Habitat suitability and connectivity for the brown bear (Ursus arctos) along the Iran-Iraq border
Kamran Almasieh authorHaydar RouhiShahram Kaboodvandpour
Introduction
Habitat selection and movement of brown bears, Ursus arctos Linnaeus, 1758, driven by the need to meet ecological requirements (e.g. feeding and shelter), are affected by landscape structure (Almpanidou et al. 2014; Ashrafzadeh et al. 2018). In fragmented landscapes, species persistence often depends on connectivity among habitat patches where the linkages among subpopulations are affected by landscape characteristics and movement capacity of the species. Population viability of large mammals depends on large, undisturbed habitat patches (Ziółkowska et al. 2016). Wide-ranging predators such as brown bears that move across large regions are particularly vulnerable to the effects of habitat fragmentation and to the loss of habitat connectivity (Sergio et al. 2008; Almpanidou et al. 2014; Ashrafzadeh et al. 2018). Identifying and preserving a permeable matrix especially for the carnivores is crucial (Kunkel et al. 2013; Dickson et al. 2013; Gray et al. 2016). Therefore, detecting and conserving vital linkages or corridors, which play an important role in maintaining viable populations, should be a priority for conservation (Dickson et al. 2013).
The brown bears have a Holarctic distribution and live in different habitats, from tundra to temperate forests and even semi-deserts (Ziółkowska et al. 2016). They prefer areas covered with high hard mast trees, sufficiently connected forest and high topographic complexity (Martin et al. 2012; Ziółkowska et al. 2016). The brown bears tend to avoid regions. with intensive human activities (Ziółkowska et al. 2016; Piédallu et al. 2017). For example, brown bear presence negatively correlates with high road density and agricultural areas (Martin et al. 2012; Whiteman et al. 2017; Ashrafzadeh et al. 2018). Overall, brown bears need large contiguous areas of habitat patches with sufficient availability of their preferred foods, escape cover and den sites. Thus, it is essential to take the availability of these structures and resources into consideration when analysing the possibility of conservation of the species at a landscape level.
Historically, brown bears were present in large areas of the Middle East, from the Sinaï Desert to Iran (McLellan et al. 2017). Yet, during few centuries, the species has declined dramatically in numbers and geographic distribution owing to habitat destruction and poaching (Can and Togan 2004; McLellan et al. 2017). In Iran, brown bears have disappeared from most of their former range (Ashrafzadeh et al. 2016). The Iranian populations of brown bear are currently distributed over north, north-west and west of Iran but with relatively low densities and decreasing population size because of deforestation and heavy human pressure (Gutleb and Ziaie 1999). The national law protects the brown bears in Iran; nevertheless, they are sometimes illegally killed (Ashrafzadeh et al. 2016).
The brown bear population in Iran is estimated at 1200–1800 individuals (Sathyakumar and Can 2007; Ashrafzadeh et al. 2016). Because of habitat fragmentation, they are divided into three divergent populations: (1) the Hyrcanian forests in northern Iran; (2) the Central Zagros Mountainous range in western Iran; and (3) the Iranian Caucasus in the north-west of Iran (Sathyakumar and Can 2007).
The brown bears in the Iranian Caucasus in the north-west of the country are mainly found in protected areas, but they also extend beyond the boundaries of these protected areas. The prevalent ecosystems of the Iranian Caucasus are mountains and semiarid steppes. Agricultural lands, subalpine meadows and grasslands are intermixed with forest patches. The Iranian Caucasus is widely utilised for livestock grazing, agricultural activities and collecting of firewood and medicinal plants.
In the present study, we investigated whether the existing conservation areas safeguard the connectivity of brown bear habitat across the Iranian Caucasus. Because forest fragmentation probably increases the vulnerability of bears when moving among patches (Caussimont and Herrero 1997), it is essential to map forested areas at a landscape level and evaluate their degree of connectivity and fragmentation. Developing an understanding of brown bear distribution, as well as the quality and connectedness of its habitat can contribute to improving brown bear management and, thus, help in the long-term conservation of its populations in the Iranian Caucasus, in the face of increasing human activities.
Materials and methods
Study area
Our study was conducted in the north-west of Iran, in an area of ~47 830 km2 (East Azerbaijan province) mainly in the Caucasian part of Iran (Fig. 1). This area includes the following terrestrial ecoregions (Olson et al. 2001): Caspian Hyrcanian mixed forests, eastern Anatolian montane steppe, Elburz Range forest steppe, Azerbaijan shrub desert and steppe, and Zagros Mountains forest steppe. Fifty per cent of the province is covered by mountain ranges, of which the three major ones are the Qara Daq (Arasbaran), the Sahand-Bozqoosh and the Qaflan Kooh. The most prevalent climates are warm humid continental climate and cold semiarid climate. Mean annual temperature and precipitation are 10.2 C and 315.2 mm in East Azerbaijan province respectively (Zarghami et al. 2011).
Data collection
We used 91 point localities of brown bears recorded during our fieldwork and by the Iranian Department of the Environment (DoE), collected using both direct observations and camera- trap photographs of bears, between 2010 and 2016. Thus, we used only presence points that were validated by our study team and the wildlife experts of DoE. So as to reduce spatial autocorrelation, we excluded all sites of multiple occurrences within a minimum distance of 2.5 km (corresponding to the mean maximum distances moved (MMDM) by brown bears, i.e. 5 km (reported ranges between 0.55 km (Naves et al. 2001) and 10.4 km (Huber and Roth 1993)). For our analysis, we used
65 unique records of brown bear presence.
Habitat variables
We selected four sets of environmental variables commonly used to describe the conditions within the brown bear habitat (land cover, bioclimatic, topographic and anthropogenic variables; Table 1). These variables are relevant to the ecology and distribution of the species (Favilli et al. 2013; Ashrafzadeh et al. 2018). We used land-cover data from the Iranian Forests, Ranges and Watershed Management Organisation (IFRWMO 2014). The land-cover dataset was reclassified into eight classes (forest, agriculture, rangeland, bare land, water, urban and other). These classes were chosen because brown bears are a generalist, opportunistic species with a high potential to adapt to different habitats and to human activities (Kanellopoulos et al. 2006; Ghoddousi 2010; Favilli et al. 2013). Land cover is considered the most important factor because it is related to food resources, hiding and thermal refuge, and human disturbances (Kanellopoulos et al.
2006; Favilli et al. 2013). We used neighbourhood analysis for calculating the proportion of each cover type (Table 1)
within a radius of 2.5 km (based on the brown bear MMDM) centred in each cell. We also considered the normalised difference vegetation index (NDVI) as a reliable measure of the spatial and temporal patterns of primary productivity, also for predicting herbivore and non-herbivore distribution (Mueller et al. 2008; Pettorelli et al. 2011; Farhadinia et al. 2015). NDVI correlates with the seasonal average energy availability in brown bear habitats (Wiegand et al. 2008). Moderate resolution imaging spectro-radiometer (MODIS) data (MOD13A2 16-day 1-km VI, available at https://lpdaac. usgs.gov, accessed 8 December 2015; Didan and Huete 2015) were used to derive the NDVI layer. Annual mean temperature and the annual precipitation sum, as well as their variation throughout the year (i.e. temperature seasonality and precipitation seasonality), were extracted from the WorldClim database (Hijmans et al. 2005).