2017 ESA Annual Meeting (August 6 -- 11)

COS 189-9 - Scaling from demography to the landscape: Predicting occurrence of a rare plant in a pyrogenic world

Friday, August 11, 2017: 10:50 AM
D138, Oregon Convention Center
Eric S. Menges1, Pedro F. Quintana-Ascencio2, Stephanie M. Koontz1, Stacy A. Smith3 and Vivienne Sclater4, (1)Plant Ecology Program, Archbold Biological Station, Venus, FL, (2)Biology, University of Central Florida, Orlando, FL, (3)Department of Agronomy, University of Florida, Gainesville, FL, (4)GIS Lab, Archbold Biological Station, Venus, FL
Background/Question/Methods

Demography often varies among populations, but most scientists have looked at these patterns in a piecemeal way. In this study, we scale up demographic data to the landscape level by building models and comparing outcomes to observed spatial occupancy patterns. Demographic data on the perennial herb Hypericum cumulicola were gathered from marked plants and density plots (1994-2014, 14 populations, 11K plants, 34K cases of annual vital rates), and combined with experiments on seed dormancy, germination and seedling survival. We used four landscape level variables, time-since-fire, patch area, patch isolation, and elevation above wetlands as predictors of population variation. We evaluated generalized linear mixed models (with random effects by population and year) to assess the effects of stage (first year plant or adult), plant height, and the landscape variables on annual survival, plant growth, probability of reproduction and fecundity. Seed survival, dormancy, and germination were estimated with general additive models. Models were then combined in integral projection models to estimate population growth rates and other parameters. We compared our predictions with data from an independent dataset of 30 patches and assessed the effect of frequent and infrequent fire regimes on population persistence.

Results/Conclusions

Both fire and elevation had significant and often interacting effects on germination, seed dormancy, plant survival, plant growth, probability of reproduction, and fecundity. Models correctly predicted 88 % of the presences (16/18) and 50 % of the absences (6/12) and explained 51% of the variance in abundance in occupied patches. These results broadly agree with prior models on the primacy of time-since-fire in driving the population dynamics of H. cumulicola, but add detail on how the landscape affects plant demography. Modeling across a landscape gradient can help inform management. In this case, prescribed burns can occur less often in drier habitats than wetter parts of the landscape to support a viable metapopulation of H. cumulicola.