Spatial distribution patterns of rural settlements in the multi-ethnic gathering areas, southwest China: ethnic inter-embeddedness perspective

ABSTRACT This study examines the distribution patterns of rural settlements from a new perspective of ethnic inter-embeddedness. The Shannon-Wiener Index (SWI) and Moran’s I methods are employed to identify the spatial distribution characteristics of rural settlements in a typical multi-ethnic gathering area. The study results are as follows: Firstly, the distribution patterns of rural settlements of different ethnic groups exhibit significant heterogeneity and remarkable differences in ethnicity and region. Each ethnic group has its core settlement areas in geography. Secondly, the results of Moran’s I represent clustering characteristics and gradient variations of spatial distribution. All ethnic groups exhibit significant aggregation characteristics except for Hui. The Global Moran’s I of Han, Yi, Tibetan, Mongolian, Qiang, and Hui ethnic groups are 0.771, 0.726, 0.646, 0.225, 0.123, and 0.037, respectively. Thirdly, the SWI results are classified into five categories: very high, high, medium, low, and very low, accounting for 23.53%, 23.53%, 19.60%, 15.69%, and 17.65%, respectively. Finally, the study find a negative correlation between the equilibrium of the minority population and the degree of inter-embeddedness of rural settlements. The study provides significant implications for promoting ethnic integration and connection, as well as optimizing the spatial patterns of rural settlements in multi-ethnic gathering areas.


Introduction
Previous studies on rural settlements primarily focus on their spatial distribution (Chen et al. 2022;Jia et al. 2020), evolution (Li and Song 2020;Lu, Song, and Lyu 2022;Qu et al. 2021), thermal environment (Wei et al. 2022b), landscape (Amani-Beni, Khalilnezhad, and Mahdizadeh 2022), and mechanisms (Ji et al. 2022;Li et al. 2022).However, most of these studies focus on Han settlements (Pan and Liu 2021), with few studies exploring multi-ethnic gathering areas (Luk 2009).China is a multi-ethnic country consisting mainly of Han and 55 minority groups (Maurer-Fazio and Hasmath 2015).According to the findings of the seventh national census, the population of ethnic minorities in China was 125.47 million, which accounts for 8.89% of the total population (www.stats.gov.cn).Remarkably, these individuals are widely distributed across 60% of the country's land area, indicating a significant dispersion in the spatial distribution of ethnic minorities.Due to historical and geographical factors, most ethnic minorities in China live in remote mountainous areas and have distinct settlements, languages, and cultures that differ from those of the Han (Ouyang and Pinstrup-Andersen 2012).The spatial inter-embedding pattern reflects exchange and intermingling of ethnic groups and has contributed to the stability and optimization of the spatial structure of rural settlements.Therefore, identifying the distribution characteristics of ethnic minorities is important for optimizing the spatial patterns of rural settlements.
Few studies have focused on the distribution patterns of rural settlements in multi-ethnic gathering areas (Jordan, Krivokapic-Skoko, and Collins 2009).We searched the relevant literature on Google Scholar, ScienceDirect, and Taylor & Francis Online but found only a limited number of relevant papers.The literature covers various topics such as spatial form (Jia, Meng, and Zhou 2021;Li et al. 2018;McGovern and Frazier 2015), distribution (Gao, Wang, and Zhu 2017;Xu et al. 2020), and sustainable development (Weitz 1967;Zhao et al. 2021).However, only a few studies have focused on the spatial inter-embedding pattern.Fortunately, many studies have examined diversity (Churchill 2020;Habyarimana et al. 2007;Koomson, Afoakwah, and Ampofo 2022;Li et al. 2021), segregation (Agyei-Mensah and Owusu 2012;Asibey, Poku-Boansi, and Adutwum 2021;Musterd 2005), and concentration (Agyei-Mensah and Owusu 2010) in multi-ethnic groups gathering areas.These studies provide an excellent theoretical reference for our research.
The paper makes the following contributions: Firstly, it focuses on the distribution patterns of rural settlements in multi-ethnic areas and measures their characteristics.Secondly, it attempts to reveal different patterns of spatial inter-embedding.Lastly, it analyzes from the ethnic inter-embeddedness perspective, thereby enhancing the study of rural settlements.The study aims to provide new insights into the optimization of spatial patterns of rural settlements in multiethnic areas.
Garze Tibetan Autonomous Prefecture, located in southwest China, is an exemplary multi-ethnic gathering area.We have chosen this region as our study area for the following reasons: Firstly, the region is situated in the transitional zone from the plains to the Tibetan plateau and is characterized by a diverse range of landforms, including mountain plains, grasslands, and canyons.Secondly, the area is culturally diverse, with various indigenous minority ethnic groups living together in this region.As a result, the social structure of rural settlements is more complex than in previous studies.Lastly, the spatial pattern of rural settlements is undergoing reconfiguration and transformation.The results suggest optimizing the distribution patterns of settlements in multi-ethnic gathering areas.
The remainder of this paper is organized as follows: Section 2 constructs the theoretical framework, providing a comprehensive analysis of the key concepts and relevant literature.Section 3 introduces the study area, methods, and data source, highlighting the research design and data collection procedures.Section 5 presents the findings of the study.Section 6 discusses the results in detail and provides recommendations for future research.Finally, Section 7 provides the conclusion drawn from the research.

Theoretical framework
The concept of "Ethnic Inter-embeddedness" reflected that two or more ethnic minorities lived together, forming a settlement (community) with equality and coexistence.Ethnic distribution significantly impacts the spatial patterns of rural settlements in multiethnic gathering areas (Raitz 1973).
The spatial inter-embeddedness of rural settlements is closely related to the spatial distribution of ethnic populations and ethnic diversity (Jordan, Krivokapic-Skoko, and Collins 2009).In general, ethnic groups that are highly spatially clustered are less likely to intermingle with other ethnic groups.In China, ethnic minority settlements are often located in remote, economically backward areas.Han rural settlements formed a spatial pattern of "scattered on the marco scale, clustered on the microscale" with other ethnic groups (Li et al. 2018).The distribution pattern is of great significance in promoting inter-ethnic communication, and it also supports the settlement's social structure to be stable and balanced.In 2014, Chinese President Xi Jinping proposed to encourage the establishment of a social network and community environment in which ethnic groups are interembedded with each other.As a result, the multiethnic gathering regions will face more significant ethnic integration exchanges.
This paper proposes a theoretical framework for the spatial interdependence of rural settlements in mixed minority areas.It aims to clarify the spatial patterns of rural settlements under the influence of different ethnic cultures and social structures and to propose a new perspective for studying rural settlements.In this section, we attempt to classify the spatial inter-embedding patterns of rural settlements into three types (Figure 1) based on the previous study (Brown and Chung 2006).The theoretical assumptions are based on the premise that: (1) Two or more ethnic minorities are interembedded and maintain a balanced state.(2) One dominant ethnic minority, or two dominant ethnic minorities, existed above.If the number of minorities is large, perhaps the number of dominant ones is even larger.(3) There is no dominant minority, and each minority remains in an equilibrium or random state.To our best knowledge, 1 and 2 above are more common.

Study area
The study area is located in the east of Garze Tibetan Autonomous Prefecture (GZTAP), China (101°07'-102°10' E, 28°19'-29°20' N) (Figure 2).It covers approximately 26,950 km 2 , consists of 3 counties (county-level cities), namely Kangding City (KD), Luding County (LD), and Jiulong County (JL).The study area is located on the southeastern edge of the Tibetan Plateau, with elevations of 956-7473 m; the average height is 3855 m.The topography can be divided into three major types: plateau, mountain plain, and alpine valley (https://www.sc.gov.cn/).The unique geographical and sociocultural environment has accumulated settlement forms with distinctive regional cultural characteristics.Several ethnic minorities inhabit the study area, including Han, Tibetan, Yi, Hui, Qiang, Mongolian, and others (Fang et al. 2014), which have formed a unique natural landscape, traditional settlements, and cultural characteristics (Fan, Wang, and Yang 2022).In this study, townships and towns are used as research units, so there are 51 research units (Figure 2 and Table 1).

Measuring spatial inter-embeddedness
The Shannon-Wiener Index (SWI) "H" was used to measure the diversity characteristics of cluster interembeddedness (Lin et al. 2011), which is calculated as follow: Where n is SWI, the higher the value, the higher the diversity of ethnic minorities, and the higher the degree of interdependence of rural settlements.P i is the number of ethnic minorities in the study unit, and P i is the proportion of the population of the i th ethnic minority to the total.

Spatial auto-correlation analysis
Using Moran's index to analyze the heterogeneity and degree of agglomeration in the spatial distribution of ethnic minority rural settlements and determine whether they belong to aggregation, dispersion, or random patterns (Jia et al. 2020;Shi et al. 2022).The spatial auto-correlation analysis mainly includes the global Moran's I and the local Moran's I.
(1) Global Moran's index.The Global Moran's I index (Yang et al. 2020;Chen, et al. 2023b) was used to analyze the distribution characteristics of rural settlements in multi-ethnic areas.The equation is as follows:  Where X ij is the number of minority populations in study unit i, j, and � X is the average number of minority populations in the study area.W ij is the spatial weight, generally calculated using the spatial weight matrix of the proximity or distance criterion, and S is the sum of the spatial weights.The value of i ranges between (−1, 1).When i is 0, it indicates a random distribution of minorities, and when i <0, it indicates that the minorities are spatially negatively correlated and show aggregation in spatially similar regions; conversely, i >0 indicates that the minorities are spatially positively correlated and show aggregation in spatially distinct regions.n is the number of study units, and there are 51 units in this study.
(2) Local Moran's index.In this study, Local Moran's I index was used as a local indicator of spatial association (LISA) (Yuan, Cave, and Zhang 2018), which could represent a set of localized statistical approaches that typically measure the relationship between individual locations and their surrounding neighbors to uncover patterns of spatial clustering (Bone et al. 2013). where: The meanings of X i;j , � X, and W i;j are the same as Eq 3 and Eq 4. When Z i >0 and I i >0, it indicates that the attribute value of study unit i and the surrounding units are high, which is a high-high aggregation type; when Z i >0 and I i <0, it indicates that the attribute value of study unit i is high and the attribute value of surrounding adjacent study units is low, which is a high-low aggregation type; when Z i <0 and I i >0, it indicates that the attribute value of research unit i is low and the attribute value of surrounding adjacent research units is high, which is low-high aggregation type; when Z i <0 and I i <0, it indicates that the attribute values of both research unit i and surrounding adjacent units are low, which is low-low aggregation type.

Data sources
(1) The vector data were obtained from the national 1:1 million public versions of topographic data (2021), which was provided by the National Catalogue Service for Geographic Information in China (https://www.webmap.cn/).(2) Multi-ethnic population data were obtained from the sixth national census of Garze Tibetan Autonomous Prefecture (http://tjj.gzz.gov.cn/).( 3) The digital elevation model (DEM) data was obtained from the geospatial data cloud (http://www.gscloud.cn/),the gross domestic product (GDP) data was obtained from the resource and environmental science and data center (https://www.resdc.cn/).

Spatial distribution characteristics of multi-ethnic rural settlements
Figure 3 shows that multi-ethnic rural settlements have significant heterogeneous characteristics in spatial distribution.In terms of spatial distribution, they have their geographic core settlement areas.This indicates specific segregation in the spatial distribution of multi-ethnic rural settlements.Specifically, (1) Han rural settlements are mainly distributed on the east side of the study area, forming a gathering core area (Figure 3a).( 2) Tibetan rural settlements are mainly distributed on the west side of the study area, occupying a large area, and are more balanced in spatial distribution.They formed two core areas in the north and south (south side, GE; north side, LC, TG, and XDQ) (Figure 3b).(3) Yi rural settlements are mainly located in HB, XJ, TK, and SY on the south side (Figure 3c), which is adjacent to Liangshan Yi Autonomous Prefecture, (3) the rural settlements of Hui, Qiang, and Mongolian have a significant aggregation in the spatial distribution, mainly located in LC (Figure 3d-f).
(4) LC is the town with the highest ethnic diversity, showing the characteristics of a large mix of multiple ethnic minorities with Han and Tibetans as the main groups (Figure 3 and Figure 4).The circle in Figure 5 was created using the Direct Distribution (Standard Deviation Ellipse) tool in ArcGIS.This tool is specifically designed to generate standard deviation ellipses, which summarize the spatial characteristics of geographic features in a useful and informative way. Figure 5a shows the influence of topography and geomorphology on the spatial distribution characteristics of minority village settlements.The elevation of the study area is divided into ten classes at 500 m intervals.It is not difficult to find that each minority rural settlement has significant characteristics in geospatial distribution.(1) Han mainly lives in the low altitude region 793-3500 m, represented by  The level of economic development (GDP) also has an essential influence on the spatial distribution characteristics of ethnic minority rural settlements (Figure 5b).We divided the GDP raster data of the study area into nine classes with an interval of 2 × 104 CNY/km 2 .It is found that: (1) the two core areas of the Tibetan agglomeration have lower GDP.(2) GDP is higher in the main areas of Han rural settlements.(3) Ethnic diversity and GDP have specific positive correlation characteristics in spatial distribution, and places with higher GDP have higher ethnic diversity.For example, LC, GZ, LQ, and other townships have more ethnic minority populations than other regions.

Spatial correlation and clustering characteristics of the distribution
Table 2 shows that the spatial distribution of rural settlements of multi-ethnic shows a significant aggregation characteristic except for Hui.Moran's I of Hui is 0.037, and the P-value is 0.341, which indicates that Hui ethnic group distribution has no spatial correlation.For the other ethnic minorities, all the values of Moran's I are above zero, displaying that rural settlement distribution showed positive spatial autocorrelation.If a region with a higher (lower) minority population also tends to have a higher (lower) minority population in its neighboring area.In terms of the degree of aggregation, Han (Moran's I = 0.770796, P < 0.01) has the most significant degree, while Qiang (Moran's I = 0.123270, P < 0.05) has the lowest.
Figure 5 shows the LISA agglomeration maps of the distribution of multi-ethnic rural settlements.
(1) Similar to the results in Table 2, the spatial distribution of hot spots (H-H) and cold spots (L-L) in ethnic minority rural clusters also showed significant differences. (2) The spatial

Inter-embeddedness patterns of rural settlements
Table 3 , Figures 6, and 7 show the inter-embeddedness pattern of multiple ethnic minority rural settlements.We divide the SWI into five classes according to the Natural Breaks method (Jenks) (ArcGIS, version 10.5), which are very high (SWI >0.80), high (0.60 < SWI ≤ 0.80), medium (0.40 < SWI ≤ 0.60), low (0.20 < SWI ≤ 0.40), and very low (SWI ≤0.20).The results reveal that: (1) The equilibrium of the minority population has a negative correlation with the degree of inter-embeddedness of clusters.Suppose there is a dominant ethnic minority in a study unit.In that case, its spatial inter-embeddedness is low (e.g., PBX, EE, SY, etc.), and if there are equal numbers of different ethnic minorities, its spatial interembeddedness is high (e.g., GE, HB, NQ, etc.).Therefore, improving the balance of ethnic population distribution is essential to promote the inter-embedding of clusters.(2) Rural settlements in very high and very low areas of spatial inter-embeddedness have the characteristics of significant clustering in spatial distribution and contiguous aggregation.In particular, the highvalue interval forms two core areas in the north and south.
Figure 8 shows the inter-embeddedness pattern and types of minority rural settlements.(1) Spatial interembedding of settlements at a fine scale is a complex process.(2) The pattern of spatial inter-embedding of settlements with high SWI has many types, such as two dominant types (e.g., LC, LuQ, XL, KY, etc.), three dominant types (e.g., GE, ZE, NQ, etc.).( 2) The uniform population size of ethnic minorities in the region positively promotes inter-embeddedness.This finding suggests a practical and theoretical guide for the future development of rural settlements in multi-ethnic regions.(3) If the population of one ethnic minority is in the majority in the region, the degree of inter-embeddedness is low,  and the spatial distribution pattern of settlements is dominated by one ethnic group (e.g., PBX, EE, GB, etc.).

Discussion
In this study, we aimed to investigate the spatial distribution of rural settlements in a multi-ethnic region at a finer scale, representing a novel attempt to explore the characteristics rural settlement patterns.This innovative approach diverges from previous research, which has primarily focused on urban areas (Chen, et al. 2023a;Wei et al. 2022a;Yang et al. 2023aYang et al. , 2023b;;Yu et al. 2022a;Yu et al. 2022b).We analyzed the spatial distribution and interaction characteristics of different ethnic minority settlements.Our study provides a theoretical reference for reconfiguring and  optimizing rural settlement space in the context of ethnic and cultural intermingling.

Spatial patterns of rural settlements in multi-ethnic areas
Rural settlement is the spatial carrier of rural production and life (Li et al. 2020).The spatial distribution characteristics of rural settlements in multi-ethnic areas reflect the pattern of exchange and integration between the populations of various ethnic minorities.
Most previous studies have focused on the spatial patterns of rural settlements (Chen et al. 2022;Jia et al. 2020;Xu et al. 2020;Yang, Xu, and Long 2016), but few focused on the spatial interaction characteristics in multi-ethnic areas.We found that multiethnic rural settlements show significant clustering characteristics in the spatial distribution and always have their core gathering areas in geography.The ethnic inter-embeddedness pattern was formed over a long period, resulting from the continuous migration, dispersal, and clustering of various ethnic minorities due to particular historical reasons.We also found that the higher the ethnic diversity of the area, the more economically developed the township (e.g., LC, LQ, etc.).This differs from the study of Barba and Jaimovich (2022) argued that ethnic diversity negatively impacts socioeconomic status.However, according to our research, the higher the ethnic diversity, the better the level of socio-economic development.Therefore, the study area's spatial distribution of rural settlements is unique.This result also further indicates that the spatial distribution of rural settlements is closely related to social, economic, and ethnocultural factors in addition to the region's physical and geographical environmental characteristics.In addition, ethnic diversity has a negative correlation with identifiability; rural settlements with low ethnic diversity have high identifiability, and we can quickly determine which ethnic minority settlement it is, while the opposite is true for high ethnic diversity.Figure 9 shows the cultural landscape characteristics of ethnic rural settlements in the study area.The payments with high ethnic diversity reflect the cultural aspects of the dominant ethnic group Tibetan.Furthermore, we found that the ethnic minority with the largest population usually held the leadership of the region (e.g., LC was named in Chinese, while WZ was in Tibetan.), which is consistent with the findings of Lin et al. (2011).In traditional Chinese society, the central government imposed a segregation system on ethnic minorities, restricting exchanges between ethnic minorities and Han (Li and Zhang 2017).Therefore, clustering is a distinctive feature of ethnic minority rural settlements, and blood relations are the basis of the social structure and become an essential support for social cohesion.Regarding spatial distribution, rural settlements of multi-ethnic areas have significant agglomeration characteristics.Living together is a typical feature in China, which formed a similar pattern of "scattered on the macro-scale and clustered on the micro-scale" during the long history (Tao et al. 2021).In recent years, with the socio-economic development, the mobility of the population and the intermingling of ethnic cultures have increased, and the patterns have changed to "inter-embedded and mixed."This pattern was further confirmed in our study.Currently, China promotes preserving the cultural heritage of multi-ethnic and ethnic integration.We believe the inter-embeddedness pattern will be a new type of rural settlement in multi-ethnic areas in the near future (Li and Zhang 2017).Thus, examining the spatial pattern characteristics of rural settlements in multi-ethnic areas is important.

Influencing factors of distribution patterns
Due to different natural environments and socialeconomic conditions, rural settlements in multiethnic gathering areas have obvious heterogeneity (Lobo, Flores, and Salvo 2007).As spatial carriers of ethnic culture, rural settlements map out distinct regional cultural characteristics.Most previous studies confirmed that socioeconomic is an important influencing factor in the social structure of rural settlements (Li et al. 2018;Zhang, Druijven, and Strijker 2017), leading to population migration and the intermingling of ethnic groups.In our study, elevation significantly impacts the distribution patterns of multi-ethnic rural settlements, consistent with the findings of Xu et al. (2020) and Li and Zhang (2017).In addition, we found that Tibetan lives in high altitude region, with an altitude range of 3500-4500 m, which is in consistent with the finding of Paik and Shawa (2013) that the highest number of Tibetan towns found at 4500 m.The high altitude, sparse vegetation, and more complex climate in these places have limited traditional farming production, and nomadic herding is the main mode of production in this place, which has become one of the main reasons for the slow economic development in minority areas.Furthermore, the GDP has a certain positive correlation with ethnic diversity in spatial distribution, promoting to form the of inter-embedding patterns, which was relatively close to a study in Dehong County that reported high GDP areas have dense multi-ethnic settlements and good development capabilities (Zhao et al. 2021), but Churchill and Smyth (2017) argued that ethnic diversity increased poverty, probably the correlation analysis should consider the difference of region.From the results of our study, the unique topographical features and socioeconomic conditions of the study area determine the uniqueness of the spatial distribution of minority populations and settlements.However, rural settlement spatial distribution is influenced not only by the natural environment, and socioeconomic development but also by ethnic culture, religion, and other factors (Li et al. 2018).In terms of the mechanism of action, the influence of the above-mentioned influencing factors on the spatial distribution and interaction pattern of multi-ethnic rural settlements is comprehensive and not only influenced by one factor.In addition, we found that this influence mechanism has both spatial and temporal properties, i.e., there is spatial variability in the influence with different regional characteristics, and secondly, there are also dynamic differences in the influence in the temporal dimension.Studying the spatial distribution and interaction characteristics of rural settlements in multi-ethnic areas can provide positive references for the optimization of the spatial pattern of rural settlements and the harmonious co-existence of multi-ethnic groups by exploring the heterogeneity of settlement distribution patterns due to socioeconomic and ethnic culture in the region (see Figure 9).

Theoretical perceptions and recommendations on rural settlements distribution patterns in multi-ethnic areas
The concept of "Community" first came from the German sociologist Ferdinand (1887; Community and Society) (Ferdinand 2012), which aimed to show that societies are communities of the closely connected and interdependent public.In this study, we have dedicated to exploring the positive impact of ethnic diversity on distribution patterns of rural settlements through the lens of ethnic interembeddedness, consistent with the conceptual connotation of "ethnic communities" (Che, Zhou, and Liu 2022;Slater 2022).In addition, we found that although multi-ethnic have their core agglomerations geographically, they also have significant spatial inter-embeddedness characteristics, which is consistent with the findings of Gao, Wang, and Zhu (2017).Moreover, we quantitatively assessed the spatial inter-embeddedness based on this finding.
In this study, we found that the equilibrium of the minority population negatively correlates with the degree of inter-embeddedness of clusters.Hence, it is important to improve the balance of ethnic population distribution to promote inter-embedding in the future.Since 2000, the Chinese government has implemented the "West Development" (Xibu Da Kaifa) policy (Han and Paik 2017;Li et al. 2018), which has greatly promoted socioeconomic development in rural areas, especially in remote multi-ethnic gathering areas.The increasing migration and integration among multiethnics have contributed to the heterogeneity of the spatial distribution.In recent years, local governments have extensively implemented relocation projects to guide residents from high-altitude areas to lowaltitude town centers, promoting access to better production resources, convenient transportation accessibility for rural residents.The implementation of these policy measures has effectively alleviated the uneven distribution of minority populations, allowing multiple minority populations to come together and promoting the intermingling and penetration of ethnic groups.In addition, the problem of uninhabitable high-altitude areas has been solved.Furthermore, most previous studies confirmed that many countries such as the UK, France, and Germany, et al. are now facing an increase in ethnic diversity (Huang et al. 2020;Li et al. 2021;Schmid, Hewstone, and Ramiah 2013), which will ensure the resulting spatial interembeddedness patterns in multi-ethnic areas (Buch, Meister, and Niebuhr 2021).
In addition, we also found that spatial interembeddedness has a significant positive correlation with the socio-economic of settlements.The better the socio-economic development of townships, the higher the degree of spatial inter-embedding.

Limitation
An interesting phenomenon that we neglected in this study, which was the existence of inter-ethnic transformation (e.g., the ethnic transition from Han to Yi, or transition from Qiang to Tibetan, etc.) in rural settlements where ethnic minorities have lived for generations, although this phenomenon is a rare presence and does not have a significant impact on the experimental study.However, it is material that cannot be ignored in the study of ethnic integration in rural settlements, yet this data is not easily obtained according to no statistical records.In the future, we will obtain data on a fine scale through fieldwork and analyze the mechanism of this transformation in depth.
Additionally, it is too difficult to obtain ethnic population data at fine scales, which involve significant human and material resources, as a result, this study only used data from one time period to analyze the spatial inter-embeddedness of the clusters and lacked dynamic change analysis.In the future, we will obtain data based on field surveys to study further the evolution of settlement patterns and the driving mechanisms.

Conclusion
Compared with previous studies, the paper analyzed the distribution patterns of multi-ethnic rural settlements from a new perspective.We focused on the spatial inter-embeddedness characteristics of specific regions and the impact of this characteristic on the spatial distribution of rural settlements.In this research, we found that multi-ethnic rural settlements have significant heterogeneous characteristics in spatial distribution, which have core clustered areas, respectively.Firstly, the eastern side of the study area shows an inter-embeddedness pattern with Han settlements as the main group mixed with other ethnic minorities, low altitude, and high economic development.Secondly, Tibetan settlements are mainly distributed on the western and northern sides, forming an inter-embeddedness pattern with Han and other ethnic minorities.Tibetan settlements occupy most of the study area, with high altitude and low economic development.Finally, on the south side of the study area, Yi is the main group, followed by Han, and other ethnic minorities live in a large mix, low altitude, medium economic development.We also found that the ethnic minority with the largest population normally held the leadership of the region.Furthermore, we confirmed that the elevation and social-economic (GDP) impact the distribution of ethnic rural settlements.In terms of clustering characteristics of spatial distribution.
In addition, we tried to divide the interembeddedness patterns into five classes and found that the equilibrium of the minority population has a negative correlation with the degree of interembeddedness of clusters.Our research would provide a theoretical reference for the optimization of distribution patterns of multi-ethnic rural settlements in the context of "Ethnic Integration" in China, which may also apply to other multi-ethnic countries.

Figure 1 .
Figure 1.Schematic representation of inter-embeddedness patterns of rural settlements.(The schematic does not mean there are only three prototypes, as the following empirical analysis will demonstrate this).

Figure 2 .
Figure 2. Location of the study area.

Figure 3 .
Figure 3. Spatial distribution of multi-ethnic rural settlements.

Figure 4 .
Figure 4. Statistics of an ethnic population of towns/townships.

Figure 5 .
Figure 5.The impact of Elevation and GDP on the distribution of multi-ethnic rural settlements.

Figure 6 .
Figure 6.LISA maps of multi-ethnic rural settlements.

Figure 7 .
Figure 7. Statistics of ethnic population and SWI.

Figure 9 .
Figure 9. Natural and cultural landscapes of rural settlements (Source: Captured by author).

Table 1 .
The names of counties and abbreviations in Figure2.

Table 2 .
Moran's I of ethnic minorities.

Table 3 .
SWI of ethnic minorities.