Indigenous mountain people’s risk perception to environmental hazards in border conflict areas

This study aims to understand community risk perception to environmental hazards in a border conflict zone context in high-mountain areas. Participatory rural appraisal (PRA) tools were applied by the social science team. The results were validated with a hazard map prepared by a separate team comprised of geologists. Turtuk, the northernmost village in Ladakh, India located near the line of control with Pakistan was undertaken as a case study. Turtuk represents a high mountain indigenous rural community which has experienced several catastrophic disasters (flash flooding and landslides in 2010, 2014, and 2015) and territorial armed conflicts (wars in 1971 and 1999 with Pakistan) in recent times. The villagers were able to identify various environmental hazards and associated risk zones through participatory timeline diagram, and hazard and dream mapping exercises. The PRA maps matched the geological hazard map of Turtuk, demonstrating that community people are highly aware of surrounding hazards regardless of differences in age, sex, education, occupation, and re-ligion. They apply indigenous knowledge to deal with the adverse climate and calamities. The technique, of analysing community vulnerability in the context of conflict and disasters by applying qualitative PRA tools and validating the mapping results, as piloted in this study is novel and replicable in any disaster setting.

and intense conflict can trigger disasters. The mainstream DRR community has long neglected the issue of confronting disasters in fragile and conflict-affected areas, although 58% of disaster deaths occur in those states [46]. As per the UN Environment Programme (UNEP), more than two billion people have been affected by disasters and conflicts since the start of the new millennium, as the world has witnessed more than 2500 disasters and 40 major conflicts. Only in 2017, there were 30.6 million new displacements associated with conflict and disasters globally. 39% of them were triggered by conflict and 61% by disasters [33]. The World Bank projects, climate change induced extreme disasters could force over 140 million people to migrate within countries by 2050 [54]. These crises destroy livelihood, ecosystem and infrastructure, and displace people.
Apart from environmental hazards, armed violence or war is considered as the leading social or man-made agent for triggering disasters in the mountains or high-hill areas [30]. Vulnerable mountain populations are most affected and contribute to the rapidly growing number of refugees and displaced people; some displaced for generations. In conflict zones, disaster-prone communities have shifting and intensified vulnerabilities as they face restrictions on free-movement that hinders emergency relief and evacuation and finding safer places for relocation. E.g., the landslide risk profile of Lebanon has increased by 75% due to the arrival of over 1.5 million displaced Syrians fleeing armed conflict [49]. One contemporary example is the recent (since August 2017) mass exodus of a million Rohingya refugees in the highly landslideprone hilly district of Bangladesh -Cox's Bazar. The Rohingyas fled from Myanmar's Rakhine State to escape armed conflict, genocide, and serious violations of their human rights by the Burmese military. Around 200 km 2 of hill forests have been destroyed (and continuing) to build makeshift huts and arrange fuel for cooking for them. They are forced to live on dangerous hill-slopes and flood-prone areas, which is now hosting the densest concentrations of refugees [1,4]. It demonstrates how conflict in a neighbouring country can destabilise regional peace and security, and increase threats of environmental hazards in another country. Community vulnerability is dynamic, varies across temporal and spatial scales [34,70]; and depends on social, economic, demographic, cultural, political, institutional, and ecological factors [51,68]. Risk perception refers to people's subjective awareness of risks, i.e. the likelihood of an adverse effect resulting from the occurrence of a hazard in a vulnerable setting [47]. How people judge the risks posed by environmental hazards in a given geographical area depends on a combination of influencing factors at both individual and collective levels. Natural scientists might rely on technical processes to collect and interpret data from past disasters to better anticipate the probability and magnitude of future events. Social scientists might examine a range of historical, socio-economic, cultural or political factors to better identify vulnerable settings. Local authorities might consider risks in relation to other development priorities, whereas local inhabitants might rely on their own observation, experience and understanding of risks that they are exposed to [48,52,57].
It is important to assess how people understand risks, why people decide (or not) to deploy protective strategies (e.g. [29]) or why some disaster preparedness measures have been implemented more than others. In Trinidad, Martin et al. [41] found that people's past experience of floods influenced their high perception of flooding risk. [28] found that the marginalised mountain communities in the Gilgit-Baltistan Province of Pakistan adopted strategies such as livelihood diversification and male labour migration to cope with cooler summer and erratic precipitation. In the same region, Azhar-Hewitt and Hewitt [10] revealed that social conditions are responsible for making the local women vulnerable to mountain hazards. In the case of New Zealand, the relationships between personal, social and civic agencies were essential in managing volcanic disasters [43].
Social scientists have also argued that a high perception of risk does not necessarily lead to protective behaviours [67]. In the Philippines, Gaillard [23] shows that the socio-economic pressures that people face in their daily life (e.g. poverty, lack of access to basic services) weighed heavier than the threat posed by recurring natural hazards such as mudflows. People's behaviour towards risk also depend on non-hazardrelated factors such as structural constraints [24], and trust in governance [67]. Overall, risk perception studies have produced an extensive literature, including in the context of mountain hazards [12,38,40]. However, the evaluation of risks in multi-hazard mountain environments where conflicts and border disputes also constitute a threat has received much less attention.
Indigenous people usually live within geographically distinct ancestral territories, self-identify as indigenous or tribal, and typically aspire to remain distinct culturally, geographically and institutionally [64]. Indigenous knowledge is characterised by originating within the community, maintaining a non-formal means of dissemination, collectively owned, developed over several generations and subject to adaptation, and embedded in a community's way of life as a means of survival [56]. In the Chittagong Hill Tracts of Bangladesh, traditional forest management system, housing design, beliefs and values, social structure and land tenure system help the tribal communities to tackle landslides [2]. The nomadic people of the central region of Mongolia apply traditional knowledge and practices such as seasonal migration, taboos and rituals, and maintain variations in livestock for grazing to combat pasture degradation and desertification due to climate change [69]. Despite some contemporary works, another critical research gap has been identified as the role of cultural knowledge and indigenous people's risk perception in DRR [15,6].
Risk perception research related to environmental hazards is primarily dominated by random sampling-based household questionnaire surveying [5,7] and quantifying the results by applying multivariate statistical and Pearson correction methods [50], and linear regressions [8]. Kelman et al. [36] argue -DRR research is overlooked by quantitative statistical approaches that are not able to incorporate all the various dimensions of social vulnerability, and fail to distinguish between the people and the natural world. In contrast, participatory rural appraisal (PRA) technique, developed in the 1990s [17], has enabled people's realities to be captured by shifting a closed system (e.g. structured interviews, self-administered questionnaires or controlled experiments) to an open system [13]. In PRA, local people determine what goes into a diagram or map. By ensuring group work and democracy on the ground, PRA increases the level of interaction with local people and builds rapport with externals [19,37,53,58]. PRA methods stress more on capturing how communities as a whole interpret their social world, compared to pre-set structured-questionnaires [42,45]. PRA emphasises on the contextual understanding of social behaviour and social system (i.e. heritage, values, beliefs and culture) as a process that can be investigated in real time [13,2]. PRA tools should be prioritised into vulnerability assessments to enable understanding of the full scenario [3,61,66].
In summary, findings from the literature review suggest that there are specific gaps in assessing indigenous communities in the Himalayas, and their overall social vulnerability to disasters in the confluence of environmental hazards and conflicts. This article intends to overcome such limitations by giving voice to the mountain people through PRA method to depict their vulnerability and propose solutions based on indigenous knowledge.

Kashmir and Turtuk -a history of conflict and disasters
Kashmir (the former 'Princely State of Jammu and Kashmir') has a long history of border conflict as the territory had been administrated by a number of rulers notably the Mughal Empire, the Durrani Empire, the Sikh Empire, and the Dogra Dynasty [11]. Kashmir was politically an independent territory even after the independence and partition of India from the British Empire on 15 August 1947. The last ruling Maharaja of Kashmir from the Dogra Dynasty, Maharaja Hari Singh, agreed to the 'Dominion of India' by signing the 'Instrument of Accession' document on 27 October 1947 to defend a tribal rebellion attack backed by Pakistani troops [26]. Since then major border disputes followed by wars have been persistent between India and Pakistan. Both countries fought five major wars (i.e. 1947-48, 1965, 1971, 1984, and 1999) since their partition from the British Raj in 1947. Out of these, four were directly related to taking possession of Kashmir, and the 1971 war was linked to the creation of Bangladesh because of the internal conflict between the East and West Pakistan [20,26,65]. The Aksai Chin territory of Kashmir that separates China from India is administered by China despite India's claim upon it.
At present, the line of control (LOC) divides the northern areas including the Pakistan Occupied Jammu and Kashmir, and the Indian State of Jammu and Kashmir. India has been occupying the Siachen Glacier since 1984 and yet there is no internationally defined boundary between India and Pakistan (Fig. A1). Minor armed conflicts are common in areas adjacent to the LOC. For instance, on 23 December 2017, Pakistani soldiers attacked an Indian Army post near the LOC and killed four Indian Army personnel including a Major. The Indian Army retaliated by crossing the LOC and killing three Pakistani soldiers [63]. Eventually, these conflicts mean that the indigenous mountain communities residing near the bordering areas suffer hugely.
An example of such an indigenous mountain community from Kashmir, where it is attempted to study community risk perception to environmental hazards, is Turtuk (also known as Turtok). Turtuk is the northern-most village in Leh District in Jammu and Kashmir located at 34.847°north latitude and 76.827°east longitude ( Figs. 1 and A1). The line of control is just 6 km away from the village (Fig. 2a). A permanent Indian Army base camp is located in between Turtuk and Tyakshi. India took control of this village from Pakistan during the 'Indo-Pakistani War of 1971' which is also known as the '1971 Liberation War of Bangladesh'. The villagers in Turtuk witnessed another major conflict with Pakistan during the Kargil War in 1999. As it is forbidden to cross the border, the community is isolated and can be accessible by only one motor road. It impacts tourism activities, health care facilities, and other social and humanitarian works that can be considered as barriers to sustainable development.
Turtuk is predominantly a Shia Muslim village where 3371 people (from 384 households) live in an area of 1.53 sq.km. The residents mainly speak Balti and Urdu, and they belong to the Baltistani ethnic group of Tibetan descent [16]. Turtuk is located at 3000 m above sea level on the banks of the mighty Shyok River which flows in the eastwest direction. A tributary stream, locally known as 'Turtuk Lungpa', flows in the south-north direction from the hills to the north (Fig. 2b).
Turtuk is characterised by cold-arid climate and is bounded by the Karakoram Range to the north and the Ladakh Range to the south. The Ladakh region is categorised into four seasons: winter (December-March), pre-monsoon (April-June), monsoon (June-August), and post-monsoon (October-November). Regionally, the temperature can rise to 34.8°C in summers and drop to as low as −27.9°C in winters. The average precipitation ranges from 0.5 to 1.5 mm/day throughout the monsoon and cloudbursts occur during the same period. Roads are hardly accessible during the winter season and the village becomes isolated. The region is experiencing a rapid increase in temperature and varied precipitation patterns in recent decades [18], which is adversely impacting the mountain communities.
Being surrounded by high altitude mountains and glacial melt rivers, flooding, rock fall and landslides frequently hit Turtuk. The region is also vulnerable to earthquakes [22]. As narrated by the local people, Turtuk was severely damaged during the 2010 and 2015 flash floods. The Indian Army base camp was previously located just beside the intersection point between the Turtuk Lungpa and Shyok River. During the July 2010 flash floods the old camp was devastated (Fig. 3c) and 18 soldiers were killed. Pakistani soldiers bombarded the same area during the Kargil War. Just opposite to the old Indian camp, a memorial is erected (Fig. 3a) where an Indian soldier (Sapper Satish Kumar) was killed due to enemy shelling on 26 July 1999. Therefore, Turtuk village sets the ideal context for investigating an indigenous mountain community that face harsh climatic conditions and is concurrently affected by environmental hazards and border conflicts.

Methodology
The authors of this article were divided into two groups: the social scientist group was responsible for conducting the community-based forensic workshop, and the natural scientist group prepared the geological hazard map of Turtuk. Before starting the fieldwork, activities such as ethical approval, risk assessments and security checks were undertaken. They included UCL ethical approval (project ID: 6141/ 001), UCL data protection registration (ID: Z6364106/2017/05/83), UCL risk assessment, authorization from the University of Jammu, Indian visa approval, security inspection at various Indian Army checkposts from Leh to Turtuk, and lastly fieldwork permissions from the Turtuk Army Camp and the local people. The work was conducted as part of the project -"Increasing Resilience to Environmental Hazards in Border Conflict Zones" funded by the Natural Environment Research Council (NERC Reference: NE/P016138/1).

Participatory forensic workshop
A two-day participatory 'forensic' workshop was held from 6 to 7 July 2017 in the Higher Secondary School, Turtuk that is located near the Shyok River (Fig. 2b). A forensic workshop is defined as a series of community-based activities in focus groups that helps to understand the root causes of community vulnerability and the best possible alternatives in building resilience. The term 'community' is used synonymously as a 'case study area' that represents the Turtuk village. The workshop focused on historic hazard events and their association with social capital, vulnerabilities and their development. The community people and concerned experts or key informants shared ideas, defined problems, mapped hazard and vulnerability, set priorities, suggested solutions within their remit and talked about future aspirations through collaborative activities.
The workshop hosted 36 participants, who along with other villagers were invited beforehand, covering local adults and representatives from the local government/autonomous/private institutions, school teachers, health care professionals (doctor and nurses), Army officer, Imam of the Mosque, politicians, drivers, farmers, members from the Gram Panchayat or village council, tourist hotel owners, and university and college students, etc. The workshop participants were purposefully divided into four groups: political (total 7 participants), administrative (10), local women (11), and local men (8). In order to understand the differences or similarities in their risk perceptions, each group was assigned similar tasks. The project team members took the role of facilitators and a translator accompanied each group. The facilitators explained the workshop activities, PRA exercises, ethical issues, and oral consent was taken from the participants as a group for participating in the sessions, taking photographs and audio-recording some parts of the event.
PRA-based qualitative research helps to obtain information on local knowledge, people's risk perceptions and experiences over time, social processes and contextual factors, and local responses in mitigating disaster risks [59]. Timeline, hazard mapping (after preparing social and resource maps), and dream mapping PRA tools were applied [3]. A 'timeline' helps to identify landmark events (war or disasters) and capture their history as recalled by the local people. This helps us to learn what the community consider to be important past events and their historical perspective on current issues [19,37]. Visualisation through participatory mapping helps participants to see and understand the inter-connections between various issues related to disaster vulnerability in a local context. Social and resource mapping is used to illustrate the overall spatial dimensions and natural and physical exposure to hazards. The maps included locations of settlements and infrastructure such as roads, health centre, schools, bazaar, shops, water points, and playgrounds; and natural resources such as land, hills/ mountains, rivers, streams and irrigation canals, agricultural fields, apricot orchards, and forests. After participants had prepared their social and resource maps, they were asked to delineate the areas, houses, community facilities and infrastructure that they perceive are vulnerable to various environmental hazards. This is known as 'hazard mapping' [19,53]. On the following day, participants were invited to undertake 'dream mapping' where they re-drew their social maps to depict their hopes and aspirations regarding building a resilient community concerning the hazards depicted earlier focusing on what would ensure sustainable development for the community.
The workshop venue was the Higher Secondary School Hall, which was well equipped with large tables and chairs. Each community group was provided with large A1 sized papers and necessary stationery. Participants were encouraged to discuss within each group and only basic guidelines were provided to initiate the mapping process. The participants drew the maps as per their local and indigenous cultural knowledge, and the authors mostly observed. After completing each task, facilitators asked the participants to explain the diagram or maps (Fig. 4) in group presentations. Field notes were taken in as much detail as possible and were later incorporated in this paper. The large maps were then scanned using a drum-scanner and digitalised (no alteration was made).
The PRA maps and diagrams were triangulated through field verification and validation. For this purpose, the PRA maps were compared with the actual hazard map prepared by the natural scientist team. The activities were performed in separate groups who did not communicate results to one another at the time to ensure the legitimacy and triangulation of the results. Researchers have applied PRA methods to understand community risk perception to environmental hazards [19,3,61], however, comparing and validating the results with actual geological hazard map, as demonstrated in this work, is unique and novel.

Geological hazard mapping
A geological hazard map of Turtuk was prepared by integrating the field-based investigations and remote-sensing data (satellite imagery) and using ArcGIS software. The field studies involved geological and geo-hazard mapping using the base maps (Survey of India topographic sheet numbers: 52F/1, 52F/2, 52F/5, 52F/6, 52F/9, 52F/10, 52F/11, 52F/13, 52F/14, and 52F/15), and ground checks of the data obtained from the high and moderate resolution panchromatic Landsat 8 images of the region. Additional data (land use and land cover, population census 2011, geology, and boundaries, etc.) were obtained from Bhuvan -the Indian Geo-Platform of Indian Space Research Organization.

Results from geo-hazard specialists' hazard mapping
The natural scientists observed that in Turtuk, continuous tectonics and geological processes, coupled with anthropogenic activities and climate change had intensified numerous hazards viz. earthquakes, landslides and other down-slope movements, floods and cloudbursts (Fig. A2). The seismicity in the area associated with the regional and local faults is poorly understood, and the seismic hazards have not been investigated thoroughly. Intense rock deformation has made the mountainous terrain very fragile, weak and highly susceptible to downslope movements. Slopes that are susceptible to these hazards have been identified towards North, South and Southwest, where the slopes range from being steep (~25°inclination) to very steep (up tõ 45°inclination). The rocks depict three prominent closely spaced and dense fractures dipping towards Southwest, Northeast and Northwest.
With the elevation of 2500-3700 m, the drainage in the area ranges from being immature (first order) to mature (sixth order), depicting an evolving topography and dynamic landscape. The occurrences and interactions of the rock layering, rock fractures and escarpment slope (inclination) are suggestive of a serious landslide hazard in the area. Additionally, three boulder terraces with steep and vertical scarps pose a significant threat of rock-fall (Fig. 5).
Flood hazard in Turtuk has been underestimated; here floods (resulting from cloudbursts and rapid glacial melt) can be very severe and powerful, exhibiting unpredictable flow paths and high velocities. These floods may cause considerable erosion in some areas while depositing large amounts of sediment and debris in others. This entire area is covered with intensely fractured and deformed brittle rocks and very loose eroded material that is highly susceptible to downslope movement and debris flow. The Shyok River flood-plain covered with a recent accumulation of sediments has been demarcated as highly prone to flooding. The south-eastern zone, covered by dissected, deformed and fractured rock slopes, is categorised as highly vulnerable to mass movements. Most of the households are located in these two zones and are highly vulnerable to hazards as mentioned above. Settlements on the western side of the 'Turtuk Lungpa' are identified as highly susceptible to flooding, debris flow and flash flooding (Fig. 5). In Turtuk, detailed hazard analysis and mapping are required for micro-zonation of hazards to delineate the safer zones.

Findings from the historical timelines
Past events in Turtuk were traced back to Kumdan glacial lake outburst floods (GLOFs) in 1926, 1929 and the 1930's somewhere between 1935 and 1938. Respondents were unsure of the exact date but the Kumdan flood was well known, especially to the older people. According to some respondents, some foreigners intervened to break the blockage and drain some of the water. Kumdan Lake, situated in Tibet, caused widespread damage at that time and impacted the whole Nubra valley from Khardong village to Turtuk and beyond in Abbottabad (now in Pakistan). According to some older people, the Kumdan outburst was the second one and that another GLOF had occurred earlier, but they did not have any knowledge about that event. In 2010, floods/cloudbursts damaged many agricultural fields especially in the valley where they graze livestock or grow crops. In that cloudburst a local man and 22 Indian Army soldiers were killed. At that time the community received assistance from the Army, but the blockage of roads from Leh to Turtuk meant that the government could not help. At that time both men and women of Turtuk and Tyakshi helped to rescue affected people. Being situated on the LOC between India and Pakistan, people of Turtuk, Thang and Tyakshi feel that their lives are at risk.
With the increase in environmental disasters in Ladakh and especially in the Nubra Valley potential losses from these disasters include agriculture, damage to access roads, communications, water supply, livestock and damage to infrastructures like homes, hospitals, schools and government offices. In 2009, the Indian government opened this area for tourism, since then it has become a top tourist attraction in Ladakh. The influx of tourists has created income opportunities for the villagers and they are converting homesteads to guesthouses. The new settlements/hotels/resorts are being constructed in hazardous areas due to lack of safer places in Turtuk.

Findings from the hazard and dream mapping
The workshop participants expressed that they are concerned about floods, cloudbursts and landslides (Figs. 6-10). They were aware that the settlements below the mountains in the Kharmang area are landslide-prone and that those who were settled near the Shyok River and stream are prone to flood. The area near the school and primary health centre are locally known as 'Chu-thang' which means 'the water area'. These are at high risk of flooding from the Shyok River because they are situated beside the riverbed. A stream divides Turtuk village into two parts named Yul and Farol and a wooden bridge connects them. This Turtuk stream is one of fastest flowing streams in Ladakh, so Rantakchan area has a high risk of floods. The dream maps of all the groups said that they need embankments to prevent flooding of the stream and the Shyok River.
A summary of the hazards discussed, community awareness of vulnerabilities to the hazards, proposed risk mitigation strategies and other developmental changes proposed by each group is provided in Table 1 (for more details please see Appendix, Tables A1-A4). Each of the four stakeholder groups of community men, community women, political leaders, and administrative officials identified similar kinds of hazards and mitigation strategies. All four groups identified cloudburst/ flash flood, large-scale flooding of the Shyok river, rock fall and landslide hazards and all suggested building of levees or embankments, upgrading of health facilities and road building as risk mitigation strategies. Other hazards included earthquakes (though there had been no large earthquake in living memory) identified by all except political leaders, and war (men and officials). Other important risk mitigation strategies mentioned by all groups except political leaders included building fences or walls to prevent falling boulders from destroying settlements and orchards, and improved mobile phone communications (towers/4G connectivity). Community men and women both identified improving bridges so that they are higher above the rivers, community men suggested stabilising slopes with cement, and building disasterresistant buildings and improving government policy for peace-building with Pakistan. Community members felt that communication problems increased their vulnerability. The mobile network is severely restricted with no Internet connection, and phone calls to relatives in Pakistan are prohibited (though incoming calls from Pakistan are permitted). Specific non-structural mitigation measures included peace talk with Pakistan, drawing tourists for economic benefit, shift in agricultural practice, creating a DRR taskforce, preparing risk sensitive plans, sustainable tourism activities, and undertaking awareness programs on disaster management and climate change adaptation. The administrative officials and the community women showed more awareness of vulnerabilities than the political leaders and community men. All mentioned the risky positioning of buildings along the riverbed flood zone and in the path of falling boulders. Climate change and the influence of increased rainfall was mentioned by political leaders and women, while only the administrative leaders mentioned poorly constructed buildings and building on previously agricultural land as increasing vulnerability.
During dream mapping common visions across all the four groups for developing the community included upgrading educational facilities (particularly to make a college of higher education), constructing new roads and making new recreational facilities such as sports arena, horse polo and ice hockey grounds and picnic areas. This was also associated with the development of tourism including the building of guesthouses and car parking area and developing trekking routes. People are concerned about transportation, communication, medical facilities, education, unemployment and development of tourism in the region. Generally, the administrative officials had a longer list of potential changes than others including intensifying agriculture, developing a waste-management system, reducing corruption, building animal husbandry clinics and creating a state disaster response force. In contrast, political leaders mentioned the building of a helipad, making an emergency shelter, and establishing a sub-district magistrate office.

Triangulation of the results
It was noticeable that the community awareness of hazards and their PRA maps overlapped quite precisely with those identified using geological mapping. What appears to differ between geologists'/disaster risk reduction (DRR) specialists' and community members' perspectives is the awareness of vulnerabilities and choice of the risk mitigation strategy. For example, where DRR specialists may have proposed shifting important infrastructure (health facility, secondary and middle school and police station) away from the flood plain, community members gave this less importance. Similarly, DRR specialists would have focused more on disaster-resilient buildings than the community proposed. The community's focus and apparent trust in the building of concrete levees to contain flooding seemed to be based on very little evidence of the effectiveness of this method to contain flash floods and major flooding events of the large Shyok River. However, the community's prioritisation of road and mobile phone communication and development of health system infrastructure is in line with the kind of strategies that might be proposed by a DRR specialist.

Discussion
The Turtuk mountain community is found highly aware of their surrounding disaster risks regardless of differences in age, sex, education, occupation, religion, ethnicity, and family status. Because the villagers witnessed catastrophic events, especially floods and landslides/rock-falls, and major armed conflicts (in 1971 and 1999) during their lifetime. The results contradict with the work of Azhar-Hewitt and Hewitt [10] who found the local women more vulnerable to disasters in Gilgit-Baltistan, Pakistan. Another misconception exists, i.e. a certain culture or indigenous knowledge can be a significant factor in creating higher levels of vulnerability for a particular community. For example, after the Indian Ocean tsunami in 2004, a group of people in Aceh province in Indonesia argued that the disaster was a punishment from God [44]. No such contradictions between culture, traditional knowledge and disaster risk perception were observed in this study. All the groups managed to depict the major historical disaster events and existing hazardous zones in PRA activities. In most cases, the groups prioritised similar strategies to mitigate the disaster risks, except some minor differences. The women-only group emphasised on better communication facilities by proposing a better access road through Tyakshi to connect Pakistan and another road in the northeast to Hanu, and more green spaces, parks, playgrounds, shops, apartment blocks and guest houses to promote tourism (Table A3). The men only group focused on sports facilities, specific zone for agriculture and better mobile networks (Table A4). The administrative group, included both men and women, proposed to build bunkers to protect themselves from potential wars, to open up the nearest border with Pakistan so that it would increase economic growth, reduce corruption in infrastructure development, and create a specialist disaster response team (Table A2). The political group (no woman member was involved) highlighted to build a helipad for tackling emergency crisis, an advanced recreational and industrial hub, produce apricot and lavender plantation, and a college for higher education (Table A1). To be specific, all the groups were found to be knowledgeable about potential hazards within their community and proposed a number of structural and non-structural disaster mitigation measures.
A long-lasting war in the border region is having adverse impacts on the community. Most families in Turtuk have close family members on the other side of the LOC. Nevertheless, they are not permitted to cross the nearest border point, which is only a few kilometres away. They are forced to take long routes, travel for several days in the other way around and need special permissions to visit their relatives in Pakistan. Even they have a specialised and controlled mobile network (with no internet services) in Turtuk that hinders the villagers to communicate frequently. It is also difficult to enter Turtuk, as foreigners and tourists require special permissions and need to cross multiple Army check posts from Leh to Turtuk due to security reasons. The village gets isolated during the snowy winter season. Any environmental disaster occurring during the winter has the potential to impact the community on a catastrophic scale, whereas a conflict or war at the same time can worsen the scenario. Communities in border conflict areas always deal with intensified vulnerabilities, in compared to non-conflict zones. For example, the villagers in Turtuk face restrictions on free-movement, communications, accessibility to modern technology and basic facilities and services. Whereas communities vulnerable to earthquakes in Nepal give more importance on social cohesion, relationships, gender inclusiveness in risk governance, neighbourhood, violence against women and girls, and festivals for sustainable disaster recovery [60,62,9]. Turtuk is now considered as a top tourist destination in India for its natural beauty and picturesque journey through the Nubra Valley. The village was not open to outsiders or tourists until 2009. Before that, the villagers' primary occupation was agriculture and some local men used to work for the Army to carry goods using donkeys. Now, apart from agriculture, they are focused on providing tourism services. Local men mostly drive sport-utility vehicles (SUVs) from Leh to Turtuk to transport the tourists and work as tourist guides. Women are involved in maintaining the guest houses. People are converting their homesteads to guesthouses, and some are currently under construction in hazardprone areas. The expansion of tourism could be considered as a new threat to the villagers, but in this case, activities related to livelihood and economic opportunities are getting more importance. One reason is that during winter they heavily rely on food and savings from summer.
There is also growing threats of damage to crops (Table A4) due to climate change and insect attacks (Table A3) and shift in agricultural practices (Table A1). Consequently, dependence on summer tourism activities is becoming an emerging trend in Turtuk. It supplements the findings of Wachinger and Renn [67] who argued that disaster mitigation measures are not always associated with local people's high perception of risk. Similar findings were verified by Le Masson [38] in the touristic Leh city where people started rebuilding houses in the same disaster-hit areas after the 2010 Ladakh Floods. In La Paz, Bolivia [14] and in Chittagong, Bangladesh [2] hundreds of thousand low-income people live on dangerous hill-slopes that are at risk of landslides and mudslides. In these contexts, people believe that moving to a safer place is less beneficial than the loss of their livelihoods over a longer term [14].
Findings from group presentations suggest that training in the construction of disaster-resilient buildings and safe site selection is needed for local engineers and contractors. Where possible construction should be undertaken with advice from geologists and engineers. Use of cement or iron sheet roofing would protect from heavy rainfall or snow and structures with appropriate reinforced concrete tie beams and pillars would be less likely to fail during floods, landslides and earthquakes. Retrofitting of existing building structures for disaster resilience needs to be initiated. Disaster resilient public buildings need to be built in safe areas to keep essential tools and rations for emergencies and provide shelter for vulnerable members of society like the elderly population, pregnant, and lactating women and children. A road from Turtuk to Hanu is needed to help connect Turtuk and provide support during emergencies. There is a need for emergency supplies of medicines and food.
In a conflict zone such as Turtuk, the interdependence of the Indian Army and the local population is deeply embedded for survival. Most, small towns and villages are cut off from the main supply routes for up to six months in a year. Even during the regular winter season, the Army provides all emergency care to the local population, and it is the same Army which shows up first in any disaster incident. The Army is well equipped and prepared for handling emergencies that can never be matched by the civilian authorities both financially or logistically. Whether it is jobs, medical, food supplies, school education for young children or medical emergencies, it is the Indian Army which remains at the forefront.
The Indian Government's policy on compensation for disasters needs reviewing, as people are dissatisfied with current levels. Local people must be trained in rescue and relief work. The religious leaders can also be instrumental because most of the people in Turtuk are religious. They perform many rituals (Muslims and Buddhist) for protection from environmental disasters, and people have faith in them. People can also use their indigenous knowledge and traditional methods along with modern equipment to enhance DRR (one example is shown in Fig. A3). The need of the hour is to prepare a risk-sensitive land use plan for the Turtuk village to control the spontaneous growth and tackle the impacts of tourism and global climate change.
Future research should cover other mountain communities on the other side of the LOC, building on the work of Azhar-Hewitt and Hewitt [10]. Findings from such research would be useful for the development of an integrated disaster and conflict resilient master plan for the HKH region and could contribute to achieving the UN Sustainable Development Goals and the Sendai Framework for Disaster Risk Reduction.

Conclusion
Indigenous mountain communities in the Hindu-Kush Himalaya (HKH) region suffer frequent disasters and economic/political marginalisation, especially in border conflict zones. This paper aimed to understand community perception of risk and vulnerability to environmental hazards in a remote border conflict zone in Ladakh.
Turtuk lies beside the Shyok River and near the LOC between India and Pakistan. The case study area is frequently affected by flooding, rock falls, landslides, border conflict and has high earthquake risk. India took control of Turtuk from Pakistan in 1971. Restricted movement across the LOC further isolates the community. This work is entirely relying on primary data. Using participatory rural appraisal (PRA) tools, community stakeholder groups of local men, women and girls, administrative officials and political/religious leaders drew maps. Hazard maps depicted the location of settlements, fields/orchards, amenities (schools, health facilities), rivers/streams and mountains and indicated flood, rock-fall and landslide hazard areas. Dream maps depicted groups' aspirations to decrease vulnerability to hazards and improve their lives. Meanwhile, specialists prepared a geological hazard map by integrating field-based investigations with remote sensing data.
Specialist-and community-produced hazard maps matched in the location of high-risk areas for flood, landslide and rock-fall and potential for damage to settlements, infrastructure (schools, health facility, government offices, roads and bridges) and agriculture, though community awareness of risk from earthquakes and poor-quality construction was low. Community members were aware that the positioning of essential government facilities on the flood-prone river plain, poorly constructed bridges close to the water, and climate change (especially increased rainfall and pests/diseases) increased their vulnerability. However, belief in the effectiveness of proposed mitigation strategies such as concrete levees for flood control and fences for prevention of rock-fall damage probably surpassed their capacity to prevent damage. Lack of availability of flat, lower-risk land for building, resource limitation and lack of awareness of locally appropriate lowcost options prevent the implementation of disaster risk mitigation.
A disaster preparedness plan is needed which should cover: monitoring hazards and climate change, shifting emergency infrastructure (hospital, school, police station) to lower risk areas, and training engineers and masons in-and implementation of -appropriate regulations on disaster resilient building. It should also include awareness-raising on appropriate low-cost community DRR strategies, sustainable tourism development, search and rescue training, road building and improved phone/internet connections, initiate peace talks to resolve border conflicts, and building of levees, slope stabilisation and protective barriers to rock-fall where appropriate and feasible.
At the local scale, this research demonstrates how a particular community deals with extreme hazards and conflicts in a mountainous environment. At the national scale, it promotes awareness of the value of risk perception studies by incorporating participatory maps into the gazetted land-use master plans, and traditional cultural knowledge in DRR initiatives. At the regional and global scales, this work provides an understanding of the root causes of disaster vulnerability and the characteristics required by a community to tackle them. Scrutinising various components of environmental disasters applying the proposed method represents an advancement and original contribution to the existing body of knowledge in DRR field. Explicitly, this paper fills gaps linked to risk communication solutions between the indigenous mountain people and decision makers, cultures and disasters, and tackling catastrophe in fragile and conflict-affected contexts.

Role of the funding source
The Natural Environment Research Council (NERC), the Arts & Humanities Research Council (AHRC), and the Economic & Social Research Council (ESRC) funded this work (NERC Reference: NE/ P016138/1). The UK based funding bodies had no influence or involvement in study design, in collection, analysis and interpretation of data, in writing the paper, or in the decision to submit the article for publication.

Declaration of interest
The authors declare no conflict of interest.    Other changes requested (not necessarily related to particular hazards)

Cloudburst/ Flash floods
First thing respondents drew on the dream map was a concrete levee/embankment on Shyok River and on Turtuk stream, which shows its importance.
Construct an Industrial area at Chu-thang to use available resources and create employment opportunities for local people.

Shyok river flooding
Whole Chu-thang area is vulnerable to the flood from the Shyok River. Earlier there was no settlement at Chu-thang.
Built an emergency shelter above the police station and near the apricot plantation.
As construction of an industrial area and new buildings will decrease the area under agricultural production, intensify/increase agricultural productivity in upper areas.
Build an alternative road from Turtuk to Hanu as a national highway (90 km only) for use during emergencies.
Chu-thang itself means water area. 'Chu' means water and 'thang' is land.  The river that separates the two villages used to be very narrow but the riverbed has widened after the fields along the river have been washed away.
Despite this debate the participants proposed to build embankments along all rivers on their dream mapping The new road to the highway could connect them with Leh in a few hours. Building bridges on different streams and upgrading the existing ones with cement and metal materials, for cars to use the bridges.
In 2010, "we survived, only thanks to the Army".
The Shyok river flooding In 2015, after heavy rains, the river diverted from its usual course, damaged the middle school and inundated the police camp.
Not enough land available (land dispute with Bogdan) If they win the lawsuit and obtain land rights they could move critical infrastructure to safer ground.
Build a concrete wall along the Shyok river

Table A3
Results from the local women group's participatory mapping exercises. Source: Community people -Turtuk, fieldwork, July 2017.

Hazard outlined Explanation (from timeline, and hazard mapping)
Vulnerabilities that community was aware about Risk mitigation suggested (Dream mapping) Other changes requested (not necessarily related to particular hazards)

Cloudburst/Flash floods
Build embankment but keep the course of the river as it is otherwise "it is dangerous".
Improve road connectivity (to the hospital), to the upper part of the villages and to Hanu to encourage tourists to go all the way. Create a bus stand.

The Shyok river flooding
Discussed where to locate the primary health centre. Decided to place it in the Chu-thang area (near the Shyok River) because it is closer to the main road and so it will become more accessible to people from other villages. But community members discussed the issue of the location in a vulnerable area, and the relative merits of move the buildings or to building an embankment. They finally decided to build the building 100m away from the riverbed and to build a wall along the river.
> "Because it is a sign of development" Upgrading bridges (higher above the stream) Upgrade the school system > For students to stay in Turtuk and because the population is increasing Increase the number of health centres on both sides of villages "It must be big" Create picnic areas and playgrounds, to increase "green areas"

Rock-falls
Building a concrete wall to prevent the boulders from falling Planting trees

Landslides
Planting more trees to protect to prevent the soil to erode and to protect areas from falling rocks > To accommodate the increase of tourists Pests destroying agricultural production (apricots) Insects destroyed apricot trees in 2016.
They used insecticide to deal with the risk this year but now the apricots are very small Climate change that increases heavy rainfall and snowfall that damage tree and drive insect infestation Develop houses (multi-storey, in cement) with numerous flats to accommodate the increase of population but to also to optimise space and avoid using too much agriculture land.

Earthquake
This did get discussed as they had not observed a major earthquake event yet Also, build more houses along one of the rivers in the neighbouring valley because there will be new roads in a couple of decades. Build more shops -that can be located right above or right below the main cliff above the Chu-thang area. Build mobile towers to create 4G coverage and have more electricity and Internet access so need to build hydroelectric power.

Table A4
Results from the local men group's participatory mapping exercises. Source: Community people -Turtuk, fieldwork, July 2017.

Hazard outlined Explanation (from timeline, and hazard mapping) Vulnerabilities that community was aware about Risk mitigation suggested (Dream mapping)
Other changes requested (not necessarily related to particular hazards) Cloudburst/Flash floods The Army camp was washed away in 2015 flooding and 22 army men died. People are afraid of flooding of the irrigation canal (Nala). Agricultural fields were destroyed.
Army hospital and trees damaged. Rainwater caused flooding and the roads were blocked.
Last 5-10 years, there has been an increase in rainfall, both volume and frequency. People are afraid of flooding of the irrigation canal (Nala).
Embankments of wall on the both sides and deepening of the canal (Nala). Cement canal to prevent water seepage.
Tiled footpaths for ease of walking and beautification Upgrade and increase the number of schools and a full-fledged hospital.

Instalment of new factories and shops
The Shyok river flooding The Shyok River has been flooded in past It is impossible to change the course of the river so the risk remains constant. However, the construction of an embankment along the river could reduce the risk to some extent. Landslides associated with the road The road has been constructed by cutting the slope but the vertical slope created by that has been left untreated. This has led to many slope failures in recent years.

Treatment of vertical slope by cement grouting
Playground and ice-hockey ground for recreational activities for children

Windstorms Destruction of crops
War This village has witnessed two major wars in 1971 and 1999 between India and Pakistan. People suffered a lot, and even they were forced to work as porters by the Indian Army during the war. Single casualty of shepherd during the 1999 war.
Fully aware that they live in a frontier zone between two countries. This problem is never-ending. Only government policy can stop the war in the future Earthquake Locals had witnessed minor earthquakes.
Not too much aware as no major earthquake is observed in recent times.