Pennsylvania Cooperative Fish and Wildlife Research Unit, Pennsylvania State University, University Park, PA, USA.
U.S. Geological Survey, Pennsylvania Cooperative Fish and Wildlife Research Unit, Pennsylvania State University, University Park, PA, USA.
Glob Chang Biol. 2018 Jun;24(6):2735-2748. doi: 10.1111/gcb.14101. Epub 2018 Mar 27.
Predictions of the projected changes in species distributions and potential adaptation action benefits can help guide conservation actions. There is substantial uncertainty in projecting species distributions into an unknown future, however, which can undermine confidence in predictions or misdirect conservation actions if not properly considered. Recent studies have shown that the selection of alternative climate metrics describing very different climatic aspects (e.g., mean air temperature vs. mean precipitation) can be a substantial source of projection uncertainty. It is unclear, however, how much projection uncertainty might stem from selecting among highly correlated, ecologically similar climate metrics (e.g., maximum temperature in July, maximum 30-day temperature) describing the same climatic aspect (e.g., maximum temperatures) known to limit a species' distribution. It is also unclear how projection uncertainty might propagate into predictions of the potential benefits of adaptation actions that might lessen climate change effects. We provide probabilistic measures of climate change vulnerability, adaptation action benefits, and related uncertainty stemming from the selection of four maximum temperature metrics for brook trout (Salvelinus fontinalis), a cold-water salmonid of conservation concern in the eastern United States. Projected losses in suitable stream length varied by as much as 20% among alternative maximum temperature metrics for mid-century climate projections, which was similar to variation among three climate models. Similarly, the regional average predicted increase in brook trout occurrence probability under an adaptation action scenario of full riparian forest restoration varied by as much as .2 among metrics. Our use of Bayesian inference provides probabilistic measures of vulnerability and adaptation action benefits for individual stream reaches that properly address statistical uncertainty and can help guide conservation actions. Our study demonstrates that even relatively small differences in the definitions of climate metrics can result in very different projections and reveal high uncertainty in predicted climate change effects.
对物种分布的预测变化和潜在适应行动的效益进行预测可以帮助指导保护行动。然而，将物种分布预测到未知的未来存在很大的不确定性，如果不加以适当考虑，这可能会降低预测的可信度或误导保护行动。最近的研究表明，选择描述非常不同气候方面的替代气候指标（例如，平均空气温度与平均降水量）可能是预测不确定性的主要来源。然而，不清楚从描述相同气候方面（例如，限制物种分布的最高温度）的高度相关且生态相似的气候指标（例如，7 月最高温度，30 天最高温度）中选择会产生多少预测不确定性。也不清楚预测不确定性如何传播到适应行动的潜在效益预测中，这些行动可能会减轻气候变化的影响。我们为布罗克鳟鱼（Salvelinus fontinalis）提供了气候变化脆弱性、适应行动效益以及与选择四个最高温度指标相关的不确定性的概率度量，布罗克鳟鱼是美国东部受到保护关注的冷水性鲑鱼。在中世纪气候预测的替代最高温度指标中，适宜溪流长度的预计损失相差高达 20%，与三种气候模型之间的差异相似。同样，在完全河岸森林恢复的适应行动情景下，鳜鱼出现概率的区域平均预测增加幅度在指标之间相差高达 0.2。我们使用贝叶斯推断为单个溪流流域提供了脆弱性和适应行动效益的概率度量，这些度量正确地解决了统计不确定性问题，并有助于指导保护行动。我们的研究表明，即使气候指标的定义略有不同，也会导致非常不同的预测，并揭示出预测气候变化影响的高度不确定性。
