United States
Environmental Protection
Agency	
Environmental Monitoring
Systems Laboratory
Las Vegas. NV 89193-3478
Research and Development
EPA/600/S4-88/040 May 1989
 Project  Summary
 Evaluation of Control Chart
 Methodologies for  RCRA  Waste
 Sites
 Thomas H. Starks
  This report is a discussion of
decision rules  relating  to  the
monitoring  of  ground  water at
hazardous  waste sites that  are
subject to  regulation under  the
Resource Conservation and Recovery
Act of 1976 (RCRA). The final rule for
RCRA regulations 40CFR part 264 was
published  October  11, 1988
(53FR39720). Understanding  the
complexity of the monitoring problem
and the diversity of RCRA sites, the
final rule  wisely   allows  the
owner/operator  to   choose,
conditioned  on EPA approval, a site-
specific "statistical  procedure."
Analysls-of-varlance,  tolerance
intervals, prediction  intervals,  and
control charts  are   included as
acceptable methods for "statistical
procedures." These  methods  are
discussed to facilitate  the choice of
decision rules. A nested random-
effects  model for ground-water
quality  parameter  measurement is
suggested and decision procedures
are developed In terms of that model.
Particular attention is  paid  to the
possible application  of  industrial
quality control  strategies  to  the
ground-water monitoring problem. A
decision procedure that changes
over  time as more information about
well  and aquifer characteristics
accumulate is  proposed. This
procedure involves the use of outlier
tests and  of  Shewhart-CUSUM
quality control strategies.
   This Project  Summary was
developed by EPA's Environmental
Monitoring Systems Laboratory,  Las
Vegas , NV, to announce key  findings
of the  research  project that Is fully
documented In a separate report of
the same  title (see  Project Report
ordering Information at back).

Introduction
   Under the Resource Conservation and
Recovery Act of 1976 (RCRA), the U.S.
Environmental  Protection Agency  has
developed  regulations for landfills,
surface  impoundments, waste piles, and
land treatment  units that are used to
treat, store, or  dispose of  hazardous
wastes. The  regulations include
requirements for the monitoring  of
ground water in  the top aquifer below the
hazardous  waste site (HWS).  This
monitoring  involves  the  drilling  of
background  well(s) and compliance wells
at the  HWS,  and the sampling and
analysis of  well water  at regular time
intervals to help  determine  whether
leachate from the  HWS has  entered the
aquifer. There are  several as  yet
unsolved problems in this  monitoring
program. They  include determination of
appropriate methods for obtaining
accurate measurements of  some
constituents such  as volatile organics,
specifications  for well construction,
detection and accommodation of shifting
direction and rate  of aquifer flow, and
development of good decision rules
based  on  measurements  of  water
samples drawn from wells near the HWS
for determining when  additional
regulatory action may be required. This
paper  discusses the  problem  of
developing  good  decision  rules and
recommends that  the development be
based on a realistic  model for  the
ground-water measurements. A nested

-------
random-effects model is suggested and
slatisticaf  procedures  based  on  that
model  are formulated  and  criticized.
Industrial quality control strategies are
considered in  terms of their  possible
application  to  the  ground-water
monitoring decision problem,

Procedure
   During  the first year  this project
developed the appropriate components
of variation   model  for the RCRA
ground-water  tesi problem.  Such  a
model is the evaluation  criterion for any
proposed RCRA ground water test. The
second year the implicit variance models
assumed by the various  proposed test
strategies  for  the  RCRA  problem  were
explicitly derived  and  compared.  The
third year the most  promising  test
procedure,  control  chart  strategy,  was
evaluated against  simulated   data
representing the most  frequent RCRA
data  problems and different values of
parameters critical to the test procedure.
   Control  chart strategy is evaluated on
simulated   "reaJ  world"  data.  Several
available sets of "real world" data were
examined for evaluating proposed RCRA
statistical decision procedures.  The sets
an had one flaw for this use  in that the
state of the site (in-control or  no leak,
out-gf-contfol  or  leak) was not known
or recorded.  Thus only  simulated  data
sets  could represent the desired  state
and  relevant RCRA-type  probiems
(multiple wetls, correlated samples, etc.).
The contro! chart strategy was evaluated
for two factors related to  the algorithms
and  five  factors  rotated to  frequent
statistical problems with environmental
monitoring data.  Each factor   was
simulated  for both states (in-control,
out-of-control).  The factors  evaluated
in the  design  of the simulation
experiments  were  (1) parameter
estimates  and  (2)  length of  learning
period for  the Snewhart-CUSUM control
chart strategy and  RCRA-type  data
complications such as: (3) multiple  wells,
(4) correlated  samples.  (5}  negatively
skewed  data (e.g., data overcorrected by
transformation), (6) positively skewed
data (e.g., monitoring  data  is   often
positively  skewed,  requiring trans-
formation), and (7)  multiple  ground-
water quality parameter data. The criteria
of evaluation are  long  average-run-
length for  the in-control state and  short
average-run-length for  the out-of-
control state.

Summary and  Conclusions
   All statistical decision procedures are
based  on assumed  measurement
models. Decision  procedures based on
unrealistic models will  not succeed  in
providing answers  to ground-water
monitoring decision problems, no matter
how simple  or elegant  the  procedures
may be. It  is essential that  a  realistic
workable model for the measurements
be  formulated  and  used both  in
construction  and  evaluation  of decision
procedure.  A nested  random-effects
model  is presented to illustrate a model
approach and to  indicate the difficulties
inherent in  developing  good statistical
procedures   for monitoring  of ground-
water quality. Obviously,  no statistical
measurement  mode!  is  as complex as
the  system in nature, but the  model for a
decision  procedure  should  be as
reasonable and as simple as possible.
  Any  decision  procedure,  based on
measurements of the  quality of the
ground  water taken  in  each sampling
period, where decisions are made at the
end of each sampling period as  to
whether of  not  additional  regulatory
actions  are  required, is by  definition a
quality control strategy. In addition,  for a
quality control strategy, one is interested
in the distribution  of run lengths in both
in-control   and  out-of-control
situations.  That  is,  a  good decision
procedure {quality control  strategy)  is
one with large average in-control run
lengths  and  small  average  out-of-
control run lengths. Hence, consideration
in  comparing  decision  procedures for
RCRA  sites should  be given to the
distributions  of their  run lengths rather
than to their probabilities of  Type I and
Type (I errors on  individual applications
of  the decision rule  in each sampling
period. (However,  the two types  of
criteria  are  obviously  not  unrelated).  In
choosing a  quality contro! scheme for
use at  RCRA sites it  is reasonable  to
consider quality control schemes that
have  been  used  successfully in  other
settings, particularly in industrial settings,
  The  formulation of  good decision
procedures for   determining  when
increased  monitoring activity  is needed
at   a  hazardous  waste site  (HWS)  is
extremely difficult because of the  high
cost, slow acquisition, low precision, and
multfvariate nature  of ground-water
monitoring  data  along with  system
instability due to intrusions  on the aquifer
caused by man outside the HWS. The
first three of these problems force the
initiation ol  quality  control  strategies
before good (highly precise) estimates of
measurement distribution  parameters
can be obtained. With good estimates of
ths  measurement distribution  param-
eters,  it is  possible  to mathematically
derive  the  run-length  distributions f
various  quality  control  strategie
However, without these good estimates,
is  necessary  to  employ  Monte  Car
techniques  to  estimate  the distribute
properties  of  run  lengths  when  tr
process is  in-control  (i.e., site is  n
leaking into aquifer and  plume is passir
through one or more  weii sites). Tr
Monte  Carlo analysis  of the  Shewhar
CUSUM  quality  control  technique
indicates  the type of results that can  t
obtained  with  this  method  and  ah
indicates  that the  method  is reasonab
robust with respect  to  left-skewei
non-normal probability  distributions  •
measurements.  However, the techniqi
is  not  robust  with  respect  to  lack  >
independence  between  measurement
in  particular,  its  in-control  run-lengf
characteristics  are shortened by positK
serial correlations.
   The Monte  Carlo simulation resul
and  methods  discussed in  this repo
provide a  basis  for comparison an
evaluation of  all other  decisio
procedures,  since similar  Monte  Car
simulations can be  performed  on ar
decision procedure to obtain estimates «
the run length distribution of such pn
cedures.

-------
  Thomas H. Starks is with the University of Nevada, Las Vegas. NV89154.
  George T. Flatman is the EPA Project Officer (see below).
  The complete report, entitled "Evaluation  of Control Chart Methodologies for
       RCRA Waste Sites" (Order No. PB 89-138 4161 AS; Cost: $13.95, subject
       to change) will be available only from:
           National Technical Information Service
           5285 Port Royal Road
           Springfield, VA 22161
           Telephone:  703-487-4650
  The EPA Project Officer can be contacted at:
           Environmental Monitoring Systems Laboratory
           U.S. Environmental Protection Agency
           Las Vegas, NV 89193-3478
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
                                                               t
-------