP efforts. The sampling plan was designed to 947

answer three questions: (1) What is the concentration of 4-MCHM in people’s 948

residences? (2) Is the average concentration observed in homes below a level of concern? 949

(3) What proportion of the homes has 4-MCHM concentrations below a level of concern? 950

WVTAP concludes only 30 homes in each of 24 different regions of the water 951

distribution system would need to be sampled to answer these questions and this result 952

would be statistically representative. 953

954

JUNE 5 955 West Virginia Senate Bill 373, The Water Resources Protection Act takes effect. 956

957

JUNE 9 958 Professor Whelton presents new chemical sorption data at the American Water Works 959

Association (AWWA) Conference in Boston, MA. Results show that two components of 960

the spilled solvent, 4-MCHM and cyclohexanemethanol (CHM), had some, but limited 961

sorption into crosslinked polyethylene (PEX) plumbing pipes. PEX pipes were observed 962

in West Virginia plumbing systems. 963

964

JUNE 12 965

WVAW completes activated carbon replacement for all of their filters. No 4-MCHM is 966

found in filter effluent at an MDL of 0.38 µg/L 967

4:30 pm – A DEP inspector observes contaminated water running from a trench located 968 at the Freedom Industries, Inc. spill site into the Elk River. The overflow is suspected 969

to be caused by heavy rain coupled with the failed operation of a sump pump located in a 970

trench meant to capture contaminated stormwater. WVAW shutsdown their raw water 971

intake in response to this discovery and conducts testing at the drinking water intake for 972

4-MCHM. No 4-MCHM was found above MDLs used by three different laboratories (10, 973

2, and 0.38 µg/L). 974


JUNE 13 975

5:00 pm – A DEP inspector observes contaminated stomwater again running into the 976 Elk River from the Freedom Industries, Inc. spill site. WVAW shutsdown their raw 977

water intake again as a precaution. 978

The DEP issues two notice of violations (NOV) to Freedom Industries, Inc. for the June 12 979

incident. 980

981

JUNE 14 982 8:30 am – WVAW releases 4-MCHM water testing results for the drinking water intake and 983

treated water conducted in response to June 13 Freedom Industries site spill. No 4-984

MCHM was found above MDLs used by three different laboratories (10, 2, and 0.38 985

µg/L). 986

987

JUNE 17 988

11:00 am – WVAW releases results of water distribution system testing for 4-MCHM. 989

No 4-MCHM was found in any sample at 0.38 ppb MDL. Locations included select 990

schools, city halls, hydrants, and blowoffs. 991 992

JUNE 26 993

WVTAP releases the final report summarizing prior efforts and recommends future 994

research. 995

996

JUNE 27 997 Freedom Industries, Inc. settles for $3,000,000 with its insurance company AIG Specialty for 998

the Elk River spill. 999

The DHHR releases a survey of area doctors who saw patients outside of hospital 1000 emergency departments between Jan. 9 and May 31. The most common symptoms 1001

associated with tap water exposure were itching, irritation, rashes, eye irritation or pain, 1002

vomiting and cough. 1003

1004

JULY 8 1005

The CDC releases results from their Community Assessment for Public Health 1006

Emergency Response (CASPER) in-home survey of households affected by the 1007 contaminated water incident. The survey was carried-out in April and revealed 1008

approximately 1 of 5 households had individuals who experienced symptoms consistent 1009

with those reported in the results in the emergency department record review, non-1010

emergency health providers’ reports, and the West Virginia Poison Center. 1011

1012

JULY 10 1013 Professor Whelton presents new preliminary data at NACCHO annual conference indicating 1014

Crude MCHM provided to his team by Eastman Chemical Company in February 2014 1015

was more toxic to an aquatic organism than reported by Eastman Chemical Company in 1016

1998. Daphnia magna was the bioindicator organism. Whelton and graduate student 1017

Caroline Novy found a 48 hour EC50 of Crude MCHM of approximately 50 mg/L while 1018

the 1998 value was 98.1 mg/L. The NOEC found by the University team was 6.25 mg/L 1019

while the 1998 value was 50 mg/L. 1020


JULY 11 1021 The DEP states air monitoring will not be conducted when Freedom Industries, Inc. tanks are 1022

dismantled because the State does not have the technology to do so. The DEP also states 1023

that there is no air screening level for inhalation exposures. 1024

1025

JULY 16 1026

The US Chemical Safety and Hazard Investigation Board present findings of their 1027

investigation at a public meeting in Charleston, WV. The team reports multiple 1028

holes and pitting in multiple tanks on the Freedom Industries, Inc. site including 1029

tank number 396. The team also indicates they believe industrial solvent was leaking 1030

from multiple tanks before the catastrophic release on January 9. The CSB also 1031

comments that the public continues to distrust that the water is safe to drink. 1032 1033

JULY 23 1034 The National Toxicology Program pledges to conduct toxicity studies. Researchers will 1035

examine fish, worms, and rats and apply computer modeling. 1036

1037

AUGUST 25 1038 The West Virginia Division of Natural Resources begins collecting samples of fish from the 1039

Elk River and the Kanawha River. The samples will be analyzed by the U.S. Geological 1040

Survey (USGS). 1041

1042

OCTOBER 2 1043

The US EPA announces they have developed a 30-day, 4-MCHM air screening level of 1044

0.010 ppmv and 4-MCHM air sampling plan for the Freedom Industries, Inc. site. 1045 Air sampling equipment is deployed in October 17 at the Freedom Industries, Inc. site. 1046

1047

NOVEMBER 12 1048 The USGS and West Virginia University researchers present new river and tap water 1049

monitoring data at the Society of Environmental Toxicology and Chemistry (SETAC) 1050

conference in Vancouver, Canada. Dr. William Foreman indicates USGS developed a 1051

new analytical purge and trap GC method for 4-MCHM detection in river and tap water 1052

samples. Elk River water samples collected < 3.2 km downstream by the USGS six days 1053

after the spill had ≤ 2.9 µg/L 4-MCHM. Eight days after the spill 4.8 µg/L 4-MCHM was 1054

found in the Ohio River at Louisville, Kentucky, (est. 590 km downstream). 4-MCHM 1055

found at the USGS office drinking water taps in Charleston, West Virginia was 129 µg/L 1056

18 days after the spill and 2.2 µg/L 25 days after the spill. 4-MCHM remained detectable 1057

(MDL=0.40 µg/L) at the Charleston office through Feb. 25, 47 days after the spill. The 1058

USGS estimated that the trans- and cis- isomer concentrations in the spilled liquid were 1059

491 g/L and 277 g/L, respectively. These constituted 84% of the source material. Methyl 1060

4-methylcyclohexanecarboxylate was also tentatively identified in diluted spilled liquid 1061

and this compound one of the Crude MCHM ingredients listed on the product’s safety 1062

datasheet. 1063