Abstract: Exploring the allocation fairness and effectiveness of spatial accessibility of urban parks and green spaces is of great significance to urban planning and layout, as well as the livability of urban residents. From the perspective of park green space accessibility, we integrated multiple accessibility assessment models that consider varying travel modes. We used an improved kernel density function based twostep moving search model (KD2SFCA) that combines multilevel reachable time thresholds for different park green space types, followed by comparing its assessment results with those of network analysis method and Gaussian two-step moving search model (Ga2SFCA), to quantify the distribution characteristics of population and settlements and the accessibility characteristics of different park types under different travel modes. We evaluated the accessibility of green space of parks within the ring road of Hefei City bypass. The differentiation characteristics of assessment results and the optimization recommendations for urban planning were proposed. The park green spaces showed a clustering distribution pattern in the southern part of the study area, with a high overlap between park green spaces and the areas with high population density. The Ga2SFCA model and the improved KD2SFCA model gave spatially similar assessment results of parkland accessibility in most areas, but substantial differences existed in few limited areas. The spatial accessibility of parklands in Changfeng district (6.54%, 9.57%) and Luyang district (1.24%, 1.57%) were lowest in terms of the accessible area ratio and population ratio. The accessibility calculation results of the improved KD2SFCA and their spatial distributions were more realistic. This study identifies areas with insufficient park green space resources and provides data support and theoretical reference for the planning and layout of urban public spaces.

Key words: park green space, public space, accessibility model, resident demand, Hefei urban area.