From turtle2@u.washington.edu Mon Nov  3 09:03:41 1997
Received: from jason01.u.washington.edu (root@jason01.u.washington.edu [140.142.70.24])
          by lists.u.washington.edu (8.8.4+UW97.07/8.8.4+UW97.05) with ESMTP
	  id JAA53970 for <cegrads@lists.u.washington.edu>; Mon, 3 Nov 1997 09:03:39 -0800
Received: from homer14.u.washington.edu (turtle2@homer14.u.washington.edu [140.142.70.15])
          by jason01.u.washington.edu (8.8.4+UW97.07/8.8.4+UW97.05) with ESMTP
	  id JAA32064; Mon, 3 Nov 1997 09:03:37 -0800
Received: from localhost (turtle2@localhost)
          by homer14.u.washington.edu (8.8.4+UW97.07/8.8.4+UW97.04) with SMTP
	  id JAA43974; Mon, 3 Nov 1997 09:03:35 -0800
Date: Mon, 3 Nov 1997 09:03:34 -0800 (PST)
From: Lorna Fitzpatrick <turtle2@u.washington.edu>
Reply-To: Lorna Fitzpatrick <turtle2@u.washington.edu>
To: "C.E. Graduate Students" <cegrads@u.washington.edu>
cc: Mariabeth Silkey <marias@stat.washington.edu>
Subject: NRCSE ENVIRONMENTAL SEMINAR TUESDAY, NOV. 4 (fwd)
Message-ID: <Pine.A41.3.95b.971103085606.75160C-100000@homer14.u.washington.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


    The National Research Center for Statistics and the Environment 

               ENVIRONMENTAL SEMINAR SERIES PRESENTS: 

                       James Hilden-Minton
             National Institute of Statistical Sciences


         "Multilevel Monitoring of Drinking Water Systems"


      Tuesday, November 4th, 3:30 - 5:00, SMITH HALL, Room 211



State and federal laws require drinking water treatment facilities to
sample periodically for specific contaminants and report findings to state
authorities who are responsible for enforcing federal and state standards. 
The regulatory goal of this program is to prevent any state resident from
receiving unsafe drinking water. This goal is not always achieved; for at
least one contaminant the proportion of residents receiving unlawful
levels of contaminant may have been as high as 3% over a recent three-year
period. 

Our multilevel model tracks contaminant levels at individual facilities. 
Sampling is usually sparse at any one facility, but the model allows us
to share information across facilities, strengthening facility-specific
inference.  The model accommodates differing sources of uncertainty and
heterogeneous facility process variance. We are therefore able to quantify
the uncertainty in our population estimates and identify facilities for
which current monitoring is insufficient for reliable inference. The
modeling effort involves more that the determination of empirical Bayes
estimators; it also provided an information tool for regulators to gain
insight into a statewide monitoring program.




(To request disability accommodations, contact the Office of the ADA
Coordinator, at least 10 days in advance of the event.  543-6452(voice);
543-6452 (TDD); 685-3885(FAX);access@u.washington.edu)





.