Correspondence among

Alb: Hare is the final draft for your review. Please let me know as soon as possible if you have comments or edits, i would greatly appreciate thase by 2 om, sa i’rn able to finalize the draft today. i'm not sure of the MDEO 1997 park reference. Deb, is there a formal rapart? {will add headers to the modeling output sheets and figures. Thank you, Jennifer xxxEND_PAGE:deq20_b10_223_431_043 DRAFT — deliberative not subject to FOIA Justification of the drinking water no “corrective action” lead level for use in the Flint school and daycare drinking water sample decision tree The Michigan Department of Health and Human Services (MDHHS) requested that the Toxic Steering Group (TSG) provide endorsement of a health-based drinking water lead levels that would require no “corrective action” to put back into service drinking water faucets and fountains to be used by Flint school and daycare children. Selection of the “corrective actions” was not included in the charge question and is not discussed in this document. A TSG subcommittee was convened to examine this issue. Subcommittee members are Deb Mackenzie-Taylor (Michigan Department of Environmental Quality [MDEQ]), Robert Sills (MDEQ), Eric Wildfang (MDEQ), Jennifer Gray (MDHHS), and Kory Groetsch (MDHHS). Neither the Centers for Disease Control and Prevention’s Agency for Toxic Substances and Disease Registry (ATSDR) nor the U.S. Environmental Protection Agency (EPA) have derived potency values (e.g. reference dose, minimal risk level) for lead to be used in risk assessment methods. Both agencies have stated that no blood lead level has been found to be safe and precautionary actions to reduce exposure to lead, such as running tapwater briefly before use, should always be followed. Due to the lack of potency values, the EPA developed a biokinetic model for evaluating lead exposure from multiple media. Lead levels in air, soil, house dust, diet, drinking water, and maternal blood are used to estimate geometric mean blood lead levels for a population of children and the probability that a child would be above a specified blood lead level’. The model results are compared to an EPA population health protection goal for young children exposed to lead at residential properties of 5% or lower risk of a child having a blood level greater than the CDC reference value of 5 g/dL”. The model, however, does not include inputs to account for direct paint chip ingestion’. Use of the EPA’s Integrated Exposure Uptake Biokinetic (IEUBK) Model The EPA’s Integrated Exposure Uptake Biokinetic (IEUBK) Model for Lead in Children version 1.1 Build 11 was used to evaluate exposure to children one year old or younger and children seven years old or younger. Default values were retained unless otherwise noted below. See Table 2-1 in the “User’s Guide for the Integrated Exposure Uptake Biokinetic Model for Lead in Children (IEUBK) Windows®” for the listing of all the default values’. Values selected that are more appropriate for a Michigan urban area, Flint, are described briefly below. Air data MDEQ performs monitoring for lead in ambient air (as measured in Total Suspended Particulates; lead [TSP]) at a number of fixed stations in Michigan. Some of these monitoring stations (e.g., in Belding, Port Huron, and Vassar) were established to intentionally capture the influence of specific nearby industrial * The modeling output includes the percent of children in a population that are above a specified blood lead level, but that value can be also be viewed as the probability (risk) that a certain child would be above the specified blood lead level. * As stated on the EPA’s Lead at Superfund Sites: Frequent Questions from Risk Assessors on the Integrated Exposure Uptake Biokinetic (IEUBK) Model webpage, found at > The model itself and documentation, including the user’s guide for the model can be found at Page 1 of 7 xxxEND_PAGE:deq20_b10_223_431_044 DRAFT — deliberative not subject to FOIA emissions, while others were established to provide data reflecting the general ambient air lead levels in urban areas. The TSG Subcommittee reviewed the ambient air monitoring data, focusing on urban ambient air lead data that would be representative of contemporary Flint conditions. The most recent MDEQ annual air monitoring report is for 2014". The summary statistics provided in the MDEQ annual air quality reports include 3-month average (mean) levels, the highest value (24-hour sample), and the 2™ highest value (24-hour sample). The National Ambient Air Quality Standard (NAAQS) for lead (0.15 ug/m?) has a 3-month averaging time. The Flint (Whaley Park; 3610 lowa Ave.) monitor, measuring lead, operated until early 2007. The 2005 calendar quarter arithmetic mean values were 0.008, 0.010, 0.011, and 0.012 micrograms per cubic meter (ug/m?). The 2006 calendar quarter arithmetic mean values were 0.005, 0.007, 0.012, and 0.006 g/m’. Although Flint data more recent than 2007 are not available, MDEQ staff considered other urban lead monitoring sites to be fairly representative of Flint conditions. Grand Rapids, Allen Park, and Dearborn each had a highest rolling 3-month arithmetic mean lead (TSP) level of 0.01 ug/m? in 2013; for 2014, these values were 0.01 g/m? for Grand Rapids and Allen Park and 0.02 g/m? for Dearborn. These data indicate that the IEUBK model default value of 0.1 g/m? is not the most appropriate air lead level for Flint, and support a representative value of 0.01 g/m? for air data model input. Water consumption The default drinking water intakes are age-dependent and based on national averages’. These were updated to match the mean drinking water intakes (direct and indirect) in Table 3-1 from the U.S. EPA Exposure Factors Handbook: 2011 Edition. e 0-1 years old - 0.32 liter/day (L/d) (age-adjusted from the values in Table 3-1) e 1-2 years old - 0.271 L/d e 2-3 years old - 0.317 L/d e 3-4 years old - 0.327 L/d (3 to 6 years old in Table 3-1) e 4-5 years old - 0.327 L/d (3 to 6 years old in Table 3-1) e 5-6 years old - 0.327 L/d (3 to 6 years old in Table 3-1) © 6-7 years old - 0.414 L/d (6 to 11 years old in Table 3-1) Drinking water data for alternate drinking water sources The alternate drinking water sources used for these modeling runs are listed below with a brief description. e “Percent of Total Consumed as First Draw” — This was set to 0% as Genesee County issued an Public Health Emergency Advisory® recommending that no one drink the City of Flint water until it has been tested or unless it is being filtered through a filter meeting the National Sanitation Foundation (NSF)/American National Standards Institute (ANSI) standard 53. “The MDEQ 2014 Annual Air Quality Report can be found at amu-2014 Annual Air Quality Report 492732 7.pdf?20151014123800. > Per the “User’s Guide for the Integrated Exposure Uptake Biokinetic Model for Lead in Children (IEUBK) Windows®”, “The default consumption rates are age-dependent and based on national averages. These consumption rates should be changed only when valid site-specific monitoring data is available.” ® The Genesee County Public Health Emergency Declaration can be found at health emergency announcement _10 1 15.pdf. Page 2 of 7 xxxEND_PAGE:deq20_b10_223_431_045 DRAFT — deliberative not subject to FOIA e “Concentration of Lead in First Draw (ug/L)” — With the “Percent of Total Consumed as First Draw” set to 0%, this value will not alter the model run output so the default of 4 micrograms per liter (ug/L or parts per billion [ppb]) was retained. e “Concentration of Lead in Flushed (ug/L)” — This was set to 1 ug/L (ppb) as filters meeting NSF/ANSI standard 53 are recommended for use in all schools and daycares. The filters distributed remove greater than 99% of the lead and would result in filtered water containing lead levels of approximately 1 ug/L (ppb) or less. e “Percentage of Total Consumed from Fountains” — This was set to 50%, as children in school and daycare setting spend around eight hours or more in those locations. Eight hours could be expected to be approximately half of their waking hours in a day. e “Concentration of Lead in Fountain Water (ug/L)” — Modeling runs were carried out for every 1 pg/L (ppb) between zero and 15, 30, and 100 pg/L (ppb). Modeling runs for the zero to 15 pg/L (ppb) were carried out to perform a sensitivity analysis to evaluate how a one ug/L (ppb) change in lead water concentrations alters the outputs of the model. The 30 and 100 pg/L (ppb) were included as preliminary representatives of high-end lead levels obtained from school drinking water fountains. Site-specific soil data Soil lead levels on residential properties vary greatly’. The primary influence is from pre-1978 use of lead paint on the outside of buildings®*"°. After 1978 lead-paint was banned. In addition, the age of housing”®, paint condition®, and distance from the building”® can greatly impact the concentration of lead in soil. The soil lead concentrations near the foundation of houses with lead paint are frequently over 1,000 ppm®****?_ Historic vehicle emissions from leaded gasoline are another source of residential soil lead, so distance from road and traffic volume at the time leaded gasoline was in use are also 210.12 Other potential influences are fill material (e.g., foundry sand, smelter slag), and/or local industrial emission sources. The highest soil lead influential factors on residential soil concentrations concentrations are typically found in urban areas with high population density’. Soil lead data for residential soil in Flint are not currently available. Other Flint park data and relevant residential soil lead studies were reviewed to evaluate appropriate surrogate concentrations for this assessment. 7 Datko-Williams, L., A. Wilkie, et al. (2014). "Analysis of U.S. soil lead (Pb) studies from 1970 to 2012." Science of The Total Environment 468-469: 854-863. 5 Erancek, M. A. (1992). "Soil lead levels in a small town environment: a case study from Mt Pleasant, Michigan." Environ Pollut 76(3): 251-257. ° Francek, M. A., B. Makimaa, et al. (1994). "Small town lead levels: a case study from the homes of pre-schoolers in Mt. Pleasant, Michigan." Environ Pollut 84(2): 159-166. ” schwarz, K., S. T. Pickett, et al. (2012). "The effects of the urban built environment on the spatial distribution of lead in residential soils." Environ Pollut 163; 32-39. ! Andra, S. S., D. Sarkar, et al. (2006). "Lead in Soils in Paint Contaminated Residential Sites at San Antonio, Texas, and Baltimore, Maryland." Bulletin of Environmental Contamination and Toxicology 77(5): 643-650. ® Yesilonis, |. D., R. V. Pouyat, et al. (2008). "Spatial distribution of metals in soils in Baltimore, Maryland: role of native parent material, proximity to major roads, housing age and screening guidelines." Environ Pollut 156(3): 723-731. Page 3 of 7 xxxEND_PAGE:deq20_b10_223_431_046 DRAFT — deliberative not subject to FOIA The MDEQ”™ reported soil lead concentrations from urban parks including Flint. This data set included 12 sample locations from 10 Flint parks. Surface soil samples were collected at 0-2 inches and subsurface samples were collected at 4-6 inches. This dataset may not adequately represent residential soil samples with a lead paint contribution from buildings. Other studies have shown that open spaces may have much lower soil lead concentrations than areas near residential buildings that were built prior to 1978°°"*, The geometric mean and geometric mean plus two standard deviation surface soil concentrations (see below) from this Flint park dataset were used for this assessment. The geometric mean plus two standard deviation surface soil concentration represents an upper percentile value that may partially account for the higher soil lead levels expected around buildings with lead paint. Another data source is from soil samples collected from the Rouge River watershed between 1992 and 2002". This dataset was compiled from MDEQ files and Clayton Group Services files for the Rouge River watershed area in southeast Michigan. Although some data is from Part 201 sites, most of the samples were from sites were collected as a result of other regulatory or real estate due diligence requirements. Sites with known industrial metal sources or that had near surface soil with unknown fill material were excluded from the dataset. The dataset included 28 residential sites that had 535 surface soil samples with data reported for lead. The geometric mean value for residential surface soils reported from this study (see below) was used for this assessment. The current Part 201 residential direct contact risk-based soil level was also used for this assessment for comparison. This value was generated from the IEUBK model using an acceptable blood lead concentration of 10 g/dL, a drinking water concentration of 4 ug/L (ppb) and a default air concentration used in the model at the time of criteria development. Soil levels examined in the model runs were: e 47.5 parts per million (ppm) — based on the geometric mean of 0-2 inch soil concentrations in Flint urban parks*®. e 155 ppm — based on the geometric mean + two standard deviations of 0-2 inch soil concentrations in Flint urban parks”. e 160 ppm — based on the geometric mean for residential surface soil from Rouge River water shed*®. e 400 ppm - the MDEQ Residential Direct Contact Risk-Based Screening Level for soil. Both the residential soil geometric mean from the Rouge River watershed and the upper percentile from the Flint park data appear to represent the best surrogates for residential soil concentrations in Flint for this assessment. Although not from Flint, the Rouge River watershed residential soil data appears to be a fairly robust data set for urban residential soils in Michigan. The Flint park data upper percentile ™ Zahran, S., H. W. Mielke, et al. (2013). "Determining the relative importance of soil sample locations to predict risk of child lead exposure." Environ Int 60: 7-14. * Murray, K. S., D. T. Rogers, et al. (2004). "Heavy metals in an urban watershed in southeastern Michigan." J Environ Qual 33(1): 163-172. 1 The value is from Table 5 in Murray, K. S., D. T. Rogers, et al. (2004). "Heavy metals in an urban watershed in southeastern Michigan." J Environ Qual 33(1): 163-172. Page 4 of 7 xxxEND_PAGE:deq20_b10_223_431_047 DRAFT — deliberative not subject to FOIA (geometric mean plus two standard deviations is approximately equivalent to the 97.5 percentile) may better account for potential building paint contributions to soil concentrations for urban residential soils than the geometric mean concentration of this dataset. The Flint park upper percentile (155 ppm) reported as two significant digits is equivalent to the Rouge River watershed residential soil lead geometric mean (160 ppm). The soil lead concentration of 160 ppm was selected for use in the calculation of the TSG Subcommittee’s recommended health-based drinking water lead levels. Maternal data Maternal blood lead levels were changed to 0.8 g/dL to match the NHANES (2011-2012) geometric mean lead level for women’’. Blood level of concern (cutoff) for risk estimation A cutoff value of 5 micrograms per deciliter (ug/dL) was selected as that is the current Centers for Disease Control and Prevention blood reference level to identify children with blood lead levels that require follow-up actions. It is based on the blood lead level for the 97.5" percentile for the U.S. population of children, ages one to five™®. Outputs of the modeling runs and recommendation of drinking water lead levels for use during the evaluation of Flint school and daycare drinking water samples The soil lead level of 160 ppm was selected as being the best surrogate for soil lead levels in the city of Flint. In the event that site-specific soil data is collected, the drinking water lead levels may need to be reevaluated. The entire set of modeling runs and figures can be found in Attachment 1. The recommended drinking water lead levels resulted in a 5% risk or lower that a child had a blood lead level above 5 yg/dL. The 5% is an EPA health protection goal for young children exposed to lead at residential properties’. Evaluating school and daycare drinking water samples together The recommended lead screening level for use while evaluating both school and daycare drinking water samples is 2 pg/L (ppb). This lead level in drinking water samples results in no more than a 5% risk that children age 7 or younger would have blood lead levels above 5 ug/dL. Aperange Blood lead geometric mean (in Risk that a child could have a blood lead micrograms per deciliter [yg/dL]) level above 5 pg/dL 0-1 year olds 2.308 5.004% 0-7 year olds 2.056 2.934% Assumptions integral to the use of the health-based school and daycare drinking water lead level ” The Fourth National Report on Human Exposure to Environmental Chemicals, dated February 2015, can be found at UpdatedTables Feb2015.pdf. *8 The blood reference level and additional information can be found at lead levels.ntm. Page 5 of 7 xxxEND_PAGE:deq20_b10_223_431_048 DRAFT — deliberative not subject to FOIA This calculation includes two assumptions. One is that people have lead levels in home drinking water of no more than 1 ug/L (ppb) or are using filters to reduce lead levels to 1 ug/L (ppb) or below. The second is that children do not live in homes with highly elevated levels of lead in home dust or soil. The modeled lead dust levels do not account for lead coming from pre-1978 paint. Eighty-two percent of city of Flint houses were built before 1970”. Soil highly impacted with lead paint may have lead levels higher than the 160 ppm used in this calculation. If children are not drinking filtered water at home, are ingesting lead paint chips, or have elevated levels of lead in home dust or soil, a lead level lower than the one recommended above may need to be selected for the evaluation of school and daycare drinking water samples. Evaluating school and daycare drinking water samples separately The recommended lead screening level for use while evaluating daycare drinking water samples is 2 ug/L (ppb). This lead level in drinking water samples results in no more than a 5% risk that children age 1 or younger would have blood lead levels above 5 yg/dL. Rpetaage Blood lead geometric mean (in Risk that a child could have a blood lead micrograms per deciliter [yg/dL]) level above 5 pg/dL 0-1 year olds 2.308 5.004% 0-7 year olds 2.056 2.934% Assumptions integral to the use of the health-based daycare drinking water lead level This calculation includes two assumptions. One is that people have lead levels in home drinking water of no more than 1 ug/L (ppb) or are using filters to reduce lead levels to 1 ug/L (ppb) or below. The second is that children do not live in homes with highly elevated levels of lead in home dust or soil. The modeled lead dust levels do not account for lead coming from pre-1978 paint. Eighty-two percent of city of Flint houses were built before 1970”. Soil highly impacted with lead paint may have lead levels higher than the 160 ppm used in this calculation. If children are not drinking filtered water at home, are ingesting lead paint chips, or have elevated levels of lead in home dust or soil, a lead level lower than the one recommended above may need to be selected for the evaluation of school and daycare drinking water samples. The recommended lead screening level for use while evaluating K-12 school drinking water samples is 11 ug/L (ppb). This lead level in drinking water samples results in no more than a 5% risk that children age 7 or younger would have blood lead levels above 5 wg/dL. However, when evaluating a population of children age 1 or younger, their risk of having a blood lead level above 5ug/dL was 8.335%. The 11 pg/L (ppb) screening level is not adequately protective for children under the age of 1. Age range Blood lead geometric mean (in Risk that a child could have a blood lead 8 8 micrograms per deciliter [yg/dL]) level above 5 pg/dL 0-1 year olds 2.61 8.335% 9 Ns determined by the U.S. Census and can be found at 14 1YR S110. &prodType stable. *° ‘As determined by the U.S. Census and can be found at 14 1YR $1101&prodType stable. Page 6 of 7 xxxEND_PAGE:deq20_b10_223_431_049 DRAFT — deliberative not subject to FOIA 0-7 year olds 2.299 4.917% Assumptions integral to the use of the health-based school drinking water lead level This calculation includes two assumptions. One is that people have lead levels in home drinking water of no more than 1 ug/L (ppb) or are using filters to reduce lead levels to 1 ug/L (ppb) or below. The second is that children do not live in homes with highly elevated levels of lead in home dust or soil. The modeled lead dust levels do not account for lead coming from pre-1978 paint. Eighty-two percent of city of Flint houses were built before 1970”7. Soil highly impacted with lead paint may have lead levels higher than the 160 ppm used in this calculation. If children are not drinking filtered water at home, are ingesting lead paint chips, or have elevated levels of lead in home dust or soil, a lead level lower than the one recommended above may need to be selected for the evaluation of school and daycare drinking water samples. *1 As documented by the U.S. Census and can be found at 14 1YR $1101&prodType =table. Page 7 of 7 xxxEND_PAGE:deq20_b10_223_431_050 0-1 year olds 0-7 year olds Geometric mean Geometric mean Blood lead level Blood lead level Soil lead levels (ppm) Fountain water (ppb) —_|(ug/dL) %>5 ug/dL |(ug/dL) %>5 ug/dL 77s TS) CT [assaf SSS ———S—at| acon 2.958 2.300] tof of za aaa 2.002 2.573 [00 of 2] 36.82) ———S—S—~— S| 3.4.30 [—______2000f of 3.43] 98.22a[ ____i4.sas| 98.807] arsf siz] zea] 1.034] 0.04 [assaf S~—S 82a] sre’ ————2.067| 3.007 | 7 7 | SE] [00 af asc] 39.009 4.23 36.161 [—______3000f a a6] 98.260[ ————ita i] 9.87 arsf | S—=it ae] ose] 1.003] 0.06 assaf SSC anf sco} ian. Cs 2 7 | CZ) [00 of as] ania] 202] 37.082] [______3000f | 13529] 98.29] ___ia.cai] 98.887] arf S—iara _oace] ———st.aan| 0.087 assaf SS SCSC~« aa] ssa] ————S—~—i as | 3.827 tof fst] raza 2.29 4.29 [00 af 50a an aze[ azn 37.997] xxxEND_PAGE:deq20_b10_223_431_051 0-1 year olds 0-7 year olds Geometric mean Geometric mean Blood lead level Blood lead level Soil lead levels (ppm) Fountain water (ppb) —_|(ug/dL) %>5 ug/dL |(ug/dL) %>5 ug/dL PC ——~C‘ Ta] Pars sisal saa] 1.276] 0.10 |) eC) 2 | | 2 [oof faa] ana] ~~ |__| -______2000f of 98.906] xxxEND_PAGE:deq20_b10_223_431_052 0-1 year olds 0-7 year olds Geometric mean Geometric mean Blood lead level Blood lead level Soil lead levels (ppm) Fountain water (ppb) —_|(ug/dL) %>5 ug/dL |(ug/dL) %>5 ug/dL ee) 6.967| 75.99 6.463| 70.746 2000 14.992 99.026 16.034] 99.342 xxxEND_PAGE:deq20_b10_223_431_053 0-1 year olds 0-7 year olds Geometric mean Geometric mean Soil lead levels Fountain water Blood lead level Blood lead level {ppm) (ppb) (ug/dL) (ug/dL) +e Se Po ao tattoo] .92] 0.026] 2) | >| | 7) KY Po area .034f 0.04) Po ae tort zai .o9f 0.06] Poof tara aes]. saa] 0.087 Po aes sazt a 76f 0.104) [tsa] of _2198| ao1al __1.958| 2.302] [155.3] al __2334[ 5.2aal 2.067) 3.007] [155.3] of 2.401] 5.932] 2.121] 3.402] [155.3] _s|__2.468| ese] ___2.175| 3.827] [155.3] 9] 2.502] 7.036] 2.202] 4.05] po tof zit 2.002] 2.573] Po two aa 38f 2165] 3.743) Po tof saz]. 209] 4.291 Po tof Sa 7 52at 2.245] 4.426) xxxEND_PAGE:deq20_b10_223_431_054 0-1 year olds 0-7 year olds Geometric mean Geometric mean Soil lead levels Fountain water Blood lead level Blood lead level {ppm) {ppb) (ug/dL) %>5 ug/dh Po oof 36.82] 4.135] 34.306] Poo 9st 7.376[ 46] 34.772] Po ao 327] 7.929] ta] 35.236] Po too 857] 8a] 4.209] 35.699] Po too 36 39.029] 4.233] 36.161] Po too ate 39.575[ 4.258] 36.622] Po too as] gota] 4.282] 37.082] Po too ara goes] 4.307] 37.5a Po oof Saat agef 4.3834] 37.997] Po ofa 73af 4855] 38.453] Po aot 62] 4.263] 4.38] 38.907] Po toot soa. 792[ 4.04] 39.359] Po ott. 3a7] 28] 39.81] [oof taf ag] aa.za] 4.453] 40.259] [oo] 1a] 4.77] _4a.asa| 4.477] 40.707] [oo] 154.706] _4a.a7s| 4.501] _a1.t5a| [oo] 30] _5.aa| 52.202] 4.1] 47.608 [oo] 100] 6.967] 75.93] 6.463 _70.746| [3000 of _13.431| 98.224| 14.54] _98.847| [2000 6] 13529] 98.29] __14.6a1| _9a.8a7| [20008] 13561] sa.ara] 14.671] 989] [2000] 9] 13577] 98.323] 14.686] 98.906] xxxEND_PAGE:deq20_b10_223_431_055 pi) = a2 2 a 2 Qa Qa Qa a om a a a oO 2 Huo mn Oo 8 N a ot nm od fS ©£ ef Ee oO oO oO oO oO £eees c = i=] Cc c = | =) a a a fe} fe} fe} fe} o eeeee g A FA L Uy | 2 @ @ poy } j Poy i i ,oo| i j ,oo| i j ,oo| i j pf i i , 4 i i ; Z : ig , 4 i 18 Hi ; i Hi ; i , 4 i i y t £ t ,o| i y i Hi Fi £ 4 i i i i i i i i Citi ke tiny ; ; < Cig ed, i : 8 Wain ty, t i ot co) Cag, i ; = “hn, “Uy i i is} og —_ H 4 ° ‘c Pky, j = tin, bo Yen, A = o a , & sc P| el, y 5 Cd Hi in, | eat H 2 r3} } “oY H i i “dh, i a y i Pn, i <= S_ } f fey Z q ye) i 4 i i _ oc t f ing i, i $ { Hi A ti, Yi Hi o f Ga; Vee as > i i i 4 ge | = y i i g Gee | = a i , ie Gg 7 8 H Hi g i 1 ; i i y WV , 4 a o } : i y ae) a j f H @ 44 3 4 ro} j i 1 g GG 4 | @ i 4 i g G44 | o } i i Bi gag | = i i i g Gig | = j i 1 i G% G3 | 3° ae te 2 g i i i Y 424 im A i iB i iA i i i 4: OG] ip j i 4 448 1 A i 4 i %¢ 4% 1% i i j 441% | i i i Gale | y i i Yea | j f i 2 434 ; j i i , BY | j i i 4 944 i i i i b Wey) A H i% 1469 3 i i { % 164} ai i i i 44 G4) i | i 4 y © i ; i bg 4 £ i i j i i i : f t i H f j i H f j i : f t i oun OH OH GO 4H GO WW Kn Oo Oo Hn Hn FT FTF MH HN 105 100 95 20 15 10 Es 0 1p/3n g aaoge % xxxEND_PAGE:deq20_b10_223_431_056 Fountain 100 ppb 2 2 a aa Qa N a £ ¢£ cs Ss £ $s Ss c as 3 o 8 GE g g @ sssoosFOUNtaiN 15 ppb sswssss FOUNTAIN 30 ppb i i i fp i i j i H t 4 i i i j j i i j j i i j j i i j i i i 8 i y i 3 i j j i i i i i i i i i i i i i i i ; i A 4 i i i i i i Lo £c ; i ; 16 o ; i i ;~ = | i i j < coc | i i 6 xo | i i y fs i i | i ” i t 4 Hi ¢ Oo - yi as an i tod gi f 4 4 Z 4 of i to 49 | a oO ; i poy He | 2 g i i po Bel = i Hi ; 4 geg or o i ' po aa o i i to Bat 2 i H , Ba i = i | ee, 6% i | ee 8 i to By i i i PooY Be i i j po gym i 1 po G45 i i , 4 Baia i i pf Arn Hi i ; ; ae i i to Gat | i poof Ge i i po a i f i ; ; GG i i i to By) i A 4 4 i GG i. / i po 4G: i : y ; % Get Hi f i j 2 Gi i i po | 4G: i 1 i i 1G i i ; ; 4g i i Po “Y i i i “4 i i po i i i pf i i H , | ; i H , | ; i H , | ; t cE i oun oOo HM Oo Hh oH Go wh oH oH 8D ye Oo Ob nH FTF FTF H HN N A 105 100 95 1p/3n g aaoge % xxxEND_PAGE:deq20_b10_223_431_057