Statistical Methods: Landscape Survival Analysis (LSA)

Despite high dimensionality in both space and time, common methods and techniques fall short of effectively characterizing the temporal dimension of many landscape processes. One aim of the CHES group is to develop models and metrics that help shed light upon the temporal patterns of landscape changes. Inspired by the beauty and power of traditional survival analysis (see the note below), we have started a journey to "borrow" and extend survival analysis to landscape analysis/modeling and spatial sciences, generating a unique research frontier named landscape survival analysis (LSA; a term by Li An).

In LSA, the event of interest is development or change of a certain landscape unit from one status to another. Our efforts in LSA are characterized by (1) extending basic concepts in survival analysis—hazard function, survival probability—to landscape change applications, (2) using survival metrics to describe temporal patterns that are not easily detected by other methods, and (3) applying survival analysis methods to disclose effects of varying temporal patterns and uncertainties. Below is a list of exemplar, stimulating articles that illustrate the uniqueness and power of LSA and generic survival analysis.

Readings and References:

Allison, P.D. (1995). Survival Analysis Using SAS?: A Practical Guide. Cary, NC: SAS Institute Inc.

An, L., D.G. Brown, J.I. Nassauer, and B. Low (2011). Variations in development of exurban residential landscapes: timing, location, and driving forces. Journal of Land Use Science 6(1):13-32.

An, L., and D.G. Brown (2008). Survival analysis in land-change science: integrating with GIScience to address temporal complexities. Annals of Association of American Geographers 98(2):323-344.

Irwin, E., and J. Geoghegan (2001). Theory, data, methods: developing spatially explicit economic models of land use change. Agriculture, Ecosystems & Environment 85 (1-3): 7-23.

Klein, J.P., and M.L. Moeschberger (1997). Survival Analysis: Techniques for Censored and Truncated Data. New York. Springer-Verlag.

Vance C., and J. Geoghegan (2002). Temporal and spatial modeling of tropical deforestation: a survival analysis linking satellite and household survey data. Agricultural Economics 27 (2002): 317-332.

Note: Traditional survival analysis is a collection of statistical methods used to characterize the occurrence and timing of events that represent any qualitative changes in state during some time-series processes, such as birth, divorce, or failure of some equipment. Survival analysis is also termed event history analysis in sociology, reliability or failure time analysis in engineering, and duration or transition analysis in economics.