Dimitri Giunchi

Bird flocks frequently generate human–wildlife conflicts, resulting in safety and health hazards, and economic losses. Remotely controlled, raptor‐like robots have recently emerged as potentially effective deterrence tools, exploiting the innate collective escape response of birds to their predators and hence avoiding the emergence of habituation, but field tests are still scanty. We tested the effectiveness of a robotic falcon (RobotFalcon [RF]) as a deterrence tool for feral pigeons Columba livia breeding and feeding in a cattle farm (~500 individuals) and black‐headed gulls Chroicocephalus ridibundus night‐roosting in a wastewater treatment plant (~8000 individuals). We analysed collective behavioural responses during 10‐min RF exposure trials mimicking natural predator attacks and assessed deterrence effectiveness (change in the number of birds) at different temporal scales (immediate response—within‐trials and returning post‐trials; short‐term response—across 5–8 successive daily exposures). We also evaluated medium‐term deterrence effectiveness by quantifying gull flight activity using a marine radar for ~10 days after RF exposure trials. The RF induced consistent collective behavioural responses, similar to real avian predators. The frequency of collective behaviours decreased both within‐ and across trials, mainly because flocks rapidly abandoned the target sites. Pigeons decreased both during and increasingly across trials, abandoning the target site soon after trial onset. However, most pigeons returned within 30 min after the end of trials, and individuals returning post‐trials decreased only slightly (13%) across trials. The RF triggered massive gull displacements during trials, with a faster reaction in the last trial days compared to initial ones, and individuals returning post‐trials decreased markedly (94%) across trials. Night flight activity showed a large (40%) decrease post‐ compared to pre‐trial days, suggesting persistent deterrence effects. Practical implication. The RF is a valuable tool to deter birds from gathering in flocks in contexts where human–wildlife conflict may arise, based on robust evidence and field tests under different conditions. Yet, its effectiveness may depend on the importance of resources used by the target flocking species at the exposure site. Predator‐like robots provide unique advantages related to animal welfare, consistency of deterrence protocols and effective management of human–wildlife conflicts.

Homing pigeons rely on familiar landscape features in learning a visual map, which is orchestrated by the forebrain visual Wulst and hippocampus. Recent global positioning system tracking studies showed that pigeons with damage to the visual Wulst or hippocampus exhibited a still poorly understood, persistent oscillatory flight behaviour, unlike intact pigeons whose oscillations decrease with experience. To evaluate whether landscape heterogeneity influences the extent of these oscillations, we compared the flight behaviour of both intact and Wulst-lesioned pigeons when flying over the sea versus land. Regardless of treatment, pigeons exhibited less oscillatory flight behaviour over the homogeneous landscape of the sea. Further releases from familiar and unfamiliar sites tested whether oscillatory flight behaviour may be influenced by the level of familiarity with the landscape. Indeed, intact pigeons reduced their oscillatory behaviour as landscape familiarity increased. In contrast, Wulst-damaged pigeons persisted in displaying robust oscillatory flight behaviour regardless of the level of landscape familiarity, suggesting that previously experienced landscapes remained relatively unfamiliar to them. The data support the hypothesis that oscillatory flight behaviour reflects active visual scanning, most pronounced over novel, heterogeneous landscapes, contributing to visual map learning. Further, the data suggest a crucial role of the forebrain Wulst for such learning.

Although ex situ conservation programs are increasingly valuable support tools for in situ conservation measures, success depends on these captive individuals to be genetically representative of the recipient population. The Eurasian stone-curlew (Burhinus oedicnemus) inhabit steppes that represent some of the most degraded and exploited habitats worldwide. A captive breeding program was implemented in Morocco as a pre-emptive effort for the conservation of the North African subspecies Burhinus oedicnemus saharae. However, the genetic origins of the founders of the captive flock were unknown. We applied a multi-locus approach to characterize the genetic ancestry of the current captive breeding flock by comparing it to wild populations from both Western and Eastern Morocco. Mitochondrial DNA and microsatellite markers were employed to assess levels of genetic diversity and relatedness within each sample, as well as potential genetic differentiation between wild and captive samples through PCA and admixture analyses. We recovered similar genetic diversity estimates, low levels of relatedness, and little differentiation between captive and wild samples. These results confirmed the Moroccan origin of the founders. We provide recommendations for the optimization of the Eurasian Stone Curlew conservation breeding program but also for future conservation breeding programs to ensure the effective conservation of genetic diversity and wild populations.

Climate change is having dramatic impacts on the distribution of animals. Birds, and especially steppe-land birds, are particularly sensitive to climate change and identifying areas that are critical for their conservation is pivotal, as well as estimating the expected impact on these areas under different climate and land use change scenarios. In-situ climate refugia (areas suitable under both current and future climates) are especially valuable for the conservation of climate-sensitive species, and is therefore important to identify them and evaluate their coverage by protected areas. Via species distribution modelling, we aimed to identify in-situ climate refugia in the Western Palearctic for the Eurasian Stone-curlew Burhinus oedicnemus , an umbrella steppic species of conservation concern. We used a comprehensive dataset of occurrences in the breeding period to fine-tune a Maxent species distribution model and project it under three carbon emission scenarios of increasing severity for the year 2050. We then identified in-situ climate refugia and performed a gap analysis estimating the percentage of refugia falling within the network of currently protected areas. In all modelled future scenarios a northward expansion of suitable breeding habitats was predicted, and suitable areas had similar extents, with a slight increase of the overall suitability under more severe scenarios. According to our results, the Eurasian Stone-curlew has the potential to maintain viable populations in the Western Palearctic, even though dispersal limitations might hinder the colonization of newly suitable breeding areas. In-situ climate refugia were mainly identified outside protected areas, particularly in Northern Africa and the Middle East. Therefore, we advocate targeted actions in climate refugia to promote the conservation of this and other steppe-land species under global environmental change.

Hair tubes are one of the most effective tools for sampling small‐mammal assemblages. Despite their efficiency, they can be damaged by wildlife. We quantified wildlife‐induced disturbance of hair‐tube sampling in the Monte Pisano mountain system in Italy. At each site we tied 2 tubes together to form a hair trap and placed them in 3 different areas. We measured disturbances by counting the number of hair traps disturbed and identified the species that caused damage with cameras. Although approximately 27% (n = 164/600) of hair traps were disturbed, 2 hair traps/site allowed us to collect data from 97% of site checks from undisturbed and retrieved hair traps. Wild boars (Sus scrofa) were attracted to hair traps by olfactory and acoustic signals and caused the most disturbances. Displaced tubes detected a similar number of hairs and number of species as undisturbed tubes; species richness estimates that include retrieved displaced tubes should provide reliable data. To avoid data loss when conducting hair‐trap monitoring, we suggest using ≥2 hair traps per site.