Large carnivores, once widespread across much of Europe, between the nineteenth and twentieth centuries have suffered a dramatic decline that brought them close to extinction in many parts of Europe. Since the '70s, factors such as legal protection, improved habitats quality and mountains depopulation enabled the recovery of several predator populations. Wolf recovery in areas from which was eradicated, or was occasionally present, has been followed by an intensification of the conflict with human activities, in particular with animal husbandry. Many farmers were unprepared to deal with this new situation having abandoned over the years the use of fences, guard dogs and the practice of monitor the stock. This research was carried out within the territory of Grosseto province, where, following the recent expansion of the wolf population, livestock activities and predator range have overlapped again leading to a situation of great conflict. The same problem occurred in different parts of the world, as demonstrated by the growing number of scientific publications on this topic. Although wolf-livestock conflict is a complex issue, its mitigation is partially fostered by damages reduction. This research shows how the analysis of the ecological context helps in preventing livestock losses. In recent years the use of models to predict the depredation risk has grown dramatically, suggesting how this technique will be increasingly applied to take management information to mitigate the human-carnivore conflict. Therefore I proposed a new three-step method to predict wild canid (wolves and wolf-dog hybrids) depredation risk using presence-only data on wild canid detections and confirmed depredation events in the study area. As a first step, wild canids probability of occurrence was predict; second, I made a prediction on where depredation events were more likely to occur; third I performed an ensemble model integrating the two previous models following an ad-hoc procedure. Models’ outputs obtained from two different approaches to species distribution modeling: Maximum Entropy (Maxent), widely used, and Bayesian for Presence Only Data (BPOD), recently proposed, were compared testing their ability to predict the occurrence of events. The ecological niche factor analysis (ENFA) was used to assess the importance of each environmental variable in the description of the presence points. Results showed that the presence of wild canids was mainly related to forests (M = 0.78). Whereas depredation events were most likely to occur close to farms (M = -0.83) where sheep densities were higher (M = 8.1) and more accessible (M = -1.46). Higher depredation risk zones were characterized by proximity to forested areas and the presence of landscape features that allowed wild canids to reach pastures with minimum effort such as the network of smaller watercourses. Although the majority of livestock within Grosseto province graze extensively and is thus potentially available for predators, only 15% sheep farms fall within higher risk areas. This suggests that at the provincial level, depredation was facilitated by environmental conditions (e.g. closeness to the woods or steams) rather than the availability of domestic prey. Overall BPOD performed better than Maxent in terms of sensitivity, suggesting that BPOD could be a promising approach to predict probability of occurrence using presence-only data. In many parts of the world, livestock guarding dogs (LGDs) are considered one of the most powerful prevention tools against carnivore depredation on domestic animals. As wolf populations are recovering their use is expected to increase. Although LGDs defend livestock against predators, they could negatively impact on some wild species and in some situations could be even a potential hazard to humans. Therefore how these dogs behave when left unsupervised with their flock on pastures is of utmost importance. 29 LGDs with GPS collars were monitored in order to investigate their space use and association with their livestock, analyzing two parameters: the dog-sheep distance and the overlap between dog and sheep movement ranges. The first parameter was evaluated by measuring the real distance between pairs of dog-sheep locations taken in less than five minutes apart to ensure the simultaneity of the two events. In addition linear mixed models were implemented to evaluate how dog-sheep distance was influenced by environmental, dog-related, and farming-related variables. UDOI (Utilization Distribution Overlap Index) and the VI (Volume of Intersection) Index for 50% and 95% kernel isopleths were calculated to quantify the overlap and the similarity in the use of space for the core area and for the whole movement range of sheep and dogs. Finally the usefulness of GPS pet collars in dogs and sheep husbandry was tested. LGDs did not leave the flock unattended when left unsupervised. They spent the majority of their time close to livestock, sharing the same areas but using the space in a different way (mean VI 95% = 0.65 ± 0.16; mean UDOI 95%= 1.31 ± 0.56). Dog-sheep distance was mostly influenced by environmental variables and the age of the dog. Dogs and sheep tended to separate more in pastures surrounded by woods (β = 1.669, p <2.2e-16) or located in heterogeneous agricultural areas (β = 1.204, p = 1.33e-05), and less in pastures close to inhabited areas (β = -1.730, p = 2.34e-07). Older dogs were more associated to the flock compared to younger individuals (β = -0.438, p = 0.002). Some of the variability linked to the dog-sheep distance was explained by the importance of the random components of the models, namely: the differences among individual dogs working in pastures with different extension (p<2.2e-16); the day when the sampling was done (p<2.2e-16); and the differences among farms (p=4.87e-07). The effectiveness of guarding dogs as a prevention tool is not only affected by the environmental features or by LGD’s characteristics and training. In fact, to be effective, livestock guarding dogs should work in conditions that allow them to protect the entire livestock. Comparing 79 sheep farms with at least one adult (> 1.5 years old) guarding dog, were highlighted the conditions that decrease the efficacy of these animals in reducing depredations. For each farm were measured: 1) the number of adult livestock guarding dogs; 2) the distance between the farmer's house and the night shelter; 3) night shelter permeability to predators; 4) flock size; 5) shepherd presence; 6) the number of depredation events over the last six months; 7) the depredation risk. Farms were classified on whether or not they experienced depredation over the last six months. The two groups were then compared using non-parametric tests and logistic regressions. Depredated and non-depredated farms differed only by the night shelter-farmer’s home distance value (W = 455, p-value = 0.005). The model averaging showed a significant positive correlation between damage occurrence and night shelter-farmer’s home distance length (β = 4.695 e-04, p-value = 0.0218). These results suggest that in environmental conditions that determine a similar depredation risk, human presence is the main feature that enhances the effectiveness of guarding dogs as a tool against canid attacks on flocks. Investigating the role of some of the ecological variables involved in depredation events helps to ensure that the wolf-livestock interactions occur in a sustainable manner. Indeed depredation risk maps could be a useful tool for farmers and manager for the timely apply prevention techniques that reduce depredation and for policymaker could be a support to allocate financial resources. Additionally conservation projects may benefit from these maps to select areas of intervention. Moreover results from this work provided some hints for farmers and conservationists to improve the use of LGDs for an effective livestock protection: some of the recommendations affected the dog management, while other the livestock husbandry practices. Finally this research introduced a new way to manage LGDs using GPS pet collars. With these devices farmers could be able to check the position of their dogs and their flock at any time, preventing wrong dog behaviors, conflicts with neighbors and accidents.

Exploring the importance of the ecological context in wolf-livestock interaction management: the case of Grosseto province / Zingaro, Margherita. - (2017 Feb 28).

Exploring the importance of the ecological context in wolf-livestock interaction management: the case of Grosseto province

ZINGARO, MARGHERITA
28/02/2017

Abstract

Large carnivores, once widespread across much of Europe, between the nineteenth and twentieth centuries have suffered a dramatic decline that brought them close to extinction in many parts of Europe. Since the '70s, factors such as legal protection, improved habitats quality and mountains depopulation enabled the recovery of several predator populations. Wolf recovery in areas from which was eradicated, or was occasionally present, has been followed by an intensification of the conflict with human activities, in particular with animal husbandry. Many farmers were unprepared to deal with this new situation having abandoned over the years the use of fences, guard dogs and the practice of monitor the stock. This research was carried out within the territory of Grosseto province, where, following the recent expansion of the wolf population, livestock activities and predator range have overlapped again leading to a situation of great conflict. The same problem occurred in different parts of the world, as demonstrated by the growing number of scientific publications on this topic. Although wolf-livestock conflict is a complex issue, its mitigation is partially fostered by damages reduction. This research shows how the analysis of the ecological context helps in preventing livestock losses. In recent years the use of models to predict the depredation risk has grown dramatically, suggesting how this technique will be increasingly applied to take management information to mitigate the human-carnivore conflict. Therefore I proposed a new three-step method to predict wild canid (wolves and wolf-dog hybrids) depredation risk using presence-only data on wild canid detections and confirmed depredation events in the study area. As a first step, wild canids probability of occurrence was predict; second, I made a prediction on where depredation events were more likely to occur; third I performed an ensemble model integrating the two previous models following an ad-hoc procedure. Models’ outputs obtained from two different approaches to species distribution modeling: Maximum Entropy (Maxent), widely used, and Bayesian for Presence Only Data (BPOD), recently proposed, were compared testing their ability to predict the occurrence of events. The ecological niche factor analysis (ENFA) was used to assess the importance of each environmental variable in the description of the presence points. Results showed that the presence of wild canids was mainly related to forests (M = 0.78). Whereas depredation events were most likely to occur close to farms (M = -0.83) where sheep densities were higher (M = 8.1) and more accessible (M = -1.46). Higher depredation risk zones were characterized by proximity to forested areas and the presence of landscape features that allowed wild canids to reach pastures with minimum effort such as the network of smaller watercourses. Although the majority of livestock within Grosseto province graze extensively and is thus potentially available for predators, only 15% sheep farms fall within higher risk areas. This suggests that at the provincial level, depredation was facilitated by environmental conditions (e.g. closeness to the woods or steams) rather than the availability of domestic prey. Overall BPOD performed better than Maxent in terms of sensitivity, suggesting that BPOD could be a promising approach to predict probability of occurrence using presence-only data. In many parts of the world, livestock guarding dogs (LGDs) are considered one of the most powerful prevention tools against carnivore depredation on domestic animals. As wolf populations are recovering their use is expected to increase. Although LGDs defend livestock against predators, they could negatively impact on some wild species and in some situations could be even a potential hazard to humans. Therefore how these dogs behave when left unsupervised with their flock on pastures is of utmost importance. 29 LGDs with GPS collars were monitored in order to investigate their space use and association with their livestock, analyzing two parameters: the dog-sheep distance and the overlap between dog and sheep movement ranges. The first parameter was evaluated by measuring the real distance between pairs of dog-sheep locations taken in less than five minutes apart to ensure the simultaneity of the two events. In addition linear mixed models were implemented to evaluate how dog-sheep distance was influenced by environmental, dog-related, and farming-related variables. UDOI (Utilization Distribution Overlap Index) and the VI (Volume of Intersection) Index for 50% and 95% kernel isopleths were calculated to quantify the overlap and the similarity in the use of space for the core area and for the whole movement range of sheep and dogs. Finally the usefulness of GPS pet collars in dogs and sheep husbandry was tested. LGDs did not leave the flock unattended when left unsupervised. They spent the majority of their time close to livestock, sharing the same areas but using the space in a different way (mean VI 95% = 0.65 ± 0.16; mean UDOI 95%= 1.31 ± 0.56). Dog-sheep distance was mostly influenced by environmental variables and the age of the dog. Dogs and sheep tended to separate more in pastures surrounded by woods (β = 1.669, p <2.2e-16) or located in heterogeneous agricultural areas (β = 1.204, p = 1.33e-05), and less in pastures close to inhabited areas (β = -1.730, p = 2.34e-07). Older dogs were more associated to the flock compared to younger individuals (β = -0.438, p = 0.002). Some of the variability linked to the dog-sheep distance was explained by the importance of the random components of the models, namely: the differences among individual dogs working in pastures with different extension (p<2.2e-16); the day when the sampling was done (p<2.2e-16); and the differences among farms (p=4.87e-07). The effectiveness of guarding dogs as a prevention tool is not only affected by the environmental features or by LGD’s characteristics and training. In fact, to be effective, livestock guarding dogs should work in conditions that allow them to protect the entire livestock. Comparing 79 sheep farms with at least one adult (> 1.5 years old) guarding dog, were highlighted the conditions that decrease the efficacy of these animals in reducing depredations. For each farm were measured: 1) the number of adult livestock guarding dogs; 2) the distance between the farmer's house and the night shelter; 3) night shelter permeability to predators; 4) flock size; 5) shepherd presence; 6) the number of depredation events over the last six months; 7) the depredation risk. Farms were classified on whether or not they experienced depredation over the last six months. The two groups were then compared using non-parametric tests and logistic regressions. Depredated and non-depredated farms differed only by the night shelter-farmer’s home distance value (W = 455, p-value = 0.005). The model averaging showed a significant positive correlation between damage occurrence and night shelter-farmer’s home distance length (β = 4.695 e-04, p-value = 0.0218). These results suggest that in environmental conditions that determine a similar depredation risk, human presence is the main feature that enhances the effectiveness of guarding dogs as a tool against canid attacks on flocks. Investigating the role of some of the ecological variables involved in depredation events helps to ensure that the wolf-livestock interactions occur in a sustainable manner. Indeed depredation risk maps could be a useful tool for farmers and manager for the timely apply prevention techniques that reduce depredation and for policymaker could be a support to allocate financial resources. Additionally conservation projects may benefit from these maps to select areas of intervention. Moreover results from this work provided some hints for farmers and conservationists to improve the use of LGDs for an effective livestock protection: some of the recommendations affected the dog management, while other the livestock husbandry practices. Finally this research introduced a new way to manage LGDs using GPS pet collars. With these devices farmers could be able to check the position of their dogs and their flock at any time, preventing wrong dog behaviors, conflicts with neighbors and accidents.
28-feb-2017
File allegati a questo prodotto
File Dimensione Formato  
Tesi dottorato Zingaro.pdf

accesso aperto

Note: tesi di dottorato Margherita Zingaro
Tipologia: Tesi di dottorato
Licenza: Creative commons
Dimensione 1.63 MB
Formato Adobe PDF
1.63 MB Adobe PDF

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/930371
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact