Seminars

May 2 2012

12:00 LSB 2320

Helen Regan
Dept of Biology, UC Riverside

Implications of Uncertainty for Species Management Under Global Change

Summary

As threats to biodiversity increase there is a growing need to synthesize and utilize a variety of tools to investigate the impacts of multiple threats and to make useful recommendations for management. In global change biology there has been increased reliance on species distribution models, global climate models and population models to highlight the impacts of different threats on species and to test management and climate change adaptation strategies. However, each model comes with its own set of uncertainties which has the potential to compound when models are linked or when the output of one model serves as the input for another. Hence, modeling frameworks that account for these uncertainties need to be established. Moreover, management recommendations need to be made in the context of multiple threats and multiple sources of uncertainty. In this talk, I present two case studies that highlight the challenges of making climate change adaptation recommendations in the face of multiple threats and uncertainties. First, I will present an evaluation of assisted colonization strategies under climate change for a rare, fire-dependent plant, Tecate cypress. The success or failure of assisted colonization will depend on a range of population-level factors on which the climate change literature has been relatively quiet—the quality of the recipient habitat, the number and life stages of translocated individuals, the establishment of translocated individuals in their new habitat and whether the recipient habitat is subject to ongoing threats all will play an important role in population persistence. We show that, under a fairly strict set of conditions, assisted colonization may be a risk-minimizing adaptation strategy. However, when ongoing threats exist, assisted colonization is ineffective. Second, I will present an uncertainty analysis of a framework that links species distribution and population models to predict the impacts of climate change, land use change, and altered fire regimes on an annual herb, Acanthomintha ilicifolia, or San Diego thornmint. Each modeling component introduces its own source of uncertainty that can potentially compound in results. We estimated the sensitivities of long-run population predictions to different model choices, ecological assumptions and parameter settings and show that model sensitivity to these uncertainties has major implications for management.

 

 

 

 

 

 



 

 

 

 

 

 

 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



this is idtest: