[ Order This Book ]

More Info:


Table of Contents

Statistical Procedures of Environmental Monitoring Data and Risk Assessment
Author: McBean, Edward A. / Rovers, Frank

Cover: Hard cover
List Price: $76.00
Published by Prentice Hall
Date Published: 06/1998
ISBN: 0136750184


The intent in this book is to carefully explain features of various statistical analysis procedures. Unlike much of the professional literature, however, this text makes special effort to describe statistical techniques in terms comprehensible to the nonstatistician. This is accomplished by downplaying mathematical notation, comprehensively explaining the development of equations, and emphasizing example applications. Thus, as successive problems of environmental-monitoring interpretation are developed, the text describes through use of simple examples how each procedure is utilized. References are provided, with particular emphasis on works describing applications reported in the technical literature. Problems included at the end of the chapters stress fundamentals and increase the usefulness of this book as a classroom text. The collection and laboratory analyses of samples needed to characterize environmental quality are already expensive. Further, as society expresses increasing concern for environmental protection and as instrumentation technology can detect contaminants at ever lower concentrations, expenditures for monitoring environmental quality will rise with time.

As a direct consequence of the rising costs of environmental monitoring, it is essential to use available environmental quality data effectively. Effective utilization involves answering questions such as, "Is the environmental quality acceptable?" and, "Is the environmental quality improving or deteriorating?" Responding to these types of questions requires interpretation of data, and this stage of assessment is beset with difficulties. Some difficulties with interpreting environmental-monitoring results:

    (i) Since the data are frequently expensive to accumulate, the data sets being interpreted are usually very modest in size.

    (ii) The data may involve a vector of chemical and biological constituent measurements because consideration must typically be given to a range of constituents. Correlation between the constituents may help the infilling of missing data or the identification of outlier data.

    (iii) Early detection of any deterioration in environmental quality is highly desirable because early detection may provide the opportunity for controlling the problem at a lower cost before the problem magnifies. Any procedure for identifying early warning signals must not, however, falsely identify a problem of apparent environmental deterioration when one does not actually exist; nor should it fail to identify a problem when one does exist.

    (iv) The vagaries of nature introduce significant noise and sources of variability such as seasonality effects. This can make the identification of trends more difficult.

    (v) The derivation of quantitative risk assessments is in many ways data dependent. But will the information returned by these risk estimates be worth additional data collection efforts?

The net result of difficulties such as the five mentioned is that making sense of environmental quality data necessarily involves statistical interpretation. Statistical interpretation procedures must be sensitive to small changes in environmental quality and yet recognize the potentially substantial costs of any additional data collection requirement. The need for the statistical interpretation of environmental quality data is widespread. The range of concerns for each environmental media-air quality, surface water quality, groundwater quality, and soil contamination-are similar in many respects. Yet there is no single statistical analysis procedure universally applicable to the variety of problems associated with environmental quality data. Instead, the practitioner needs to have an array of statistical procedures available. A multitude of statistical analysis tests are available, but each of the tests possess assumptions that may or may not be appropriate for specific circumstances. Computer programs now becoming widely available facilitate use of various procedures. The difficulty remains for the student and the practitioner to learn which conditions dictate a particular procedure and which conditions render it highly inappropriate.
Following the introduction (Chapter 1), the book is organized into four parts as follows:

  • Part I Chapters 2 through 5 develop the fundamental measures used to describe data and the distributions employed to describe the data.
  • Part II Chapters 6 and 7 describe procedures commonly utilized to detect changes occurring over time, the detection of outliers and the mathematical procedures for quantifying coincidental behavior in data sets.
  • Part III Chapters 8 through 11 describe the bases used in hypothesis testing to determine when there are differences in environmental quality at various locations. Problems of censored data are considered as they influence the utilization of alternative tests.
  • Part IV Chapter 12 focuses on the interrelationships between risk assessment and the data upon which the risk characterization procedures rely. Simulation procedures for risk characterization using sampling methodologies from probability distributions of data are described.

Authors of the Book
Edward A. McBean (B.A.Sc. from the University of British Columbia and S.M., C.E., and Ph.D. from the Massachusetts Institute of Technology) is an associate of Conestoga-Rovers and Associates and president of CRA Engineering Inc. Dr. McBean's experience includes more than 20 years as a faculty member at the University of Waterloo and the University of California. Much of the focus of Dr. McBean's research and professional work has been on the specific problem of interpretating environmental quality data. He is the author of more than 300 technical articles and has authored or edited eight books. Frank A. Rovers (B.A.Sc. and M.A.Sc. from the University of Waterloo) is president of Conestoga-Rovers and Associates, an environmental engineering company with more than 850 employees located in 29 offices. Mr. Rovers has been involved for more than 25 years in a very large number of environmental engineering problems dealing with the complete spectrum of environmental quality issues. Frank is the author of numerous technical journal articles dealing with the interpretation of environmental quality data.
Both authors have been heavily involved in the teaching of professional development courses, including those at the University of Wisconsin-Madison, University of Toronto, Nova Scotia Technical College, and UCLA.

We are under no delusion that the work reported in this book is just our work. Clearly the material is the product of many people's efforts. Our intent was to assemble and organize the considerable range of experience and understanding culled from literature about statistical evaluation of environmental quality data.
In addition to the literature, we have drawn upon the experience and efforts of many individuals, and for this assistance we are grateful. During the years preceding publication, the authors worked closely with many colleagues and students, among them:

  • The employees of CRA who so generously provided examples. The advice and assistance of many is acknowledged, with special mention of John Donald, Klaus Schmidtke, Darrell O'Donnell, Mark Schwark, and Wes Dyck.
  • All the people who examined drafts of the book and whose comments for improvements were valuable. In particular, useful comments by Bill Lennox, Aditya Tyagi, and many students are gratefully acknowledged.
  • The secretarial staff at CRA who so obligingly "revised the last revision." In this respect, special acknowledgement must be given to Maria Manoli, who continued to remain cheerful in the face of numerous rewrites.
  • Melissa McBean, whose preparation of figures for this book is also gratefully acknowledged.

To all of the above, we owe our sincere thanks for their assistance. In an undertaking of the magnitude of this text, it is not possible to avoid errors, and for this we apologize in advance. Any corrections, criticisms, or suggestions for improvements will be greatly appreciated by the authors. We would also welcome any additional information and data that would make future editions of the book more complete.

Edward A. McBean
Frank A. Rovers