Department of Integrative Biology, University of Guelph, 50 Stone Road East, Guelph, ON N1G 2W1 Canada.
Conserv Biol. 2010 Oct;24(5):1398-406. doi: 10.1111/j.1523-1739.2010.01555.x.
Information required to evaluate the extent to which species are at risk of extinction is usually limited and characterized as highly uncertain. In this context, we define information availability as the presence or absence of information used to determine the value of an ecological variable. We examined which of three hypothetical approaches best matched how levels of risk are assigned to species: (1) precautionary approach in which analysts designate levels of risk regardless of the amount of information available, (2) worst-case approach in which analysts assign the maximum level of risk possible from the criteria, and (3) insurance approach in which analysts assign poorly known species to a high-risk category when little information is available. We used the quantitative assessment criteria of the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) as a case study. We created a binary (0/1) matrix of all 2.4192 × 10(7) logical combinations of available information for the 14 ecological variables included in the quantitative criteria. We processed each combination of information availability represented in the matrix with a computer algorithm designed to emulate COSEWIC decision-making rules. Low information availability was associated with a relatively high frequency of not being able to assign a candidate taxon to a risk category, which does not follow the precautionary principle. Information availability and the level of risk assigned to species were directly related, which is associated with the worst-case approach, and counter to the insurance approach. Our results suggest that information availability can have a major effect on the level of risk assigned to a species. We recommend a conscious determination of whether such effects are desired, and we recommend the development of methods to explicitly characterize and incorporate information availability and other sources of uncertainty in decision-making processes.
评估物种灭绝风险程度所需的信息通常是有限的，并且具有高度不确定性。在这种情况下，我们将信息可获得性定义为用于确定生态变量值的信息的存在或缺失。我们研究了三种假设方法中的哪一种最符合将风险级别分配给物种的方式：（1）无论可用信息量如何，分析人员都指定风险级别，这是一种预防方法；（2）分析人员根据标准分配最大风险级别，这是一种最坏情况方法；（3）当可用信息量很少时，分析人员将知之甚少的物种分配到高风险类别，这是一种保险方法。我们使用加拿大濒危物种状况委员会（COSEWIC）的定量评估标准作为案例研究。我们创建了一个二进制（0/1）矩阵，其中包含了 14 个定量标准中包含的所有 2.4192×10^7 个逻辑组合的可用信息。我们使用一种计算机算法来处理矩阵中表示的每种信息可用性组合，该算法旨在模拟 COSEWIC 的决策规则。低信息量与无法将候选分类单元分配到风险类别的相对高频率有关，这不符合预防原则。信息可用性和分配给物种的风险级别之间存在直接关系，这与最坏情况方法有关，与保险方法相反。我们的结果表明，信息可用性会对分配给物种的风险水平产生重大影响。我们建议有意识地确定是否需要这种影响，并建议开发方法来明确描述和纳入决策过程中的信息可用性和其他不确定性来源。
