SSES
Overview
Spatial
Environmental
Research
Preprints
Teaching
Web-Projects
Events
People
Archive
Links
THE SSES PROGRAM
Overview


The Program in Spatial Statistics and Environmental Statistics is involved in teaching, research, and science collaboration. Graduate courses in Spatial Statistics (MS and PhD level) and Environmental Statistics (MS level) have been established, and Brown Bag Seminars, discussion groups, student projects, and dissertations make up the rest of the teaching component. Regarding research, the Program has emphasized computational aspects of spatial and spatio-temporal statistics applied to areas of "big science," such as remote sensing of the earth on a global scale, regional climate modeling in space and time, and Bayesian statistical exposure modeling from sources to biomarkers. Other research areas include spatial command and control, disease mapping, medical imaging, ice-stream dynamics, and spatial data fusion. Science collaboration has been fostered through SSES-organized interdisciplinary seminars and personal contacts with scientists and engineers at OSU and elsewhere. The SSES Program's expertise is also available to OSU faculty and staff for consulting on an occasional basis, in areas that are substantially environmental or spatial.

Director
Noel Cressie, 408B Cockins Hall, 614-292-5194 ( sses@stat.osu.edu)

Teaching

Introduction to Spatial Statistics (3-credit MS-level course): Stat 631
The course will be taught next in the Spring 2011 quarter. The lectures present an introduction to statistical methods for geostatistical data, regional data, and spatial point patterns. Theoretical properties are illustrated with simulation, and applications are emphasized in take-home tutorials.

Spatial Statistics (3-credit PhD-level course): Stat 829
The course is taught every other Spring quarter (taught next in the Spring 2012 quarter). The lectures present topics that include exploratory spatial data analysis, (multivariate) spatial prediction, spatial hierarchical modeling (empirical Bayesian and fully Bayesian), and the incorporation of a temporal component in spatial models.

Environmental Statistics (3-credit MS-level course): Stat 662
The course is taught every other Spring quarter (taught next in the Spring 2012 quarter). The lectures present topics that include standard statistics used in environmental settings, sampling design, causality, limits of detection, toxicology, risk analysis, time series, spatial statistics, and hierarchical modeling.

Discussions In Spatial and Environmental Statistics (DISES)
The SSES Program organizes DISES on an occasional basis. It consists of Brown Bag Seminars and of discussion groups in spatial, spatio-temporal, or environmental statistics. If you would like more information, please e-mail us at sses@stat.osu.edu.

Tutorial on Bayesian Statistics for Geophysicists
A web-based instructional tool, Tutorial on Bayesian Statistics for Geophysicists, has been developed in conjunction with members of the Byrd Polar Research Center; it is part of the Web-Project, Ice Streams.

Tutorial on Hierarchical Bayesian Modeling for Exposure to Arsenic
A web-based instructional tool, Tutorial on Hierarchical Bayesian Modeling for Exposure to Arsenic, has been developed in conjunction with a research project joint with Battelle; it is part of the Web-Project, Sources to Biomarkers.

Tutorial on Fixed Rank Kriging (FRK) of CO2 Data
This instructional tool is part of the Web-Project, CO2.

Research
A statistical approach to uncertainty in science can be found in: Cressie, N., Calder, C., Clark, J., Ver Hoef, J., and Wikle, C. (2008). "Accounting for uncertainty in ecological analysis: The strengths and limitations of hierarchical statistical modeling (with discussion)". Ecological Applications, 19, 553-570.

Some of the science and engineering problems to which this approach has been applied are given below.

Remote Sensing
Enormous amounts of time and money have been spent on obtaining multi-resolution satellite data with pinpoint accuracy. Data-production algorithms currently use rudimentary statistical methods, such as the sample mean and the sample variance, to interpolate these "level 2" data (irregularly distributed in space and time), in order to obtain regularly distributed "level 3" data. In doing so, the obvious spatial and temporal dependencies present in remote-sensing data gathered from satellites have been largely ignored. We address this problem through an empirical Bayesian methodology that yields a dynamic, spatial, change-of-resolution Kalman filter. The methodology has been applied to Total Column Ozone (TCO) data. Further details are available on the Web-Projects page; see TCO. A more recent empirical Bayesian methodology, called Fixed Rank Kriging, is being applied to large remote-sensing datasets in research funded by the National Aeronautics and Space Administration (NASA). Further details are available on the Web-Projects page; see CO2. Bayesian versions of this methodology are also being developed.

Geophysical Modeling
The dramatic world-wide shifts in weather patterns and associated societal impacts caused by the 1997-98 El Nino brought unprecedented attention to the role of inter-annual climate variability in our daily lives. Climate processes exhibit substantial spatio-temporal variability that is often non-stationary, non-linear, and dependent. Although it can be quite difficult to model such variability with traditional spatio-temporal approaches, we have demonstrated that it is possible using highly structured (hierarchical), Bayesian, spatial-temporal models. Physical knowledge of the system is used wherever possible in constructing the models, resulting in what has been called physical statistical modeling. Two projects that take this approach are the study of sea surface temperature in the equatorial Pacific Ocean in order to forecast El Nino Southern Oscillation (ENSO) events, and a study of ice-stream dynamics in Antarctica and Greenland. Further details are available on the Web-Projects page; see ENSO and Ice Streams, respectively.

Human Exposure to Metals
The project, "From Sources to Biomarkers: A Hierarchical Bayesian Approach for Human Exposure Modeling," seeks to characterize multi-pollutant (arsenic, lead, cadmium, and chromium) human exposures by linking sources to biomarkers using a multi-scale hierarchical Bayesian statistical model. Further details are available on the Web-Projects page; see STB.

Collaboration
Part of the SSES Program's mission is to collaborate with The Ohio State University's spatial and environmental science/engineering communities; for example, physical statistical modeling of ice-stream dynamics in Greenland was done in collaboration with scientists from the Byrd Polar Research Center at OSU. The SSES Program's expertise is also available to OSU faculty and staff for consulting on an occasional basis, in areas that are substantially environmental or spatial. The SSES Program organizes an occasional themed interdisciplinary seminar series (e.g., environmental data, remote-sensing data, environmental exposure and health data, statistics and climate change). If you would like more information, please e-mail us at sses@stat.osu.edu.

Support
Support for the Program comes from several sources. The first is a 1998 Academic Enrichment Award granted to the Department of Statistics as the result of a competition across all units of The Ohio State University. The Department renovated space in Cockins Hall to house the SSES Program and appointed a Director. An Assistant Director helps administer the Program.

Past and current extramural support has come from the Environmental Protection Agency, the National Science Foundation, the Office of Naval Research, the National Aeronautics and Space Administration, the American Chemistry Council, and the National Center for Atmospheric Research.

Location
The Program's Administrative Office is in 408 Cockins Hall, adjacent to the Main Office of the Department of Statistics.

The SSES Program
Department of Statistics
408 Cockins Hall
1958 Neil Avenue
Columbus, OH 43210-1247
e-mail: sses@stat.osu.edu
telephone: 614-292-5194
fax: 614-292-2096


Photo Gallery
Photo Gallery