Introduction

Rapid urbanization, the dominance of economic relations over ecological values, and the transformation of lifestyles have confronted citizens with challenges such as mental exhaustion, disconnection from nature, and the weakening of sense of belonging. In such circumstances, the concept of "urban refuge" has gained increasing importance as a need of urban dwellers for spaces to restore tranquility, reflect, and rebuild connection with themselves and nature. This concept is rooted in Appleton's "prospect-refuge theory," according to which environments that simultaneously provide opportunities for seeing (prospect) and shelter (refuge) help reduce stress. From an environmental psychology perspective, these spaces are also consistent with Kaplan and Kaplan's "attention restoration theory." The garden cities surrounding Zanjan, as valuable examples of nature-compatible settlements, possess unique qualities that can inspire the creation of refuge spaces for Zanjan. However, in recent decades, these garden cities have experienced physical and functional disconnection from the city and a process of decline and destruction due to reasons such as unplanned urban physical development, land use changes, and neglect of their inherent values. This research aims to explain the effective components in creating urban refuge spaces through the garden cities around Zanjan.

Methodology

The present study is applied in purpose and mixed-method (qualitative-quantitative) in nature. The qualitative section includes identifying effective components in creating urban refuge spaces in the garden cities surrounding Zanjan and reviewing theoretical foundations through library studies and conducting fuzzy Delphi analysis with the participation of 35 experts in the fields of hazards and urban planning. In the quantitative section, using structural equation modeling and path analysis method in PLS software, the type of components was determined according to their influence and effectiveness on other components. For validity and reliability of the questionnaire, face validity and construct validity (factorial) were used, and Cronbach's alpha test was employed for reliability assessment. The investigated components included six main dimensions: access and connectivity, safety and comfort, sense of place belonging and perception, ecological and social, management and maintenance, and urban refuge spaces, each with numerous sub-components.

Results and discussion

The results of the fuzzy Delphi in two stages showed that out of 36 initial components, 30 components with a fuzzified value above 0.7 were approved by the experts. The factor analysis results showed that most items had factor loadings above 0.7, indicating appropriate convergent validity of the model. In the "access and connectivity" dimension, factor loading values ranged from 0.722 to 0.896, with the highest values related to safe pedestrian paths (0.896) and access to public transportation 0.877. In the "safety and comfort" dimension, factor loadings ranged from 0.755 to 0.873, with the indicators of appropriate boundary definition 0.873 and lighting 0.853 having the greatest impact. In the "ecological and social" dimension, values ranged from 0.702 to 0.856, with the highest values assigned to natural smells (0.856) and natural materials (0.828).In the "management and maintenance" dimension, factor loadings ranged from 0.669 to 0.768, with water management and native species (0.768) and transparency of rules (0.730) being the most important indicators. In the "sense of place belonging" dimension, equitable access (0.886) was the strongest indicator. In the "urban refuge spaces" dimension, values ranged from 0.750 to 0.898, with distance from high-risk land uses (0.898) and the ability to change use in crises (0.863) playing the most significant roles.

The structural model results showed that the component of "sense of place belonging and perception" with a path coefficient of 0.477 and a T-value of 7.869 had the most significant positive effect on urban refuge spaces. The "ecological and social" component with a path coefficient of 0.358 and a T-value of 4.584 had a positive and significant effect. In contrast, the "access and connectivity" component with a negative path coefficient of -0.353 had a significant but negative effect on urban refuge spaces. The "safety and comfort" component with a significance level of 0.892 had no significant effect. The "management and maintenance" component, although having a positive effect, was not significant at the 95% level due to a significance level of 0.053. The coefficient of determination (R²) for the dependent variable was 0.763, indicating the high explanatory power of the model. The goodness-of-fit indices (SRMR, d_ULS, d_G, and NFI) were all within acceptable ranges, and the model had a good fit.

Conclusion

The findings of the present study showed that the creation of urban refuge spaces in the garden cities surrounding Zanjan is influenced more by perceptual, identity, ecological, and social dimensions than by purely physical and functional components. The most important finding is the prominent role of the "sense of place belonging and perception" component. This result is consistent with Norberg-Schulz's view on "genius loci" and the theories of Relph and Tuan on place attachment. This finding can also be explained by Kaplan and Kaplan's "attention restoration theory." The negative and significant effect of the "access and connectivity" component indicates that excessive increase in permeability and accessibility can reduce the refuge quality of garden cities. This finding is consistent with Appleton's "prospect-refuge theory," which emphasizes the necessity of simultaneous prospect and refuge. The "safety and comfort" component, despite having relatively high factor loadings, had no significant effect in the structural model, indicating that safety and comfort are necessary conditions for presence in space but not sufficient conditions for the formation of a sense of refuge

The research results show that planning for creating urban refuge spaces in Zanjan should not focus solely on developing green spaces or improving physical indicators, but should emphasize strengthening the perceptual, identity, and ecological qualities of garden cities. Protecting the natural and historical structure of garden cities, strengthening the sense of place belonging, controlling unplanned physical development, enhancing the sensory qualities of the environment, and creating a balance between accessibility and tranquility can pave the way for the formation of spaces that play a restorative role in normal conditions and a refuge role in crisis conditions. Such an approach can open new horizons in contemporary urban planning, especially in Iran's transitioning cities