Agro-Pastoralists’ Perception of Climate Change and Adaptation in the Qilian Mountains, China

Global climate change affects all aspects of human society, especially agricultural and animal husbandry production. Northwest China has detrimentally affected by the climatic variations due to its high exposure to extreme climatic events. A number of studies have reported on agro-pastoralists' perceptions, adaptation and mitigation of climate change, but the current knowledge of agro-pastoralists' perceptions of climate change in China is insucient. To ll this research gap, this study aims to investigate the cognition level of agro-pastoralists in Northwest China on climate change and related factors. Data were collected using a structured questionnaire based on household surveys of 554 study participants in four counties in Gansu Province, China. Raw data were collected using stratied random sampling. A probit model was used to analyze the respondents' understanding of climate change and its related socioeconomic and demographic variables. Our results show that more than 70% of agro-pastoralists are perception of the changes of temperature and precipitation. Socioeconomic and demographic variables such as gender, farming experience, education level, cultivated land size, agricultural income, livestock, village cadre experience, access to weather information of agro-pastoralists are pertinently related to agro-pastoralists' awareness of climate change. Farming experience, education level, household size, grassland size, agricultural income, association membership, village cadre experience has a high impact on agro-pastoralists' adaptation to climate change. The results of this study will help guide government agencies and decision makers, and help arid and semi-arid areas to build sustainable adaptation measures under the framework of climate change.


Introduction
In recent decades, with more and more scholars studying, global warming seems to be an indisputable The study area is located in the Qilian Mountains in northwest China, which is a typical area of fragile ecological environment, including the Qilian Mountain National Nature Reserve, also known as the Qilian Mountain National Park. After the establishment of the reserve, the government strictly controlled the number of livestock and strengthened the balance management of grass and livestock, which forced agro-pastoralists to make adaptive changes and reduce the impact of this change on themselves. Based on the data of Participatory Rural Appraisal (PRA), this paper takes four counties involved in the protected area as the research area, and uses qualitative and quantitative methods. The main research contents are as follows :(1) assess the perception of climate change by agro-pastoralists in the research area, and whether the perception of climate change is accurate? (2) Analyze the factors affecting agro-pastoralists' perception of climate change; (3) understand the perception of the impact of climate change on agropastoralists' livelihood; (4) explore the impact of agro-pastoralists' characteristics on adaptation to climate change.
sampling strategy of the Household Income and Expenditure Survey (HIES) and interviewed fteen to twenty farm households from each village as a sample (RSID, 2008).
The formal investigation took the form of household face-to-face interviews, focusing on interviews with principals and village cadres of specialized agricultural and animal husbandry cooperatives. A total of 34 village-level questionnaires were completed, including 565 sample households, of which 554 households had effective questionnaires, with an effective rate of 98.05%.
The initial research permit was approved by the head of research, including the school, the head of the county government o ce, the head of the agriculture and Rural Affairs Bureau, the head of the village and the head of the reserve administration, paving the way to seek the consent of the respondents and participants. Seek the oral informed consent of agro-pastoralists and typical gures in the village. Voluntary participation was also insisted upon and respondents/participants received detailed information about the study objectives to avoid deception. In the case of interview, participants choose the place and way of interview, and try to avoid the presence of many people or persons in charge, so that the interviewees can show their true self and be willing to give truthful answers. Face-to-face interviews and telephone interviews were conducted according to the participants' own reasons.
All the agro-pastoralists conducted face-to-face interviews at their homes or designated places. In the case of village leaders, face-to-face and telephone interviews were used. Face-to-face interviews took place in homes, shops, workplaces and farms. Interviews lasted about 25 to 35 minutes and were recorded with the consent of a large group of participants. All the survey interviews agro-pastoralists were conducted in localdialects or minority languages, eventhough some o cers used both Mandarin and local dialects.
In addition, before the formal survey, we also selected Huangcheng Town of Sunan Yugur Autonomous County and Haxi Town of Tianzhu Tibetan Autonomous County as the survey areas for the preliminary survey (Table 1), so as to ensure the scienti c and practical design of the questionnaire. In addition, in the process of questionnaire survey, centralized inspection and review of each questionnaire should be carried out in time to ensure the completeness and accuracy of questionnaire data.
The structured questionnaire used to collect respondents' information mainly includes ve parts: the basic characteristics of agro-pastoralists, awareness of climate change, climate change and disaster effects, and cognition and adoption of coping behaviors of agro-pastoralists; village information structured questionnaire mainly includes three parts, including the basic characteristics of the village, climate change and adaptation measures to support and use situation.

Data Analysis
The coding and analysis of qualitative and quantitative data used Microsoft Excel and the Statistical Product and Service Solutions (SPSS version 24.0).The meteorological data include diary records to calculate the average climatic values of seasons, months and years. The normality and outlier test are analyzed by using line diagram, box diagram and scatter diagram. In addition, frequency and percentage were used for descriptive statistics, and demographic independent variables (such as age, farming experience, years of education) and dependent variables (perception) were used for reasoning statistics, such as independent sample t-test and one-way analysis of variance (ANOVA)(Asare-Nuamah and Botchway2019).Based on the hypothesis, agro-pastoralists' perception of climate change and its impacts, and their socio-economic characteristics are the main factors in uencing agro-pastoralists' adaptation choices. Probit model was used to evaluate the impact of selection variables on the respondents' perception and adaptation level. When the independent variables are composed of continuous variables and dummy variables, probit model is particularly suitable for the analysis of binary dependent variables. Assuming that it can predict the impact of independent variables on dependent variables, it is used to evaluate the impact of perception and socio-economic characteristics on family adaptation.

Ethics approval
The

Results And Discussion
3.1. Basic information of interviewees Table 2 shows the demographic characteristics of agro-pastoralists. More than half of the respondents were males (69%). The respondents were on average 41.3 years old while more than 32 years of farming experience. The study area consists of minorities nationality (Tibetan, Yugur, Mongolian, Hui, etc.) and Han. In most cases, the main livelihood activity of minorities nationality are livestock, while Han main livelihood activities are farming. The majority of respondents (64%) were minority nationality. The vast majority of the agro-pastoralists (86%) have a primary school education or above, even though only 1% of them have Undergraduate education or Above. The results also reveal that 92% of respondents have access to weather information. The averages of 10.23 Mu of cultivatedland and 156.21 Mu Grassland, respectively. The average per family income is RMB78000, and agricultural income s RMB52000.
Due to their long-term farming experience, the agro-pastoralists were expected to have a high-level of understanding of local climate knowledge. Also contributing to this could be the information they receive about climate change and for some, the associated training through agro-pastoralists' associations. Therefore, they also have a propensity to adapt to adverse conditions resulting from climate change impacts. In addition, the high-level of farming experience, the cultivated-land size, grassland size, Credit loan, Insurance, Village cadres all have a positive impact on the level of agro-pastoralists' adaptation to new climate scenarios.
However, the education level and cadres experience may be the major limiting factors for adopting speci c long-term adaptation strategies. Ethnicity and gender are also expected to be key factors in uencing awareness and adaptation to climate change. Minorities and Han may observe and understand the change in climate and related impacts differently because of their cultural ecology. (the main livelihood activity of minorities nationality are livestock, while Han main livelihood activities are farming.) In terms of gender, women in rural areas are less mobile and have less access to information and rights. They are also heavily involved in domestic work. However, men may have easier access to information (socializing, going out to work, etc.) Therefore, male headed households are expected to be more likely to adapt to the impact of climate change.  Figure 2 shows the trend of annual precipitation, annual rainfall and annual snow at different meteorological stations in the study area. As shown in the gure 2, precipitation, rainfall and snow show an increasing trend, but the increase range of snow (0.0325-0.375/a) is signi cantly lower than that of precipitation(1.22-3.1/a) and rainfall (1.04-2.81/a). Similarly, through the inspection, it is found that the Multicollinearity among precipitation, rainfall and snow at each meteorological station is obvious (most R 2 > 0.5, and p < 0.01 or 0.05). The oscillation mode of rainfall shows that most of the highest rainfall over the past 32 years occurs in 2019, Indicating the wettest year , while 1991 was the driest year over the same period (see Figure 2). IPCC AR5 pointed out that the global climate system will continue to warm in the future, global precipitation will increase, and water cycles will accelerate, but there are signi cant regional differences ( 2018) pointed out that climate is the limiting factor affecting agricultural production, especially in inland areas. On the one hand, the temperature uctuates and rises continuously, the sowing date of crops is advanced, and the growth period of crops becomes longer; On the other hand, the annual uctuation rate of precipitation increases, and the uncertainty of "rainy year" and "dry year" increases the change of crop yield.
In Figure 3,

Respondents' perception of climate change
The climate change perception of agro-pastoralists are presented in Table 3. Among the 565 agropastoralists, 554 had heard about climate change and about 500 believed that climate is changing. Agropastoralists understood climate change by different indicators. In the case of temperature, a vast majority of the respondents (83.03%, summer /68.95%, winter) perceived that there have been changes in temperature in the district. Similarly, there have been changes in rainfall in the district as reported by 77.08%(rainfall)/72.74%(snow) of the respondents. In terms of quantity, the majority of the respondents(39.53%) perceive a decrease in rainfall, the next majority of respondents (37.55%) feel that rainfall is increasing . and about 4.69%/7.58% respondents feel that rainfall is unpredictable in terms of quantity (sometimes high,sometimes low).
Regarding the changes in temperature, the majority of respondents have noticed the rising summer temperature (71.3%), while 11.73% of the respondents perceive that the decreasing summer temperature. For the winter temperature, nearly 32.31% perceive that winter is becoming colder while nearly equal percentage of the respondents (36.64%) perceive that winter is getting warmer. In our study, there are 3.25% -7.58% of respondents who do not perceive any changes in temperature; yet this is nearly equal compared to those who did not perceive any changes in rainfall. Climate variables particularly rainfall and temperature have been extensively studied as they are perceived to be signi   Fig. 6. Figure 6 reveals that the respondents received information about climate change mainly from personal experience (73.37%), internet(53.25%), television(42.06%). Other sources such as radio(12.27%), family/friends (15.34%), association/extension (5.78%), newspaper/magazine(3.79%) and Other(1.44%) were far less important. Recently, communication devices (e.g., mobile phones, computers) that provide access to the internet play an important role for socioeconomic development in agro-pastoralist areas and may be relevant to spread information about climate change among agro-pastoralists.

Determinants of agro-pastoralists' perception
In order to further understand the relationship between Agro-pastoralist population characteristics and climate change perception, multiple regression (probit) analysis was used (table 4).With the increase of experience, education, cultivatedland size, agricultural income, livestock, village cadres, access to weather information, agro-pastoralists' awareness of climate change has increased signi cantly. The results also show that female are more sensitive to climate change than male.
The gender of respondents is a signi cant explanatory determinant(p-value is 0.02, coe cient is 0.325).
In other words, women have more information and experience than men. This may be because in agricultural production, particularly in the planting and livestock sectors, women work signi cantly longer hours than men and are more sensitive to climate change (Ding et Table 4(where 5 =High, 4=Medium, 3= low,2=No and 1=Don't know).
Respondents have different views on the impact of climate change (table 5). More than 50% of respondents said that climate change had a medium or above impact on pasture quality housing security livestock loss crop / livestock diseases agricultural income, and these variables were greatly affected by climate change. As for the planting area, the agro-pastoralists who are mainly planting are the most affected. Among the respondents, nearly 15% disclosed that they have abandoned part of the land or are ready to abandon it; With regard to harvest time and harvesting / calving time, more than 70% said that the impact was small or not, but the survey showed that nearly 10 people said that the seeding time was ahead of schedule, and the calving time also changed, which may be related to the introduction of new varieties. Education is considered an important factor in adapting to climate change. Table 6 shows that education has a signi cant positive impact on respondents' adaptation to climate change impacts. An increase in formal education by one level was associated with a 8.1% increase in the likelihood of family adaptation.
Our  In order to reduce the negative impacts of climate change and maintain the livelihoods of rural households, it is necessary to improve the education level of agro-pastoralists, strengthen the capacity of community organization building and information dissemination services, and help agro-pastoralists better implement appropriate climate change management strategies. agricultural income, association membership, Village cadre experience is closely related to the adaptability of agricultural and pastoral personnel. In addition, our ndings indicate that agro-pastoralists surveyed indicated that they had adopted aggressive adaptation strategies. In addition, our analysis of respondents' adaptive and perceived behavior provides a comprehensive understanding of the importance of climate in determining the effectiveness of agricultural programs. Moreover, our ndings could help government o cials and agricultural extension workers expand their services and raise awareness of climate change in areas most vulnerable to climate hazards.
While the study was limited to Gansu province in northwest China, the results of the study could be applied to other regions where adaptation to climate change remains ineffective. The study's recommendations help policy makers and researchers develop the necessary strategies and provide respondents with facilities to deal with climate risks. However, necessary action is needed to raise awareness of climate change. These efforts could focus on deep-seated adaptation measures, such as the use of high-yielding varieties and diversi cation of agro-pastoral operations. To strengthen such actions, policies and adequately supported approaches are needed to strengthen the capacity of agropastoralists to invest in adaptation strategies. It is suggested that agro-pastoralists' understanding of climate change and adaptation measures should be supported by mass media and information technology through agricultural and animal husbandry associations.

Declarations Declaration of competing interest
The authors declare no con ict of interest.  Trends of summer mean temperature Figure 5 Trends of winter mean temperature Agro-pastoralists' sources of information, about climate change