This study is the first in Tibet to investigate public perceptions, experience, behaviour and coping resources regarding high temperatures and future climate change. The findings will inform subsequent quantitative studies including a map-based human health risk assessment and management plan. Overall, we found widespread awareness in the local population of rising temperatures and their effects on health. Some segments of the population are likely to be at greater risk due to environmental differences, demographic factors and limited coping resources.
Previous studies of this kind have been undertaken mainly in developed countries including the USA [11, 12, 15, 17, 18], Australia [19, 20], and Canada [12, 13]. Such studies can provide important information to policy makers and assist in developing and promoting practical adaptation measures. Our study has added a valuable dimension to the existing literature as little has been written previously about perceptions in relation to climate change in developing countries [14, 21]. In many developed country studies there are high levels of knowledge and awareness about climate change  and heat warnings [11, 15]. In Tibet, despite the lack of a heat plan or any educational program, we found that survey respondents were well aware of the rise in temperatures that has occurred and the effects this might have on health. This is consistent with another study conducted in Bangladesh, a developing country which has been identified as one of the most vulnerable to climate change . In Bangladesh there was widespread awareness of global warming among respondents who were mostly rural people with little formal education. We surmise that people in low-income settings may be more likely to regard themself as vulnerable to heat than those in developed countries who live in more comfortable circumstances, are protected by advanced public health infrastructure and are informed via pervasive media coverage. If this is indeed the case, then there is both a need and an opportunity to introduce community-based heat adaption program in regions such as Tibet.
In our study, age and health status were found to be good indicators of perceived risk to heat. Older respondents, along with people with pre-existing chronic diseases saw themselves at greater risk from heat. Those with high incomes were also more likely to consider themselves to be sensitive to the heat, although it is low income groups who in fact are likely to be more vulnerable because of limited coping resources and access to information . However, perception of risk does not mean that rising temperatures are the issue that most concerns people. Earthquakes were named as the biggest threat by most respondents which is understandable given local circumstances. In China’s recent history earthquakes have killed hundreds of thousands of people, and these tragedies have made a big impact on the Chinese population. On the other hand, in Phoenix, already one of the hottest cities in the US, over half of respondents treated increasing heat as the biggest threat, and fewer than 10% of them named earthquakes .
Previous papers addressing the effects of heat on health have concentrated on persons aged 65 years and over [22, 23], and other age-groups have received much less attention. Our findings suggest that middle aged persons in Lhasa may be significantly affected the high temperatures: self-reported heat-related illness was most common in persons aged 42–65. There may be two reasons: this is an age group commonly affected by chronic medical conditions particularly cardiovascular diseases; and, unlike the elderly, most middle age people are still at work and thus may be exposed to more outdoor heat. Our finding is consistent with a model-based quantitative study which reported a marked effect of heat on mortality among person aged 45–64 . The finding that people with chronic diseases reported a higher incidence of heat-related illness than those who do not have chronic medical conditions is in agreement with results from a recent Canadian survey . Over 60% of respondents in our survey with chronic conditions felt their symptoms had been worsened because of the heat. In Tibet, there are unique environmental factors such as extreme altitude and low oxygen pressure that create special challenges to the health of local people. Chronic mountain sickness (CMS) is common, and is marked by hypoxia, high red cell count, and associated heart diseases [25–27], all of which reduce heat tolerance and make Tibetan people even more vulnerable to the rising temperatures forecast under climate change.
A most surprising result in this study is that the vast majority of respondents reported modifying their behaviour because of high summer temperatures. This is totally contrary to a previous study in North America suggesting that few people altered their behaviour during heat events, even if a heat warning or a heat advisory was issued . We further observed that the widespread behaviour changes may exist on a common sense basis since measures most cited by respondents to deal with the heat are drinking more fluids, along with wearing light clothing and staying indoors. These measures are more convenient, easier and somewhat cheaper than moving out of town to a cooler place, going swimming, using a home air-conditioner or seeking an air-conditioned location. It is important in developing regions such as Tibet to take into account cost, convenience, accessibility to adaptive facilities, and the common behaviours of local people when developing and implementing education programs. Finding out about these enabling factors may require in-depth field studies. Moreover, given the demographic differences in behaviour that showed up in our survey, it is equally important to develop specific coping instructions in relation to those at most risks to heat including babies, the elderly, outdoor workers, and people with medical conditions. As expected, a correlation between behaviour changes and perceived risk is observed in our study, and this has been identified by previous research as well . Those who reported not modifying their behaviour on hot days were prone to perceive themselves as immune to the heat and have low or no sense of worry about rising temperatures in Lhasa.
Populations living in urban areas where temperatures tend to be higher than in suburban and rural locations are at greater risk of heat stress. Lack of adequate social and material resources further adds to their vulnerability. Harlan et al.  evaluated residents’ vulnerability to thermal environments of urban neighborhoods by analyzing the relationships between outdoor human thermal comfort index (HTCI), population characteristics and coping resources. They reported that neighborhoods with high population density, sparse vegetation and no open space were exposed to more extreme temperatures. Moreover, local residents, particularly low socio-economic groups, in those warmer neighborhoods, are more vulnerable due to limited resources. In order to understand whether or not people living in urban sub-districts in Lhasa face the same issues, we looked for inequalities in resources across our four study locations. Bakuo, located at the urban core, which is the most densely populated, crowded and noisy sub-district with few green and open spaces is expected to be the hottest area in Lhasa. Bakuo may also be most vulnerable because of the higher proportions of poor people and old dwellings with poor ventilation, small living spaces and low comfort level. The prevalence of home air conditioning is low in urban Lhasa, and lowest of all in Bakuo. We note that air conditioning at home was not found to be a strong protector against heat-related illness in our survey, and this is consistent with at least one previous study .
However, the strongest neighborhood social ties were seen in Bakuo sub-district. The great majority of respondents reported having a close relationship with their neighbors. Other studies in developed countries found that people on a low income suffer from poorer health than richer groups in a society because of either material poverty or a lack of social relationships and support [28–30], while our results indicated the neighbour bonding was weakest in those upper-income people. A possible explanation is that relatively well-off people are prone to isolate themselves from others in their neighborhood and make their social connections with people of similar socio-economic status in other parts of the city, or further afield. Importantly, Perry and Lindell  note that opinions based on personal experience of natural disaster tend to be enhanced through neighbors, friends and strong social networks. The strong neighbour ties that we observed in all sub-districts in Lhasa probably serve to promote information exchange and may be one explanation for the widespread perception of risk due to heat and the high level of reported behaviour changes. Again, for developing regions including Tibet, this indicates a heat adaptation program may be most effective in communities where people have a sense of competence and good neighbor bonding.
Apart from close relationships among neighbors, strong family connections were also observed in Lhasa. Social isolation or living alone, which is identified as a high risk factor in heat events , is not common in Tibet. Most of the households in this study included two or three generations, with an average size of four persons. Unlike other Chinese inland cities with a rapid increasing number of elderly persons living alone, old people in Lhasa normally live with their children and help to look after grandchildren. However, the small proportion (9.4%) of single person households in this study may be an under-estimate, since older people with chronic disease or other conditions that restrict their mobility are less likely to come out and participate in surveys. A more in-depth understanding of local circumstances and vulnerable populations is most important for developing practical adaptive measures and prioritizing actions.
Several limitations of the study should be noted. The study was conducted on a convenience basis due to local factors that prevented strictly random sampling. Older individuals and other subjects at high risk may be underrepresented in our sample. The study was only conducted within four urban sub-districts in Lhasa, so caution should be applied in generalizing the results to other parts of Tibet, particularly rural and remote areas. Additionally, as with all studies of this kind, the answers may be affected by a social desirability bias, meaning that participants are prone to report what they think to be the socially accepted answers, not necessarily their real perceptions.