Landscape Ecology of Grey Wolf (Canis lupus) in Northern Pakistan

Abstract

Background: The grey wolf (Canis lupus) has been extirpated from most of its historic range in Pakistan primarily due to its effect on livestock and livelihoods. Lack of scientific knowledge is likely to be an important issue in Asian countries. The scientific information is necessary to inform practical conservation strategies, which are unfortunately lacking across the distribution range of the grey wolf in Pakistan. Field research on the grey wolf is challenging because it lives in remote areas, and its low-density population makes it difficult to monitor. Such scarcity of scientific information hinders effective conservation planning of the grey wolf. Current management efforts are ad-hoc, which may or may not deliver conservation targets. Effective conservation of the grey wolf, therefore, requires a detailed understanding of factors that govern the specie’s spatial distribution and habitat use. Information about resource availability and selection, conflict with humans, and feeding ecology are important requisites to devise a conservation strategy for the species. The grey wolf has been extirpated from most of its historic range in Pakistan primarily due to its effect on livestock and livelihood of rural people. It is often considered a problem animal in mountain ecosystems in northern Pakistan. Methodology: The first objective of the study was to understand the habitat requirements of the wolf and update its geographical distribution in Pakistan. Habitat suitability models are useful to understand species distribution. The study used non-invasive survey data from camera traps and genetic sampling to develop a habitat suitability model for the grey wolf in northern Pakistan and explored the extent of connectivity among populations. The suitable habitat was investigated through Maximum Entropy (Maxent) approach (Maxent ver. 3.4.0) and movement corridors were identified using the Circuitscape 4.0 tool (McRae and Shah, 2011). After the identification of a suitable habitat, the study investigated local sites used by wolves through occupancy modelling using camera trap data. The probability of wolves encountering camera traps is a function of the micro-habitat where the camera traps were installed. Different combinations of survey and site covariates were tested using the Presence software (version 12.7_170921, Hines 2006) to describe variation in the probability of detection and markability and occupancy of wolves at the site level. The third aspect of the study related to understanding the socio-ecological factors that contribute to human-wolf conflicts. Interviews with 2,317 local people were held and a mixed-effect nested negative binomial regression model was constructed using the “lme4” package and the “pscl” package in R (software) (https://cran.r-project.org/) to determine the influence of socio-ecological factors on livestock losses due to grey wolf predation. Once the average rate of livestock predation was estimated for different valleys, it was extrapolated to entire study area using geostatistical analysis in ArcGIS (10.2). Determining the diet profile of a predator is essential to understand its role in regulating given ecosystem and rationalizing public claims of livestock depredations. I analyzed wolves’ scats to document diet composition and prey selection of the species. We searched and collected 1,186 unidentified scats samples belonging to different carnivore species found in the suspected grey wolf range, in both protected and non-protected areas, from 2009 to 2017. Upon genetic validation of the samples through mitochondrial DNA analyses, confirmed wolf samples were further processed for physical analysis. We estimated biomass consumption rate by following the linear relationship developed by Ackerman et al. (1984). Results: The habit suitability model predicted ca. 23,129 km2 (15% of study area) of area to be potentially suitable for the grey wolf, though most of the suitable habitat was in remote and inaccessible areas that appeared to be connected through movement corridors. The Maxent model demonstrated high levels of predictive performances (area under curve: 0.971±0.002, and true skill statistics: 0.886±0.021). The main predictors for habitat suitability included the distance to road and rivers and the mean temperature of the wettest quarter. The corridor modeling generated estimates of habitat connectivity among scattered wolf populations in northern Pakistan. Four patches of suitable habitats were identified within the Himalayas, Pamirs, Hindukush, and Karakoram mountains ranges of northern Pakistan. These habitat patches have strong but unprotected connectivity and corridors of movement exist between all major habitats. The model identified weak linkages between populations found at lower altitudes with high disturbance rates. The mean occupancy of wolves across the sampling sites was 0.43 (SE = ±0.09). Detection probability for wolves varied across different types of terrain and the effect of efforts and plateau had a positive influence on it, whereas the cliff base had a negative effect on probability of detecting wolves on cameras. In the top three models, three site covariates including average distance to road, average distance to settlements, and available habitat were found to collectively affect the probability of wolf occupancy. The human-wolf interaction analysis revealed that 7,583 livestock were killed from 2011 to 2015. Small ruminants were more likely to be killed by wolves as compared to large livestock. Wolves preferred to attack small ruminants in open pastures and large ruminants in corrals. Several factors were positively correlated with livestock predation, including the size of livestock holding, mortality due to diseases, and occupation of the owner (livestock rearing). The negative binomial model suggests that predation incidences tend to decline as the education level of respondents increases. In contrast, predation counts doubled when livestock holding increased. Livestock disease was a stronger cause of loss per household with an average of 3.5 animals dying annually. Diet analysis, confirmed that wolves consumed 12 different prey species, including wild and domestic ungulates. Domestic ungulates were a substantial part of the wolf diet (49%) and goats were the most common domestic prey (19%), followed by sheep and yak (12% each). Wild prey constituted 51% of the wolf diet with a major contribution of Himalayan ibex (18%), followed by blue sheep (7%) and Himalayan marmots (7%). A minor fraction of small mammals and birds were also consumed. In terms of biomass consumption, livestock dominated (62.2%), followed by wild ungulates (28.7%). Conservation of the research species: The study demonstrates that non-invasive study techniques such as camera trapping, genetics sampling, and interviews in combination with statistical and analytical tools is a promising approach to understand landscape ecology of threatened carnivore species. The study represents the first large-scale assessment of wolf distribution, habitat suitability, and movement in Pakistan. The identified movement corridors suggest that wolves can potentially expand their range in northern Pakistan in the presence of a conducive environment. Management of protected areas with rigid restrictions is not a practical solution in northern Pakistan, partly due to heavy dependence of people on natural resources. Habitat suitability map generated in the study can help authorities in identifying key conservation areas where management efforts should be focused. The occupancy estimates, being the first effort in Pakistan, set a credible baseline against which change in wolf population and distribution could be gauged in the future in response to conservation efforts. The outcome of this study has policy implications for managing human-wolf conflicts. Increasing population of livestock is fostering conflicts and undermining support for wolf conservation in Pakistan. The study highlights livestock loss by the wolf as a real and serious problem and this issue needs to be addressed to promote wolf acceptance in the rural communities of Pakistan. Better guarding could prevent wolf attacks, particularly in a high-depredation season. I recommend establishing veterinary centers with the allocation of substantial budgets to reduce livestock mortality due to diseases. A major obstacle for wolf conservation and extensive killing is livestock depredation in the area. The study suggests multi-pronged conservation management programs that include sustained compensation schemes, livestock vaccination, and awareness campaigns to sustain the co-existence of wolf populations with rural communities.