Dispersal evolves as an adaptive mechanism to optimize individual fitness across the landscape. Specifically, dispersal represents a mechanism to escape fitness costs resulting from changes in environmental conditions. Decades of empirical work suggest that individuals use local habitat cues to make movement decisions, but theory predicts that dispersal can also evolve as a fixed trait - independent of local conditions - in environments characterized by a history of stochastic spatiotemporal variation. Until now, however, both conditional and fixed models of dispersal evolution have primarily been evaluated using emigration data (stay vs. leave), and not dispersal distances - a more comprehensive measure of dispersal. Our goal was to test whether conditional or fixed models of dispersal evolution predict variation in dispersal distance in the stream salamander Gyrinophilus porphyriticus. We quantified variation in habitat conditions using measures of salamander performance from 4 years of spatially explicit, capture-mark-recapture (CMR) data across 3 headwater streams in the Hubbard Brook Experimental Forest in central New Hampshire, USA. We used body condition as an index of local habitat quality that individuals may use to make dispersal decisions, and survival probability estimated from multistate CMR models as an index of mortality risk resulting from the long-term history of environmental variation. We found that dispersal distances increased with declining survival probability, indicating that salamanders disperse further in risky environments. Dispersal distances were unrelated to spatial variation in body condition, suggesting that salamanders do not base dispersal distance decisions on local habitat quality. Our study provides the first empirical support for fixed models of dispersal evolution which predict that dispersal evolves in response to a history of spatiotemporal environmental variation, rather than as a conditional response to current habitat conditions. More broadly, this study underscores the value of assessing alternative scales of environmental variation to gain a more complete and balanced understanding of dispersal evolution.

中文翻译：

---

长期生存概率，而不是当前栖息地质量，预测蝾螈的扩散距离

分散作为一种自适应机制进化，以优化整个景观中的个体适应度。具体来说，分散代表了一种逃避环境条件变化导致的健身成本的机制。数十年的实证研究表明，个体使用当地栖息地线索来做出运动决定，但理论预测，在以随机时空变化为特征的环境中，散布也可以作为一种固定特征进化——独立于当地条件。然而，直到现在，扩散演化的条件模型和固定模型都主要使用移民数据（停留与离开）而不是扩散距离进行评估——这是一种更全面的扩散度量。我们的目标是测试扩散进化的条件模型或固定模型是否能预测流蝾螈 Gyrinophilus porphyriticus 的扩散距离变化。我们使用来自美国新罕布什尔州中部哈伯德布鲁克实验森林 3 个源头溪流 4 年空间明确的捕获标记重新捕获 (CMR) 数据的蝾螈性能测量来量化栖息地条件的变化。我们使用身体状况作为当地栖息地质量的指标，个人可以用它来做出分散决策，并将根据多状态 CMR 模型估计的生存概率作为长期环境变化历史导致的死亡风险指数。我们发现扩散距离随着生存概率的下降而增加，这表明蝾螈在危险环境中的扩散距离更远。散布距离与身体状况的空间变化无关​​，这表明蝾螈不会根据当地栖息地质量来决定散布距离。我们的研究为扩散演化的固定模型提供了第一个实证支持，该模型预测扩散演化是对时空环境变化历史的响应，而不是作为对当前栖息地条件的条件响应。更广泛地说，这项研究强调了评估环境变化的替代尺度的价值，以获得对扩散演变的更完整和平衡的理解。我们的研究为扩散演化的固定模型提供了第一个实证支持，该模型预测扩散演化是对时空环境变化历史的响应，而不是作为对当前栖息地条件的条件响应。更广泛地说，这项研究强调了评估环境变化的替代尺度的价值，以获得对扩散演变的更完整和平衡的理解。我们的研究为扩散演化的固定模型提供了第一个实证支持，该模型预测扩散演化是对时空环境变化历史的响应，而不是作为对当前栖息地条件的条件响应。更广泛地说，这项研究强调了评估环境变化的替代尺度的价值，以获得对扩散演变的更完整和平衡的理解。