Defining Post-Flood Survival of Planted Oak Seedlings Using Odds Ratios

For many planting scenarios, it is desirable to estimate the number of seedlings that need to be planted on flood prone sites to achieve management goals, based upon species flood tolerance ratings and desired stocking levels.  However, previous assessments of flood tolerance for a range of oak species used in such plantings in Missouri have been inconsistent.  Therefore, the objective of this study was to assess the flood tolerance of seven different oak species by examining the survival rates of newly planted seedlings in response to four flood treatments over two sampling dates. A total of 2713 one-year-old seedlings of seven native oak species were planted in an outdoor, multi-channel flood tolerance laboratory located at the University of Missouri Horticulture and Agroforestry Research Center in New Franklin, MO in March 2005.  End of season survival was determined in September 2005 (15 weeks post-flood) and in June 2006 (45 weeks post-flood).  Survival data were analyzed using SAS PROC GENMOD, which provided for an analysis of species and flood treatment differences in survival rates based upon calculated odds ratios over two sampling dates. In this study, an odds ratio compared the probabilities of seedling survival rates between two species by calculating the antilog (e^x) of an estimate (= Logit P_i) value that represented the difference between two species least square means.  These values were interpreted as the likelihood of one species to survive 15 weeks, or 45 weeks, post-flood versus the alternate species.  Conversely, such odds ratios indicated the number of seedlings that would need to be planted of the less flood tolerant species for every seedling representing the more flood tolerant species to obtain the same survival at either 15 or 45 weeks post-flood.  Survival rates based on odds ratios differed for non-flooded and flooded seedlings in both a flood tolerant (swamp white oak) and flood intolerant species (northern red oak) at 45 weeks post-flood, which clearly demonstrated the longer term impact of flooding on survival rates in response to flooding.  It is recommended that an odds ratio approach be used as a management tool when making decisions on deploying specific species in adequate numbers across flood prone sites.