Data from: Predation risk shaped by habitat and landscape complexity in urban environments

1. Habitat loss and modification are hallmarks of anthropogenic ecosystems, but the consequences for ecosystem functions and services often remain unclear. Understanding these links in cities is complicated by strong but fine-scale differences in habitat structure among green space patches, and a high variance in habitat amount across urban landscapes. 2. We used airborne laser scanning (ALS) data to disentangle the effects of 3D woody habitat heterogeneity of urban home gardens, and woody habitat amount at four landscape spatial scales (50, 100, 250, 500 m), on the predation risk of artificial sentinel prey by birds and arthropods. 3. In both predator groups, and at all the investigated spatial scales, cross-scale interactive effects between garden habitat heterogeneity and habitat amount in the urban landscape were the main drivers of predation. Risk of bird predation by birds was highest in heterogeneous garden habitats, but only in densely built urban landscapes where habitat amount was low to intermediate (10-20%) at large spatial scales (250-500 m). It dropped independently of garden habitat heterogeneity when habitat amount became too low (<10%) at small (50-100 m) spatial scales. In contrast, risk of predation by arthropods mostly peaked in homogenous garden habitats when habitat amount was intermediate (20%) at large spatial scales. 4. Our findings show that the ability of urban green space patches to sustain ecosystem functions in cities mainly depends on cross-scale interactive effects with larger-scale habitat amount. In birds, predation risk can increase when high patch-scale habitat heterogeneity contrasts with reduced larger-scale habitat amount, suggesting concentration effects. Yet thresholds exist under which ecosystem functioning drops independently of habitat structure. 5. Synthesis and applications. The potential of small-scale interventions to enhance habitat heterogeneity (e.g. by planting native trees with understory shrubs) for restoring ecosystem functions such as bird predation in urban areas is dependent on wider landscape habitat structure. Urban planning should therefore adopt a multi-scale approach to sustain and restore ecosystem functions and services; a simple but still not broadly recognized finding. Airborne laser scanning is a useful tool to infer habitat structure across a hierarchy of scales in spatially heterogeneous anthropogenic ecosystems.1. 生境丧失与改造是人为生态系统（anthropogenic ecosystems）的典型特征，但其对生态系统功能与服务的影响往往尚不明确。城市中这类关联的认知因绿地斑块间生境结构存在显著且精细尺度的差异，以及城市景观中生境总量的高度变异性而变得复杂。 2. 本研究借助机载激光扫描（airborne laser scanning, ALS）数据，解析了城市住宅花园的三维木本生境异质性，以及4种景观空间尺度（50、100、250、500米）下的木本生境总量，对鸟类与节肢动物捕食人工哨兵猎物的风险的影响。 3. 两类捕食者群体在所有调查的空间尺度下，城市景观中花园生境异质性与生境总量的跨尺度交互效应均为捕食过程的主要驱动因素。鸟类捕食风险在异质型花园生境中最高，但仅出现在大空间尺度（250-500米）下生境总量处于低至中等水平（10%-20%）的高密度建成城市景观中。当小空间尺度（50-100米）下生境总量过低（<10%）时，无论花园生境异质性如何，捕食风险均会下降。与之相反，节肢动物的捕食风险大多在大空间尺度下生境总量处于中等水平（20%）时，于同质型花园生境中达到峰值。 4. 本研究结果表明，城市绿地斑块维持生态系统功能的能力，主要取决于其与更大尺度生境总量的跨尺度交互效应。对于鸟类而言，当斑块尺度的生境异质性较高，而更大尺度的生境总量降低时，捕食风险会上升，这提示存在浓度效应。但存在临界阈值，当生境总量低于该阈值时，生态系统功能会脱离生境结构独立下降。 5. 综合与应用。旨在提升生境异质性的小规模干预措施（例如通过栽植本土乔木搭配林下灌丛），在恢复城市地区鸟类捕食等生态系统功能方面的潜力，取决于更广泛的景观生境结构。因此，城市规划应采用多尺度方法来维持并恢复生态系统功能与服务——这是一个简单但尚未被广泛认知的结论。机载激光扫描是在空间异质的人为生态系统中，跨层级尺度推断生境结构的有效工具。