American Statistical Association (ASA) Section on Statistics and the Environment (ENVR) Last Modified: 2001-Apr-29 Send comments, corrections, additions to: ENVR Web Editor or ENVR Webmaster

In this month's ENVR section news, we first highlight a call for ideas forJSM 2002 invited sessions. In addition, we highlight the statisticalactivities of one of our section members. This is the first in what wehope will be a series of "stuff we do" in environmental statistics pieces.If you have any interest in contributing to this series, please send me anote (ajbailer@muohio.edu).

It is time to develop proposals for invited sessions for the 2002 JSMmeetings in New York. Please send complete proposals (with theme, speakersand titles in place) to me (Jay Ver Hoef) as soon as possible, in time forthe 2002 Program Committee's initial meeting at the 2001 JSM in Atlanta.The ENVR Section will have at least three invited sessions, but we cancompete for additional slots.

An invited session can take one of many forms. Examples include: a sessioncontaining one or more paper presentations, perhaps followed bydiscussions; a panel discussion; a tutorial; a workshop; a demonstration; amock trial; or a debate. The most popular type of invited session is aseries of presentations of papers and (possibly) discussions, which mayinclude up to four presenters (including the discussants) plus a chair.The second most popular is a panel discussion, which may include up to fivepresenters plus a chair. Other formats are possible as well, and mayinclude up to five participants plus a chair. Typical formats consist of 3speakers giving 25 minute talks and one discussant with 15 minutes at theend; or 3 speakers giving 30 minute talks; or 4 speakers giving 25 minutetalks. There is usually a little time at the end for floor discussion. Afull session lasts 1 hour and 50 minutes.

Please keep in mind that the more organized proposals will have betterchances competing for the limited number of slots, so your ideas will haveto be in final form by July at the latest. Also keep in mind that everyJSM participant is entitled to only one appearance as a speaker or formaldiscussant.

I look forward to working with you so that we may present interestingsessions for our members and all ASA members. Please contact me if youhave any questions or suggestions.

The ENVR and its members are involved in a broad range of statisticalissues and activities linked by a common focus on the development andapplication of statistical methods to environmental research and policy.The activities performed under the general heading of "environmentalstatistics" can be quite diverse. This is the first of a series ofarticles that will describe specific examples of the contributions ofenvironmental statisticians.

The Pacific Northwest National Laboratory (PNNL) is operated by BattelleMemorial Institute for the U.S. Department of Energy. Since its inceptionin 1965 there has been a contingent of statisticians involved withenvironmental applications. In this article two of their currentenvironmental statistics activities will be described. The first involvesdeveloping user-friendly software, Visual Sample Plan (VSP), that allowsthe user to generate statistically sound environmental sampling plans intwo-dimensions through a visual interface. The second is an effortunderway to develop methodology useful in efforts to delineate the locationof unexploded ordnance (UXO).

Visual Sample Plan (VSP) is an effort to make statistical planningmethodology for the purposes of environmental characterization easilyavailable to non-statisticians . It strives to put the U.S. EnvironmentalProtection Agency's Data Quality Objectives (DQO) Process intouser-friendly software. VSP does not analyze data but aids samplingpersonnel in developing statistically defensible sampling plans when theyare not entirely familiar with the statistical underpinnings of the DQOprocess and do not have a statistician available. The user enters a map ofthe site they are interested in and designates the general areas they needto sample and characterize. The user then selects from a menu of projectobjectives that describe the intent of their sampling effort. Thecurrently available objectives are: comparing central tendency to athreshold, comparing the central tendency of one population with that ofanother, and looking for hot spots. Once the objectives have been selectedthe user needs to enter the performance criteria for the sampling effort,i.e., how good does the answer to the question they are asking need to be.Given this information VSP proceeds to answer the ever present question of"How many samples do I need?". VSP then generates the following information:

• the map of the site with the sampling locations plotted on it,
• a text file that contains the geographic coordinates of the sampling locations,
• a Decision Performance Goal Diagram describing the user-specifiedperformance of the sample,
• a report of the input to and output from VSP to document the sampling plan.

Unexploded Ordnance (UXO) in soils at many Formerly-Used Defense Sites(FUDS) that are slated for clean-up and transfer to the public for otheruses are a significant concern. The risks associated with UXO at thesesites must be managed to acceptable levels commensurate with hazardousclean-up activities and intended future use. There are two types of risksthat must be considered:

1. Risk of unidentified UXO that could be inadvertently discovered duringor after clean-up resulting in hazards and lawsuits
2. Risk of spending too many tax dollars on characterization with anon-optimal characterization scheme or when there is no real threat present.

The Strategic Environmental Research and Development Program (SERDP) is theDepartment of Defense's (DoD) corporate environmental research anddevelopment (R&D) program, planned and executed in partnership with theDepartment of Energy (DOE) and the Environmental Protection Agency (EPA).SERDP is presently funding three related efforts that will evaluate anddevelop statistical methods and tools that can be used for characterizationand verification plans and data evaluation schemes that will appropriatelymanage the two types of risks mentioned above. These research efforts arebeing conducted by statisticians at three DOE National Laboratories asfollows:

Statistical Methods and Tools for UXO Site Characterization

Pacific Northwest National Laboratory

Bayesian Approach to UXO Site Characterization with Incorporation ofGeophysical Information

Sandia National Laboratory

Spatial Statistical Models and Optimal Survey Design for RapidGeophysical Characterization of UXO Sites

Oak Ridge National Laboratory.