U.S. Geological Survey, National Wetlands Research Center, 700 Cajundome Boulevard, Lafayette, Louisiana 70506, USA.
Ecol Appl. 2010 Jan;20(1):192-204. doi: 10.1890/08-1443.1.
The success of population-based ecological restoration relies on the growth and reproductive performance of selected donor materials, whether consisting of whole plants or seed. Accurately predicting performance requires an understanding of a variety of underlying processes, particularly gene flow and selection, which can be measured, at least in part, using surrogates such as neutral marker genetic distances and simple latitudinal effects. Here we apply a structural equation modeling approach to understanding and predicting performance in a widespread salt marsh grass, Spartina alterniflora, commonly used for ecological restoration throughout its native range in North America. We collected source materials from throughout this range, consisting of eight clones each from 23 populations, for transplantation to a common garden site in coastal Louisiana and monitored their performance. We modeled performance as a latent process described by multiple indicator variables (e.g., clone diameter, stem number) and estimated direct and indirect influences of geographic and genetic distances on performance. Genetic distances were determined by comparison of neutral molecular markers with those from a local population at the common garden site. Geographic distance metrics included dispersal distance (the minimum distance over water between donor and experimental sites) and latitude. Model results indicate direct effects of genetic distance and latitude on performance variation among the donor sites. Standardized effect strengths indicate that performance was roughly twice as sensitive to variation in genetic distance as to latitudinal variation. Dispersal distance had an indirect influence on performance through effects on genetic distance, indicating a typical pattern of genetic isolation by distance. Latitude also had an indirect effect on genetic distance through its linear relationship with dispersal distance. Three performance indicators had significant loadings on performance alone (mean clone diameter, mean number of stems, mean number of inflorescences), while the performance indicators mean stem height and mean stem width were also influenced by latitude. We suggest that dispersal distance and latitude should provide an adequate means of predicting performance in future S. alterniflora restorations and propose a maximum sampling distance of 300 km (holding latitude constant) to avoid the sampling of inappropriate ecotypes.
基于种群的生态恢复的成功依赖于所选供体材料（无论是完整植物还是种子）的生长和繁殖表现。准确预测性能需要了解各种潜在过程，特别是基因流和选择，这些过程至少可以部分通过中性标记遗传距离和简单的纬度效应等替代物来衡量。在这里，我们应用结构方程模型方法来理解和预测广泛分布的盐沼草——互花米草（Spartina alterniflora）的性能，这种草在北美的原生范围内被广泛用于生态恢复。我们从整个范围内收集供体材料，包括来自 23 个种群的每个种群的 8 个克隆，将其移植到路易斯安那州沿海的一个共同花园，并监测其性能。我们将性能建模为一个由多个指标变量（例如，克隆直径、茎数）描述的潜在过程，并估计地理和遗传距离对性能的直接和间接影响。遗传距离通过与共同花园地点的本地种群的中性分子标记进行比较来确定。地理距离指标包括扩散距离（供体和实验地点之间的水上最小距离）和纬度。模型结果表明遗传距离和纬度对供体站点之间的性能变化有直接影响。标准化效应强度表明，性能对遗传距离变化的敏感性大致是纬度变化的两倍。扩散距离通过对遗传距离的影响对性能产生间接影响，表明典型的遗传隔离与距离模式。纬度也通过与扩散距离的线性关系对遗传距离产生间接影响。三个性能指标在性能上有显著的负荷（平均克隆直径、平均茎数、平均花序数），而性能指标平均茎高和平均茎宽也受纬度影响。我们建议扩散距离和纬度应该能够充分预测未来互花米草恢复中的性能，并提出最大采样距离为 300 公里（保持纬度不变），以避免采样不适当的生态型。
