Incorporating parametric uncertainty into population viability analysis models
Incorporating parametric uncertainty into population viability analysis models
复制标题
DOI：
10.1016/j.biocon.2011.01.005
复制
发表时间：
2011-05-01
影响因子：
5.9
通讯作者：
Larson, Michael A.
中科院分区：
文献类型：
--
作者：
McGowan, Conor P.;Runge, Michael C.;Larson, Michael A.
来自抽样变异或专家判断的参数估计的不确定性会给基于这些估计的生态预测带来很大的不确定性。然而，在标准的种群生存力分析，最广泛使用的工具之一，用于管理植物，鱼类和野生动物种群，参数的不确定性往往被忽略或丢弃的模型预测。我们提出了一种方法，明确将这种来源的不确定性纳入人口模型，以充分考虑管理和决策环境中的风险。我们的方法涉及两步模拟过程，其中参数不确定性被纳入模型的复制循环中，时间方差被纳入模型中的时间步长循环中。使用管道普洛弗，在美国和加拿大的联邦受威胁的滨鸟，作为一个例子，我们比较丰富的预测和灭绝概率的模拟，排除和包括参数的不确定性。虽然所有模拟集的最终丰度都非常低，但在复制循环中纳入参数不确定性的模拟中，估计灭绝风险要大得多。关于物种保护的决定（例如，列入名单、从名单上除名和处于危险之中）可能会有很大的不同，这取决于对总体模型中参数不确定性的处理。爱思唯尔有限公司出版
Uncertainty in parameter estimates from sampling variation or expert judgment can introduce substantial uncertainty into ecological predictions based on those estimates. However, in standard population viability analyses, one of the most widely used tools for managing plant, fish and wildlife populations, parametric uncertainty is often ignored in or discarded from model projections. We present a method for explicitly incorporating this source of uncertainty into population models to fully account for risk in management and decision contexts. Our method involves a two-step simulation process where parametric uncertainty is incorporated into the replication loop of the model and temporal variance is incorporated into the loop for time steps in the model. Using the piping plover, a federally threatened shorebird in the USA and Canada, as an example, we compare abundance projections and extinction probabilities from simulations that exclude and include parametric uncertainty. Although final abundance was very low for all sets of simulations, estimated extinction risk was much greater for the simulation that incorporated parametric uncertainty in the replication loop. Decisions about species conservation (e.g., listing, delisting, and jeopardy) might differ greatly depending on the treatment of parametric uncertainty in population models. Published by Elsevier Ltd.