Dangers of predicting bird species distributions in response to land-cover changes.
Dangers of predicting bird species distributions in response to land-cover changes.
复制标题
预测鸟类物种分布对土地覆盖变化的响应的危险。
DOI：
10.1890/08-0348.1
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
W. Thuiller
中科院分区：
文献类型：
--
作者：
S. Vallecillo;L. Brotóns;W. Thuiller
过去几十年的土地覆盖变化导致生境质量严重下降，导致世界各地鸟类物种分布发生变化。然而，土地覆盖变化的结果，从各种不同的过程，目前还不清楚如何有效的物种分布模型是在捕捉物种对这些变化的反应。在这项研究中，我们评估了我们的能力，预测土地覆盖变化的影响，在大的空间和时间尺度上的物种分布的变化，使用地中海景观和早期演替，开放栖息地的鸟类作为研究模型。基于第二个加泰罗尼亚繁殖鸟类地图集（1999-2002年）的存在-不存在数据，我们采用了6种不同的物种分布建模技术，10种鸟类，气候，地形和土地覆盖数据作为预测变量。然后，我们从1980年的土地覆盖条件的模型进行了反向预测，以评估预测与实地观测数据从第一加泰罗尼亚繁殖鸟类地图集（1975-1983）。最后，我们评估，如果除了土地覆盖变化导致的栖息地适宜性的变化，火灾影响的描述有助于进一步解释物种分布动态：殖民化和局部灭绝。我们开发了准确的模型预测当前和过去的全球物种分布模式，但我们预测物种分布动态的能力降低了。殖民化动态一般更强烈地与火灾描述比来自土地覆盖变化的整体栖息地适宜性的变化。我们的研究结果警告的危险，预测物种分布模型到未来的条件下，如果没有明确包括物种分布动态背后的过程。考虑对物种具有生态意义的过程（即，火干扰）时，模拟物种的分布可能有助于更好地解释物种的殖民动态。
Land-cover changes from the last decades are leading to important declines in habitat quality, giving rise to changes in bird species distribution all over the world. However, land-cover changes result from a variety of different processes, and it is not clear how effective species distribution models are in capturing species responses to these changes. In this study, we evaluated our ability to predict the effects of land-cover changes on shifts in species distributions at large spatial and temporal scales using Mediterranean landscapes and early-successional, open-habitat birds as study models. Based on presence-absence data from the second Catalan Breeding Bird Atlas (1999-2002), we applied six different species distribution modeling techniques for 10 bird species using climate, topographic, and land-cover data as predictor variables. Then we back-projected the models on land-cover conditions from 1980 to evaluate the projections with field observation data from the first Catalan Breeding Bird Atlas (1975-1983). Finally, we assessed if, in addition to changes in habitat suitability resulting from land-cover shifts, descriptors of fire impact contributed to further explain species distribution dynamics: colonization and local extinction. We developed accurate model projections of current and past global patterns of species distribution, but our ability to predict species distribution dynamics was reduced. Colonization dynamics were generally more strongly related to fire descriptors than to changes in overall habitat suitability derived from land-cover changes. Our results warn of the dangers of projecting species distribution models onto future conditions if processes behind species distribution dynamics are not explicitly included. Consideration of ecologically meaningful processes for species (i.e., fire disturbance) when modeling species' distribution might contribute to a better explanation of species' colonization dynamics.