Habitat variables related to vegetation type and structure are routinely identified as important components of nest‐site selection for birds. For ground‐nesting birds, small‐scale (< 0.5 m) microtopography may also play a role in nest‐site selection through its effects on nest concealment and microclimate. Manual measurements of microtopography are challenging, time‐consuming, and subject to user error. Ultrahigh‐density point clouds generated using structure‐from‐motion (SfM) algorithms and photomosaics captured during Unmanned Aerial Vehicle (UAV) surveys can potentially provide detailed topographical measurements with less error. We used multiple indices of surface roughness to examine the effect of microtopography on nest site selection by White‐rumped Sandpipers (Calidris fuscicollis) at East Bay Migratory Bird Sanctuary, Nunavut, Canada. We measured microtopography manually using the relative height index within a 1‐m radius of used and unused sites. We generated three digital indices of surface roughness for 1‐m and 5‐m scales around nests and unused sites, including standard deviation of slope, slope variability, and 2D:3D area ratio. We compared the digital indices of used and unused sites at both spatial scales. Relative height of nearby hummocks was significantly lower at nests than at unused sites, indicating selection for flatter sites. Similarly, two of three digitally calculated terrain roughness indices were significantly lower at nests than unused sites at both the 1‐m and 5‐m scales. The 2D:3D ratio did not differ between nests and unused sites. Manual and UAV‐derived measures were significantly correlated, but with substantial unexplained variation (Pearson’s r = 0.35–0.46). Our results suggest that White‐rumped Sandpipers select nest sites in areas with low terrain roughness, potentially to increase their field of view to monitor approaching predators. We also demonstrate the applicability of SfM habitat reconstruction for testing hypotheses of small‐scale habitat variables that may be impossible to study using traditional methods.

中文翻译：

---

应用从动结构栖息地重建和GIS地形分析来检验关于水鸟选择巢穴的假设

通常将与植被类型和结构有关的栖息地变量确定为鸟类筑巢地点选择的重要组成部分。对于地面筑巢的鸟类，小规模（<0.5 m）的微形貌还可能通过影响巢藏和微气候而在巢址选择中发挥作用。手动测量微观形貌具有挑战性，耗时且容易出错。使用运动结构（SfM）算法生成的超高密度点云和无人飞行器（UAV）调查中捕获的光电马赛克可以潜在地提供详细的地形测量结果，而误差较小。我们使用表面粗糙度的多个指标来检验微观形貌对白腰Sand（Calidris fuscicollis）选择巢位的影响）在加拿大努纳武特的东湾候鸟保护区。我们使用已使用和未使用位置的1米半径内的相对高度指数手动测量了微形貌。我们在巢和未使用位置周围生成了1m和5m尺度的三个表面粗糙度数字索引，包括坡度的标准偏差，坡度变异性和2D：3D面积比。我们在两个空间尺度上比较了已使用和未使用站点的数字索引。与未使用的地点相比，燕窝附近的山岗相对高度明显较低，表明选择了平坦的地点。同样，在1m和5m尺度上，三个通过数字计算得出的地形粗糙度指数中的两个在巢上都显着低于未使用的地点。嵌套和未使用的站点之间的2D：3D比率没有差异。手动和无人机衍生措施之间存在显着相关性，r  = 0.35-0.46）。我们的结果表明，白腰矶pi在地形粗糙度较低的地区选择巢穴，可能会增加它们的视野以监视接近的掠食者。我们还证明了SfM生境重建可用于测试小规模生境变量的假设，这些假设可能无法使用传统方法研究。