**1. Introduction**

The sixth assessment report (AR6) released by the Intergovernmental Panel on Climate Change (IPCC) pointed out that extreme climate events have occurred more fre-

**Citation:** Li, J.; Zou, Y.; Zhang, Y.; Sun, S.; Dong, X. Risk Assessment of Snow Disasters for Animal Husbandry on the Qinghai–Tibetan Plateau and Influences of Snow Disasters on the Well-Being of Farmers and Pastoralists. *Remote Sens.* **2022**, *14*, 3358. https://doi.org/ 10.3390/rs14143358

Academic Editors: Massimo Menenti, Yaoming Ma, Li Jia and Lei Zhong

Received: 26 May 2022 Accepted: 8 July 2022 Published: 12 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

quently in the context of global warming compared with pre-industrial-revolution levels [1]. The meteorological disasters and derived disasters triggered by extreme climate events have caused increasingly large losses and higher disaster risks which seriously affect society, the economy, and human life, so the topic has attracted wide attention [2]. The Qinghai–Tibetan Plateau (QTP), located in a mid-to-low-latitude region in the northern hemisphere, is the highest plateau in the world. At an average altitude of above 4000 m, animal husbandry is one of the economic pillars for residents on the QTP [3]. However, husbandry on the QTP is highly susceptible to natural disasters due to the special weather conditions and types of vegetation [4]. Among the natural disasters, snow disasters have become one of the leading meteorological disasters in winter and spring in alpine pastoral regions due to their long duration and wide range of influence. Continuous snowfall is frequent in alpine pastoral regions in winter and spring, and, at the same time, the snow persists for a long time due to the low temperature and readily covers short forage grass [4]. As a result, livestock feeding on forage grass may die of frost and starvation, which greatly threatens the livelihoods and property of local farmers and pastoralists and influences the productivity of husbandry. Meanwhile, previous research showed that the frequency and hazard of snow disasters on the QTP have also risen in the context of global climate change [5,6]. Therefore, assessing the risk of snow disasters based on determining the spatio-temporal distribution characteristics of snow disasters on the QTP is of significance for disaster prevention and protection of the husbandry on the plateau.

In recent years, much research into the spatio-temporal variation characteristics of snow disasters on the QTP has been performed using various data and technological means. Based on occurrence records and observations at meteorological stations, previous studies found that snow disasters on the QTP during winter and spring are mainly caused by abnormal snow accumulation from November to the following March [7]. In addition, snow disasters showed obvious interdecadal variations and a significant variation in the early 1990s. The frequency of snow disasters has shown an increasing trend since the 1990s [8,9], and the Lhoka City in the Tibetan Autonomous Region (Lhoka) in the south-west and the border between southern Qinghai Province and Sichuan Province are two centers with high frequencies of snow disasters [10]. With the climate warming over the plateau, the snow depth and the number of snow cover days for the majority of the QTP show a decreasing trend [11], and the decrease in the snow depth is more significant in the high-altitude areas [12]. Although the above studies achieved certain goals, the research conclusions were quite different due to the sparsely distributed nature of the meteorological stations on the QTP and differences in the selected meteorological stations and research areas [11,13–15]; because remote sensing data can provide snow information with high spatio-temporal resolution, they are widely used for the inversion and monitoring of snow [16–18], assessment of snow disasters [19,20], and early warning of snow disasters [21] in areas with sparse meteorological stations. These works greatly improved popular perception of the variations in, and possible drivers of, snowfall, and some scholars also used remote sensing data to explore the occurrence of snow disasters on the QTP. For example, Yin et al. [4] used AVHRR archival reflectance products to find that the grade of snow disaster on the QTP reduced from 1982 to 2012. No matter which data were used, most studies focused on the spatio-temporal variation characteristics of snow disasters. However, snow disasters, as one of the natural disasters that most greatly affects animal husbandry on the QTP, exert remarkable influences on all aspects of society, the economy, and people's livelihood. Therefore, snow disasters need to be comprehensively studied from the perspective of risk assessment, in addition to the existing studies which discussed the influences of snow disasters on livestock in typical regions of the QTP [22,23]. Meanwhile, researchers have used human well-being to characterize the material and spiritual satisfaction of residents in recent years. Because of the close relations between the people's livelihood and governmental decisions, human well-being has recently been paid much heed by many researchers [24,25]. Numerous studies were

conducted on human well-being from multiple perspectives, including studies of variation characteristics and influence factors [26–30]; however, most studies focused on social and ecological topics. Natural disasters may affect material supply, living environment, and even life and property security of residents, so their influences on residents' physical and mental health cannot be underestimated. However, there are few studies on the influences of natural disasters on human well-being.

Therefore, this research mainly aims to determine regional differences in the comprehensive risk degree of snow disasters for husbandry on the QTP by comprehensively considering the risk of hazard factors, sensitivity of hazard-inducing environments, vulnerability of hazard-affected bodies, and disaster prevention and mitigation capacity from the perspective of the risk assessment of snow disasters. This is based on analysis of spatiotemporal variation characteristics of snow disasters on the QTP. Then, the farmer and pastoralist well-being (FPWB) index is constructed to evaluate possible influences of snow disasters on FPWB on the QTP since the beginning of the 21st century. The research results provide a theoretical basis for making policies to prevent snow disasters and selecting policies for FPWB on the QTP.

### **2. Materials and Methods**

### *2.1. Definition of Snow Disasters*

Snow disasters on the QTP mainly occur from October to the following May, so this time period was selected for calculating snow disasters. For time recording, the period from October 1979 to May 1980 was used as a statistical time period, recorded as of the year of 1980, which was divided into last winter (from October 1979 to February 1980) and this spring (from March to May 1980). Other years were recorded in the same way, thus, obtaining snow disasters over 40 years from 1980 to 2019. According to previous research and relevant meteorological standards [4,31,32], the snow disasters of last winter and this spring on the QTP were graded following criteria in Tables 1 and 2. Based on the criteria, the grade, duration, and average snow depth of snow disasters were summarized. Therein, the highest grade of snow disaster was taken as the annual grade of snow disaster. For example, if three snow disasters occurred in a year, including a slight, a moderate, and an extremely heavy event, then the year was recorded as having had an extremely heavy snow disaster. The sum of durations of several snow disasters in a year was recorded as the duration of snow disasters. The average snow depth was the average value during the snow disasters.


**Table 1.** Division criteria for snow disasters of last winter.


**Table 2.** Division criteria for snow disasters of this spring.

### *2.2. Risk Assessment Method of Snow Disasters*

Snow disasters are a type of natural disaster. In risk assessment, the comprehensive risk of snow disasters is reflected by the risk of hazard factors, sensitivity of hazardinducing environments, vulnerability of hazard-affected bodies, and disaster prevention and mitigation capacity according to the risk-forming theory of relevant natural disasters. The disaster risk is expressed as follows:

$$D = f(H, S, V, R) \tag{1}$$

where *D*, *H*, *S*, *V*, and *R* separately represent the disaster risk, risk of hazard factors, sensitivity of hazard-inducing environments, vulnerability of hazard-affected bodies, and disaster prevention and mitigation capacity; *f* is the function relationship.

When assessing the risk of snow disasters on the QTP, the following equation was used:

$$FDVI = \left(E^{\rm WE}\right) V^{\rm WV} \left(S^{\rm WS}\right) (10-R)^{\rm WR} \tag{2}$$

where *FDVI* represents the comprehensive risk index of snow disasters, and its value can be used to characterize the risk degree of snow disasters for husbandry on the QTP; the larger its value, the higher the risk of snow disasters. *E*, *V*, *S*, and *R* separately denote indices of various assessment factors, including the hazard factor, hazard-inducing environment, hazard-affected body, and disaster prevention and mitigation capacity; *WE*, *WV*, *WS*, and *WR* represent weights of various assessment factors, which are determined using the analytic hierarchy process (AHP). Weights of various factors are listed in Table 3.

In the calculation, various factors contain several different indexes, each of which has a different dimension and order of magnitude. Therefore, Equation (3) is used to normalize the various indices to ensure the comparability of various indices; thereafter, the indices lie within the range 0.5–1.

$$A\_{ij} = 0.5 + 0.5 \times \frac{a\_{ij} - min\_i}{max\_i - min\_i} \tag{3}$$

where *Aij* denotes the normalized value of the *i*th index at the *j*th station (or grid); *aij* is the value of the *i*th index at the *j*th station (or grid); *maxi* and *mini* separately represent the maximum and minimum of the *i*th index.

Finally, the natural breaks method was adopted to grade the comprehensive indices of snow disasters for husbandry as high-risk, sub-high-risk, medium-risk, low-risk, and sub-low-risk zones.


**Table 3.** Risk assessment indices for snow disasters on the QTP and their weights.

### *2.3. Establishment of the FPWB Index*

Human well-being is used to characterize the living conditions of people, involving health, happiness, and affluence of materials. Early research on human well-being was mainly dedicated to economics and sociology. In recent years, research on human wellbeing has been gradually heeded by scholars in ecology and geology with the promotion of the idea of sustainable development. Meanwhile, characterization of human wellbeing has also gradually expanded from a single economic index to the ecological system. According to differences in research foci, human well-being is also divided into objective and subjective dimensions. This research focused on well-being of farmers and pastoralists (shorted as FPWB) according to sources of income, living styles, and the factors influencing the economy of residents on the QTP. To characterize FPWB, the FPWB index on the QTP was established by combining the conceptual framework of objective well-being and the concept of livelihood capital.

The FPWB index is composed of various factors. This research selected key factors that are closely related to the life of farmers and pastoralists from the agricultural part in provincial statistical yearbooks. These factors can be grouped into the following four aspects: natural resources, human resources, material resources, and social and financial resources, and indices contained in each level are listed in Table 4. The indices are quantified using the weighted comprehensive evaluation method, and their weights are determined by the AHP. In this way, the FPWB index can be expressed by Equation (4):

$$FP\% \text{V} = V\_1 \text{V}\_1 + V\_2 \text{V}\_2 + V\_3 \text{V}\_3 + V\_4 \text{V}\_4 \tag{4}$$

where *FPWB* represents the farmer and pastoralist well-being; *W*1, *W*2, *W*3, and *W*4 separately denote the four aspects that constitute the *FPWB* index, namely, natural resources, human resources, material resources, and social and financial resources; and *V*1, *V*2, *V*3, and *V*4 are weights of each level of assessment, which are determined using the AHP. The final weights are listed in Table 4. Likewise, each index is also normalized because each level of assessment involves different indices that are in different units and dimensions and must be normalized to reach the goal of eliminating differences and making the indices comparable.


**Table 4.** Components of the FPWB index on the QTP.

### *2.4. Data Sources*

Snow data: the snow depth long time-series data set in China (1979–2019) was provided by the National Tibetan Plateau Data Center (TPDC). The data set was obtained by inversion of SMMR (1979–1987), SSM/I (1987–2007), and SSMI/S (2008–2019) daily EASE-Grid brightness temperature data processed by the National Snow and Ice Data Center of the United States with a spatial resolution of 25 km. The data set has been widely proved to be reliable, and its development is described elsewhere [33–35].

Socio-economic data: socio-economic data, including the number of rural households, the number of employees in farming, forestry, animal husbandry, and fishery, GDP, and net income of rural residents, were extracted from statistical yearbooks of Qinghai Province, the Tibetan Autonomous Region, Sichuan Province, Gansu Province, and the Xinjiang Uygur Autonomous Region.
