UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effectiveness of pH and alkalinity adjustments in reducing lead and copper levels in rechlorinated.. Chan, Kenneth C. H. 1994-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


ubc_1994-0397.pdf [ 2.27MB ]
JSON: 1.0050406.json
JSON-LD: 1.0050406+ld.json
RDF/XML (Pretty): 1.0050406.xml
RDF/JSON: 1.0050406+rdf.json
Turtle: 1.0050406+rdf-turtle.txt
N-Triples: 1.0050406+rdf-ntriples.txt
Original Record: 1.0050406 +original-record.json
Full Text

Full Text

THE EFFECTIVENESS OF pH AND ALKALINITY ADJUSThIENTS INREDUCING LEAD AND COPPER LEVELS IN RECHLORINATED AM)CHLORAMINATED TAP WATERbyKenneth C.H. ChanB.A.Sc.(Honours), U.B.C., 1992A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Civil EngineeringWe accept this thesis as conformingto the equired standardTHE UNIVERSITY OF BRITISH COLUMBIASEPTEMBER1994© Kenneth C.H. Chan, 1994In presenting this thesis inpartial fulfillment of therequirements for an advanced degreeat the University of BritishColumbia, I agree that the Library shall makeit freely availablefor reference and study. I furtheragree that permission forextensive copying of this thesisfor scholarly purposes may begranted by the head of my departmentor by his or herrepresentatives. It is understoodthat copying or publication ofthis thesis for financial gainshall not be allowed without mywritten permission.Department ofThe University of British ColumbiaVancouver, CanadaDate_____________(Signature)ABSTRACTThe effectiveness of pH and alkalinity adjustments in reducing copper and leadlevels in rechiorinated and chioraminated tap water was assessed for a study areainGreater Vancouver. Standing cold water, running hot and cold sampleswere collectedfrom 105 houses that were located in the study areas. The samples wereanalyzed inthe laboratory for lead and copper concentrations, pH, and alkalinity.After following an complex scheme of data manipulation, sorting, and statisticaltesting, comparisons of the copper and lead levels between the study areas weremade.This research study found that pH and alkalinity adjustments were definitelyeffective in reducing the copper levels in rechlorinated and chioraminated tapwater.Lead levels were also reduced, but the magnitude of the reduction was statisticallyinsignificant. The effects of pH and alkalinity adjustments on copper and lead levelsdid not appear to be different for rechlorinated and chloraminated tap water.Compared to houses with copper plumbing, houses with plastic plumbing and no-leadsolder had extremely low levels of copper or lead in the tap water. This study alsofound that the age of a house, which was fitted with copper plumbing and was less than15 years old, was not a significant factor in affecting lead concentrations.11TABLE OF CONTENTSABSTRACT iiLIST OF TABLES vfflLIST OF FEGURES ixACKNOWLEDGMENTS x1. INTRODUCTION 11.1 Greater Vancouver’s Drinking Water 11.2 GVRD Initiatives 11.3 Study Area & Preliminary Results 21.4 Objective and Scope of Study 32. BACKGROUND AND LITERATURE SEARCH 42.1 Health Impacts of Lead 42.2 Health Impacts of Copper 42.3 Copper and Lead Regulations 52.4 Principles of Metallic Corrosion 52.4.1 Uniform Corrosion 62.4.2 Galvanic Corrosion 62.4.3 Crevice Corrosion 62.4.4 Pitting Corrosion 72.4.5 Concentration Cell Corrosion 72.4.6 Selective Leaching 72.4.7 Erosion Corrosion 72.4.8 Stress Corrosion 82.4.9 Microbiologically Induced Corrosion 82.4.10 Chemical Corrosion Mitigation 81112.5 Previous Studies in Copper and Lead Levels 92.5.1 Seattle, WA 92.5.2 Boston, MA 102.5.3 Glasgow, Scotland 102.5.4 Portland, ME 102.5.5 Covewood Lodge, NY112.6 Tap Water Sampling vs. Laboratory Simulations 112.7 Sources of Copper in Tap Water122.8 Sources of Lead in Tap Water122.9 Factors Affecting Tap Water Sampling Results 132.10 Factors Affecting Lead Leaching 162.11 Factors Affecting Copper Leaching192.12 Plastic Pipes vs. Copper Pipes 192.13 Quality Control in Laboratory Tests 202.14 Reporting of Low Level Data 212.15 Considerations in Experimental Design 212.16 Sampling 232.17 Common Assumptions Made in Statistics 252.18 Measuring Central Tendency262.19 Measuring Spread 272.20 Estimate 282.21 Coefficient of Variation282.22 Skewness 282.23 Central Limit Theorem 282.24 Hypothesis Testing 292.25 Constructing Power Curves 302.26 Statistics of the Maximum 30iv2.27 Systems of Measurement312.28 Nonparametric Testing312.29 Kolmogoroff-Smirnoff Comparisonof TwoIndependent Samples322.30 Median Test322.31 Kruskal-Wallis H-Test322.32 Confidence Intervals for the DifferenceBetween 2 Means342.33 Test for Equal Variance342.34 Spearman Rank Correlations352.35 Regression362.36 One-way ANOVA362.37 Theory of Causation383. EXPERIMENTAL METHODS403.1 Experimental Design403.1.1 Controllable Factors413.1.1.1 Direct Control423.1.1.2 Indirect Control433.1.2 Uncontrollable Factors443.2 Arrangements with Governmental Agencies463.3 Bottle Preparation473.4 Sampling Package493.5 Sampling Procedures493.6 Participation of Homeowners503.7 Bottle Pick-up513.8 Lab Testing513.8.1 Instruments52V3.8.2 Testing Scheme533.8.3 Data Recording534. RESULTS AN]) DISCUSSION554.1 Summary of Returned Bottles554.2 Data Cross-check by GVRD574.3 Reporting of Data584.4 Data Used for Statistical Analysis584.5 Non-Parametric Testing 604.6 Other Statistical Problems614.7 Identifying Flushed Standing Samples 624.7.1 Rejected Samples for n,y,s,w 654.7.2 Rejected Samples for d 654.8 Simple Tests664.9 Kruskal-Wallis H-Test (Copper) 674.10 Confidence Intervals for the DifferenceBetween 2 Means 684.11 Median Test684.12 Test for Equal Variance684.13 Compliance with Regulations694.14 Plot of Lead vs. Alk and pH704.14.1 Spearman Rank Correlations714.14.2 Regression of Alk and pH714.14.3 Filling in Missing Data Points734.15 Copper vs. pH744.l6Leadvs.pH754.17 ANOVA (Lead)764.17.1 Using One-Way ANOVA77vi4.17.2 One-Way ANOVA grouped by area 784.17.3 One-Way ANOVA grouped by pH rangewith age as covariate 784.17.4 Residuals 794.18 Testing of Oldcopper and Plastic 824.19 Comparison of Flushed Cold and Flushed Hot 835. SIJIVIIvIARY AND CONCLUSIONS865.1 The Effects of pHIAlk Adjustment 865.2 Significance of Other Factors 865.3 Recommendations876. REFERENCES 88APPENDICESAppendix A Sampling Locations 92Appendix B Written Sampling Instructions 93Appendix C Cartoon Sampling Instructions 94Appendix D Questionnaire 95Appendix E GVRD Data Crosscheck 96Appendix F Raw Data 97viiLIST OF TABLES1 Preliminary Copper and Lead Study Results (From1988 to 1991)22 Typical Data for a Single-Factor Experiment373 The ANOVA Table for theSingle-Factor, Fixed Effects Model 384 Results of Bottle Treatment Study485 Summary Information ofReturned Bottles 556 Detection Limits of Measurements587 Summary of Simple Statistical Evaluation 668 Percentage Reduction from Delta (mean values)679 Difference Between 2 Means6810 Test for Equal Variance 6911 Regulation Compliance (as per EPA guidelines, 1991) 6912 Regression Analysis of pH versus Alkalinity 7213 Comparison of Oldcopper and Plastic with Main Data 8314 Summary of All Statistical Testings 86viiiLIST OF FIGURES1 Comparison of Data Types 602 Rejection Procedure 643 3D Plot of Lead, Alkalinity, and pH 704 Regression Plot of pH versus Alkalinity 735 Filling-in Missing Data for pH versus Alkalinity Data 746 Copper vs. pH 757 Leadvs.pH 768 Normal Probability Plot of ANOVA of Lead 809 Student t vs. Lead Estimate 8110 Cook’s Distance vs. Lead Estimate 8211 Quantile-Quantile Plot of Cold & Hot Running Copper 8412 Quantile-Quantile Plot of Cold & Hot Running Lead 85ixACKNOWLEDGMENTSThis research project was funded by a Natural Sciences and Engineering ResearchCouncil (NSERC) scholarship, by the UBC Civil Engineering Department, and by theGreater Vancouver Regional District (GVRD). I would like to thank Dr. DonMavinic, and Prof. Jim Atwater of UBC, Mr. Doug Neden, Mr. Mark Ferguson, Ms.Judy Smith, and Mr. Indergit Singh of the GVRD for their help in the planning andanalysis of this project. I want to thank the staff members at the City of Surrey and theCorporation of Delta for their assistance. My sincere appreciation to all thehomeowners in North Delta, Newton, and South Surrey, who participated in this study,for taking the time to collect the water samples. I would also like to thank Ms. SusanHarper, Ms. Paula Parkinson, and Ms. Jufang Zhou of UBC, and Mr. Peter Zadoroznyof the GVRD for their expert help in laboratory sample analysis. Soli Deo Gloria.x1. INTRODUCTION1.1 Greater Vancouver’s Drinking WaterThe water supply of the GVRD comes fromthree watersheds on the north shoreof the Burrard Inlet. These watersheds, covering a totalof 585 squared km of land, arethe Capilano, Seymour, and Coquitlam. Each watershedhas its own major storagereservoir. The Cleveland Dam is constructed on the Capilano River, theSeymour FallsDam on the Seymour River, and the Coquitlam Dam on theCoquitlam River.The drinking water supply of the study areas for thisresearch project mostlycomes from the Seymour water source. The water from theSeymour reservoir isuntreated except for the addition of chlorine. In 1992, thechlorinated source waterwas characterized by having an average pH of 6.0, dissolvedoxygen content of 10.2ppm, total chlorine residuals of 0.9 ppm as chlorine, totaldissolved solids of 16 ppmas CaCO3,and total alkalinity of 2.3 ppm as CaCO3 (GVWD,1992). The naturalcopper and lead levels of the water in the reservoir was negligible.1.2 GVRD InitiativesA water with these characteristics is considered to have ahigh corrosionpotential on metals. Since 1984, the GVRD has studied thebacteriologicaldisinfection, and disinfection by-products of the GreaterVancouver’s drinking water(GVWD, 1992). Studies have also been done onthe corrosion aspects of the waterquality improvement program of the GVRD. Corrosion canlead to green staining ofporcelain fixtures and basins in bathrooms. It can alsoincrease the copper and leadlevels of the tap water. The economic impact ofwater corrosion is substantial.Recognizing that the present water supply system isinadequate to meet futuredemands in Greater Vancouver, GVRD is planning a comprehensiveprogram that willincrease storage and transmission capabilities. Atthe same time, the GVRD’ s DrinkingWater Quality Improvement Plan (DWQIP) is seeking toimprove the quality of the1drinking water in order to consistently meet the Guidelines for Canadian DrinkingWater Quality.1.3 Study Area & Preliminary ResultsIn 1988, a monitoring study was initiated in South Surrey, Newton, and Deltato study the effects of secondary disinfectants on the water supply. Newton receivedsecondary chlorine treatment, while South Surrey received chioramination. Theeffectiveness of the secondary treatments on controlling bacterial growth werecompared to the results from the control area in Delta. Besides the disinfectioninitiatives, the study areas were also used to study the dynamics of corrosion control.The pH and alkalinity of the Newton and South Surrey areas were raised in order tomake the water less corrosive. From 1988 onwards, the pH and alkalinity adjustmentshave been in continuous operation. It was hoped that collecting tap water samples fromvarious locations in the three study areas would reveal whether or not pH and alkalinityadjustments were effective in reducing the lead levels of the drinking water.About 15 tap water samples were collected in 1988, just prior to theintroduction of pH and alkalinity adjustments. Three years later, in 1991, about thesame number of samples were again collected. Of the samples, 6 (2 in Newton, 2 inSouth Surrey, and 2 in Delta) were collected from the same houses both in 1988 and in1991. The data almost uniformly supported the conclusion that copper levels in the 1liter standing samples were lowered after the introduction of pHand alkalinityadjustments. However, the lead data yielded inconclusive results(Table 1).Table 1 Preliminary Copper and Lead Study Results (From 1988 to 1991)Sample # - % change in copper - % change in Lead1 -78 -322 -85 +1773 -89 -384 -95 05 +16+126 -9 +612For half of the 6 houses sampled, the lead level increased in 1991 from 1988, whereasit decreased for the other half of the houses.1.4 Objective and Scope of StudyThe 6 data points that were collected in 1991 and 1988did not give enough datato do any meaningful statistical analysis on the lead data.The main objective of thisstudy, then, was to collect enough samples from the study areas sothat meaningfulstatistical statements could be made on whether or not the pH andalkalinity adjustmentprogram was effective in reducing lead levels in the tap water. Secondly,the copperlevels of the collected samples were also analyzed to confirm the conclusionof theprevious study that the copper level is reduced after the pH andalkalinity of the wateris raised.32. BACKGROUND AND LITERATURE SEARCH2.1 Health Impacts of LeadLead is not a required trace metal for bodily functions. When taken up,itaccumulates in the body. Lead is potentially poisonous to human beings, especiallywhen there is continuous exposure to small amounts of it. Therefore, trace levelsoflead in drinking water could become a problem to humans.Lead has the most harmful effects on infants and young children, followed byadult females, and has the least effect, comparatively, on adult males (NationalAcademy of Sciences, 1982). Upon entering the human body, some of thelead isabsorbed in the blood stream and is excreted from the bodythrough the kidneys and theintestinal tracts. Lead can also be deposited in soft tissues, hair and nails.However,most of the lead that stays in the body is found in the skeleton (Departmentof theEnvironment, 1977). Lead has a half life of 2 to 4 weeks in blood,about 4 weeks insoft tissues, and 27.5 years in bone (World Health Organization, 1984).Lead canenter the body through water, food, or air. It has been shown that the intakeand theuptake of lead in drinking water can be a significant percentage of the totalintake anduptake (Drill, 1979). When the amount of lead that has accumulatedin the body hasreached acute levels, the victim could suffer tiredness, abdominaldiscomfort,irritability, and anaemia (World Health Organization, 1984).2.2 Health Impacts of CopperIn contrast to lead, copper is an essential tracemetal that is required by thehuman for normal physiological functions. Long termexposure to copper at low levelsdoes not have documented toxic effects on humans.But, the intake of very largedosages of copper could lead to severe mucosal irritationand breakdown, capillarydamage, liver and renal damage, and central nervoussystem irritation, followed bydepression (World Health Organization, 1984). When the bodydetects that it has anexcessive amount of copper in its system, the person isinvoluntarily induced to4vomittmg. This action helps to reduce the amount of copper in the body. Hence,copper poisoning is actually rarely found in humans. Copper is also of concernbecause it is toxic to fish at relatively low levels. In addition, at high levels, theformation of copper salts can cause blue green staining of plumbing fixtures.2.3 Copper and Lead RegulationsThe USEPA Lead and Copper National Interim Primary Drinking WaterRegulations (1991) specified that lead levels in 1 liter standing samples should notexceed 15 ppb in 10% of the samples taken at the tap. Copper levels should not exceed1.3 ppm in 10% of the samples. The samples are to be taken from the kitchen taps of anumber of targeted residences. The public water utilities must show that the watersupply does not exceed the prescribed levels for both copper and lead. Depending onthe size of the distribution system, a number of samples have to be collected andanalyzed in order to determine whether the supplied water meets the copper and leadregulations. When the regulations are shown to be exceeded, then the water utilitymust implement a corrosion control program. Alternatively, the utility mustdemonstrate that the existing treatment program has already maximized the level oftreatment possible.In Canada, the Guidelines for Canadian Drinking Water Quality proposedifferent regulations for lead levels. The maximum acceptable concentration (MAC)for lead is 10 ppb (Health and Welfare Canada, 1989). However, this guideline appliesto a thoroughly flushed sample. There is no MAC for copper, but it is recommendedthat copper level be kept below 1 ppm. It is not specified whether this aestheticobjective applies to standing samples or flushed samples.2.4 Principles of Metaffic CorrosionCorrosion of metals means that the metal is gradually destroyed by chemical orelectrochemical reactions with the environment. In almost all cases, metal corrosion inan aqueous environment is caused by electrochemical processes (Obrecht and Pourbaix,51967). An electrochemical cell must have four important elementsin order for thecorrosion reaction to take place. First, there must bean anode; this is the place wherethe metal is oxidized. Electrons are generated fromthis place and are passed unto thecathode, which receives the electrons. At the cathode, corrosivesubstances, such asdissolved oxygen, chlorine, and hydrogen ionsare reduced. Between the anode and thecathode, a conductor must exist. This is usually themetal pipe, which allows theelectron to move from the anode to the cathode. Fourthly, an electrolytemust bepresent to provide a medium for moving the various ions involved inthe oxidizing andreducing half reactions.2.4.1 Uniform CorrosionUniform corrosion usually occurs in copperand lead pipes. In this form ofcorrosion, the corrosion uniformly penetrates the entire metal surfacefor a certaindepth. Any given site on the metal surface could be anodic at one moment and becathodic at another (Snoeyink and Kuch, 1985). This phenomenon is promotedwhenthe metal being corroded is immersed in acid solutions or in waterwith high totaldissolved solids (TDS) and with high electrical conductivity.2.4.2 GalvanIc CorrosionGalvanic corrosion takes place when two metals of differentelectrode potentialscome into contact with each other. The metal withthe more positive electrodepotential is the sacrificial anode and the metal with themore negative electrodepotential becomes the cathode. Galvanic corrosion is aproblem when copper pipeconnections are soldered by lead solders. The lead solderis the anode, where the leadis converted into soluble ionic species (Oliphant, 1983).2.4.3 Crevice CorrosionThreaded junctions, screwed joints and inverted seamsare places where crevicecorrosion could occur. There is poor circulation andoxygen depletion at thesecrevices. Halides and sulfates can migrate into thesecrevices and combine with6dissolved metals to form strong acids. After the local corrosion process is initiated, thecrevice becomes bigger, and the rate of corrosion accelerates.2.4.4 Pitting CorrosionPitting corrosion or localized corrosion is a process where the anode remains ata fixed location. This type of corrosion generally occurs on ferrous surfaces. Thefixed anodes could be imperfections in the metal or its oxide film. Regions of morehighly stressed metal can also serve as permanent anodes. Because the anode isconfined to a fixed spot, localized corrosion at that spot will eventually lead to a pinhole leak in the pipe. Pitting corrosion also occurs where sediment has built up onbottom of copper pipe.2.4.5 Concentration Cell CorrosionIn a concentration cell corrosion, different parts of the same metal are subject todifferent environmental conditions and have, therefore, different electrode potentials.The difference of electrode potential between different parts of the metal could becaused by minute differences in pH or concentration of dissolved oxygen or hydrogenions.2.4.6 Selective LeachingWhere the supply water is soft and aggressive, the lead component in brass canbe selectively leached away. This leaves the copper with more pores and makes it softand brittle. A similar phenomenon happens when tin is selectively leached from bronzein soft and aggressive water.2.4.7 Erosion CorrosionErosion corrosion occurs when there is cavitation or impingement attacks. Atthe entrance to pipes, and sharp bends, such as joints and elbows of pipes, the velocityof the travelling water changes according to Bernoulli’s Law. The resulting changes inwater pressure at these locations may cause cavitation, where gas bubbles trapped in thewater collapse and produces high localized pressure on the pipe surface. Impingement7attacks occur when the gas bubbles directly strike the metal surface. The energythat isreleased is sometimes sufficient to breakup protection films that have built up onthepipe surfaces.2.4.8 Stress CorrosionDuring the threading of pipe ends, the cold working of the metal often causesithave to dissimilar stress on different parts of the metal. This could lead tolocalizedcorrosion. The threaded ends of galvanized steel pipes are especially susceptible tothiskind of corrosion mechanism.2.4.9 Microbiologically Induced CorrosionMicrobiologically induced corrosion is, as yet, a poorly understood, butimportant form of corrosion. Bacteria can live on the surface of pipes. Nitrifiers,forexample, use the ammonia in the water as a source of energy. During biochemcialreactions, oxygen is used to oxidize the ammonia. Hydrogen ions are produced inthese reactions. The resulting acidic environment that is produced encourages thereaction rate of the other forms of corrosion mentioned. Where chloramination isthechoice of disinfectant, the availability of ammonia in the water system clearlyencourages the activities of nitrifying bacteria. Microbiological corrosion can also beinduced by iron and sulfur reducing bacteria.2.4.10 Chemical Corrosion MitigationThe majority of the corrosion mechanisms are influenced by suchparameters aspH, dissolved oxygen, standing time, the buffering capacityof water, and the mineralcontent of the water. In general, a soft and aggressive waterpromotes corrosion, and aharder, less aggressive water decreases the potential for corrosion in the watersupplysystem. Controlling corrosion, by making the water harder, is termed “neutralization”.Alternatively, a passivation approach may be employed, where inhibitors, such asphosphate and silicate inhibitors, help to form a protective ifim on the pipesurface toblock the contacts between the electrolytes, the anodes, and thecathodes. The8passivation approach is more expensive to operate than the neutralization approach.Using Zinc orthophosphate inhibitor appears to decrease copper corrosion but mayactually increase lead corrosion (MacQuarrie, 1993).The neutralization approach primarily involves adjusting the pH and alkalinityof the water to make it less aggressive. pH adjustment should always be accompaniedby alkalinity adjustment, in order to provide enough buffering capacity to limit pHfluctuations. Due to the lowering of pH from carbon dioxide absorption, the pH of thewater can drop after the neutralization step. The side effect of the neutralizationapproach are that raising the pH also increases the rate of THM formation anddecreases the effectiveness of chlorine disinfection.2.5 Previous Studies in Copper and Lead LevelsIn previous years, other cities have tried to solve their copper and lead corrosionproblems with pH and alkalinity adjustment programs. Based on the following studies,it has been shown that pH and alkalinity adjustments can reduce the lead levels at thetap by 35% to 75%.2.5.1 Seattle, WAThe Seattle Water Department (SWD) provides drinking water to over onemillion people living in the Seattle Metropolitan Area. The water supplies come fromthe Cedar and the Tolt Rivers, which are characterized by low pH and alkalinity. In1982, a corrosion control program was initiated to reduce corrosion and relatedaesthetic and economic problems. To the Cedar and the Tolt water supplies, 2 mg/L ofcalcium oxide was added. In addition, 9 mg/L of sodium carbonate was added to theTolt supply (AWWARF, 1990). The effectiveness of the treatment program wasextensively monitored throughout the areas serviced by the water supplies. Of thesampling sites, about half were randomly chosen, while the other half comprised ofhouses that made complaints about the water concerning rust stains, yellow water, and9metallic taste. Both standing and flushed samples were taken. The results of themonitoring program showed a lead reduction of 61 % to 68%.2.5.2 Boston, MABoston purchases its drinking water from the Metropolitan District Commission,which operates a surface water supply system. The water delivered to Boston had apHbetween 5.9 to 6.8, alkalinity of 8 mg/L as calcium carbonate, and hardnessof 12mg/L as calcium carbonate (Karalekas et. al., 1983). In 1977, after unsuccessfullytrying to control the lead corrosion problem by adding inhibitors, the pH of thewaterwas raised to 8.5 by adding 14 mg/L of sodium hydroxide. Residences from areasthathad lead service lines were chosen for monitoring the lead levels at the tap. Someofthese residences also had interior lead plumbing. The average lead levels were reducedby 73%, and the variability of the lead measurements also decreased significantlyafterthe pH adjustment. The flushed samples exhibited highly skewed concentrationfrequency distributions.2.5.3 Glasgow, ScotlandThe water supply of Glasgow comes from Loch Katrine. The water ischaracterized by low alkalinity and a pH of 6.2. Due to thefact that there wereextensive lead domestic plumbings, 50% of random tap samples registered leadconcentrations in excess of 100 ppb (Richards et. al., 1984).Beginning in 1978, limewas added to the Milngavie Treatment Works. The pH of thewater supply wasincreased from 6.3 to 7.8. Random tap samples collected after thepH adjustmentshowed a decrease in lead concentrations to the point wheremore than 80% of thesamples had less than 100 ppb of lead. In another part of the distribution system,Kings Park (which is close to the end of the distribution system), the pH wasincreasedto 9.0. The pH adjustments resulted in more than 83%of the samples from this areahaving lead levels less than 100 ppb.2.5.4 Portland, ME10The major source of water supply to the Portland Water District is the SebagoLake. In addition, three well systems supplement this water supply. In 1986, the pHof the water was raised to 8.3 through the addition of sodium hydroxide (AWWARF,1990). The reduction of copper levels in standing samples was 61% to 85%. Thereduction of lead levels was about 69%.2.5.5 Covewood Lodge, NYCovewood Lodge is a resort in upstate New York, and contains severalcabins.A spring that has low pH and low alkalinity serves as the majorwater supply for thisresort. In 1981, a baffled limestone contactor was constructed to treat the water of afew cabins at this resort. The dissolved calcium carbonate raises the pHof the water to7.3 and the alkalinity to 28.5 mg/L (AWWARF, 1990). Over the next2 years, 23samples were taken and analyzed. It was found that the average lead level for theuntreated cabins was 46 ppb and the average for the treated cabinsfell to 18 ppb.2.6 Tap Water Sampling vs. Laboratory SimulationsIn order to assess whether a corrosion control strategy will improve the qualityof supplied water so that it complies with the USEPA copper and lead rule, a fieldstudy or a lab simulation can be conducted. A properly designed tapwater samplingsurvey can determine the level of lead and copperin the distribution system, and theeffectiveness of a corrosion control program in reducing the levelof lead and copper atthe tap. The significant disadvantage to this approachis that any corrosion controlexperiment could actually adversely affect the qualityof the supply water. A saferapproach of determining the effectiveness of a corrosion controlprogram is to do alaboratory simulation of the various corrosion control options. Pipe loops,and coupontesting have been used for a long time to assess corrosion rates andthe relative impactsof various water treatment options. The major problemwith using a laboratorysimulation is that we do not know enough aboutthe sources of lead, and the11mechanisms for lead mobilization, in order to accurately simulate an actualwaterdistribution system.2.7 Sources of Copper in Tap WaterCopper in tap water comes mostly from the leaching of copper from thewatersupply system. Leaching of copper can occur in the transmission or thedistributionsystem. However, most of the copper leaching takes place inthe household plumbings.The copper can be leached from the copper pipes and faucets. Copper leachingfrombrass faucets is the major source of trace copper in the tap water.Copper might also be found in significant concentrations in the source water.Copper can be introduced into water from natural and anthropogenic sources.Throughnatural erosion, copper particles can be picked up and deposited in water sources bywind. Infrequent natural phenomena, such as volcanic eruptions, can also extrude agreat deal of copper dust into the environment. Of the anthropogenic sources,copperis introduced into the environment in metal production, wood andfossil fuelcombustion, and waste incineration activities (Environment Canada, 1981).2.8 Sources of Lead in Tap WaterThe lead found in tap water could be from the water supply source, the watertreatment process, the transmission and the distribution system, orthe service andhousehold plumbings. The following is a list of potential leadsources (adapted fromAWWARF, 1990).Water source and water treatment:• lead containing air pollutions, emitted by industriesthat are located near raw watersources• deposits of lead-bearing materials that are naturally foundin the watershed• point sources of domestic wastewater or industrial discharges, located upstreamfrom the water intake, that might contain leadTransmission and distribution system:12• lead pipes in the distribution system• large water meters or flow detector checks that have a leadcounter weight inside• lead-caulking compounds used to seal joints in the water mains• lead gaskets used as flanges to join large valvesService piping and household plumbing:• resetters for meters that used 50:50 lead/tin solder asthe joining material• lead or lead-lined iron service lines or premise piping• lead goosenecks or lead pigtails• lead solder used to join copper service lines or copperpremise piping• brass fixtures and fittings or pipes with high leadcontents• water coolers that have lead components2.9 Factors Affecting Tap Water Sampling ResultsThe variation of tap water results is caused by a number of analytical, chemical,and physical factors. Uncertainties of instrument calibration, andrandom instrumentresponse errors are two of the most important analytical errorsthat could be made. Inaddition, there could be procedural errors, such as errors in dilution, andsamplemanipulations. The presence of potential interferences in thesamples could alsoincrease the analytical variability of data.The lead levels at the tap are affected by the alkalinity,pH, and dissolvedinorganic carbonate (DIC) of the supply water. Forexample, the equilibrium leadconcentration in the pH range of 6 to 8 could vary by afactor of 5 to 10 per pH unit(Schock, 1980, 1985). The actual amount by which theequilibrium concentrationvaries is dependent on the type of solid that formson the pipe surface.Changes in any of the chemical characteristics of thewater or in the chlorinationpractice, such as the chlorination dosage or the relativeproportion of free andcombined chlorine, could cause a subsequent changein the corrosivity of the water.13The initiation of corrosion treatment to the water supply could result in changes to thesolubility and the adherability of the corrosion products to the pipe walls.Likewise, the standing time of a sample is important. The amountof time thewater is left standing in the pipe can affect the pH, chlorine residuals, dissolved oxygenlevel, temperature, calcium and magnesium hardness, and total andcarbonate hardnessof the water.Physical factors can also influence the outcome of the lead levelsat the tap.There are normally interconnecting lines within a plumbing systemin a house. Thewater in the kitchen faucet is connected in some way to the faucetsin the bathroom,utility room, and the exterior of the house. Any water usage fromany of these faucetswill cause some mixing of the water in the plumbing system.When the faucet, from which the sample is taken, is turned on,a plug flowcondition is created. As the water is drawn through the pipe, the shape and the lengthof the original plug of water flowing through the pipe is altered by the friction of thewater against the pipe wall and by mixing due to turbulent eddies generated in thewater. In almost all houses, the pipe diameter changes from the point where theservice line connects with the house plumbing to where the interior plumbingconnectswith the faucet. At these junctures, there could be alterations tothe shape of theoriginal plug flow. The extent to which this occurs depends onthe flow rate, thedistance between joints, and the size and interiorcondition of the plumbing system(AWWARF, 1990). The amount of a plug ofstanding water that can be recovered insample collection depends on the volume of samples takenrelative to the diameter ofthe pipe, the size of the plug of standing water, andthe degree of water mixing in thesystem. Therefore, even if the faucet from which the samplesare taken has not beentouched, the sampled water could still be effected bythe mixed water.The contribution of metals from the various parts of theplumbing system can beisolated by varying the volume of water sampled. The precisionin identifying the14contribution from the faucets, for example, increases as the volume of water collectedis decreased. The internal volume of the faucets differs according to the faucet design.Most kitchen faucets are usually about 90 to 120 mL (AWWARF, 1990). Bathroomfaucets are smaller. Since standing samples are usually 250 to 1000 mL, measuring thecontribution of the metals from the faucets requires collecting samples that have lessvolume than the volume for the standing samples. In comparing data from onesampling study to another, it is important to note the size of the samples collected.It is important to consider the physical state of the lead which we want tomeasure. The popular mass transfer model of Kuch and Wagner (1983) deals with leadin its various aqueous forms of free ion, ionic complexes, and uncharged complexes.One study examined the size distribution of lead in tap water (AWWARF, 1990). Thestudy found that 65% to 84% of the lead was less than or equal to 0.4jim. Thedissolved lead species are of most concern to sampling studies because they are easilytaken up into the human body after being ingested.The Kuch and Wagner model predicts the amount of lead picked up duringsteady-state turbulent flow through lead pipe. It can also calculate the concentration oflead in the water under no flow conditions in the water. To use the model forpredicting lead concentrations in steady flow conditions, the experimenter must knowthe diameter of the pipe, the pipe length, the water temperature, and the volume rate ofwater flow. As well, an observed equilibrium lead level must be ascertained ahead oftime by analyzing field samples. An estimate for the mass transfer coefficientand thediffusion coefficient of the Pb+ +must also made. However, the mass transfercoefficient is only important if the pipes have thick coats of scales that inhibit thediffusion of lead into the water.In addition to the dissolved forms of lead, particulate lead is present in waterdistribution systems as well. Lead can be adsorbed onto foreign particles, such as ironoxides, corrosion products, or calcium carbonate particles. There can also be15adsorption or ion-exchange with sediment materials, colloidal hydrous ferric, andmanganese oxides (AWWARF, 1990). The complicated reactions involving leadwithother substances is probably regulated and influenced by the chemical characteristics ofthe supply water.2.10 Factors Affecting Lead LeachingThe contact of water with lead soldered joints makes the water nearby the solderenriched in lead. The contact between the solder and the water allows the lead in anylead containing solder to migrate into the water by the process of simple diffusion.Depending on the amount of time the water is in contact with the solder, and theamount of lead containing solder that is present in the plumbing system, variableamounts of lead can be leached into the water.Small differences in the amount of time water stands in a pipe containing leadcan contribute to considerable differences in the degree of lead leaching. Differencesof 10% to 30% in lead concentration is achieved in standing times that differ only 30 to60 minutes. Lead leaching occurs at a faster rate in pipes of small diameters (USEPAProject Report).Brass faucets contain lead that can leach out into the water. The amount of leadleaching from brass faucets depend on the stagnation (standing) time of the waterin thefaucet. The leachability of lead from brass probably decreases with the age ofthefaucet (Neff, 1987). The phenomenon is due to the fact that the leachable zone atthemetal surface is depleted of lead, or because a passivating film is deposited onthemetal surface (Sharrett, 1982; Britton et. aL, 1981). A study that showed veryhighlead concentrations from standing samples implicated the brass faucet asan importantlead source (Murrell, 1985).The liberation of lead from brass faucets is due to dezincification. In areaswhere the free energy of the alloy is above the average value, dezincification canoccur(Oliphant, 1978). These abnormal areas are the resultof crystallographic dislocations,16distortion of the normal atomic array due to casting or drawing of the alloy,inhomogeneities in the alloy, or the presence of impurities. At these places, the zincselectively dissolves and leaves the copper and any lead impurities behind. The contactbetween the lead impurities and the copper matrix can result in galvanic coupling andsubsequent lead oxidation.Numerous studies (Moore, 1973; Schaut, 1942; Gregory et. al., 1984) haveshown that there is an approximately 2 to 3 fold increase in lead solubility when thetemperature of the water is increased from 5 to 25 °C. Because temperature affectsvarious dissociation, solubility, and complexation reactions, the actual amount ofincrease in lead solubility will also depend upon the pH of the water and its carbonatecontent.Particulate lead can be formed through the precipitation of lead solids, whichmay deposit on the pipe surface as passivation film. However, if the velocity of thewater flow through the pipe is too high, these lead deposits might slough off.Relatively large pieces of solder can also be dislodged from the solder mass. Thesolder will travel down the pipe until it is stopped by a bend, elbow, restriction, or thescreen in the faucet. Through the mechanism of leaching, the continuous exposure ofthese particulate lead in the plumbing system can elevate the lead levelsof tap watersamples.Many plumbing systems use components that are made of a variety of differentmaterials. The service lines could be made of lead, the interiorplumbing of copper,and the solder of lead and tin. In these situations, when thedifferent materials comeinto contact with each other, such as when copper pipes are soldered togetherwithlead/tin solder, a galvanic corrosion current is produced (Lyon et. al., 1977).Thisleads to the dissolution of the metals. Solder is shown to be anodicrelative to copperpipes. Water acts as a bridge between the solder and the copper, the two polesof thecorrosion cell. Galvanic corrosion at the solder joints can be a problem even ifthe17capillary joints are well made (Oliphant, 1983). The presence of chloride, andnitratecan increase the galvanic corrosion rates of soldered joints. The chloride penetratesand breaks down the protective ifims on the pipe surface. Nitrate stimulatescorrosionactivities at places where the protective film is exposed. The shift in pH changesthesolubiity constants, hence, the degree of protection, of the protective films.Furthermore, the overall corrosion rate at a soldered joint is determined by theagressivity of the water supply.Not all galvanic corrosions are due to the plumbing systemitself. At least onestudy has shown that copper in the water supplycan deposit on lead pipes and create alocalized galvanic electro-potential cell (Britton et. al., 1981).A newer house, with predominantly newer copperplumbing systems, gives riseto higher lead levels than do older houses. Age has amarginal effect on cold first-flushlead concentrations. However, hot water lead levels appear to be unaffected byage.Cold first-flush lead levels appear to be the same for copperand plastic plumbingsystems. This result is the same for running hot water samples (Singh,1990).Lead exists in its elemental form in plumbing materials. In alloys, such asbrass, the lead component is spread throughout thealloy matrix in particle form. Theelemental lead is oxidized to the 2 + valence state when itcomes into contact withwater. This oxidation process enables the lead to becomemobile and transportable intothe water. These oxidation processes usually take place atthe anodic areas of corrosioncells.In potable water systems, dissolved oxygen andvarious chlorine speciesintroduced through disinfection are the most commonkinds of oxidizing agents forlead. The oxidation reaction is promoted byincreases in the dissolved oxygen content,by decreases in pH, and by the complexation offree lead ions by ligands such ascarbonate, hydroxyl, sulfate, and chloride (AWWARF,1990). The effects of chlorinespecies on lead oxidation depends on the activities ofhydrochiorous acid, hypochiorite18ion, chloramine species and chloride ion. One study showed that, under somecircumstances, chioramination can solubilize more lead than chlorinationwith freechlorine. However, the rate of corrosion due to chloramination is slower thanchlorination (Treweek, 1985). Also, the complexation of lead byhydroxyl andcarbonate ions predominate in normal situations (Schock, 1985).The rate of oxidation reactions goes up with increasing temperature;however,at the same time, the solubility of many film-forming solids goesdown. As a result,the rate of lead oxidation might increase, while the diffusion of the oxidation productsto the surface of the pipe scales could face inhibitions. The scouringaction of water onthe pipe scales can expose certain parts of the pipe to oxidation processes. This servesas a new source of solubilized lead and film-forming components.2.11 Factors Affecting Copper LeachingA newer house, with predominantly newer copper plumbing systems, gives riseto higher copper levels than do older houses. Age strongly influences cold first-flushcopper concentrations; however, hot water copper levels appears to be unaffected byage. Cold first-flush has low copper concentrations for plastic pipe compared tothatfor copper pipes. This result also holds true for running hot water samples (Singh,1990).2.12 Plastic Pipes vs. Copper PipesIn contrast to copper pipes, plastic pipes should have less problemswithcorrosion and metal leaching. For this reason, plastic pipes are nowcommonlyinstalled in new homes as the preferred plumbing material (Economicand EngineeringServices Inc., 1990).Plastic pipe is made primarily of polymerized organic compounds.Someresidual unpolymerized monomers may be present. PVC pipes aremade by extrudingthermoplastic PVC at temperatures between 150 to 200 °C. In orderto make theprocess more stable, lead or tin compounds are commonly added.The presence of this19lead in the PVC pipes could contribute, but not necessarily, to minor levels of leadleaching. As well, there are other problems with using plastic pipes, such as the risk ofleaching carcinogenic and other organic compounds.2.13 Quality Control in Laboratory TestsA sample validation process should be in place to make sure that a measurementis correctly reported for the sample from which it came. Where the samples are takenfrom a targeted population with known water characteristics, the samples should beanalyzed for those characteristics to verify that they correspond to the expected values.If a sample does not meet the criteria for a good sample, it should be excluded from thedatabase.For analyses involving metal ions, samples should be acidified to pH of lessthan 2. The acidification of the samples minimizes the possibility of the metal ionsprecipitating or adsorbing onto the walls of the containers (Mancy, 1971).The quality of sampling data depends heavily on the precision and the bias ofthe measurement methods and instruments used. If the samples are tested by severallaboratories, the variability of measurements between the laboratories will be affectedby the type of instruments and reagents used, the sampling handling techniques,thediffering abilities of analysts, and the quality of laboratory support facilities.Both quality control and quality assurance programs should be employed.Quality control refers to those activities, such as spiking, and calibration,that are usedto assess the quality of the measurements. Qualityassurance is the larger, overallmanagement system that ensures that the quality control programis working effectively(Keith, 1991). Normally, quality control charts are drawn tomeasure the stability ofthe measurement instruments. Standards are periodically tested. Themeasurementprocess is out of control when a measurement of the standard is abovethe upper (UCL)and lower than the lower (LCL) control limits; these aredefmed to be plus and minus 320sigma around the sample mean. When the process is out of control,the instrument hasto be recalibrated and the samples reanalyzed.2.14 Reporting of Low Level DataZero or negative values in measurements are usually considered to beoutliers.This presents a problem when most of the measurements in a sampling program,suchas found in this study, have true values that are expected to beclose to zero.Every measurement instrument has its detection limits. There are threebasicexpressions of the detection limit. Firstly, there is the limit of detection(LOD). Thisis the lowest concentration level that can be statistically determined to be differentfroma blank at a specified level of confidence. Secondly,there is the reliable detection level(RDL). This is the concentration at which we can say it is extremelylikely that thereis detection. Thirdly, there is the limit of quantitation (LOQ). This is defmedas thelevel above which concentrations can be specified with a certain degreeof confidence(Keith, 1991). The LOD is usually set at three times the standarddeviation of theinstrument precision (3a). This ensures that we encounter “false”positives only 0.1%of the time, i.e., 99.9% level of confidence. If RDL is chosen to be6cy, as is theconvention, then the chances of having false negatives is also0.1 %. The LOQ isusually recommended to be set at lOa. Values at the LOQ havean uncertainty of plusand minus 30% at the 99% confidence level.A measurement that is lower than the LOD is sometimesnot included in thedata analysis because we can not be sure about the actualvalues of these very low levelmeasurements. These measurements theoreticallyhave finite, and positive values.Ignoring all data less than LOD might result in a left-censored dataset. This meansthat the resultant database might be biased to the right. Somepeople prefer to retain allof the data as is, including all the values that are lessthan LOD (ASTM, 1984).2.15 Considerations in Experimental Design21Most of the previous studies that looked at the metal levels in buildings orhouses only reported the mean values of the metal levels. However, almost noneofthese same studies examined the variability of the metal levels in the system.It hasbeen well documented (Bailey, 1986) that the mean lead level in one house can besignificantly different from the mean lead level in another house of the samearea. Thisresult shows that the distribution of lead in the water supply system is a highlyvariableprocess. Not only is there a great deal of variability between the samplingsites, thereis also a high degree of variability within the same site. Published field andlaboratorystudies of lead, solder, and brass corrosion (AWWARF, 1990) indicate thattheequilibrium condition is usually not achieved in most samples taken. According totheKuch and Wagner model of the lead stagnation curve, there is a sharp rate of changefor lead before the equilibrium level is reached. This means that at concentrationsmuch below the equilibrium, the level of lead in a standing sample could vary by awide margin. In this case, it is clear that sampling should be repeated several times,for each house, in order to reflect the variability within each site.To ascertain the effectiveness of a corrosion control strategy on reducing metallevels at the tap, detailed statistical analyzes should be performed. Some problemsarecommon to statistical reporting (Study Group on Environmental Monitoring, 1977).These include: lack of statistical sophistication, no calculations of the precisionofestimates, no statement of the test hypothesis, insufficient sample size,no descriptionof the method of sample collection, no non-response mechanism,and no respondentbias mechanism.Much effort should be paid to the selection of sampling sites toensure that thesampling program yields the desired results. In general, the moresampling sites areincluded in the study, the more accurate will be the results. However, itis not easy topersuade some homeowners to provide standing samples. Due to budgets, anymonitoring program will have constraints on how many samples can be collectedand22analyzed. A crucial question to be considered at the start ofthe experimental design iswhat kind of gain in accuracy is there for every marginal increasein the samplingeffort (Gilbert, 1987).2.16 SamplingSampling means to select a few houses and to measure their metal levels,instead of doing the measurements for all the houses in the study area. Theresults weobtain from the few selected houses can also tell us something about themetal levels ofthe entire study area, through the process of statistical inference.In a sampling survey, attention must be paid to all aspects and phasesof thesurvey. Poor work in just one phase of the survey may ruin theresults even if all theother phases have been done well.Sampling survey theories have been developed extensively for normallydistributed distributions. A large part of these theories is concernedwith finding aformula for the means and variances of the distributions. The sample survey theorydiffers from the classical theory of sampling in that a population group insurvey workcontains a finite number of units, whereas the classical theory assumes an infinitelylarge population. For practical purposes, the differencebetween the two theories areseldom important.Nonprobability sampling refers to those sampling methodsthat are not amenableto sampling survey theories because the selection ofsamples is not random. Forexample, the samples might be selected haphazardly, or theselection process mightinvolve human judgment. In judgment, or purposive,selection, the sampler inspects aheterogeneous population and selects a typical unit, whichthe sampler deems to beclose to the average of the population. Nonprobability sampling canyield useful resultsif good judgment is employed.If the population has an underlying normal distribution, goodsampling tends tomake the sample distribution more normal. Bad samplingpractice usually results in the23presence of many outliers, which potentially not only skews the sample distribution,but also increases the sample variance and decreases the precision. When possible, thecause of the outliers should be identified and the design of the sampling survey shouldbe adjusted accordingly.Nonresponse refers to not being able to measure some of the units in a selectedsample group. As a consequence of nonresponse, the statistical estimates may bebiased, since the nonresponse part of the population may be different from the part thatdid respond. Evidences from previous studies suggest that the magnitude ofnonresponse bias varies widely from survey to survey (Cochran, 1977). Because thesample actually obtained is smaller than the size of the targeted sample, the variancesof the estimates are increased.Stratification sampling is an efficient way to sample a heterogenous population,especially when it is possible to divide the population into subpopulations, each ofwhich is internally homogeneous. Each subpopulation is called a “strata”. Since eachstrata is homogeneous, an estimate of any stratum mean can be obtained from a smallsample size. The estimates for the strata can be combined to form an estimate for thewhole population group. When used properly, stratification almost always results in asmaller variance for the estimates than is given by a simple random sampling.Quota sampling is basically stratified sampling with nonrandomselection ofunits within a stratum. Because the sample selection is not strictly random, the usualstatistical formulae do not apply. Sampling continues until a targeted number ofsamples is attained.Single-stage cluster sampling is type of stratified sampling. Here, insteadofrandomly selecting units within each stratum, only one cluster is randomly selected.All the units that are defined to be within the cluster will be sampled. Clustersamplingis the most economical way to do sampling if it is expensive to travel between theunits, and the distance between the units is long.24Before any of the stratification sampling methods can be employed, thefrequency distributions of the population must be known. This kind ofinformation iseither available from previous studies done on the same population group, or ifnot, canbe obtained by doing a preliminary sampling. This techniqueis known as the doublesampling or two-phase sampling. This method is useful only if the behaviour oftheparameter to be measured in the population group does not change with respect totime;otherwise, the results of the preliminary sampling can not be used to definethe stratafor the later sampling stage. The accuracy of multiple-stage sampling improves as weincrease the number of samplings. However, respondents who are repeatedlyasked forthe same information may not be willing to cooperate after a certain time.2.17 Common Assumptions Made in StatisticsMost of the standard statistical tests that have engineering applicationsaredeveloped based on certain assumptions. First, the underlying distribution isusuallyassumed to be normally distributed. This assumption is often made because theanalysis of non-normal distributions is highly difficult to compute. In caseswhere theunderlying distribution is not normal, transformation routines can be applied totransform the original distribution into a normal one. Sometimes, non-parametricstatistics have to be employed if the transformation routines are unsuccessful.In theoretical statistics associated with the normal distribution, thepopulation isconsidered to be infinite. In sampling statistics, the population is almost alwaysfinite.As long as the sampling population is large in comparisonwith the size of the samples,there is not much difference between theoretical and samplingstatistics.The selection of the samples from the population is usually assumed to beindependent and random. In other words, the probabilityof selecting a certain sampleis equal to, and not influenced by, the selection of anothersample. To ensure that thisassumption holds true, the population size must be large.Also, the selection ofsamples must follow a random scheme. In real life,the “random scheme” is actually25computed by a random number generator that gives numbers, within a predeterminedrange, with a equal degree of probability.2.18 Measuring Central TendencyThe measurement error is composed of the random error plus the systematicerror. The random error is related to the precision of the measurement process and thesystematic error reflects the accuracy of the measurement. The systematic error is alsocalled the bias, which is the difference between the expected value of the distributionand the population mean.Of the systematic errors, blunders is one type. A blunder is a technical termmeaning that a wrong sample is measured, or there is a misreading of the measurementscale, or a mistake in transcribing or transposing measured values. If large, a blundermay show up as an outlier. But once a blunder is made, it is next to impossible torecover the true value.Due to the central limit theorem, we know that the magnitude of the randomerror decreases as the number of measurement increases. As n gets larger, the sampledistribution also approaches the normal distribution. There is no safe rule available topredict the sample size needed to approximate the normal distribution. However, forsamples that primarily deviate from normality due to positive skewness, a cruderule isn>25G12where G1 is Fisher’s measure of skewness, and is equal toG1This rule is designed so that a 95 % confidence probability statementwill be correct94% of the time (Fischer, 1932).For statistically independent variables, the expected value, E, has the property:E(ax+by+cz+ . ..)=aE(x)+bE(y)+cE(z)+Also, if U is defined as26U=ax+by+cz, then the bias of U isB(u) =aB(x) +bB(y) + cB(z) +... (Mandel, 1964).In statistical language, the central tendency is measured by a parameter,themost common of which are the mean, median, and the mode. Themean, or arithmeticmean, is defined as the sum of all the observations divided by the number ofobservations, i.e.The median, in a set of observations that is ranked or arranged in orderof magnitude,is the middle observation. If the number of observations is even,then the median is theaverage of the two middle observations. As a measure of the central tendency, themedian is not as prone to be affected by outliers as the mean measurement. Themedian is also easier to calculate than the mean, but, the mean has an advantagethat itis always an unbiased estimate of the population mean (Kennedy, 1986). Modeis thevalue that occurs most frequently in a set of observations, or, in a continuousdistribution, the value with the highest frequency.2.19 Measuring SpreadFor statistically independent variables, the variance, V,has the property:V(ax +by + cz + .. .)=a2V(x)+b2V(y) +c2V(z) +In statistics, the spread of a distribution is measured by thevariance, which isdefined asJ2=_______Because the variance has units of the square of the units of thevariate, the standarddeviation is often used in place of the variance. The standard deviation,a, is definedas the square root of the variance, and has unitsthe same as those of the variate.27When the mean value, ji, of the population is not known, an estimator can beused to estimate the standard deviation. The estimator is defined as=V n—iThe estimator is almost the same as the standard deviation, except thatthe denominatoris n-i instead of n.2.20 EstimateThe statistic that estimates a parameter of a population is calledthe estimate. Agood estimate should be unbiased, consistent, efficient, and sufficient. Unbiased meansthat the estimate is the same value as the true value for the population. Theestimate isconsistent if it approaches the population value as the sample size increases. Efficiencyrefers to the variance of the estimate. A highly efficient estimate has a low estimatevariance. Sufficient means the estimate has used all the information a sample containsabout the parameter to be estimated.2.21 Coefficient of VariationSupposing that x can only be positive, the ratio of the standard deviation to themean value is called the coefficient of variation, or coefficient of variability, c.v.forx>O.xThe convenience of this quantity is that it is a dimensionless measureof dispersion,with the mean value as the measuring unit.2.22 SkewnessThere are many ways to measure skewness. In general,skewness is a measureof the asymmetry of the distribution, as opposed to the perfect symmetryof the normaldistribution.2.23 Central Limit Theorem28The central limit theorem proves that if we take independentsamples from apopulation with a finite variance, all of size N, then the averagesof these samples willresult in a sample distribution that tends toward normal.This result is true regardlessof the original population from which the samples were taken.Also, the larger the N,the greater will be this tendency towards normality.2.24 Hypothesis TestingHypothesis testing has to do with making inferences about thepopulation, basedon the information from the samples. The hypothesis that acertain parameter of twopopulation distributions agree with each other is calledthe null hypothesis, whichassumes that the difference between the parametersis zero. The null hypothesis istested against an alternative hypothesis. If the null hypothesisis tested to be not true,then the alternative hypothesis is said to be true, providedthat the null hypothesis plusthe alternative hypothesis cover all the possible outcomes.One problem with almost all hypothesis testing is that we do notknow thepopulation distributions. Instead, we only have sample distributionsthat approximatethe populations. The samples from any one population will have samplingvariations.Therefore, hypothesis testing of two sample distributions will practicallyalways yield adifference.Hypothesis testing is meaningless unless we knowthe accuracy and theconfidence intervals of the conclusions from thehypothesis tests. To determinewhether or not the differences are due to the differencesof the populations or thesampling variations, the probability of the conclusionsbeing right must be stated. Thenull hypothesis is rejected when the testing indicates that thereis only a small chancethat the populations are the same.The level of significance, x, of a hypothesis test is themaximum probability ofrejecting a true null hypothesis. For most statistical tests,a. is chosen to be 0.05. The29power, 1-f3, of a hypothesis test is defined as the probability of rejecting a false nullhypothesis.A test is unbiased if the probability of rejecting the null hypothesis, Ho, whenHo is false, is always greater than or equal to the probability of rejecting Ho when Hois true, i.e., f3cL (Conover, 1980).A test is conservative if the actual level of significance of the hypothesis test issmaller than the stated level of significance.Conclusions that are drawn from hypothesis testing are never sure. At best,they give us an indication of what might be true about the population. The conclusionsfrom one hypothesis test can give us clues as to what other kinds of hypothesis testsought to be performed. By properly structuring the hypothesis testings, we canhopefully find out what we want to know from the data that have been collected.2.25 Constructing Power CurvesA power curve is a plot between the power of the test, 1-3, and the sample size.It is very useful to be able to know the power of the test associated with the sampledata being tested. However, the construction of power curves requires prior knowledgeof the population means, which are usually not available for most studies (Yamane,1964).From the theoretical results of power curve analyzes, we know that, for thesame level of significance, a one-tail test is always more powerful than a two-tail test.2.26 Statistics of the MaximumWhen we are sampling a population for the maximum values obtained over adefined time interval, we end up with a set of extreme values. The classof distributionfunctions that can describe this kind of distribution is called the “double exponentialdistribution function”. Theoretically, it can be shown that the double exponentialdistribution is the limiting distribution of extreme values of large samples takenfrompopulations such as Gaussian (Kinnison, 1985). Many types of environmental pollution30problems, including tap water sampling, can be viewed as an extreme valueproblem.The “maximum value” statistics should be used if we want to know themost likelymaximum values to be obtained in a sampling program that spans acrossvarious timesand locations.2.27 Systems of MeasurementWe normally measure things using the system of realnumbers. For example,we say that a certain pen is eight inches long or a chairacross the room weighs fivepounds. This type of measurement is sometimes also called theratio scale ofmeasurement. The main characteristic of this measurement system is that thereis atrue zero point. In systems where the zero point is only arbitrarily set, such asin athermometer measurement, the measurement is defined in terms of an interval scale.In such a system, the intervals between the numbers have an empirical meaning,but theratio between numbers do not. Certain other measurements are not as easilyquantifiable. Licence plate numbers and zip codes of addresses, for example, aredefmed rather arbitrarily. These measurements are said to be nominal. In a taste test,a panel of judges are asked to rank the flavor of chlorinein drinking water samples.On a scale from 1 to 5, one being no chlorine detectable and five being strongchlorinetaste detected, only the relative order or position of theparameter of interest isimportant. This system of measurement is called “ordinal”.2.28 Nonparametric TestingIt is hard to handle sample distributions that are notnormal or log normalbecause most statistical tests are designed on the basis ofthe normal distribution.Whenever a nonnormal distribution is encountered, it canpotentially be transformedinto a normal distribution by the log function or thefunctionf(x)=where q belongs to the set of real numbers. The most commontransforming functions of this form are the square root, thenegative reciprocal, and thenegative reciprocal square root (Tukey, 1977).31However, in situations where even the transformation procedure can nottransform the actual distribution into one that is normal or log normal, thenonparametric testing method should be used. Nonparametric statistical methodsarethose that can be used on data based on almost any measurement system, such asnominal, ordinal, interval, and ratio data. It can also be employed when thedistribution function of the random variable producing the data is either unspecified,oris specified but has a large number of unknown parameters.2.29 Kohuogoroff-Smirnoff Comparison of Two Independent SamplesThe Kolmogoroff-Smimoff test can tell us if two independent samples frompopulations with continuous or discrete distributions, but both of the same type, aredrawn from the same population. This test is robust against differences in the shape ofthe distribution, especially differences in the mean, median, dispersion, and skewness.The maximum difference between the cumulative disthbution functions, F, ofthe two populations serves as a test statistic, D.D = max(— — —)fl‘2The critical D value, for an aggregate sample size of 35, can be approximated byDa = Ka/ni +n2Vi2Ifni=2,Ka=1.36 when a =0.05 (Sachs, 1982). If D> Da then there is asignificant difference between the distribution of the two populations.2.30 Median TestThe median test examines whether the sample sets being tested come frompopulations that have the same median. However, the underlyingpopulations need notbe identical when the Ho is true.2.31 Kruskal-WaHis H-Test32The Kruskal-Wallis H-Test is the extension of the Mann-Whitney Test for TwoIndependent Samples. Where there are k random samples, each of which is possiblyfrom a different population, the H-Test tests the null hypothesis of whether or not all ofthe populations are identical. The alternative hypothesis is that one or more of thepopulations tend to give larger values than the other populations, i.e., not all thepopulations have identical means.The H-Test is similar to the median test. However, the H-test uses moreinformation contained in the observations than the median test. Therefore, the H-test isusually more powerful than the median test. The disadvantage of using the H-test isthat all the observations have to be ranked in the combined sample. For a large dataset, the H-test involves more work than the median test.The assumptions of this test are that all samples are random samples from theirrespective populations. In addition to independence in each sample set, the varioussample sets must be mutually independent from each other. Also, the measurementscale of the data must be at least ordinal.The test statistic, T, is defined asT=-1-(Y——N(N+l)2)S2 ‘=‘n 4S is defined asS2= N_lR3S N_l(j)N4)N is equal to the total number of samples of all the sample groups. And R istheaggregate rank, across all the sample groups, of each observation. The null hypothesisis rejected for T>Ta.For k=3, and n >5 for each sample group,Tacan be found in atable (Conover, 1980).If, and only if, the null hypothesis is rejected, we can determine each pair of thecompared populations differ from each other. The test statistic is33jR12 N—l—T 1/2 1/2>t1_(/2)(S) ( )n. n1 N—k i n1Population i is said to be different from population j if the inequality holds true.2.32 Confidence Intervals for the Difference Between 2 MeansThe confidence intervals of the difference between 2 means can be computed bythe non-parametric method that bears the same name. In this method, it is assumedthat the two distributions being compared are identical except for a difference in thelocation of the mean. If n denotes the number of samples in distribution A, and mdenotes the number of samples in distribution B, then we calculate an intermediate,k= wa,2— n(n + 1)/2, whereWa12is a function of n and m (Conover, 1980). For allpossible pairs of (X,J.), the kth largest difference, U, and the kth smallest difference,L, are the respective upper and lower limits of the confidence interval for thedifference between 2 means, i.e.,P[LE(X)-E(Y)U]l-a2.33 Test for Equal VarianceThe test for equal variance is designed to test whether or not population X hasthe same variance as another population, Y. If the populations of X and Y have normaldistributions, the F test should be used in place of the nonparametric test for equalvariance. However, the F test is extremely sensitive to the assumption of normality.Even if the true underlying distribution is a double exponential distribution, which canresemble the normal distribution, the true level of significance may be two or threetimes as large as it is supposed to be. Therefore, the F test is not safe to use unless weare sure that the populations are normal. The A.R.E. of the Squared Ranks Test forEqual Variance is 0.76 (Conover, 1980) if it is used instead of the F test, when thepopulations are actually normal. But, for other distributions, the A.R.E. increases andapproaches unity.34In addition to the usual assumptions of independence, and randomness, the testfor equal variance also assumes that the measurement scale is at least interval. For atwo-tail test, the null hypothesis is Var(X) =Var(Y).Let n denote the number of samples in distribution X, and m denote the numberof samples in distribution Y. The absolute deviation of each observation from themean is U =X —p, i=1,...,n; and J’.The ranks 1 to n+m are assigned to the combined sample of Us and Vs. Insituations where several Us and Vs are exactly equal to each other, the average of theranks, if there are no ties, are assigned to each value. The rank of each observation isdenoted by R(U) and R(.).The test statistic is- T-nR21r ,1/2timp4nm[N(N_i)i=’N—i’ ‘where N=n+m, and,= ±{$[R(u1)]2+[R(v)]2};= $[R(U)] +[R(v,.)]4The null hypothesis for a two-tail test is rejected if i’ is less than ctI2 or greater than 1-cd2 (Conover, 1980).2.34 Spearman Rank CorrelationsWhere the data consists of pairs of numbers (x,y), a measure of correlationbetween the two numbers can be calculated. Correlation estimates the degree ofdependence between x and y. The Spearman measure of correlation in ranked data isdefined as- iJ][R(Y)- LiJ]n(n2—1)11235The correlation measure only assumes values between -1 to +1. Thecorrelation is closer to +1, if the larger values of X tend to be paired with the largervalues of Y. If the larger values of X tend to pair with the smaller values of Y, themeasure of correlation tends to -1. If there the pairing of X and Y do not follow anypredictable pattern, then the correlation should be close to zero.2.35 RegressionAs opposed to correlation methods, regression methods are used to inspect moreclosely the relationship between x and y in bivariate data sets. One important objectiveof regression methods is to predict a value of y where only the value for x is known.This is done based on the information we obtain from existing (x,y) data sets.The variables used for regression analysis should be carefully selected. Also,we should know beforehand whether the form of the fitting function is linear ornonlinear. A linear regression function is in the form y =a+bx, where a is theintercept, and b is the slope of the regression line. Both the a and b regressioncoefficients have physical meaning. Therefore, given that the data can be fitted by alinear regression line, the most important task is to determine the values for the a and bcoefficients.The best empirical fit of the data points may not necessarily be accurate. It ispossible to develop, between variables, relationships that are completely meaninglessin a practical sense (Montgomery, 1991).Regression relationships are only valid for the range of values in the originaldata. Regression models should never be extrapolated to predict values outside of therange of the original data.2.36 One-way ANOVAThe analysis of variance, ANOVA, seeks to determine if the differencesbetween groupings of data are significant. Data is classified into groups based on thedistinguishing features of the data. In the analysis of water samples, the chemical36parameters of the samples can be used to group data. Lead sampleswith low pH andlow alkalinity can be grouped together, and samples with higher pHand higheralkalinity grouped separately. In this illustration, pHand alkalinity are said to be theANOVA factors of the lead analysis. If the analysisreveals significant differences inthe lead concentrations between the various pH and alkalinity datagroupings, we cansay that the pH and alkalinity factors significantly influencethe lead data. Thisillustration is also an example of a two-way ANOVA because there are twofactors: pHand alkalinity. In an one-way ANOVA, the data is grouped basedon only one factor(Montgomery, 1991).The study of ANOVA is an extremely vast and complicatedfield of study instatistics. For the purposes of this particular research study,it will suffice to explainthe simplest case.A typical data table for a single-factor experiment is presentedin Table 2.Table 2 Typical Data for a Single-Factor ExperimentTreatment Data TotalsAveragesLevel1Yii Y12 Yin Yl2Y22 Y2n Y2aYal Ya2 YanYaYaFrom the data table, we can calculate the totalof the treatment level totals, yy=and the average of the averages is, y = yIN.i=1 j=1In addition to these, we can calculate other important quantities,such as:SsT=2(yy); sSE=SSE=(YYl);SST=SSfreat,fleflj+SSE;i=1 j=1 2=1J=I37Source of Sum of Degrees of Mean Square F0Variation Squares FreedomBetweenSStreatmentsa-itreatmentsF—°MS,Error (WithinSSEN-aMSEtreatments)TotalSSTN-iThe most important part about this whole table is the F0. If the computed Fishervalueexceeds the critical Fisher value, then we conclude that the between treatmentfactor issignificant.2.37 Theory of CausationStatistical methods can function to demonstrate the relationships, theinterrelatedness, or the correlations between variables; however, no statistical methodscan prove that there is certain causation. Suppose we find a positive relationshipbetween the number of books a high school student purchases and the gradepointaverage the student receives. We can demonstrate through the useof statistics that astudent who purchases a lot of books also receives a high grade point average.However, it is not valid to, therefore, say that buying a lot of bookswill ensure thestudent of a high grade point average. There may be other factors involvedhere. Forexample, the students who purchased a lot of books may also havespent a lot of time inreading those books. As a result of their reading efforts, their increased knowledgeofthe school subject matters has helped them to excel in examinations. We see,then, it isin fact reading books, not purchasing them, that affects grade point averages.A poorstudent who does not have the money to purchase books, but assiduously poursthroughMSat,nenttm;and MSE=. We summarize these important deriveda—i N—avariables in the one-way ANOVA analysis table (Table 3).Table 3 The ANOVA Table for the Single-Factor, Fixed Effects Model38volumes borrowed from the local library, may do better in examinations than studentswho buy a lot of books but do not read them.In certain cases, the presence of correlation between variables has no meaningin terms of causal relationships. For example, we find that there is a strong correlationbetween the length of the right and the left arm, and the height and body weightof aperson (Sachs, 1982).There is another class of noncausal relationship called the inhomogeneitycorrelation. Consider a population that has three subgroups. When we measure twoquantities: A and B, they may not be correlated in each of the subgroups, but acorrelation may be found when all the three subgroups are analyzed together asonepopulation.Although we do not know for certain if there is a causal relationship betweencorrelated variables, but we can make this conclusion once we have excluded otherpossibilities. This method of proving causality is only useful if we have an exhaustivelist of all the possible types of noncausal correlation relationships. Furthermore,wemust have a way of detecting these relationships.393. EXPERIMENTAL METHODS3.1 Experimental DesignThe study areas of this research project are located in North Delta, Newton, andSouth Surrey of the Greater Vancouver Regional District. North Delta is the controlarea. Due to bacteriological and chemical chlorine demand within the distributionsystem, the supply water from the Seymour Reservoir is stripped of chlorine residualsby the time the water reaches North Delta. In North Delta, no chemicals are added tothe water to adjust for pH and alkalinity. The Seymour water supply also feedsNewton and South Surrey. In Newton, the water is rechlorinated and adjusted for pHand alkalinity. In South Surrey, the water is chioraminated and also adjustedfor pHand alkalinity levels.The major water feeder main for Newton comes into the distribution grid at theintersection of 128th Street and 64th Avenue. Similarly, 144th Street and 32nd Avenueis the feed point for South Surrey. As the water enters into the distribution grid, thechemical characteristics of the feed water, such as residual chlorine levels, pH, andalkaiinity, change according to the physical, chemical, and biological conditions of thedistribution grid. In general, as the water travels farther away from the feedpoint, theresidual chlorine levels drop. The change for pH and alkalinity is harder to predictbecause the interactions between the supply water and the distribution pipescould leadeither to increases or decreases of pH and alkalinity.By collecting tap water samples from North Delta, Newton, and South Surrey,we can compare the copper and lead levels in Newton and South Surreyagainst those inthe control area. The comparison tells us if pH and alkalinity adjustments are keyfactors in lowering copper and lead levels in tap water. Because pH, alkalinity, andresidual chlorine levels vary in Newton and South Surrey, each area has to be studiedinternally to see if there are significant internal differences.40Samples from the control area should have fairly homogenous pH, alkalinity,and residual chlorine measurements. Therefore, random aerial sampling (takingsamples from locations randomly distributed throughout the control area) would be thebest way to assess the copper and lead levels. In contrast to North Delta,Newton andSouth Surrey will have pH, alkalinity, and residual chlorine measurements that span awide range. Simply taking random aerial samples from each of these two districtswould not reveal significant differences, if there were any, in copper and lead levelswithin each district. Assuming that water with higher residual chlorine levels also hashigher copper and lead levels, areas closer to the feed points should have the highestmetal levels in Newton and South Surrey and areas farther in distance from the feedpoints should have the lowest metal levels. By getting a representative set of samplesfrom an area closest to and another set of samples farthest from the feed points, wehave an excellent chance of finding any significant differences in metal levels betweenthe two areas. Due to the spatial variations in water quality parameters, we cannottravel too far away from one sampiing site to another before we find substantiallydifferent pH, alkalinity, and residual chlorine levels. Hence, cluster sampling is thebest way to get the two sets of representative samples from Newton and two setsfromSouth Surrey (Appendix A).3.1.1 Controllable FactorsBesides pH, alkalinity, and residual chlorine levels, there areother factors thatcould influence the copper and lead levels in tap water. Mostof these factors can becontrolled directly or indirectly. By direct control, we mean that we cancontrol thequantity or quality of the parameter of interest. For example, we can controlthe pH,alkalinity, and residual chlorine levels inside the rechiorination station. Byindirectcontrol, we mean that we cannot control the parameter of interest, but we can useothertechniques to influence the outcome of a variable. Outside of therechlorination station,the pH, alkalinity, and residual chlorine levels begin tovary in an unquantifiable way41because we do not have precise information on every part of the entire distributiongrid. But, if our objective were to collect tap water samples that have a similarlevel ofresidual chlorine, we can indirectly achieve this objective by taking our water samplesfrom houses that are very close to each other. This way, although we cannot say “apriori” that the levels of residual chlorine in all the samples will be exactly the same,the chances are that they will be similar. Direct ControlThe age of the house, from which the water sample is collected, is a factor thatcan be controlled directly. This study investigated the worst potential copper and leadproblems, which previous studies have shown to be houses plumbed with copper pipingthat are less than ten years old. Therefore, the houses selected for this study had to beless than ten years old. Some very old houses may have been renovated in the past tenyears with new copper plumbing. But, since it is nearly impossible to find out whichold house has new plumbing, it was much easier to target for houses built in the pastten years.A problem arises with using this strategy because some of these very newhouses had plastic plumbing. These houses are not likely to have a problem withmetalleaching into the tap water, and these houses were not included in this study. Theresidents of the houses were asked whether their plumbings are plastic or copper.Ifthey were not sure, we could check the water service pipes underneath the kitchensink.Most houses that have plastic plumbing inside the house will also have plasticservicepipes. However, this method of determining the predominanttype of plumbingmaterial used in the house is at best tenuous. On the other hand, this was probablytheonly method available, if the residents are not sure what type of plumbing is inthehouse.Samples were collected both from houses that have plastic and houses that havecopper plumbing. The data from houses plumbed with plastic pipes had to beanalyzed42separately from those with copper pipes. Results fromthe two separate data sets werecompared with each other to confirm that houses withplastic pipes indeed did not havea metal leaching problem.All tap water samples were collected from the kitchen tap of houses(singlefamily dwellings). The metal levels in the tapwater from the same house will vary,depending on from which tap the water is collected.This is due to the fact that the tapfaucets contribute to the metal leaching problem. Not only aredifferent faucets in ahouse of different sizes, but they may be of differentmakes and have a differentpercentage mixture of metallic components that are susceptible toleaching. Collectingwater samples only from the kitchen tap minimizes the variation in tapwater metallevels due to the variation in the faucets.This study targeted houses, rather than apartmentsor buildings, because thedifferent types of dwelling cannot be studied together. Previous studieshave shownthat the type of building will influence lead levels in tap water. Apartmentsandbuildings have, in general, higher levels of lead than houses.Apartments or buildingshave much longer internal plumbing loops than houses. The averageresidence time ofwater in the plumbing system of a house is much lessthan in an apartment. Therefore,there is less contact time between water and the leachablemetals in the plumbingsystem of a house (Singh, 1990).For the past few years, on-line water purificationdevices have become popularin the market. The house residents were asked if they haveinstalled such devices ontheir kitchen taps. In cases where they had,they were instructed to turn off thosedevices during the sampling process. Someof these water purification devices areveryeffective at removing metals in the water. The failure toconsider this variable could bedetrimental to this study. Indirect Control43The natural level of lead and copper in the supply watercould be a problem ifthey were much higher than the levels leached from thedistribution and plumbingsystems. Fortunately, GVRD data show that the naturallead and copper levels in theSeymour source water is extremely low. Hence, the backgroundmetal levels were notexpected to “mask” the effects of pH and alkalinity adjustmentson copper and leadleaching. Furthermore, there could be a problem if the drinkingwater of North Delta,Newton, and South Surrey was supplied from water sources that have differentwaterchemistry. Having anticipated this problem, the drinkingwater of all the study areaswere supplied exclusively from the Seymour reservoir.The type of solder and the lengths and sizes of pipes used in the plumbingsystem were variable in each house. Even though the old type of lead/tinsolder hasbeen prohibited in municipal plumbing codes for many years, theyare still sometimesused illegally due to the ease of handling lead/tin solder. Housesthat have more of thelead/tin type solder, rather than the newer tin/antimony solder, in the houseplumbingwill probably also have higher lead levels at the tap. Unfortunately, thereis no way toquantify the amount of lead/tin solder present inside the houseor in the distributionsystem. The technique of sampling a number of houses instead of afew houses solvesthis problem. Sampling ensures that some houses sampled havemore lead/tin solderand some houses have less. The average of the lead concentrationof the samplinggroup will then reflect an averaged amount of lead/tinsolder.The temperature of the supply water in the variousareas of the distributionsystem could also affect the rate of metalleaching from the pipes, solders, and faucets.However, considering the close proximity of the varioussample locations, and the factthat solar heating of the supply water is retarded by the soilcover on top of the entiredistribution system, the temperature variation across thesampling sites was notexpected to significantly affect metal levels.3.1.2 Uncontrollable Factors44There is one very important factor of copper and leadleaching that iscompletely uncontrollable in a field study. This factor is the standingtime of the watersamples. This study examined standing cold water, flushed coldwater, and flushed hotwater; however, the focus of the study was on the standing cold watersamples becausethey are likely to have the highest levels of copperand lead. Standing samples are alsoimportant because they are used to determine the compliance of tracecopper and leadlevels in USEPA regulations.In a laboratory or coupon study, the standing time of water can be controlled.In a field study, however, the standing time cannot be controlledprecisely. Thestanding time is the time between the last use of water in the housebefore theoccupants go to sleep and the time the sample is collected first thingin the morning,i.e., before any water is used in the morning. There are at least fourpotentialproblems with the control of standing time. One, not everybody participatingin thisstudy would sleep the same amount of time. Two, oneor more of the residents mayhave to use the bathroom sometime during the night. When thishappens, some of thestanding water is flushed out of the plumbing system. This has the effect of decreasingmetal levels in the house plumbing, and hence, decreasing the effectivestanding time.Three, when the occupants wake up first thing in themorning, they may use thebathroom or the kitchen before they remember to dothe sampling. This is probablythe most serious problem, and this scenario is more likely tohappen if there are manyoccupants living in the house. The larger thehousehold, the higher is the chance ofhaving at least one of the occupants forgetting aboutthe sampling study. Four, somehouses have one or more leaky water taps. Thefaucets may be getting too old orsomeone may forget to shut-tight a tap, which causes continuous slowleakage of thestanding water. Depending on the degree of leakage, the effectivestanding time coulddrop insignificantly or a great deal.45The participants of this research project were instructed about these potentialproblems. But human behaviour is hard to predict and impossible to control.Therefore, we can predict that the samples come from a wide variation of standingtime. Since we wanted our standing cold water samples to have remained in the pipesfor at least six or seven hours, samples that had an effective standing time less thanthat will significantly bias the copper and lead data toward the low side. We alsoexpected that a lot of the scatter in the copper and lead data could be attributed to thestanding time problem.There are two possible ways to get around this problem. Both are highlyimperfect. Assuming that most people sleep more than six hours and that the waterleakage problem is insignificant, we can ask the participants if they used any waterbetween the time they went to sleep and the time they collected the samples. Someparticipants who know that they have made a mistake may not be willing to admit it.Also, the person who is asked this question may not know whether or not anotherperson in the household has used any water before collecting the water samples. Thismethod of determining if the standing sample from a particular house has been flushedor not is seriously flawed.The second way to solve this problem is to identify the flushed standing samplesby using statistics. This approach, while imperfect, is nevertheless more scientific thenthe first approach described earlier. This research project adopted this approach(referto section 4).3.2 Arrangements with Governmental AgenciesBefore the samples were collected, various levels of government wereconsulted. In particular, the GVRD, the Municipality of Delta, and the City of Surreygave this researcher permission to solicit houses, within their jurisdictions, forcollecting tap water samples. Assistance was also offered by these government46agencies to provide letters of recommendations, in the event that thehome owners haddoubts about the legality of this project.From the archives of these government agencies,it was possible to locate theaddresses of about two hundred houses that werenewer than ten years old, plus a fewmore houses that were between ten to fifteen years old. These houses,locatedthroughout the five sample areas (one in North Delta, two inNewton, and two in SouthSurrey), were the target houses for collecting water samples. TheGVR]) assisted inmailing out letters to these homeowners, informingthem that this researcher mightcome to their house to ask for their cooperation in this studyand to provide tap watersamples from their kitchens.3.3 Bottle PreparationAll the bottles used in this study were brand new plastic bottlesthat weredelivered to the UBC laboratory sealed in the original packaging.This was the firststep toward minimizing the potential of metal contaminationof the bottles due to airparticulates, solvents, or chemicals in the lab. Becausethe degree of existingcontamination due to copper or lead particles on the inside surfaceof the bottles wasunknown, a bench test was performed. Six different kindsof bottle preparationtechniques were tested. Three bottles were preparedfor each bottle treatment. Eachcopper measurement was repeated five times, and lead threetimes.The first set of bottles (NT) were the controls.Only deionized distilled waterwas poured into the bottles, and no pretreatment wasgiven. The second set of bottles(AiD) were pretreated by acid washing with 10%nitric acid for one minute and thenrinsing with deionized distilled water. The third set(A5D) was similar to the secondset, except that the acid wash time was five minutes;the fourth set(A9OD) had a ninetyminute acid wash time; the fifth (AOD) had the acid stayin the bottle for overnight.The sixth set of bottles (SDAOD) were pretreated bythe method recommended by theStandard Methods (17th edition). The bottles weresoaped, rinsed by deionized47distilled water, acid washed overnight (about eight hours), and then rinsed again bydeionized distilled water. This last method is extremely labor intensive. Towash allthe bottles needed for this study (about 600) by this method would have requiredapproximately 2 months for one person. If any of the first five shorter methodsprovedto be equally as effective as the method recommended bythe Standard Methods, thenthat method was adopted by this study.Table 4 Results of Bottle Treatment StudyTreatment Method Average Copper (ppm) Average Lead (ppb)NT 0.023 0AiD 0.000 0A5D*0.016 0A9OD 0.016 0AOD 0.016 0SDAOD 0.000 0The results of the bottle treatment study (Table 4) shows that the deionizeddistilled water used for this study was free of lead. Also, the zerolevel of leaddetection, regardless of the treatment method employed, means that the bottles had nolead contamination.The detected levels of copper were extremely low compared with the range ofcopper concentrations that one would expect to find in the collectedtap water samples.Therefore, it did not really matter which treatment method waschosen. The methodrecommended by Standard Methods worked the best; but, forthe purposes of thisstudy, is inconsequential. All the treatment methods resultedin a lower level of coppercontamination. A decision could have been made not to treat anyof the bottles.However, to avoid the unlikely possibility that excessive amounts of lead or copperparticulates might be found in some of the bottles due to chance,it was decided to acidwash all the bottles used in this study for five minutes, followed byrinsing withdeionized distilled water.483.4 Sampling PackageThe sampling packages that were to be delivered to the homeowners wereprepared beforehand, at the UBC lab. The package was a large ziplock bag that had,sealed within it, a 1L bottle, two 125 mL bottles, a copy of the letter that the GVRDhelped to send to the homeowners beforehand, and two instruction sheets. The oneliter bottle was marked with the number “one”. This bottle was used to collect thestanding cold water sample. The metal levels in this sample came primarily from theinterior home plumbings, soldered joints, and faucets (AWWARF, 1990). One literwas chosen to be the size of the standing sample because the one liter sample size wasrecommended by the USEPA in the May 1991 Lead and Copper National InterimPrimary Drinking Water Regulations.One of the two 125 mL bottles was marked with the number “two” and wasused to collect the flushed cold water sample. The other 125 mL bottle was markedwith the number “three” and was used to collect the flushed hot water sample. Allthree bottles were empty and capped. The flushed samples measured the metal levelsin the water distribution system. The sample volume for the flushed samples was notan important issue, assuming that the kitchen tap had been adequately flushedbeforethe sample collection. Since the metal levels in the distribution system have lessvariation compared to those inside the house plumbings (AWWARF, 1990), collectinga large or a small does not make much of a difference. A relatively smallsample sizewas chosen, making storage and transportation more convenient.The first page of the instruction sheets was a concise written explanationof thesampling procedures (Appendix B). The second page was a self explanatorycartoonthat helped the homeowner to visualize the sampling procedures, in casetheinstructions on the first page was unclear to them (Appendix C).3.5 Sampling Procedures49Two identical ziplock bag packages were delivered toeach house thatparticipated in this study. One was marked bag “A”and the other bag “B”.On the first morning after being contacted by the researcher,the homeownerwas to take the three bottles out of bag “A” andplace them, in order of theirnumbering, on the counter beside the kitchen sink.First thing in the morning, beforeany water was used in the house, the homeowner was to turn on thecold water tap andfill bottle number 1. The cold water tap was left on,until the water became cold,before filling bottle number 2. When the water from the cold water tapturned cold, itwas a sign that water came from the distributionpipes outside the house plumbing.After the flushed cold water sample was collected, thecold water tap was turned off,and the hot water tap was turned on. Bottle number 3 was filled afterthe water becamehot. All three bottles were to be capped tightly, and sealed in theziplock bag forpickup. On the next morning, the second morning after being contacted,the sameprocedure was repeated for bag “B”.3.6 Participation of HomeownersThe participation and cooperation of the homeownersin this study wasabsolutely crucial to its success. First, the homeowner had to bewilling to participatein the study. Second, the homeowner had to followthe sampling procedures correctly.In order to obtain a database of sufficient size, there was atarget of enlisting 35houses in each of North Delta, Newton, and South Surreyto participate in the study.Having 35 data points for each area allowed the researcher to dostatistical comparisonsbetween the study areas with an acceptable degreeof confidence (see Section 4), whilekeeping the sampling and lab testing efforts at a manageablelevel. Quota samplingwas employed for this study. Targeted homeowners,who had been contactedpreviously by mail, were contacted personally by door knocking.They were asked ifthey were willing to participate in the study. Thosewho were willing were given thesampling packages and were instructed about the samplingprocedures. The50homeowners were also asked a series of questions regarding their houses (Appendix D).For a variety of reasons, a portion of the homeowners who were contacted did not wantto participate in the study. Enough homeowners on the list of targeted addresses wereasked until 35 homeowners in each study area agreed to participate.By far the biggest error in this study was due to the homeowners not followingthe sampling instructions. The most common error was the use of water in the houseprior to sampling. Other less common, but equally serious, errors included: notletting the cold water flush enough when collecting bottle number 2, not letting the hotwater get hot before collecting bottle number 3, and mixing the order of collecting thesamples.3.7 Bottle Pick-upThe sample bottles, sealed in the ziplock bags “A” and “B”, were left outsidethe houses after the second morning of sampling. The bottles were picked up in theafternoon and brought back to the UBC lab for analysis. Due to the large number ofsamples that had to be delivered and picked up, the sampling process took place over athree week period between September 20th to October 10th, 1994. Each bottle wastracked so that the bottles from each house could be picked up right after the secondmorning of sampling.3.8 Lab TestingThe pH, alkalinity, and chlorine residuals in tap water deteriorated quickly. Todetermine the values of these parameters for the houses, the standing sample from bag“B” of each house was measured in the lab on the same day that it is picked up. Thecopper and lead metals did not deteriorate as quickly as the other measured parameters.Therefore, the measurement on copper and lead levels was deferred for a few weeksafter all the bottles were picked up from the houses. Copper and lead were measuredfor all the bottles from bags marked “A” and bags marked “B”.51After the bottles were brought into the UBC lab, each bottle was coded sothatone could identify later on where each bottle came from. ThepH, alkalinity, andchlorine residuals were measured before acid was added to preserve thewater samples.Preservation with acid was necessary for measuring leadand copper. Copper ions hada tendency to adsorb onto the bottle surface and, thus, theaddition of acid de-adsorbedthe copper ions. For lead analysis, the acid added served as a matrixmodifier thatreduced interferences (Standard Methods, 17th edition). Nitric acid at2.5% wasselected as the preserving acid. Normally, a concentrationof 0.3% would have beenused. To make the copper and lead testing feasible in terms oftime, the samples werenot digested before measuring the metal levels. A higher concentrationof acid wasadded to the samples in order to compensate for this deficiency.This did not mean thattotal metal was recoverable without going through the digestionstep; however, at leastall of the dissolved metals could be recovered.In essence, this procedure measured and analyzed the dissolvedcopper and lead.This study focused on the dissolved metals, since they accountfor most of the totalmetals in tap water (AWWARF, 1990). Also, dissolved copper and lead havemorepotential harmful effects on humans than the particulate counterparts,since the humanbody can uptake the dissolved forms much more easily.In the past, studies that werestrictly concerned with dissolved lead species filtered thesamples through a memberfilter of 0.4 or 0.45mm pore size. This procedure was noteffective because thedissolved lead tended to adsorb onto and was retained by thefilter material (Schock,1983).3.8.1 InstrumentsThe pH of the samples were measured by a Fisher glass electrodeand aBechman pH meter. Because tap water has a low ionic strength,an orion buffersolution was added to the water to give more stablereadings. The pH meter wascalibrated by pH buffer solutions. Total alkalinitywas determined by the titration52method (Standard Methods, 17th edition).The free and total chlorine levels weremeasured by the DPD Colorimetric Method.Copper was analyzed by a Thermo JarrelAsh Atomic AbsorptionSpectrophotometer, using lean acetylenefuel. Samples were aspirated directly outofthe sample bottles in order to avoid contaminationin the measurement stage. Standardsolutions, in the range of copper concentrationswe expected to find, were made.2.5%nitric acid was added to the standards in order tomake them equivalent to the samples.The AAS instrument was recalibrated againstthese standards for every ten samplesmeasured. All measurements wererepeated five times. Only the averages oftheserepeated measures were reported.Lead was determined by atomicabsorption graphite furnace. Leadsampleswere poured into pre-rinsed sampling cupsbefore the graphite furnaceinstrument testedthem with the autoanalyzer. The qualitycontrol procedure was the same forlead as forcopper.3.8.2 Testing SchemeFive percent of the standing samples werecross-checked by the GVRD waterquality lab. Interlaboratory testing couldhelp to determine whether or not theequipment at the UBC lab wasworking properly and was in goodcondition.Because there were roughly six hundredsamples that had to be measuredforcopper and lead, it took twomonths to complete the testingfor all the samples. Duringthis testing period, the workingconditions of the instrumentschanged according to theenvironmental conditions of thelab, on any given day. The samesample that wasmeasured on one day might have adifferent reading thanon another day. Thevariations of the instruments could havebeen purely random or mayhave hadsystematic trends. Besides thetechnique of constant recalibration,the instrument effectwas taken care of by selectingthe sample to be tested in a randomsequence.3.8.3 Data Recording53There were inevitably some errors in the process of reading, recording, andtranscribing data. The validity of the databank would be in jeopardy if the rate of thesekinds of error was high. There should not have been any reading and recording errorsfor the lead and copper measurements because the test instruments automatically printedout the data on paper. To assess the rate of transcription error made in the process oftranslating the data into the computer databank, the data in the databank was crossreferenced with the data tapes. The rate of error that could not be corrected was foundto be 0.1%.544. RESULTS AN]) DISCUSSION4.1 Siimmry of Returned BottlesOf the 105 houses that agreed to provide samples for this study, 92 ofthemreturned bottles. Thirteen houses did not return any bottles. The 88% returnrate washigh for a study of this type. Participants of this study were informed aboutthepotential dangers of consuming water with excessively high levelsof copper and lead.They were told that their participation was vital to the success of identifyingthe safetyof the water supply in their area. Those who wanted to know aboutthe outcome of thisstudy received the conclusions of this study after it wascompleted. Personalcommunication with the house owners, and linking this study totheir health and safety,were the main ingredients of obtaining a substantial number ofsamples. A summary ofthe information about the returned bottles is provided in Table 5.Table 5 Suinmar Information of Returned BottlesArea code #pH** alk**freechlorine* chloramine*samplingDelta d 19 n n .01 .01 randomNewton 1 n 31 y .28 .05 clusterNewton2 y 10 y .91 .10 clusterS.Surreyl s 19 y .29.27 clusterS.Surrey2 w 13 y .33.44 cluster*GVRD data (avg for Sept 20- Oct 10)**The presence of pH or alkalinity adjustmentsThe samples from the 92 houses were collected from thefive different samplingareas. The North Delta area had 19 bottle returns.As expected, free chlorine andchloramine residuals were almost zero becauseall of the chlorines had been oxidized inthe system by the time the supply water from Seymourtravelled to North Delta. NorthDelta was given a shortform “d” in sample identification andstatistical analysis. Thefirst of the two sample areas of Newton,“n”, registered levels of chlorine residualsmuch higher than the Delta area. Comparing thesecond of the two sample areas ofNewton, “y”, to “n”, we clearly see that the free chlorineresiduals of “y” was much55higher than “n”. This was expected since “y” was closer geographically to the watersupply feedpoint than “n”.Chioramine was added to the two South Surrey study areas. The first of thetwo sample areas of South Surrey, “s”, had higher levels of free chlorine andchloramine residuals than “n” or “y”. The second of the two sample areas of SouthSurrey, “w”, had chloramine levels that were even higher than “s” because “w” wascloser to the feedpoint than”. pH and alkalinity tests on the samples from the fivestudy areas showed the addition of chemicals had substantially raised the pH andalkalinity of the water in Newton and South Surrey. In all, measurements of the fourwater quality parameters: pH, alkalinity, free chlorine residuals, and chloramineresiduals, agreed well with the expected levels.The results of this statistical analysis will be valid only for the range of thewater quality parameters detected in the samples of this study. The conclusions of thisstudy do not apply to ranges of water quality parameters that are above or below thosein the study samples. Conclusions based on statistical statements are also specific to thestudy areas. Lead and copper levels in the other areas of the Greater Vancouver WaterDistrict have to be examined separately, because the water quality characteristics andthe water distribution systems might be significantly different from those in this study.However, the same methods and analysis techniques can be applied when studyingthese other regions of Vancouver.Although the chlorine levels of the samples were measured in the UBC lab,none of these measurements were used in the data analysis. All the measuredchlorinevalues were very low in comparison with the GVRD chlorine data.There was a time gap of several hours between the time the homeowners tookthe water samples and the time the chlorine levels were measured in the UBC lab.Because chlorine is a very volatile gas, much of it escapes from the bottles even if thebottles are tightly capped. For any chlorine measurement to be valid, the analysis must56be carried out concurrent with the time of the actual sampling. Fortunately, the GVRDregularly monitors various water quality parameters from many placesin the GreaterVancouver. There was at least one of these GVRD monitoringsites in each of the fivestudy areas. The reported chlorine residuals in this study were the time averaged (Sept20 to Oct 10) values of all the monitoring stations in each area.4.2 Data Cross-check by GVRDThe accuracy of the measurement data from the UBC lab was cross-checkedbythe GVRD lab (Appendix E). Because the standing samples constitutedthe focus ofthis research, only standing samples were selected for cross-checking interlaboratoryaccuracy. Of the 177 standing samples from bags “A” and bags “B”, fifteen percent,or 28, of the samples were randomly selected and sent to the GVRDlab. Copper andlead were measured by the GVRD lab using the same principles of measurement asthose in the UBC lab.The GVRD and the UBC lab data for the 28 samples revealed some differences.While the copper values for the two data sets were remarkably similar, 9 out of 28 ofthe lead measurements appeared to be significantly different. If theUBC and theGVRD data sets were significantly different from each other, theneither the UBC orthe GVRD lab or both labs had a problem with measurement accuracy.TheKolmogorov-Smirnov test was employed to test for a statistical differencebetween thetwo thta sets. The UBC data set was randomlydrawn from the pool of standingsamples, which was the population of values measured by theUBC instrument. Thenull hypothesis of the test was that the GVRD data set belongedto the same populationof values as the UBC data set. The rejection of thenull hypothesis meant that theGVRD data set differed significantly from the UBC data set.If this were the case,more interlaboratory testing had to be conducted to identify the sourcesof themeasurement inaccuracies.57A two sided test was performed, with n=28, c=0.O5, and Dcritical=0.36.For copper, d=0.07. Because d < <Dcrit, the null hypothesis was retained. For lead,d=0.33. The null hypothesis was also retained in this case. The resulting conclusionwas that the GVRD data set validated the accuracy of the UBC data set.4.3 Reporting of DataVarious methods of reporting laboratory data have been employed by differentgroups of people. For this study, since most of the data was around zero, the practiceof reporting and using only data that was above the LOD (level of detection) wouldhave resulted in too little information left for statistical analysis. On the other hand,the suggestion that reporting all the data as it was, including the negative values(ASTM, 1984), might have led to negative bias in the data set.For computational purposes, this study reported all negative values as zero.Also, all values below the LOD were reported and used in the statistical analyses.The sigma, standard deviation (precision), for copper was 0.Olppm and sigmafor lead was 1 ppb. So the LOD for copper was 0.03 ppm and for lead was 3 ppb.The LOQ for copper was 0. l0ppm and for lead was 10 ppb (Table 6).Table 6 Detection Limits of MeasurementsMetal a LOD LOQcopper (ppm) .01 .03 .10lead (ppb) 1 3 104.4 Data Used for Statistical AnalysisA choice had to be made as to how to analyze the raw data (AppendixF). Rawdata consisted of measurement data of the samples in bags “A”or of the samples inbags “B”. Previous studies that characterized the metal levels atthe tap used theaverage values of the raw data (Karalekas, 1983). Although this has beenthe mostwidely employed method of comparing metal levels, this study had used anothermeasure of comparison. Instead of averaging the “A” and “B” values, themaximum58value of the two was chosen for performing subsequent analyses. The rationale forusing this method was that the metal level for the sample with the highest standing timeshould be known. Due to the inevitability of human error, it was anticipated thatcertain homeowners collected the samples improperly for one of the two sets of samplesthey had to collect. In such cases, choosing the sample with the maximum metal levelwas equivalent to throwing out bad data. In a situation where both “A” and “B”samples had been collected properly, one of the two samples was likely to have alonger standing time than the other. Here, picking out the maximum value improvedthe chance that this databank reflected the highest metal levels at the tap during thesampling period. If samples from both mornings had been collected improperly, itwould not have mattered if the maximum method or if the averaging method had beenused.The method of using the maximum value had implications for the statisticalanalyzes. If, other than “A” and “B” values, many more sampling repetitions werepresent, then the metal level frequency distribution of the maximum values should haveapproached a double exponential disthbution, as the number of repetitions approachinfinity (Kinnison, 1985). In other words, it would be expected that the copper andlead histograms for the various study areas would follow the double exponentialdistributions instead of the normal or log-normal distributions. This would haveadded complications to the statistical analyses, because most of the available statisticalmethods have been developed for normal distributions. For this study, the number ofsampling repetitions was two. Since this number did not even come close toapproaching infinity, the problem of having to deal with double exponentialdistributions was only a theoretical concern. Examining the lead standing samplehistograms (data combined for study areas “s”, “n”, “w”, and “y”), it can be seen thatthe irregular distribution of the maximum value histogram is not substantially differentfrom the shape of the average value histogram (Figure 1).59Figure 1 Comparison of Data TypesIt can also be noted that the maximum value histogram has less low value datapoints and more of the higher value data points. This result was predicted bythetheory of the maximum value method.4.5 Non-Parametric TestingThe frequency distribution of lead of all the study areas exhibited highlyirregular patterns. Although some were close to being log-normal, most of themwerehighly skewed to the left and had bi- or multi-modal tendencies. These kindsofdistributions were not amenable to the common statistical teststhat were developed fornormal or log-normal distributions. To get around thisproblem, it was common totransform the existing distribution into a normal orlog-normal distribution and then toperform the usual analyzes. This researcher tried to applythe square root, logarithmic,and reciprocal transformations. Using SYSTAT,testing was carried out to see if thedistributions obtained after the transformations deviatedsignificantly from the normaldistribution, at a = 0.05. Using the Lilliefor Test, it wasfound that the transformationprocedures were not successful in converting our distributionsinto ones that wereinsignificantly different from the normal distributions(SYSTAT, 1992). As a lastmeans, non-parametric statistical tests have been used to analyzethe data.standing sample35_____________30A raw2520raw15• avg10—o——— max01.0 1.0 1.0 (.0C%J - C\J - C\i I— — c’J c\JLead (ppb)60Non-parametric testing is not only particularly suited to ill-behaveddistributions, but it is also a powerful method. In fact, the asymptotic relativeefficiency (A.R.E.), a measure of power, of a non-parametric test to its parametriccounterpart, is greater than 90% for most of the tests used in this study. In order toincrease substantially the power of the test, the number of samples collectedhas to beincreased. For a given number of samples, there is a very little loss of power by usingthe non-parametric rather than the parametric method.4.6 Other Statistical ProblemsThe statistical significance of this study has to be qualified by certaindisclaimers. First, the so-called spatially random or cluster sampling techniquesemployed in this study were not really random or cluster in the true statisticalsense. Ina truly random sampling, a population would be well defmed.A random sample wouldthen be randomly picked out of the population group. Similarly, in a clustersampling,there would be clearly identifiable sub-populations. One of the sub-populationgroupwould be picked from the whole population and all of the units in the selected sub-population group would be sampled.In this study, the governing factors of copper and leadlevels were chemicalparameters that could not “a priori” be known with certainty.Therefore, it was notpossible to clearly identify population or sub-populations groups.The samplingtechniques employed in this study were the best that could havebeen employed underthe circumstances. Engineering judgment, based ongeneral knowledge of waterchemistry in a distribution system, was used to decide, roughly, the samplepopulationand sub-population boundaries. This type of samplingbelongs to a type of nonprobability sampling called the “judgment or purposive selection”(Cochran, 1977).The quota sampling technique which was employed, inorder to obtain aminimum number of samples for analyses, violated the principleof true randomselection.61The 13 households that did not return bottlespresented a problem with non-response bias. Non-response bias wouldhave occured if the non-returnscame fromhouses that had very clean orvery contaminated water, in relation to thehouses thatreturned bottles. For the housesthat had ultra-clean water, the houseownersmay nothave bothered to return bottles, becausethey did not feel this study couldbenefit themin any way. For the housesthat had very contaminated water,the houseowners mayhave been embarrassed to disclosethe fact that they had a water qualityproblem. Non-response may have contributed to ahigh or a low bias in our database.Although modern statistical researchis progressively finding bettertechniques todeal with this type of non-idealsampling scheme, no one methodcould have accountedfor all of these statistical problemsat the same time. However, justbecause theseproblems existed did not meanthat our analyses were completelyinvalid. The resultswere still useful; but one must beextra careful in interpreting the conclusionsdrawnfrom these analyses.4.7 Identifying Flushed Standing SamplesTo obtain a valid characterizationof the copper and lead levels inthe studyareas, the researcher had to identifythe flushed standing samplesand to reject themfrom the database before performing statisticalanalyzes. Including these resultsin theanalyzes would have grossly biasedthe sample distributions tothe low side.Flushed samples were treated asoutliers that had very low valuesin comparisonto the rest of the population group.Normally, theseoutliers could have been identifiedby outlier identification proceduressuch as the Rosner’ sTest for Detecting Up to kOutliers or the Dixon Test (Gilbert,1987). Alternatively, ageneral rule in statistics isthat, given a data set with at least 10 values,a value that liesoutside 4a from the meanof the data set can be treated asan outlier (Sachs, 1982).The problem with using theseconventional methods is that they assumethat the number of outliers issmall comparedto the number of datain the data set.62In a field study, such as this one, the percentage of flushed samples was notsmall. Hence, one had to base a rejection criterion on somethingother than the set ofstanding samples. The flushed cold water samples could be used as a baselinetocompare with the standing sample values. A flushed standingsample should have hada lead value that was very similar to the value of aflushed cold water sample.Therefore, as long as one could establish the criterion of similarity between theflushedcold water samples and the flushed standing samples, there existedan objective way toassess which data point should have been rejected. Due to adifferent water chemistry,flushed cold water samples may have had higher copper concentrationsthan 1 literstanding samples. To avoid a possible conflict between arejection criterion based onlead and one based on copper, one metal was finally chosen. Sincethe emphasis ofthis study was to clarify the extent of lead problems in the studyareas, lead was chosento be the basis for establishing a rejection criterion.Using the flushed cold water samples as the baseline, the simplest way toidentify outliers was to take the difference between the flushed cold water valueand thestanding sample value for each house. A distribution of the differencesbetween thevalues could have been plotted, and the outliers could have beenidentified by theRosner or the Dixon Tests. However,this method would be invalid if there were alarge number of true outliers.The rejection procedure that was used was slightlydifferent. This rejectionprocedure could be illustrated by looking at thehistogram of the flushed cold waterlead values for the Newton and South Surrey areas(Figure 2).63Figure 2 Rejection ProcedureAbout 90% of the flushed cold water values were less than 2.5 ppb, although therewere a few values as high as 5 ppb. Suppose one wanted to know if a particularstanding sample with a lead value of 20 ppb had been flushed or not. Knowing that 20ppb was higher than the highest flushed cold water level found in the study areas, onecould confidently say that this particular standing sample had not been flushed. On theother hand, it would have been concluded that the standing sample had been flushed ifit measured 0 ppb. However, a standing sample with a value between zero to five ppbwould make it harder to decide. Therefore, one needed to fmd a cut-off value belowwhich all standing samples would be classified as flushed and, above which, classifiedas unflushed. If the cut-off was chosen to be at 5 ppb, the chances werethat onewould reject some standing samples that legitimately had low standing lead levels. Asone lowered this cut-off, the opposite problem was encountered, of not rejecting someof the flushed standing samples. Committing the error of not rejecting flushedstanding samples would be worse than rejecting legitimate standing samples, becausethe presence of a low outlier in the database would cause more bias to the databasethanthe absence of a good piece of data. Therefore, the strategy was to choose the cut-offas high as possible, while making sure that enough data remained to beanalyzedstatistically.cold-flushed(forNeMon &Souih Surrey)40_30C200‘ 10Rejectsamples less than 1 ppb0 1 2 3 4lead (ppb)564This rejection procedure was performed oncefor the samples that came from theNewton and South Surrey regions. The procedure was repeated separatelyfor thesamples that came from the Delta region, becauseDelta did not receive any corrosiontreatment. Failing to analyze the two groups of dataseparately would have resulted inrejecting too many unflushed standing samples that came fromNewton and SouthSurrey, and not rejecting enough of the flushed standing samplesfrom Delta.4.7.1 Rejected Samples for n,y,s,wThe flushed cold water samples were, just like the standingsamples, sampledfrom a larger population group. Therefore, the quantificationof the cut-off point wasitself subject to statistical uncertainties. The cut-off point could bequantified as aquantile of the set of flushed cold water samples. If one chose 5 ppb asthe cut-offpoint, then the cut-off was at 100 quantile of the flushed cold watersample set. But, inthe larger population group, from which the samples were taken,the real 100 quantilewas most probably higher than 5 ppb. Stating that 5 ppb was atthe 100th quantile ofthe population group should be qualified by a certainlevel of confidence.The statistical test that was perfectly suited to this typeof task was a non-parametric test called the “quantile test”. In theNewton and South Surrey areas, therewere 48 standing samples that had to be tested for flushing.This researcher performeda upper-tail quantile test, withn=48, a=0. 10, and choose the 65th quantile.Therewere 18 standing sample data rejected, and 1 ppb wasthe cut-off point. Choosing ahigher quantile to be the cut-off point would have resulted inrejecting almost all of thestanding samples.4.7.2 Rejected Samples for dThe rejection scheme for the standing samplesof Delta had to be the same asthat for Newton and South Surrey, i.e., ct=0.10,and p=O.65. Standardizing therejection scheme ensured that the unrejected data fromthe different areas could becompared on an equal basis. There were 12 standingsamples that had to be tested.65Given c=0. 10, and p =0.65, and n= 12, 5 ppb was calculated to be the cut-off point,and three standing samples were rejected.4.8 Simple TestsAfter the rejection procedures, there were 14 unrejected standing samples fromSouth Surrey, 16 from Newton, and 9 from Delta (Table 7).Table 7 Summary of Simple Statistical EvaluationSouth Surrey Newton Deltacopper lead copper lead copper leadmean .129 12.3 .174 10.8 1.439 16.3a .031 7.9 .127 7.9 0.736 6.2- .008 2.1 .032 2.0 .245 2.1c.v. .24 .6 .726 .7 .512 .4median .12 10.0 .135 6.5 1.780 14.0Note: Copper measured in ppm, Lead in ppbThe values had significant figures equal to the detection limit plus 1 decimalplace. For copper, the detection limit was 0.01 ppm, and the significant figures weretherefore given in units of 0.001 ppm. For lead, the detection limit was 1 ppb, and thesignificant figures were in units of 0.1 ppb. The coefficient of variations, a measure ofthe standard deviation divided by the mean, were under 1 for all the categories ofcopper and lead measurements.A coefficient of variation that is under 1 is considered to be low in comparisonwith other studies of copper and lead levels in tap water (AWWARF, 1990). A lowvariability of copper and lead measurements indicates that the sampling areas wereindeed homogeneous. Therefore, we can be sure that whatever statistical statements wemake from our analysis will not be distorted by having taken samples frominhomogeneous sampling areas. Furthermore, a low coefficient of variation excludesthe possibility that there were serious procedural or laboratory errors. In short, a lowvariability in copper and lead measurements confirms the soundness of the entiresampling strategy employed in this study.66One can also do a cursory examination for the skewness ofour data sets. Meanvalues should theoretically be higher than median values for apositively skeweddistribution (Lapin, 1983). The more skewness there is, thefarther is the distancebetween the mean and the median.Based on this reasoning, the set of lead measurements for Newton isprobablyhighly skewed in the positive direction. This is an indication thatthe samplingdistribution for this set of data, even after having rejected theflushed standing samples,is still not transformable into a normal or log-normal distribution.By the samereasoning, the other data sets are probably transformable. Because transformeddatacannot theoretically be compared with untransformed data, non-parametricstatisticalmethods should be used to analyze the data after the flushedsamples have beenrejected.Using the mean value as the measure of central tendency, it can beseen thatboth the mean copper and lead values are lower in Newton and South Surreythan inthe control Delta area (Table 8).Table 8 Percentage Reduction from Delta (mean values)lead (ppb) copper (ppm)to Newton 34% 88%to South Surrey 25%91 %The copper values are reduced dramatically, while the lead values are marginallyreduced.4.9 Kruskal-Wallis H-Test (Copper)The result of the comparison between the mean coppervalues strongly suggestsa statistically significant difference betweenthe copper levels of the different studyareas. To investigate this further, the Kruskal-WallisH-Test was conducted. Forn=39, x=O.05, the null hypothesis is that the samples fromNewton, South Surrey,and Delta all belong to the same population. Thisnull hypothesis is rejected, which67means that at least one of the three sample sets is significantly different fromthe othersets. Using the Kruskal-Wallis multiple comparisonsmethod, it was discovered thatthe copper levels in Delta were significantly higher thanSouth Surrey and Newton.Copper levels in South Surrey were statistically equivalent tothose in Newton.4.10 Confidence Intervals for the Difference Between 2MeansThe difference between the expected copperlevel of Delta and Newton or Deltaand South Surrey is subject to statistical uncertainties.The 95% confidence interval ofthe differences can be calculated by a non-parametricmethod. The differences betweenDelta and the two test areas are similar. The lower boundshover around 1 ppm andthe upper bounds around 1.8 ppm (Table 9).Table 9 Difference Between 2 Meanslower bound (ppm) upper bound (ppm)Delta minus South Surrey 1.011.82Delta minus Newton .92 1.804.11 Median TestThe median test tests the null hypothesis that the samplesfrom Delta, Newton,and South Surrey all come from population groups that havethe same median. Forn=39, x=0.05, the null hypothesis is rejected. Thisresult corroborates with theresults from the Kruskal-Wallis H-Test.A similar analysis on lead samples did notreveal any statistically significantdifferences between the study areas.4.12 Test for Equal VarianceThis non-parametric test tests if the variancesof the study areas are equal. Forn=39, a =0.05, the test shows that the variancesof the sample sets from South Surreyand Newton were less than samples from Delta(Table 10).68Table 10 Test for Equal Variancecopper leadSouth Surrey vs Delta less variance Ho okNewton vs Delta less variance Ho okThis piece of evidence further proves that pH and alkalinitytreatment in the SouthSurrey and Newton areas are changing the copper levels at the tap.In contrast, thevariances for the lead samples from the study areas are statistically equivalent.4.13 Compliance with RegulationsThe most up-to-date copper and lead guidelines from the USEPA was issuedin1991. The guidelines stipulated that 90% of the samples in a monitoring programshould have less than 15 ppb of lead and 1.3 ppm of copper. For the samplesin thisstudy, this researcher analyzed the effectiveness of pH and alkalinity adjustmentsinreducing copper and lead levels in terms of regulation compliance (Table11).Table 11 Regulation Compliance (as per EPA guidelines,1991)NewtonlSouth Surrey Deltacopper lead copper leadCold Standing ok no no noCold Flushed ok ok oknoHot Flushed ok ok ok noCopper is out of compliance in the control area, but is incompliance in areas with pHand alkalinity adjustments. Lead in the standing samplesis out of compliance bothbefore and after pH and alkalinity adjustments. The flushedcold and hot lead samplesare also out of compliance. However,these comply after the adjustments are in place.It was concluded that pH and alkalinity adjustments do notsignificantly help lead tocomply with the regulations.The copper and lead guidelines in Canada are muchmore lenient than theUSEPA guidelines. The USEPA guidelines are chosen asthe compliance benchmarkbecause most of the Canadian Drinking Water Regulationsfollow the lead of the69USEPA. Secondly, the present Canadian lead and copper regulations do not havewellspecified sampling protocols (Environment Canada, 1981).4.14 Plot of Lead vs. Alk and pHFrom the previous statistical analysis, it is abundantly clear that pH andalkalinity adjustments work to reduce copper levels at the tap. The effects onlead ismuch less clear. In order to investigate the relationships between pH and alkalinitywith lead, this researcher started the data exploration (Tukey, 1977) by making a3-Dplot of pH, alkalinity and lead (Figure 3).Figure 3 3D Plot of Lead, Alkalinity, and pHFrom this graph, we see that there is a strong correlation between pH andalkalinity. However, there does not seem to be a pattern of increaseor decrease of leadwith rising pH or rising alkalinity.3020704.14.1 Spearman Rank CorrelationsThe extent of the correlation betweenpH and alkalinity can be quantifiedbycalculating the Spearman Rank Correlationcoefficient between the two variables.Thecalculation indicates that there is a 0.917 correlation between the pH and the alkalinity,with 1.000 being perfect correlation.4.14.2 Regression of Alk and pHBasic water chemistry tells us that pH should be relatedto the amount ofalkalinity in the water. Therefore, it is not surprisingto find confirmation of thisprediction from the Spearman Rank CorrelationTest. If the regression test shows aequally strong relationship between the two parameters,then we can simplify thesubsequent statistical analyses by eliminating one variable.A three dimensionalproblem can then be reduced to a two dimensional problem: copperor lead versus pH.Because of the amount of work that had to be accomplishedduring the samplingperiod (a period of three weeks), only selected samples were pickedfor measuring pHor alkalinity or both. Of the 39 samples that were not rejected by theprocedure thatweeded out flushed standing samples, only 21 were measured forboth pH andalkalinity.Using SYSTAT, a regression analysis was undertaken,with pH as theindependent variable and alkalinity as the dependent variable(Table 12).71Table 12 RegressionAnalysis of pH versus AlkalinityDEP VAR: ALK N:21 MULTIPLER: 0.911SQUARED MULTIPLE R: 0.830ADJUSTED SQUARED MULTIPLE R:.821 STANDARD ERROR OFESTIMATE: 1.692VARIABLE COEFFICIENT STD ERROR5Th) COEF TOLERANCE T P(2TAIL)CONSTANT -16.106 2.9950.000 . -5.378 0.000PH 3.7450.389 0.911 1.0009.626 0.000ANALYSIS OF VARIANCESOURCE SUM-OF-SQUARES DFMEAN-SQUARE F-RATIO PREGRESSION 265.133 1265.133 92.658 0.000RESiDUAL 54.367 192.861The regression analysis gives us a squaredmultiple R of 0.830. This meansthat 83.0%of the total variation in alkalinitycan be accounted for by a linear equationinvolvingpH. The adjusted squared multiple R of0.82 1 is the expected value for the squaredmultiple R if the same analysis was performedon another set of 21 samples takenfromour study areas (SYSTAT, 1992). A squared multipleR of 1.000 means we can fit aperfect linear line through the alkalinity versus pHdata. Our computed squaredmultiple R of 0.830 suggests that there is a goodempirical linear relationship betweenpH and alkalinity. This linear relationshiphas a slope and an intercept on theabscissa.The slope of the linear fit is 3.745 andthe intercept is -16.106. Two-tail studentt testsat ct=0.05 show both the intercept and theslope to be statisticallysignificant. TheAnalysis of Variance Test also tells usthat the overall regression modelis significant atcL=0.05.The regression analysis also testseach piece of data individually tosee if it is anoutlier in the linear regression model.Two, one point from Delta and one pointfromstudy area “y”, of the 21 pointsare outliers (Figure 4).72Figure 4 Regression Plot of pH versus AlkalinitypHAfter these outliers are taken away, the regression modeling can be repeated forthe remaining 19 data points. The regression estimate is improved slightly, with thenew adjusted squared multiple R being 0.887. By visual inspection, the outliers do notseriously affect the linear regression line.4.14.3 Filling in Missing Data PointsThe regression analysis demonstrates that we can approximate the relationshipbetween pH and alkalinity by the equationAlkalinity=12.893+3.379*pH(1)Having demonstrated the validity of this relationship, we may analyze the copper andlead patterns with respect to only pH.2015505 6 7 8 973Having only 21 data points with pHmeasurements may not be good enoughforperforming subsequent statistical tests.To increase the number of available datapoints, one may exploit the knowledge ofthe relationship between pH and alkalinity.All 18 of the 39 standing samples that do nothave a pH measurement do have analkalinity measurement. By using Equation1, one can predict the pH of those 18standing samples by knowing their alkalinitymeasurements (Figure 5).Figure 5 Filling-rn Missing Data for pH versusAlkalinity Data20 I I I IS15fl:S10d -dddd5 -dd0d -5I I I I I3 4 5 6 7 89pH4.15 Copper vs. pHThe copper levels of the standing samples cannow be plotted against theirrespective pH values (Figure 6).74Figure 6 Copper vs. pH3 I I Id2- ddcicicijci dci0o1I,dci0pfl3 4 5 67 8 9pHA visual inspection dramatically underscores the differencein copper levelsbetween samples from Delta and the samples fromthe areas with pH and alkalinityadjustments.4.16 Lead vs. pHOne can also plot the lead levels of the standingsamples against their respectivepH values (Figure 7).75Figure 7 Lead vs. pH30 I Idd n nSd S S20d n -ncidd-Jd10-d SWdySn nnWn SSyn0 I3 4 5 6 7 8 9pHUnlike the copper vs. pH graph, this figure does not show anobvious trend inlead reduction, in sampling areas that have pH and alkalinity adjustments.However,there are three interesting observations in this figure that may beof significance.First, the lead levels are confined to a narrow interval in thesamples that have pH lessthan about 6. Secondly, from pH of 6 and above, there is awide scatter in the leadlevels. This scatter is probably attributable to the standingtime, which is the onlyuncontrollable factor in this study. Thirdly, there is a windowbetween pH of 8 to 8.5where there appears to be a drop in the sample lead levels.4.17 ANOVA (Lead)So far, the evidence has conclusively proven that pH and alkalinityadjustmentsreduce copper at the tap. For lead, there appears to be somereduction, but preliminary76tests do not indicate these reductions tobe statistically significant.Further analysis ofthe lead data can be done withANOVA (the analysis of variance) becauseit is apowerful statistical tool that can analyzethe significance of a largenumber of differentfactors at the same time.ANOVA is a technique that is highly sensitive tononnormally distributed data(Sachs, 1982). There is no readily applicableANOVA technique that can handlenonparamethc data. However, it is possible to makethe ANOVA technique morerobust, more able to accommodate distributionfree data without substantially losing thetrue level of significance or the true power (A.R.E.)of the test (Conover, 1980). Thetechnique requires transforming the measurement datainto ranked data, and then toapply the ANOVA tests on the ranked data. This techniqueis employed for theanalyzes in this research study.4.17.1 Using One-Way ANOVAThere are many types of ANOVA tests. Since we want to knowthe effects ofpH on lead levels, one-way ANOVA appears to be the most appropriatemethod. Inone-way ANOVA, the lead values are separated into groups basedon a factor. Thefactor that determines which group a data point belongs tocould be the study areas,i.e., “d”, “s”, “n”, “y”, “w”, or pH separated into differentranges of values. For bothof these methods of grouping lead data, the experimentaldesign is consideredunbalanced. This means that not all groups havethe same number of data points.Also, the ANOVA we have employed uses the randomeffects model. Although allthe data from each group are characterized by the samefactor, the factor itself is notfixed and is subject to variations. For example, the datapoints in study area “d” allhave slightly different pH, alkalinity, and chlorine residuallevels. SYSTAT hasincorporated both the unbalanced design and the randomeffects model in itsalgorithms. SYSTAT can also do split-plot analysis.This routine is usually required ifthe sequence of data measurements is not randomized.Since our database was77measured using arandomized sequence,it is not necessaryto include thesplit-plotdesign in these analyses.4.17.2 One-Way ANOVAGrouped by StudyAreaThis researcher analyzedthe lead pattern of thestanding samples byone-wayANOVA, grouped bystudy area. Thereare five groupings,namely “d”, “s”,“n”,“y”, and“w”. Although thesamples from the Deltaregion have lower pHthan thesamples from theother areas, the pH of samplesfrom the other study areashave someoverlap. The primarypurpose of this test, then,is not so much to investigatetheeffects of pH alone, as itis to investigate all ofthe factors, includingchlorine residuals,at the same time. TheFisher Test probabilityis p=O. 131 for this model.At (1=0.05,this model is not consideredto be significant.The conclusion fromthis test is that pH,alkalinity, and chlorineresiduals are not effectivein reducing lead levels atthe tap.4.17.3 One-Way ANOVAGrouped by pH Rangewith Age as CovariateThe plot of lead vs. pH gavethis researcher some cluesas to the effectsofaltering the pH of thewater on the lead levels atthe tap. Chlorine residualshave beenignored from this analysis.If the chlorine residualin the water is a significantfactor,then it would have shownup in the one-wayANOVA grouped by area.In the pH analysis,we are most interestedin examining the two zonesthatappear to be anomalies:pH less than 6 and pHbetween 8 and 8.5.For this analysis,then, the lead data has beenseparated into bins.Each bin contains datathat have pHmeasurements falling intoa certain range. Thefive bins are: pHbetween 3 to 6, 6to7, 7 to 8, 8 to 8.5, and 8.5to 9. No pH measurementsrecorded less than3 or morethan 9; thus, the reasonfor selecting pH of 3and 9 as the minimaand maxima of thisanalysis is a sound one.Besides pH, the age of thehouse from which the sampleis taken, might alsoaffect the concentrationof lead in the standing samples.With one-way ANOVAanalysis, it is possible toisolate the effects ofhouse age from theeffects of pHon the78lead values. pH, then, is the grouping parameter, andage is the covariate of theanalysis.The ANOVA analysis indicated that the covariate, p =0.685,and the pH bins,p=O.O72, are not statistically significant at cx=O.05.The Tukey HSD MultipleComparisons (SYSTAT, 1992) show that the pHrange 8 to 8.5 is not significantlydifferent from the data from the other four pH bins.4.17.4 ResidualsThe residuals for all the ANOVA analyzes have beenexamined. The resultsfully confirmed that none of the essential assumptions of the ANOVAmethod havebeen violated by the data sets. As an example of residual examination,the residuals forthe pH range ANOVA analysis will be looked at closely.Firstly, the residuals should follow a normal distribution, forthe ANOVAmethod to be valid. Although the measurement data of the lead valuesdeviate highlyfrom the normal distribution, the same data, after being ranked, do closely followthenormal distribution (Figure 8).79wD-JUiF—0Ui0><UiFigure 8 Normal Probabifity Plot for ANOVA of Lead320-2-3-20 20 30We know this is the case because the expected normally distributed valuesplotin a relatively straight line against the data residuals. Secondly, the plot of the student tversus the ANOVA estimates of the lead values (Figure 9) tells us that theerrors of theestimates are not dependent on each other.-10 0 10RESIDUAL80Figure 9 Student t vs. Lead Esthnate302-0001-0 00o 0F— 00 0z 00 0LU00-0H— o00—1-0 00 o0-2I I I10 15 20 25 30ESTIMATEWe know the errors are independent because the student t values are randomly scatteredabove and below zero in this figure. Dependent errors would show a definite patternortrend. Also, we note from this figure that all the points fall between plus and minus 3standard deviations of the student t distribution. The absence of a widening range oferrors means that the variance of the lead values is constant with respect tothemagnitude of the lead measurement.Finally, the plot of the Cook’s distance versus the lead estimate (Figure 10)shows all of the datapoints do have a Cook’s distance of much lessthan 1. Thismeans that all the datapoints used in the ANOVA analysis belong to and are adequatelydescribed by the statistical model. The residual analysis shows thatthe ANOVAmethod employed here is a valid way to analyze the lead data.81Figure 10 Cook’s Distance vs. Lead Estimate0.15 I I I000.10 -0o000.05 -8000000 00000o000.00°0 0 010 15 20 25 30ESTIMATE4.18 Testing of Oldcopper and PlasticOf the 92 houses that returned bottles to this study, 6 houses had a houseage ofbetween 11 to 15 years old (Oldcopper), and 25 houses had plastic plumbing(plastic).These data were not included in the previous analyzes. TheOldcopper and the plasticdata are divided into the Delta, South Surrey, andNewton areas and compared againstthe main data from the same study areas (Table 13).82Table 13 Comparison of Oldcopper and Plastic with Main DataOldcoDDer(N= 6) Plastic(N = 25)copper lead copper leadSouth Surrey same same lower lowerNewton same same lower lowerDelta same higher (1 pt) lower lowerAlmost all the Oldcopper copper and lead data are in the same range as the main dataset. The only exception is the lead data of the Delta area, where the Oldcopper leadvalue is higher than the main data. This particular comparison may not be validbecause there was only one Oldcopper data in the Delta area. We can say, in general,that the copper and lead levels at the tap are not significantly different between houseswith copper pipes that are 11 to 15 years old and houses with copper pipes that are 10years old or less.For the plastic data, the copper and lead values in all the study areas aresubstantially lower than the main data. This shows that houses that use plastic pipes asthe plumbing material do not have the same copper and lead problems as the housesplumbed with copper pipes.4.19 Comparison of flushed Cold and flushed HotSo far, almost all of the analyses completed apply to the standing samples. Weknow that the flushed cold water samples have very low metal values compared withthe standing samples. Previous studies (Singh, 1990) have shown that flushed hotwater samples have higher copper and lead levels than the flushed cold watersamples.This researcher has subsequently plotted the flushed cold water data versus the flushedhot water data for copper (Figure 11) and lead (Figure 12) measurements.83Figure 11 Quantile-Quantile Plot of Cold & HotRunning Copper10 II I I I0000Copper0Running05Cold0(ppm)000000—05I I I-0.1 0.0 0.1 0.2 0.3 0.4 0.5Copper. Running Hot(ppm)84Figure 12 Quantile-Quantile Plot of Cold & Hot Running Lead30I I I I020-Lead8RunnngCod (ppb)10-0000000000-00-10 I II-10 0 10 20 30 40Lead8Running Hot (ppb)These are non-parametric quantile-quantile plots, i.e., the largest flushed coldwaterdata is paired with the largest flushed hot water data, and the smallestwith the smallest,and so forth. For copper, it can be seen that the flushedhot samples have less copperconcentration than the flushed cold samples. Theflushed cold samples also exhibit alarger variability than the flushed hot samples. Forlead, although the flushed hotsamples have marginally higher lead values than theflushed cold samples, one can stillsay that they are much less than the standing samples.The variability of both the hotand cold flushed samples are roughly the same. Becausethe flushed hot samples do notgive us substantially different information from the flushed cold samples,future studiesshould not waste the effort in collecting flushed hot water samples.The effects of pHand alkalinity adjustments on running cold and hotcopper and lead levels did notappear to be different for rechlorinated and chioraminated tapwater.855. SUMMARY AND CONCLUSIONS5.1 The Effects of pH/Alk AdjustmentsThe results of the various statistical tests on theeffectiveness of pH andalkalinity adjustments in reducing lead and copper are summarizedin Table 14.Table 14 Summary of All Statistical Testingscopper leadsimple test reduction reductionH test reduction-median test reduction noreductionequal variance reduction no reductionUSEPA regulations compliance no complianceCanadian regulations compliancecomplianceANOVA(code) - no reductionANOVA(pH range) - no reductionThe pH and alkalinity adjustments are definitely effectivein reducing the copper levelsin rechiorinated and chioraminated tap water. The same adjustmentsappear also tomarginally reduce the sample lead levels. However, the reductionis so slight that it isnot statistically significant.5.2 Significance of Other FactorsIn the ANOVA analysis, age (houses fitted with copper pipesthat are 1 to 10years old) as a covariate was calculated to be insignificant as afactor on leadconcentrations at the tap. Furthermore, the analysisinvolving Oldcopper data alsoshowed that houses fitted with copper pipesthat were 10 to 15 years old yielded leadconcentrations that were indiscernible from themain data set. In general, then, it isconcluded that the age of a house, which is fittedwith copper plumbing and is less than15 years old, is not a significant factor in affectinglead concentrations at the tap insystems with or without pH and alkalinity adjustments. The effectsof pH andalkalinity adjustments on copper and lead levels didnot appear to be different forrechiorinated and chloraminated tap water.865.3 RecommendationsFor the houses in the study areas that arefitted with copper plumbing, and areless than 15 years old, pH and alkalinity adjustmentsare not effective in reducing leadlevels at the tap. More research should be done toascertain why pH and alkalinityadjustments by themselves are not effective treatmentmethods. As fundamentalresearch in the behavior of lead chemistry in the distributionsystems and houseplumbings is improved, it should be possible todevelop more effective methods toreduce the level of lead concentrations at the tap.For the same level of significance, obtaining moresamples will also improvethe power of the statistical tests that have been used toanalyze the databank of thisstudy. Because non-parametric testsare conservative, increasing the power of the testsmight make the outcome significant for certain teststhat were close to beingsignificant.One has to remember that the results of thisresearch project are valid only forthe range of chlorine residuals that were measuredfor the samples during the samplingperiod. Since the levels of the chlorine residuals changefrom day to day (GVWD,1992), the effects of the change in chlorine residualsin the system could beinvestigated to assess its impact on the lead levelsat the tap. In addition, the effects ofrechlorination could be compared with the effectsof chloramination, for all thedifferent levels of chlorine residuals.Finally, it has been mentionedthat some of the assumptions of the statisticalmethods employed in this study have been violateddue to a number of practical issues.New advances in statistical research, especiallyin quota sampling, non-responseproblems, and sampling from non-defmable populationgroups, could be applied to thestatistical analysis.876. REFERENCES1. American Society for Testing and Materials. Annual Book ofASTM Standards,vol. 11.01. Designation: D4210-83.ASTM, 1984.2. American Water Works Association Research Foundation.Lead ControlStrategies. American Water Works Association, Denver, 1990.3. APHA-AWWA-WPCF, 1989. Standard Methods,17th edition.4. Bailey, R.J., et. al. “Lead Concentration and StagnationTime in Water DrawnThrough Lead Domestic Pipes”. Water Research Center TechnicalReport,TR243, 1986.5. Britton, A., and Richards, W.N. “Factors Influencing PlumbosolvencyinScotland”. Journal of Institute of Water Engineers and Scientists,Vol. 35, Apr81.6. Cochran, W.G. Sampling Techniques. Wiley, NewYork, 1977.7. Conover, W.J. Practical Nonparametric Statistics.Wiley, New York, 1980.8. Department of the Environment. Lead in DrinkingWater. A Survey in GreatBritain, 1975-1976. Report of an Interdepartmental WorkingGroup. PollutionPaper No. 12, London, England, 1977.9. Drill, S. et. al. The Environmental Lead Problem.An Assessment ofLead inDrinking Waterfrom a Multimedia perspective. Washington, D.C.,USEPA,1979.10. Economic and Engineering Services Inc.Greater Vancouver Regional DistrictWater Quality Improvement Program: Final Corrosion ControlReport. March,1990.11. Environment Canada. Guidelines for Surface WaterQuality. Vol.]. InorganicChemical Substances. Inland Waters Directorate,Water Quality Branch,Ottawa, Canada, 1981.12. Fisher, R.A. Statistical methodsfor research workers. Oliverand Boyd,Edinburgh, 1932.13. Gilbert, R.O. Statistical Methodsfor EnvironmentalPollution Monitoring.Van Nostrand Reinhold, New York, 1987.8814. Greater Vancouver Water District, 1992. Quality Control Annual Report.15. Gregory, R. and Jackson, P.J. “Central Water Treatment to Reduce LeadSolubility”. Proceedings of AWWA Annual Conference, Dallas, TX, June1984.16. Health and Welfare Canada. Guidelinesfor Canadian DrinkingWater Quality.Federal-Provincial Advisory Committee on Environmental and OccupationalHealth, Ottawa, 1989.17. Karalekas, P.C., C.R. Ryan, and F.B. Taylor. “Control of lead,copper, andiron pipe corrosion in Boston”. Journal AWWA, Vol. 75, No. 2 (Feb 83),pp92-95.18. Keith, L.H. Environmental Sampling and Analysis: A Practical Guide. LewisPublishers Inc., Chelsea, MI, 1991.19. Kinnison, R.R. Applied Extreme Value Statistics. Battelle Press,Columbus,Ohio, 1985.20. Kuch, A., and Wagner, I. “A Mass Transfer Model to DescribeLeadConcentrations in Drinking Water”. Water Res., Vol. 17, Oct 83.21. Lapin, Lawrence L. Probability and Statisticsfor Modern Engineering.PWSPublishers, Belmont, California, 1983.22. Lyon, T.D.B., and Lenihan, J.M.A. “Corrosion in Solder Jointed CopperTubes Resulting in Lead Contamination of Drinking Water”. Br. CorrosionJournal, Vol.12, Jan 77.23. MacQuarrie, Doug M. GVWD Corrosion Control Initiative - PhaseII.Inhibitor Chemical Testing at Seymour Dam. M.A.Sc. Thesis,Department ofCivil Engineering, University of British Columbia, 1993.24. Mancy, K.H. editor. Instrumental Analysis of Water Samples.Ann ArborScience Publishers, 1971.25. Mandel, J. The Statistical Analysis ofExperimentalData. Interscience-Wiley,New York, 1964.26. Montgomery, Douglas C. Design and Analysis ofExperiments,3rd edition.John Wiley & Sons, Inc., 1991.27. Moore, M.R. “Plumbosolvency of Waters”. Nature,Vol. 243, 1973.8928. Murrell, N.E. “Summary of Impact of Metallic Solders on Water Quality,Plumbing Materials and Drinking Water Quality”. Proceedings of a Seminar,Cincinnati, Ohio, May 16-17, EPA 600/9-85/007, 1985.29. National Academy of Sciences. Drinking Water & Health. Safe DrinkingWater Committee. National Academy of Sciences, Vol. 4, Washington, D.C.,1982.30. Neff, C.H., Schock, M.R., and Marden, J.I. “Relationships between WaterQuality and Corrosion of Plumbing Materials in Buildings”. Project report forGrant No. CR8085660101, USEPA, EPAI600/S2-87/036, 1987.31. Obrecht, Malvern F. and Marcel Pourbaix. “Corrosion of Metals in PotableWater Systems”. Journal AWWA, Vol. 59, No. 8, August 1967.32. Oliphant, R.J. “Lead Contamination of Potable Water Arising from SolderedJoints”. Water Supply, Vol. 1, 1983.33. Oliphant, R.J. “Dezincification by Potable Water of Domestic PlumbingFittings: Measurement and Control”. Water Research Center Technical ReportTR88, 1978.34. Richards, W.N. and M.R. Moore. “Lead Hazard Controlled in Scottish WaterSystems”. Journal AWWA, August, 1984.35. Sachs, L. Applied Statistics: A Handbook of Techniques. Springer-Verlag,New York, 1982.36. Schaut, G.G. “The Action of a Chlorinated Water Supplyupon Lead Pipe”.American Journal of Pharm., 1942.37. Schock, Michael R. “Factors Affecting the Temporal Variability of LeadConcentrations in Domestic Plumbing Systems”. Project Report forUSEPAOrder No. 7W-7480-NASA in partial fulfillment of USEPA Order No. 7W-7425-NASX.38. Schock, Michael R. “Response of Lead Solubility toDissolved Carbonate inDrinking Water”. Journal AWWA, Vol. 72, Dec 80.39. Schock, Michael R. and Wagner, I. Internal Corrosion ofWater DistributionSystems. AWWARF/DVGW-Forschungsstelle Cooperative ResearchReport,1985.40. Schock, M.R. and Gardels, M.C. “Plumbosolvency reduction byhigh pH andlow carbonate-solubility relationships”. Journal AWWA, Vol. 75, Feb 1983.9041. Sharrett, A.R. et. al. “Daily Intake of Lead, Cadmium, Copper,and Zinc fromDrinking Water: The Seattle Study of Trace Metal Exposure”. EnvironmentalResearch, Vol. 28, 1982.42. Singh, Inderjit. Signcance ofBuilding Plumbing Specficson Trace MetalConcentrations in Drinking Water. MASc Thesis, Dept.of Civil Engineering,University of British Columbia, August, 1990.43. Snoeyink, V.L. and A. Kuch. Principles ofMetallic Corrosionin WaterDistribution Systems. AWWARF. Denver, Colorado, 1985.44. Study Group on Environmental Monitoring. Analytical Studiesfor the USEnvironmental Protection Agency: Environmental Monitoring.NationalAcademy of Sciences, 1977.45. SYSTAT for Windows, Version 5 Edition. Evanston, IL:SYSTAT, Inc., 1992.46. Treweek, G.P., et al. “Pilot Plant Simulation of Corrosionin Domestic PipeMaterials”. Journal AWWA, Vol. 77, Oct 85.47. Tukey, J.W. Exploratory Data Analysis. Addison-Wesley,Reading,Massachusetts, 1977.48. USEPA. Lead and Copper National Interim Primary DrinkingWaterRegulations. May 1991.49. World Health Organization. Guidelines for Drinking Water Quality.Vol. 2.Health Criteria and Other Supporting Information. Geneva, 1984.50. Yamane, T. Statistics: An introductory Analysis. NewYork, 196491Appendix A Sampling LocationsCoquNtlamPNttMeadowsMapDeRNdgeBuirnabyTest AreaDeiltaSouth SurreyControlTest Area92Appendix B Written Sampling InstructionsDear Resident:Thank you for your participation. You are given twoZiploc bags. Please fill thebottles in bag “A” tomorrow morning (morning 1). Andfill the bottles in bag “B”the second morning (morning 2).SAMPLING INSTRUCTIONSMorning 1 (Fill the bottles in bag “A”)1. Please take all samples from the kitchen tapfirst thing in the morning.BEFORE you use the showers, toilets, etc.If you use a water purifier, please turn it OFF during sampling.2. Put bottles #1, #2, and #3 on the kitchencounter in order.3. Take off the caps of bottles #1, #2, and #3.4. Bottle #1 has to be filled first. Place bottle #1under the kitchen tap.Turn on the cold water and fill up bottle #1 completely.5. Turn on the cold water full blast and let it rununtil the water becomesreally cold. Fill up bottle #2 completely. Turnoff cold water.6. Turn on the hot water and let it run until thewater becomes hot.Please don’t burn yourself. Fill up bottle #3completely. Turn off hot water.7. Cap bottles #1, #2, and #3 tightly and putthem in the Ziploc bag.8. Leave the Ziploc bag and everything in it outsideyour front door.Morning 2 (Fill the bottles in bag “B”)Repeat steps 1 through 8.I will come by to pick up both bagsoutside your front door.If you have any problems, please donot hesitate to call Ken Chan at430-2420. Thank you for your cooperation.93Appendix C Cartoon Sampling InstructionsU©U©NMght NDoihtNbghtMornong Mornöng 2)-o--6-pROD ag °°A”FON BagDBhOFDrt ThOng nFOrst ThUng OnThe MornUngThe MornOngCoNd Water/Cod WaterHot Water1U LIII23¶F BoMie #B Frst94Appendix D QuestionnaireWATER QUALITY STUDY QUESTIONNAIRE1. May I have your name please?2. May I have your telephone number please?3. Do you use a water purifier on your kitchen water tap? y/n4. What is the type of plumbing material in your house?Look at the water pipe (it is a thin pipe) underneath the kitchen sink.Type Colora) copper bronzeb) plastic could be any colorc) galvanized metal grayishd) other5. Has the plumbing been replaced in the past? y/n6. Would you like a summary of the study results? y/nPlease return this questionnaire with the bottles. Thank you for your cooperation.95Cl,I—I—I——)1I—0).C“00000I-’\0)C000.IxACCCCCPC0t.)L’JL’.)0CCCPCC—0iCC0CC0CC;—;_.-bP.I-C.t)JD‘OO”0O0.CQ- 000pppppppppp-ppppppppp-pb—‘—ooPco..-—PcP-.——PAAAAAAAAAAAAAI——I————‘C,——t’)—C.“0C‘.0AA ——0——I-—)‘.00OUU0000000Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl))00Q—4\..)CO00---\O00))00C.).lL’.)—)__.—)_)_-—-.—-—0’.............).l)t))t’.)t’.)L’)L’)0000—.C).‘.0...‘Jt’)aC)C)C)CCCCCCCC0C)-—‘.0‘.0‘.0‘.0‘.0‘.0‘.0‘.0a)))D))D)D)P)DD))<Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)Cl)a<<CDCDCDfl•—e—fl—*aaaaaaaa—---————--—--————--————---——II))‘)—I-I)eCt’)t)C)0.C’C’C’-——.000000‘.0‘.0—I‘.0‘.0‘.0‘.0-——-C‘)tJ‘C‘.‘‘a.UC’00‘.0‘.lc0vi)L%)00-U.C)L4.VUi.00Cl.CI-C.))4I.-‘.0—1-‘.0‘C‘.00’Ui—0’.00t’.)00C000Ui-.‘.00’.C—Ui‘.00‘.l000’.C)00‘.0‘.lUUi‘.0ii11EC)BDG--------------------------a 0 -) ‘—I—I—lIiCl)Cl)Cl)Cl)Cl)Cl)Cl)—.J1.‘D-.Ui‘0Ui0D-O.L’)00Ui.00.“00-000000000000000000001-0000000000000000000000001———0000000000c000000000000c)ce)(.e)W“0‘.oocooo?CDCDCDCDCDCDCDCDCDDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCD)))III).)I•—Cccc0000cccclU)ce000000cccccccccccccc.....1•.l-_000coocooccocoocJOoooc-C)C)C)C)C)C)C)C)C)’C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)C)I.CD C CD—00id# street house# plumbing age of house(yr)dl woodridgecres 11930c 10d2 woodridge cres 11926 c 9d3 woodgien St 6444 c 5d4 woodlynnct 11802c 12as woodlynuct 11838c 5d6 woodlynn Ct 11857 c 7d7 woodlynn Ct 11858 C 7d8 alderwood cres 11877 c 9d9 beechwood ave 11938 C 7diO woodridgecres 11916c 7dli doncaster Cr 10742 C 6d12 doncaster Cr 10714 C 5d13 doncastercr 10778 p 6d14 83aave 11115 p 4d15 83aave ill25p4d17 uppercanyon rd 11081p4dl6 uppercanyonrd 11084 p 5d18 uppercanyon rd 11116 C 6dl9 cherrylane lO8’74p 4d20 cherrylane 10847 C 6d21 cherrylane 10834 C 699cic,c.ciccic,—.>0.CUiC\O00-..C.\O00Uict-CD-00000000000000)t)C00UiOcMUiC00Ot)CUiL’JCO000,UiUL’)00—-‘0.00I-—I————————I-—I-————————)IUiCCUiCCCCOCcMCCCUiUicMUiUiCCUiCCUiCCCUiUiUiCUiUiCCCUiUiC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CUi.UiUiCDUi‘CO—00-CUi.U00C—D—00000090p090Z Ui‘O-CCoUi00CUit-).I-))I-000OCC-0CD-oL’-)pp.00.Oppoo.000o.o00..aoUi.OCCOUiUiUiOUiCCOUiCC0UiCCCCUiCCUitiUiUio,ibboCoUiUiCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCp----p---..c. L’)I-t.JC00\DpCoopCUCCCMQOCCO100000000)10000000000000000000oOCICflr,I2rjCflCICl)Ct)ClCl)Cl)Cl)Cl)Cl)CI)CflCl)Ci)Cl)Cl)Cl)Cl)—.>4>4>4>4>4>4>4>4>4>4>4>4>4>4CUi.L’.)0D00C-00—CUi.t3‘-.-.O00UiU)clQ>4>4>4>4>4>4>4>40pppopppppppc000ccceeoccc00PP0P0PoPo0*OU)L’.)0\OO-I-Ui0Vi-OL’)‘-DUi000C0CCO-%D00->4>4>4>4>4>4>40ppppppppppPPppppP000000000000000Ct000t.)00t.)o00O00000C0t000—It.)U)*>4>4>4>4>4>4>4>40ppppppppppppppppppp000000000000000000*0000U).U).Ui00U)0000U)000U)0Ui0U)U)>4>4>4>4>4>4>4000000000PPP000PPOOOPPO000000———bb.—o—o—0———0—0*O.01’.)ViU)U)40UiVi)O—0-\000ViU)>4>4>4>4>4>4>4>4>4•ppppppppppppppp00000000000000C.0000t’.)0‘00Vi-000—000—0000..——*>4>4>4>4>4>4>4000ooppppppppppppooo0000U)00000P0000000000*•L’)000.Ci-L’.)0.000C.00U)L’.)C\D000U))L’.)—aViU)———-—————>4>4>4>4>4>40 -a -Ut)iiiII—IIIIIII*0O•ViViVi.)\O00001t-.).Vi0U))OC-OC.tU)>4>4>4>4>4>4>4>40 rIIIIIIIIIIIIIII-t-.)0U)0—‘10it)—Ct’t—)t)00—L’)c’-)L’-—*>4>4>4>4>4>4>4>40) wIIIIIIIIIIIIII-vI—0•-I——0—I—I‘)i—0I—t-)—0C‘000)0U)*>4>4>4>4>4>4>4UIIIt—)11IIIiII*0L’.)L’)U)U)o.t’)C01-000—.CU)Ui00..L’-)C‘‘-Vi>4>4>4>4>4>4>4F) -U*00)00-0-’ciia0)C00CC0->4>4>4>4>4>4>4(i) -U*IIIIIIIIIIIIIIIIIU)0t)t)L’)-U)‘t-Jt’)t.)0t)0Q0U)0L’))I-CCl)Cl)Cl)CV)0Cl)Cl)—.CC00->4>4>4>4>4>4CCCCOO00CCCPP00p00ppCCC1’.)PCCCCC0t’.)1Pt.)P—c*0CJi0L’Jt)OO00CCCC00O000Cc.C..WC).C-0J100>4>4>4>4>4>4pppppppppppppppppppp0000000C0CCC0CCCPcCC*tJCCL).1CC000L’)01J.J10—C‘CC)CC-tJL’)CL’J00>4>4>4>4>4>40pppppppppPppppppppppppCc-COCO0CCCCOOCCCCC0000CC*C).C00CuCCI-0O-‘clC).C‘000CJL’JC..C•CC>4>4>4>4>4——CCCCCC00CCCC0CCCCCCCCCCccC)b.*00C)1OOcCC’00000CCOcCt)C\0CC00>4>4>4>4>4MCCCCCCpppppppOcCPC’CCCCCCCCCCCCCCCC*CCCCC.Cc)CL’.)‘CCCCCOCCL’)-CC-00CI-CCCC>4>4>4>4>4(A)oopopppppppppppocoocc-)CCCC0C0CCCCCCCCCCCCCCCC*CCc)cAJiC.lCCc)0L’JCC-‘-CC‘DC.cAo0’CCCC->4>4>4>4>4>4IIIIIIIIIIIIIIII1’.)I-I-CV—Cul-‘...Ct)c)t.).•.cv..0’C...çi.1.*>4>4>4>4>4>4r.A)IIIIIIIIIIIIIIIII-CCc)Ct.)i—CI—11C—.l1.CcC‘—IS)C—)0CCC*>4>4>4>4>4>40 (*) -)CCCC)CCJI-CC-‘a-CCCCC-CI)C)C-CCC-C-t)*>4>4>4>4>4-.IIIIIIIIIIIIIIIIIIII-i—c.c).1.cCIS.)IS)I—’I—’IS).rco0—))ci)CIS)i—0cçç*>4>4>4>4>4F’)IIIC.IIIIIIIIIIIIIIIIIIIIIIIcici)—.1CciIS)-CCCci)IS)ci)-..‘—.l.—C.—.IS.)ci)cicCC*!>4>4>4>4>4Ci)CIS)ti)Ci-IS)-CCCC-C)-CCCCid#alC* a2C* a3C* blC*b2Ctb3C* alP* a2P* a3P* blP* b2P* b3P*dl 0.24 0.15 0.18 0.29 0.11 0.2 21 1 24 5 25d2 x x x x x x x x x x x xd3 0.91 0.13 0.23 0.23 0.11 0.13 1 -1 2 -10 0d4 1.4 0.5 0.32 1.45 1.22 0.35 3 2 4 2920 5cL5 0.4 0.2 0.21 0.42 0.18 0.22 -1 1 3 45 2d6 0.17 0.11 0.17 0.28 0.36 0.15 -2 3 7 13 35d7 0.19 0.08 0.14 0.08 0.21 0.16 6 6 4 -2 34d8 1.16 0.84 0.21x x x 10 22 4x x xd9 1.35 0.59 0.27 1.32 0.34 0.27 24 2 3 4 1 3dlO 2.29 0.24 0.37 2.01 0.84 0.38 8 1 5 11 2 4dli x x x x x x x x x x x xd12 1.87 0.11 0.38 1.93 0.14 0.4 20 2 30 18 1 8d13 0.2 0.07 0.04 0.06 0.07 0.11 12 2 3 2 2 18d14 0.29 0.16 0.26 0.23 0.08 0.26 2 0 4 5 1 1diS 0.64 0.43 0.22 0.55 0.29 0.22 3 0 — 5 -1 1d17 0.08 0.09 0.27 0.07 0.52 0.48 -1 2 13 9 410d16 2.12 0.09 0.09 1.9 0.07 0.06 3 1 1 3 0 0d18 1.81 0.84 0.23 1.86 0.29 0.29 7 -2 4 14 0 -2d19 0.06 0.01 0.18 0.17 0.02 0.11 3 0 8 0 -2 2d20 1.76 0.42 0.45 2.01 0.97 0.42 9 -2 5 6 15 3d21 1.78 0.87 0.44 1.76 0.2 0.39 25 2 -2 17 3 3*a (sample collected from morning 1)b (sample collected from morning 2)1 (standing sample)2 (running cold sample)3 (running hot sample)C (copper, ppm)P (lead, ppb)105


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics

Country Views Downloads
United States 76 1
Germany 11 4
China 10 44
United Kingdom 7 0
Canada 6 0
France 5 0
Sweden 3 0
Russia 3 0
India 2 0
Japan 2 0
Hong Kong 1 0
City Views Downloads
Unknown 65 16
Washington 16 0
Ashburn 11 0
Shenzhen 9 43
Saint Petersburg 3 0
Sidney 3 0
Boardman 2 0
Stockholm 2 0
Antigonish 2 0
Tokyo 2 0
Buffalo 2 0
London 1 0
Central District 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items