American Statistical Association (ASA) Section on Statistics and the Environment (ENVR) Last Modified: 2000-Oct-14 Send comments, corrections, additions to: ENVR Web Editor or ENVR Webmaster

It was a pleasure to meet and greet so many of you at the JointStatistical Meetings in Indianapolis. I trust you had as pleasant andinformative time as I did. As a byproduct of that meeting, it is hardfor me to watch the Indianapolis Colts play football in the RCA Domewithout immediately thinking of the statistical meetings in the adjacentIndiana Convention Center. However, I am enough of a realist torecognize that the probability is quite small that anyone else isthinking of this at third and long yardage.

You will note from the article below that we are off and runningpreparing for the 2001 JSM in Atlanta. There is also a report from theAccuracy conference this past July in Amsterdam. Note that the articleon accuracy, Amsterdam and Atlanta just covers the A’s. Think ofwhat would happen if we had more space!

Woollcott Smith, wksmith@unix.temple.edu, is helping organize aspecial contributed paper session on "Species diversity in randomenvironments: models, estimation and sampling design" at the JointStatistical Meetings in Atlanta next August. It is hoped that thissession will attract papers on new technical approaches to the analysisof species diversity. If you are working in this area, we invite yourparticipation in the session. Please notify Woollcott or TimothyGregoire, Timothy.Gregoire@Yale.edu, if you would like furtherinformation or would like to participate.

The 4^th International Symposium on Spatial AccuracyAssessment in Natural Resources and Environmental Sciences (Accuracy2000) was held July 12-14, 2000, in Amsterdam, The Netherlands. Themeeting attracted over 160 participants and 126 papers. The three-dayevent presented the state-of-the-science on the assessment, modeling,visualization, and propagation of uncertainty in spatial data andspatial process models. The aim of the symposium was "Within theenvironmental sciences - whether it be hydrology, soil science,geomorphology, oceanography, forestry, climatology, geo-ecology or anyother branch that one may think of - much use is made of spatiallydistributed data. These data, frequently stored as maps in a spatialdata base or Geographical Information System (GIS), are rarely, if ever,truly free of error. Furthermore, these data are often used as inputs tovarious models of environmental processes that are themselves alsosubject to uncertainty. Spatial statistics and stochastic systems theoryoffer methodologies to handle these uncertainties, but the successfulimplementation and application of these methods requires the jointefforts of scientists from various disciplines." The meetingattracted experts from environmental science, spatial statistics, geographic information science, and other scientists who wish to furtherdevelop theory and practical application of handling spatialuncertainty.

The meeting builds on the success of previous symposia inWilliamsburg, Virginia (1994), Fort Collins, Colorado (1996) and QuebecCity, Canada (1998). The organizing committee was chaired by GeraldHeuvelink from the University of Amsterdam and co-chaired by PeterBurrough of the University of Utrecht. Organizations co-sponsoring thesymposium were Netherlands Centre for Geo-ecological Research,University of Amsterdam, International Union of Soil Science WorkingGroup on Pedometrics, Bernoulli Society, International StatisticalInstitute, International Society for Photogrammetry and Remote SensingGIS Division, National Center for Geographic Information and Analysis,IUFRO Commission 4, and The International Environmetrics Society.

The meeting featured three plenary speakers. Giles Foody, Departmentof Geography, University of Southhampton, discussed the accuracy ofthematic maps derived from remote sensing. He focused on seven broadproblem areas that limit the ability to document and use the accuracy ofsuch thematic maps. His seven areas were: accuracy measure andreporting, sampling, accuracy of ground or reference data, type oferror, error distribution, error magnitude, and use of the confusionmatrix. Keith Loague, Department of Geological and EnvironmentalSciences, Stanford University, reviewed environmental modeling: thepast, the present and the future. The underlying focus of his talk wasthe characterization of accuracy and the impact of uncertainty inenvironmental modeling. He believed that "environmental modelswill not reach their potential in the decision management arena until(i) rigorous nonsubjective model performance criteria are established,and (ii) uncertainty impacts, related to both model and data errors, areincorporated into simulation protocols." Arnold Heemink, Facultyon Information Technology and Systems, Delft University of Technology,discussed modeling and prediction of environmental data in space andtime using Kalman filtering. He viewed the Kalman filter as a frameworkfor space time data analysis to overcome some of the problems in theanalysis of spatio-temporal data. He illustrated the approach with anexample of tropospheric ozone over Europe.

A Scientific Advisory Board actively participated in organizinginvited paper sessions on digital elevation models, change detection,uncertainty propagation in GIS, spatial statistics, scale issues inenvironmental monitoring, managing and visualizing spatial uncertainty,alternative methods for spatial uncertainty assessment, data quality anduncertainty propagation, object-related uncertainty, handling spatialvariability in forestry, kriging, assessing the validity ofphysically-based models, remote sensing, spatial sampling theory,uncertainty analysis in distributed modeling, development oferror-sensitive GIS, natural resources mapping, and handling temporaland spatial variation in the soil.

A book of symposium proceedings is available: