Uncertainty in ecological simulation modeling: Applications to conservation of endangered species and forest growth.

摘要

A population viability analysis (PVA) is applied to the endangered neotropical migrant bird species black-capped vireo (Vireo atricapillus) employing the model VORTEX. Results show that this species is likely to go extinct in about 18 years if no conservation measures are taken. To meet the conservation goals, the rate of fecundity has to be increased up to 2.50 young fledged per female per year, while maintaining carrying capacity at the current level over the next century. A framework for management and monitoring of endangered species from a management-unit point of view is proposed as well. This approach is to be applied in conjunction with PVA. The black-capped vireo at Fort Hood (Texas) is employed as a case study. Results indicate that population size is expected to increase at Fort Hood from 1995 to 1996. The effect of error (uncertainty) in PVA models parameter estimates on the establishment of conservation management objectives is also analyzed. When uncertainty is accounted for in the black-capped vireo PVA, the acceptable conservation goals are only achieved under a fecundity of 3.50 young fledged per female per year.; An efficient method is presented to obtain uncertainty information and error budgets in process models. Examples are conducted employing a process forest growth model for red pine (Pinus resinosa Ait.) and VORTEX for the black-capped vireo. It is concluded that the proposed method provides a valuable means to analyze the uncertainty of large or computationally-slow process models. A sampling optimization procedure is presented as well. Uncertainty is also assessed for the management-unit conservation model applied to the black-capped vireo at Fort Hood. An example of optimization of sample size allocation is also conducted for this last case.; Finally, a framework is proposed to employ process models in hypothesis testing of ecological scenarios when external entities either interact or do not interact with treatment entities. The analysis of the effect of acidic deposition on the site index of a red pine stand is employed to illustrate this methodology. It is argued that these are valuable methods for conducting process model-based experiments.