Resilience is a key concept in both psychology and physics that was introduced into geography in the late twentieth century and has since been examined from various human and natural perspectives. This study aims to analyze the spatial pattern of the physical resilience of rural settlements in Rudbar County against earthquakes. To this end, 12 criteria and 36 sub-criteria were employed, and the Analytic Hierarchy Process (AHP) and Weighted Sum (WS) methods were applied to quantify settlement resilience. Furthermore, Local Moran’s I and (Getis-Ord G*) hot spot analysis were utilized to examine the spatial patterns of resilience. The results indicated that, using the AHP method, the resilience levels of rural settlements were 0.66% very low, 13.24% low, 59.61% moderate, 26.49% high, and 0% very high, whereas the WS method yielded 3.32% very low, 33.77% low, 55.64% moderate, 7.27% high, and 0% very high. High-High (HH) clustering patterns were observed in the plains and piedmont areas, while Low-Low (LL) clusters were located in mountainous regions. The HH clusters corresponded to villages with high resilience, whereas LL clusters coincided with villages exhibiting low resilience. Moreover, resilience hot spots were concentrated in the northern piedmont region of Rudbar County, while cold spots were located in the mountainous Khorgam area. Overall, results obtained from both AHP and WS methods were consistent with those from Local Moran’s I and (Getis-Ord G*) analyses, confirming significant spatial heterogeneity in the physical resilience of rural settlements against earthquakes. These findings highlight the importance of spatially informed resilience assessment for disaster risk reduction and rural planning.
Introduction
   The concept of resilience to natural hazards is used in the field of disaster management. Numerous models and indicators have been developed and applied to examine resilience to natural hazards. However, most models and indicators do not take into account spatial differences and patterns. Therefore, the results of most studies cannot represent a spatial perspective (Sung & Liaw, 2021). The aim of this research is to investigate the spatial pattern and differences in the resilience of rural settlements to earthquake risk. Therefore, spatial autocorrelation analysis, which is one of the most ideal statistical approaches, has been used to detect spatial distribution patterns. This research seeks to identify the spatial pattern of the level of physical resilience of rural areas of Rudbar County against earthquakes, and aims to answer the question "To what extent is the physical resilience of rural settlements in Rudbar County against earthquakes?"
Material and Methods
   This study was conducted in Gilan province and Rudbar city at geographical coordinates 36° 32′ to 37° 07′N latitude and 49° 11′ to 50° 05′E longitude. According to the 2016 census, this county has a population of 100,943 people, 30,350 households, and accounts for 4% of the total population of Gilan province. There are also 151 inhabited villages in this county, according to the general population and housing census.
   In this study, 12 criteria and 36 sub-criteria effective in the physical resilience of rural settlements were identified with the help of experts and specialists in this field. Then, the effect of each criterion on the final level of resilience was determined in order of importance, and the Analytic Hierarchy Process (AHP) and Weighted Sum (WS) models were used to determine the level of resilience of rural settlements, as well as two local Moran I indices and hot spot analysis in ArcGIS software to identify the type of spatial pattern.
Results and Discussion
   The results of spatial cluster and non-cluster analysis of the analytic hierarchy process (AHP) index values ​​using the local Moran I method showed that the high-high (HH) cluster is located in the northeast of the study area. In other words, villages with high resilience are visible in the plain and foothill areas of northern Rudbar County along the Sefidrood Valley to Rostam Abad and have a high-high (HH) cluster pattern. In this regard, most rural settlements in the Khorgam district of Rudbar county also have a low-low (LL) cluster pattern and are less resilient to earthquakes. Also, most of the villages located in the foothills of this county have moderate resilience to earthquakes and do not show significant differences with the high-high or low clusters, and have random and meaningless behavior. As a result, out of 151 villages in this county, 69 villages have random behavior (45.7%), 29 villages have a high-high pattern (19.2%), zero villages have a high-low pattern (0%), 18 villages have a low-high pattern (11.9%) and 35 villages have a low-low pattern (23.2%).
   The spatial pattern of the resilience of rural settlements using the weighted sum (WS) method with the local Moran's I statistic showed that the high-high (HH) pattern includes villages in the plain and foothills, and the low-low (LL) pattern includes villages in the mountainous region. The HH pattern includes villages located in the northern part of Rudbar County, which extends along the Sefidrood Valley to the beginning of Rudbar. Also, most of the villages located in the mountainous area of ​​Khorgam district of Rudbar county also have the LL pattern and their resilience to earthquakes is low. As a result, out of 151 villages in this county, 63 villages follow random behavior (41.7%), 46 villages follow the high-high pattern (30.5%) and 30 villages follow the low-low pattern (19.9%).
   The results of calculating the hot spot analysis (*G Getis-Ord) of the resilience of rural settlements using the AHP method show that the center of the hot spots of resilience is located in the northern area of ​​the county along the Sefidrood Valley (north-south). Also, the epicenter of the earthquake resilience cold spots includes most of the villages in the Khorgam district of this county. These points are marked with Z equal to ±3 and their hot and cold spots are significant with 95% confidence. In a part of the mountainous area of ​​this county, the meaninglessness of hot spot analysis is quite clear, indicating the mediocre resilience of rural areas in this part. A small number of villages are located in the Z range of +2 or -2, and in general, most hot and cold spots are observed in a cluster pattern in the plain and mountainous areas. The results of the weighted sum are not much different from the analytic hierarchy process model, and the rules governing hot spot analysis in the AHP model also apply to the WS model. Therefore, out of 151 villages in this county, 25 villages are among the cold spots at a 99% confidence level and 26 villages are among the hot spots at a 99% confidence level in the analytic hierarchy process model. This was also obtained for the weighted sum model, equal to 30 and 34 villages for cold and hot spots, respectively. The analysis of hot spots shows that 65 and 54 villages behaved randomly in the hierarchical analysis and weighted sum model, respectively, against earthquakes, and their resilience was average in both methods.
Conclusion
   The aim of this paper is to identify the spatial pattern governing the resilience of rural settlements in Rudbar County against earthquakes. Considering the very severe experience of the earthquake of June 21, 1980, it should be said that earthquakes are still considered a serious danger for this county due to the moderate to low resilience of rural settlements. It should also be said that topography is one of the effective factors in the spatial distribution of vulnerability and physical resilience to earthquakes in this county. In other words, the mountainous areas of this region have high vulnerability and low resilience to earthquakes, and this increases towards the foothills and plains. As a result, the level of development of areas in various dimensions (physical, economic, social, etc.) has a direct impact on the level of resilience, and this issue works completely the opposite in the studied area with altitude. Therefore, it is necessary for the responsible organizations to pay special attention to the issue of earthquakes and to strengthen residential buildings against earthquakes in the mountainous and foothill areas of the county. In addition, the construction of rural settlements appropriate to local conditions and their correct location according to various environmental criteria (including slope, slope direction, soil type, geological formation, distance from faults, etc.) are among the issues that must be considered. Another issue related to rural settlements is the use of indigenous architecture with new technology, as well as indigenous and traditional materials, in order to preserve the heritage of past rural settlements in this area, while observing engineering and architectural principles, which must be considered.