www.environmentalhealthproject.org       Community  Assessment  of  Penn  Trafford  Outdoor  Air     Monitoring  Results   June  6,  2016       In  the  fall  of  2015  air  monitoring  was  conducted  at  five  residences  located  in  the   vicinity  of  a  new  unconventional  natural  gas  well  pad  in  Penn  Township,  PA.  Speck   air  monitors,  which  measure  PM2.5,  were  located  inside  and  outside  residences   situated  between  500  –  2,500  feet  from  the  pad.  Monitoring  occurred  during  times   when  the  well  pad  operators  reported  that  drilling,  fracking  and  combustion   activities  took  place  during  the  development  of  the  first  two  wells  on  the  pad.   Volatile  Organic  Compound  (VOC)  samples  were  also  taken  with  summa  canisters   outside  two  of  the  residences  on  9  occasions.  This  report  presents  the  findings  of   outdoor  air  quality  based  on  PM2.5  and  VOC  results.       Key  Findings:     1)  The  Speck  monitors  placed  outside  the  residences  recorded     approximately  2-­‐4  peaks  per  day  of  PM2.5.1   2)  The  Speck  closest  to  the  well  pad  recorded  peaks  more  often  and  at   higher  levels  than  Specks  at  further  distances.  PM2.5  peaks  were   substantially  higher  closer  to  the  well  pad.   3)  The  short–term  peaks  and  the  accumulated  concentration  per  day  of   PM2.5  found  at  these  locations  were  high  enough  to  potentially  cause   respiratory  problems.2     4)  Sampling  for  VOCs  detected  a  range  of  chemicals  that  have  been  found   at  other  shale  gas  sites.  More  VOCs  were  identified  near  the  closest   residence  (41)  than  at  the  further  residences  (33).  VOCs  can  travel  with   PM2.5  into  the  deep  lung  and  cause  or  exacerbate  health  problems.   5)  More  VOCs  were  detected  when  the  wind  was  calm  or  light  during  the   sampling  period.                                                                                                                         1  EHP  defines  a  peak  as  two  standard  deviations  over  the  mean  of  a  moving  15  minute   average.   2  Short-­‐term  increases  of  more  than  10ug/m3  have  been  shown  to  cause  health   effects.  The  accumulated  concentrations  of  peaks  levels  over  time  are  higher  than   EPA’s  National  Ambient  Air  Quality  Standard  for  PM2.5.     SW Pennsylvania Office 2001 Waterdam Plaza Suite 201, McMurray, PA 15317   Office: 724.260.5504 • Cell: 724.249.7501   Connecticut Office 760 Chapel Street, New Haven, CT 06510 Fax: 203.691.7606 The  PM2.5  Speck  monitors  recorded  results  continuously  for  approximately  32  days   at  each  residence,  while  VOCs  were  sampled  nine  times  for  a  24-­‐hour  period.  PM2.5,   because  it  can  be  monitored  continuously,  is  considered  a  surrogate  for  other   chemicals  that  are  emitted  from  the  well  pad.  The  24-­‐hour  VOC  samples  provide  a   qualitative  assessment  of  the  chemicals  present  in  the  air  in  the  vicinity  of  the  site   and  the  monitored  residences.       PM  2.5  Outdoor  Monitoring  Results     Chart  1  shows  PM2.5  results  for  the  five  residences  over  the  entire  monitoring   period.  The  time  periods  reported  to  Public  Source  indicating  stages  of  well   development  are  shaded  blue.  The  number  of  days  for  each  stage  are  found  in  Table   1.  The  Speck  monitor  reports  minute  by  minute  data  points  and  these  were   averaged  over  15  minutes.         Table  1.  Length  of  time  in  days  for  the  reported  stages  of  well  development       Stage  of   Length  of  stage   development   in  days   Drilling   9   Lull   7   Fracking   15   Lull   14   Combustion   10       Chart  1.  Outdoor  Speck  PM2.5  results  at  five  residences,  averaged  every  15   minutes.  Arrows  show  2  points  in  time  when  all  five  Speck  monitors  showed   significant  spikes.       2   The  chart  shows  that  the  Speck  located  outside  the  residence  closest  to  the  well  pad   (gray  line)  recorded  much  larger  peak  levels  of  PM2.5  than  the  four  houses  further   to  the  southeast.  In  general,  the  four  homes  at  further  distances  show  very  similar   results  as  a  group.  There  are  times  when  all  five  outdoor  Specks  showed  increased   levels  at  the  same  time,  most  notably  just  before  fracking  reportedly  begins  and  also   early  during  the  combustion  phase.  The  arrows  in  the  chart  point  to  these  time   periods.     Exposure  levels  at  all  five  residences  showed  short-­‐term  increases  in  PM2.5  levels   that  could  cause  upper  respiratory  symptoms  in  children,  elderly,  individuals  with   chronic  illnesses  and  recovering  patients.  Short-­‐term  increases  of  more  than   10ug/m3  may  produce  wheezing,  asthma  symptoms  and  other  respiratory   problems  in  sensitive  individuals.3,4     Important  components  of  PM2.5  exposures     Charts  2-­‐5  show  the  range  of  results  for  5  components  we  measured  using  the   available  outdoor  Speck  data.  Each  chart  shows  the  calculated  mean  for  outdoor   PM2.5  for  each  stage  of  well  pad  development,  and  each  “lull”  between  stages.  The   time  periods  for  each  stage  are  found  in  Table  1.  For  this  analysis  we  combined   results  from  the  four  homes  located  in  the  same  neighborhood  to  the  southeast  and   compare  them  to  the  single  home  close  to  the  well  pad.      In  the  charts  the  orange  line  shows  results  for  the  residence  closest  to  the  well  pad   site  (500  feet  to  the  east).  The  blue  line  shows  the  combined  results  for  the  4   residences  southeast  of  the  site  located  between  2000-­‐2500  feet  from  the  well  pad.   The  standard  deviation  for  the  four  homes  is  shown  as  well  as  the  mean.  It  is   important  to  remember  that  there  may  be  activity  at  the  site  during  the  “lulls”.  In   particular,  the  volume  of  truck  traffic  likely  may  increase  during  preparation  for   each  stage.                                                                                                                                     3  Strickland  MJ,  Hao  H,  Hu  X,  Chang  HH,  Darrow  LA,  Liu  Y.  Pediatric  Emergency  Visits  and   Short-­‐Term  Changes  in  PM2.5Concentrations  in  the  U.S.  State  of  Georgia.  Environ  Health   Perspect;  http://dx.doi.org/10.1289/ehp.1509856.   4  Shi  L,  Zanobetti  A,  Kloog  I,  Coull  BA,  Koutrakis  P,  Melly  SJ,  Schwartz  JD.  2016.   Low-­‐concentration  PM2.5  and  mortality:  estimating  acute  and  chronic  effects  in  a   population-­‐based  study.  Environ  Health  Perspect  124:46–52;   http://dx.doi.org/10.1289/ehp.1409111     3   Chart  2.  Peaks  per  day:  This  graph  shows  the  mean  number  of  large  scale  changes   (peaks  at  least  2  standard  deviations  over  the  mean)  in  air  quality  per  stage  of  well   pad  development.         The  nearest  Speck  recorded  between  2-­‐4  peaks  per  day.   The  remaining  Specks  recorded  between  1.5  –  2.5  peaks  per  day.       Chart  3.  Peak  Duration:  the  average  length  of  time  peaks  lasted  per  stage  of  well   pad  development.           The  duration  of  peaks  recorded  by  the  Speck  at  the  closest  residence  were  shorter   than  those  recorded  at  further  distances,  and  peaks  occurred  more  often  closer  to   4   the  well  pad  (Chart  2).  There  is  less  dispersal  of  the  concentrated  emissions  near  the   source  and  near  the  closest  residence,  causing  the  shorter  duration  of  peaks.     Chart  4.  Time  between  peaks:  the  average  length  of  time  between  peaks.           These  results  show  that  the  closest  Speck  monitor  recorded  shorter  times  between   peaks  which  results  in  more  peaks  per  day  (Chart  2).     Chart  5.  Baseline  air  quality:  the  mean  level  of  particles  found  outside  when  peaks   are  not  occurring.         The  baseline  levels  represent  the  average  background  levels  of  PM2.5,  when  peaks   are  not  occurring.    Since  peaks  occur  about  2-­‐4  times  a  day  and  last  for  20-­‐  25   minutes,  most  of  the  time,  PM  2.5  levels  would  not  be  hazardous  to  residents’  health   but  recurring  exposures  can  have  a  cumulative  impact  on  health.  The  baseline  levels   for  all  homes  are  below  35ug/m3,  the  EPA  regulatory  standard  for  National  Ambient   5   Air  Quality  (NAAQS),  and  are  within  the  “good”  to  “moderate”  levels  of  the  EPA  Air   Quality  Index.     Chart  6.  Accumulated  particle  matter:  the  total  sum  of  particle  counts  during   peak  exposures  above  the  baseline  per  stage  of  well  development.         The  “accumulated  particle  concentration”  (APC)  shows  the  daily  average  of  the   combined  amount  of  peak  levels  for  each  stage  of  well  development.  The  APC   represents  the  additional  exposure  to  PM2.5  above  baseline  levels  (Chart  5).  The   data  shows  that  during  times  of  peak  exposures  (about  2  –  4  times  a  day)  residents   closest  to  the  source  experienced  impacts  well  above  the  NAAQS  regulatory  level  of   35ug/m3,  and  those  further  away  also  exceeded  the  recommended  level.    The   NAAQS  does  not  address  peak  exposures  from  PM2.5.       The  episodic  peak  exposures  can  impact  health.  The  health  effects  caused  by  APC   are  related  to  the  inhaled  dose  of  particles  by  an  individual.  The  inhaled  dose  would   be  approximately  1/3  of  the  accumulated  level.  The  results  found  here  could  induce   respiratory  effects  such  as  asthma  attacks.  Actual  inhaled  doses  for  individuals  vary   widely,  and  depend  upon  a  person’s  overall  health,  age,  and  activity  levels.  Medical   evaluation  is  recommended  to  determine  individual  levels  of  risk.     Summary  of  charts:     For  each  of  the  measured  components,  the  closest  residence  showed  the  highest   impact  for  PM2.5,  except  for  duration  of  peaks.  The  closest  home  experienced  more   peaks,  of  shorter  duration,  and  with  less  time  between  peak  events.  This  fits  with   the  expectation  that  the  source  of  PM2.5  is  closer  to  this  residence  than  it  is  to  the   remaining  four  homes.  The  closest  home  also  maintained  a  higher  baseline  value   and  a  higher  “accumulated  particle  concentration”  for  the  duration  of  the   monitoring  period.       6   Each  of  these  parameters  is  helpful  in  determining  the  potential  exposure  and   associated  health  impact  over  time:     Peaks  per  day:  All  homes  experienced  between  2-­‐4  peaks  per  day,  with  peaks   occurring  more  often  at  the  closest  residence.   Peak  Duration:  Peaks  lasted  between  20-­‐24  minutes  at  the  closest  residence.  Peaks   lasted  20  –  29  minutes  at  the  remaining  four  homes.   Time  between  peaks:  At  the  closest  residence  duration  was  6  to  11+  hours;  duration   was  between  8  -­‐12  hours  at  the  remaining  four  homes.   Baseline  levels:  PM2.5  ranged  from  15  -­‐17  ug/m3  at  the  closest  residence;  the  range   was  6-­‐12  ug/  m3  at  the  further  residences.  The  baseline  levels  are  lower  than  the   EPA  NAAQS  recommendation  for  24-­‐hour  averages  (35ug/m3)  and  are  within  the   “good”  to  “moderate”  levels  of  the  EPA  Air  Quality  Index.  It  is  the  episodic  peak   exposures  that  are  of  greatest  health  concern.   Accumulated  particle  concentration:  At  the  closest  home  levels  ranged  between   4,000  –  9,000  ug/m3/day;  the  range  was  between  2,000  –  4,000  ug/  m3/day  at  the   further  homes.  These  levels  show  that  during  times  of  peak  exposure  residents  are   exposed  to  unhealthy  levels  of  PM2.5.     Summa  Canister  Results     Volatile  Organic  Compounds  as  well  as  PM2.5  are  regularly  emitted  from  industrial   sources  such  as  shale  gas  development  sites.  We  sampled  for  VOCs  to  identify   chemicals  that  would  potentially  travel  with  PM2.5  from  the  well  pad  to  nearby   residences.  Some  VOCs  can  adhere  to  PM2.5  and  have  the  potential  to  adversely   affect  human  health  through  inhalation.     Nine  VOC  samples  were  taken  outside  of  two  residences  between  August  11  –   October  16  2015.    Four  were  taken  at  site  A,  2000  feet  from  the  well  pad  and  five   were  taken  at  site  B,  500  feet  from  the  well  pad  on  the  dates  indicated  in  Table  2.   The  chemicals  detected  are  the  same  chemicals  found  at  numerous  shale  gas   development  facilities  by  other  researchers.       An  assessment  of  VOC  sample  results  based  on  concurrent  weather  conditions   shows  that  wind  conditions  were  good  predictors  of  the  number  of  VOCs  found  in   ambient  air.  Sampling  times  with  calm  or  light  winds  generally  resulted  in  higher   numbers  of  contaminants  in  the  air.  Times  of  higher  wind  speed  resulted  in  fewer   chemical  detections.  These  results  hold  true  for  both  sites  A  and  B.  Sampling   locations  were  situated  southeast  and  east  of  the  well  pad,  such  that  light  west  and   northwest  winds  would  carry  air  contaminants  toward  the  sample  sites.  As  can  be   seen  in  the  tables,  some  samples  were  taken  during  the  reported  “fracking”  stage  of   well  production.  On  these  days  winds  were  variable  both  in  speed  and  direction.         7   Table  2.  Site  A  summa  canister  sample  results,  arranged  by  date  of  sample   initiation.  Industry  reported  activity  and  concurrent  local  weather  conditions  are   included.    Chemical  amounts  in  ug/  m3.  *   Site  and  date  of  24   hour  sample   overnight   Industry  reported   activity   Site  A   8/11/15   Site  A   9/21/15   Site  A   10/6/15   Site  A   10/14/15       fracking     combustion     Meteorological   conditions   Moderate   wind  5-­‐ 10;west  and   north;   scattered  to   broken   clouds   Medium  to   light  wind,   east,  11-­‐0   mph:   overcast,   scattered,   clear  clouds   Light  wind   0-­‐7mph,   North  and   west,   overcast  to   broken   clouds   Winds  3-­‐15   mph;  north,   west  and   south,  broken   to  scattered   clouds             Chemicals   Dichlorodifluoro-­‐ methane  (CFC12)   Trichlorofluoro-­‐ methane    Number  of   Detections   2.2   2.2   2.4   1.7     4   1.1   1.1   1.2   0.95     4   Propene   0.76   2   1.2   1.3   4   Acetone   12   15   8.4   ND     3   Toluene   ND     2.3   1.7   0.92   3   Ethyl  acetate   ND     3.7   3   12   3   Ethanol   8.8   12   9.5   ND     3   n-­‐Butyl  acetate   ND     15   4.3   ND     2   2-­‐propanol   ND   ND   ND   6.9   1   m,  p-­‐Xylenes   ND     ND     2.4   ND     1   o-­‐Xylene   ND     ND     1.1   ND     1   Acrolein   ND     3.4   ND     ND     1   Methylene  chloride   ND     ND     1   ND     1   alpha-­‐Pinene   ND     ND     0.77   ND     1   Trichloroethene   ND     ND     1.1   ND     5   9   13   Total  #  of  chemicals     1   Total  Det.     6   33   *  Site  A    is  2000  feet  from  the  well  pad   8   Table  3.  Site  B  summa  canister  sample  results,  arranged  by  date  of  sample  initiation.   Industry  reported  activity  and  concurrent  local  weather  conditions  are  included.   Chemical  amounts  in  ug/m3.*   Site  and  date   of  24  hour   sample   overnight   Industry   reported  activity   Site  B   9/17/15     Site  B   9/21/15     Site  B   9/26/15     Site  B   10/6/15   Site  B   10/14/15     fracking   fracking   fracking     combustion     Meteorological   conditions   South  wind   mostly   calm;   mostly   clear   East   wind,  7-­‐ 15  mph;   overcast       Light  wind   0-­‐7mph,   north  and   west,   overcast  to   broken   clouds     wind  3-­‐15   mph;  north,   west  and   south,  broken   to  scattered   clouds   Chemicals   Medium  to   light  wind,   east,  11-­‐0   mph:   overcast,   scattered,   clear         Number  of   Detections     Dichlorodifluoro -­‐methane   (CFC12)   Trichlorofluoro-­‐ methane   1.9   2.4   2.2   2.2   1.8   5   1   1.2   1.2   1.2   0.96   5   Acetone   9.7   8.7   65   30   ND     4   Propene       0.8   0.9   1.6   0.84   4   Toluene   3.5   1.1   ND     1.3   1.2   4   Ethyl  acetate   13   3.5   ND     ND     39   3   Ethanol   ND     11   ND     ND     24   2   n-­‐Butyl  acetate   ND     2.9   1   ND     ND     2   3   ND     ND     4.8   ND     2   1.2   ND     ND     2.1   ND     2   1.8   ND     ND     ND     ND     1   Benzene   0.93   ND     ND     ND     ND     1   Ethylbenzene   ND     ND     ND     1.2   ND     1   n-­‐Heptane   ND     0.72   ND     ND     ND     1   n-­‐Hexane   4-­‐Methyl-­‐2-­‐ pentanone   1.9   ND     ND     ND     ND     1   ND     1   ND     ND     ND     1   ND     ND     ND     1.3   ND     1   1.6   ND     ND     ND     ND     11   10   5   9   6   1   Total  #   Det.    41   m,  p-­‐Xylenes   o-­‐Xylene   1,2,4-­‐Trimethyl-­‐ benzene   2-­‐hexanone   Tetrachloro-­‐ ethene   #  of  chemicals   per  sample   *  Site  B  is  500  feet  from  the  well  pad       9   Table  4.  Comparison  of  canister  samples  for  Site  A  and  B,  showing  only  the  results   from  dates  when  both  sites  were  sampled.  Chemical  amounts  in  ug/m3.     Date  and   Industry   reported  activity   Meteorological   conditions   Site   9/21/2015   fracking     Medium  to  light   wind,  east,  11-­‐0   mph:  overcast,   scattered,  clear   clouds   Site  A   Site  B     10/6/2015   10/14/2015   combustion   Light  wind  0-­‐7mph,   north  and  west,   overcast  to  broken   clouds   wind  3-­‐15  mph;   north,  west  and   south,  broken  to   scattered  clouds   Site  A   Site  B   Site  A     Site  B           Chemicals       Dichlorodifluoro -­‐methane   (CFC12)   Trichlorofluoro-­‐ methane   2.2   1.9   2.4   2.2       1.8   1.1   1   1.2   1.2       0.96   Acetone   15   9.7   8.4   30           Propene   2       1.2   1.6       0.84   Toluene   2.3   3.5   1.7   1.3   0.92   1.2   Ethyl  acetate   3.7   13   3       12   39   Ethanol   12       9.5           24   n-­‐Butyl  acetate   15       4.3               m,  p-­‐Xylenes       3   2.4   4.8           o-­‐Xylene   1,2,4-­‐Trimethyl-­‐ benzene       1.2   1.1   2.1               1.8                   Acrolein   3.4                       Benzene       0.93                   Ethylbenzene               1.2           n-­‐Hexane   Methylene   chloride       1.9                           1               alpha-­‐Pinene           0.77               2-­‐hexanone   Tetrachloro-­‐ ethene               1.3               1.6                   Trichloroethene           1.1               #  of  chemicals     9   11   13   9   2   6           10   These  results  show  that  the  number  of  VOCs  in  the  air  at  any  one  time  is  variable   and  that  levels  of  VOCs  also  vary  over  time.  The  24-­‐hour  samples  cannot  show   continuous  changes  in  levels,  as  we  have  for  PM2.5  results.  It  is  assumed  that  VOC   levels  would  also  change  over  time  in  parallel  to  PM2.5.  Some  VOCs  will  adhere  to   PM2.5  and  be  carried  into  the  deep  lung  during  inhalation.     The  levels  of  reported  VOCs  were  not  high  in  terms  of  health  effects  for  a  single   chemical  exposure,  but  are  still  of  concern  if  exposures  occur  continuously  over  a   long  period  of  time  or  if  high  spikes  periodically  occur.  Exposures  to  a  mixture  of   chemicals  may  also  exacerbate  health  effects.         Conclusion     This  project  monitored  air  quality  near  five  residences  located  within  2,500  feet  of  a   new  well  pad  in  Penn  Trafford,  PA.  Monitoring  occurred  while  two  wells  were   drilled,  fracked  and  combusted  in  the  fall  of  2015.       The  results  of  the  study  show  that  the  closest  residence,  500  feet  from  the  well  pad   had  greater  exposures  to  PM2.5  and  to  a  greater  number  of  VOCs.  Peak  exposures  of   PM2.5  at  all  the  residences  occurred  approximately  two  to  four  times  a  day.  Baseline   levels  were  in  the  “good”  to  “moderate”  range  based  on  the  EPA  Air  Quality  Index  for   PM2.5.  The  average  daily  “accumulated  particle  concentration”  of  peaks  for  each   stage  of  monitoring  was  well  above  the  EPA  NAAQS,  suggesting  that  residents  are   likely  exposed  to  short-­‐term  unhealthy  levels  of  air  pollutants  two  to  four  times  a   day.  VOC  sample  results  suggest  that  times  of  low  wind  speed  will  increase  potential   exposures  to  local  air  pollution  by  reducing  the  rate  of  contaminant  dispersion.       It  is  well  understood  that  PM2.5  presents  serious  health  risks  when  levels  are  higher   than  the  regulatory  standard  of  35ug/m3.  When  a  combination  of  other  chemicals,   such  as  VOCs,  are  also  in  the  air  the  risk  of  health  impacts  are  increased  though  less   well  understood.    Based  on  the  study  results,  to  protect  residents’  health,  EHP  highly   recommends  the  use  of  indoor  air  filters  in  homes  located  near  this  site  and  other   similar  industrial  sites.  Because  the  health  risk  to  individuals  varies  widely,   depending  upon  a  person’s  overall  health,  age,  and  activity  levels,  EHP  recommends   that  patients  and  their  primary  care  physicians  discuss  environmental  exposures  to   determine  individual  levels  of  risk.               EHP  Report  Authors   Celia  Lewis,  PhD   Ken  Hamel,  MS   David  Brown,  ScD   11