Using Statistical Modeling to Identify Perturbations in Earth System Processes: Examples from Landscape Evolution and Tectonics
Organizer: Dorothy Merritts
(D_Merritts@acad.FandM.edu)
Franklin and Marshall College

Description:
A fundamental research goal in the Earth sciences is to construct numerical/stochastic models that explain the origin and evolution of common features of the earth's landscape, such as mountain belts and river networks. A further goal is to use those models in diverse applications ranging from estimating past earthquake activity (and future potential) to assessing the relative role of landsliding in hillslope evolution.

In the past there has been relatively little collaboration between geomorphologists and statisticians. In part that was due to the nature of the data available. Relatively small amounts of data collected in the field often provided information without the use of statistical techniques. With increasing amounts of low-cost, digital data available from satellite and other sources, and the need for statistical analysis to obtain useful information from those data, this area is ripe for fruitful collaboration.

Format:
This session will emphasize the collaboration between Earth scientists, mathematicians, and statistical researchers in the study of tectonics, topography, and landscape evolution. Presentations by Earth scientists will provide an introduction to relevant scientific issues and numerical models. Presentations by statisticians will indicate how some of these modeling questions can be addressed through statistical techniques. Two poster sessions---one by a statistician and one by an applied mathematician---will highlight related research problems that involve the use of topographic data.

Participants:
Peter Dodds and Daniel Rothman (presentation, River Network Scaling Laws: Deviations and Fluctuations)
Peter Dodds, Graduate Student in Applied Mathematics at the Massacusetts Institute of Technology, is a researcher interested in applications of mathematical and statistical techniques in networks (especially river systems), biology and ecology. This presentation will address the origin and importance of fluctuations and deviations in river network scaling laws.

Garry Willgoose and Greg Hancock (presentation, Quantitative Testing of Landform Evolution Models)
Garry Willgoose, Ph. D. is an Associate Professor in The Department of Civil, Surveying and Environmental Engineering at The University of Newcastle, Australia. His interests are landform evolution, soils and vegetation modelling applied to landscape geomorphology and mine site rehabilitation design. His lecture will present recent work testing his landform evolution model against experimental landform simulator and field data. Some challenges in quantitative hypothesis testing in these spatially organised and chaotic systems will be discussed.

Tim C. Hesterberg, MathSoft, Inc. and Dorothy Merritts, Franklin and Marshall College (presentation, Deviations in Slope-Area Relations that Indicate Geologically Recent Crustal Deformation) Tim Hesterberg, Ph.D. in Statistics, is a Research Scientist at MathSoft, Inc. He designs algorithms and software for bootstrap inference and diagnostics, nonparametric surface estimation, missing data, and sequential designs for clinical trials. His lecture will present estimates of co-seismic ground deformation from analysis of stream topography. (www.statsci.com/Hesterberg)

Derek Stanford (poster, Conjugate Gradient Methods for Large-Scale Sparse Regression, with Applications to Seismic Deformation Estimation)
Derek Stanford, Ph.D. in Statistics, is a Research Scientist at MathSoft, Inc. His research interests are statistical image analysis, methods for large databases, and automatic model selection. His poster presents conjugate gradient methods for large-scale sparse regression, with particular applicatioin to estimating deformation from seismic events.