Title: Survival analysis for a large scale forest health issue: Missouri oak decline

Author: Woodall, C.W.; Grambsch, P.L.; Thomas, W.; Moser, W.K.

Year: 2005

Publication: Environmental Monitoring and Assessment (2005) 108: 295-307

Abstract: Survival analysis methodologies provide novel approaches for forest mortality analysis that may aid in detecting, monitoring, and mitigating of large-scale forest health issues. This study examined survivor analysis for evaluating a regional forest health issue - Missouri oak decline. With a statewide Missouri forest inventory, log-rank tests of the effects of covariates on the survivor function and equality of the survivor function among classes were conducted for selected oak species.

Additionally, hazard functions were determined for diameter classes for damaged and undamaged oaks. Results indicate that mortality appears to vary significantly among some inventory classes such as oak species, but not among other classes such as ownership class. Indicators of individual tree vigor (i.e., crown class and ratio) were more significant predictors of oak tree mortality than site/stand attributes (i.e., density and aspect). Finally, results indicate that even fast-growing oak trees are at high risk of mortality if damaged by disease. Survival analyses, such as those applied in this study, may enable testing of forest health hypotheses using large-scale inventories. In the context of Missouri`s oak forest decline, study results suggest management efforts should focus on limiting the spread of disease damage, increasing the vigor of residual trees, and emphasizing small trees when developing stand prescriptions.