*4.3. Limitations and Future Research*

This study also has certain drawbacks. Firstly, we generate the OD matrix using centroid coordinates of the park as end point, but, people enter the park when they arrive at the entrance [36]. In the subsequent study, the OD matrix can be generated by acquiring the coordinates of the park entrance as the end point of the trip.

In addition to distance and park attraction, individual socio-economic variations play a significant role in determining individuals' inclination to visit a park [42], this article makes no distinction between individuals' travel inclinations (e.g., income, ethnoracial characteristics, age, gender, and disability). Recent research has shown that people with low socioeconomic status have less access to parks [43,44]. However, few research has examined the disparities in park accessibility across Xi'an's various demographic groups. Consequently, future research might categories the population to investigate the environmental justice of the park in greater depth.

Thirdly, the accessibility of park during working days and on weekends also deserves further discussing. People have various preferences for parks on weekdays and weekends [45]. The prevalence of weekends encourages the practice of long-distance travel. This study focuses solely on the impact of various travel modes on park accessibility. Future researchers should describe accessibility factors at various times to reinforce the findings.

#### **5. Conclusions**

At the street level, the distribution of park supply and population demand in Xi'an 3 city is uneven. Zhangbagou Street Park has the largest number, including 14 parks, although there are still nine streets without parks. The relationship between population geographical distribution and policies is strong. The ring road has a small population density. However, the high-tech zone is densely populated due to company development policy support.

The G2SFCA underestimates regional accessibility. The iG2SFCA can provide a more realistic evaluation, yet the spatial pattern of accessibility under both models is comparable. Specifically, a very accessible gathering area is located next to the park. As distance increases, the level of accessibility eventually declines.

Under three modes of transport, the reachable range expands as the travel duration increases. In walking mode, the reachable area extends from the core of the Ring Road, Tang Great Wall Heritage Park and Qujiangchi Heritage Park area. In the 5-min scenario, the area of riding accessibility is the largest. The average values of park accessibility and SD decline as time passes. At 30-min, the park service gap between communities is at its smallest, but some areas remain unreachable in walking mode.

High accessibility areas differ with travel mode. In walking and driving modes, highaccessibility area moves northward. In riding mode, high value areas of accessibility are located on both sides of the east and west third ring roads. Under the 15-min walking condition, high values areas appear near Huancheng Park, Qujiangchi Heritage Park, and Daming Palace Heritage Park, which corresponds to Huang's research findings [46].

The distribution of park resources in the study area is inequitable. Due to the fact that 14.5% of the streets are not distributed with parks, the residents' demand for parks cannot be addressed within the streets. 33.3% of the streets have a location quotient of less than 0.2, while 31.2% having a location quotient of greater than 1, indicating those streets have a higher level of equality.

**Author Contributions:** Methodology, Y.L.; Data curation, Y.L. and S.S.; Writing—original draft preparation, Y.L.; writing—review and editing, Y.X. and L.H.; Visualization, Y.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** The dataset utilized and/or analyzed during the present study are available on reasonable request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.
