Establishing a "dynamic two-step floating catchment area method" to assess the accessibility of urban green space in Shenyang based on dynamic population data and multiple modes of transportation

Abstract

As an integral part of the urban environment, urban green space (UGS) is of great significance in maintaining urban ecosystem balance and biodiversity. Spatial accessibility is an important indicator of UGS distribution and can be calculated by the two-step floating catchment area method (2SFCA). However, problems exist in previous studies using 2SFCA: (1) the dynamics in population distributions are ignored when measuring UGS demand; and (2) travel costs are calculated for only a single mode of transportation. To address these problems, this study proposes a dynamic two-step floating catchment area method (D2SFCA) based on the Baidu heatmap and direction application programming interface and compares it with the traditional 2SFCA to investigate the characteristics of and differences in UGS accessibility in the first-ring built-up area of Shenyang, China. The results show the following: (1) the dynamic population distribution data calculated by the heatmap yielded the highest population density raster (581.0–1342.0 p/hm²) in areas with dense road networks and the lowest population density raster (2.0–91.0 p/hm²) in areas with railways and rivers, thereby more accurately reflecting reality and better quantifying the UGS demand than the static population distribution data; (2) the D2SFCA and 2SFCA findings had a slightly different distribution, and the D2SFCA assessment results more accurately reflected actual patterns, especially when the road data were inaccurate. In summary, the D2SFCA is more suitable for assessing accessibility and can identify specific areas that lack UGS. This study provides a scientific basis and methodological support for improving the services level and equity of UGS.

Introduction

Urban green space (UGS) is an important part of the urban built environment (Rutt and Gulsrud, 2016) that can help to reduce the heat island effect and noise annoyance (Arghavani et al., 2020, Van Renterghem, 2019); it can also improve the urban microclimate, water quality and air quality (Cetin, 2019, Ibrahim et al., 2020, Chen et al., 2019, Matos et al., 2019), provide a comfortable spatial environment for people's recreation and communication activities (Carrus et al., 2015, Zhao et al., 2020), and provide important ecological and social services and functions. According to some research, the severity of the heat island effect is decreased by 0.058 °C for every 1% increase in urban vegetation (Park et al., 2014). Jeanjean et al. (2016) suggested that aerodynamic dispersion mechanisms in Leicester city center green areas could lower air PM2.5 concentrations by 9%. In recent years, as social and economic development and ecological civilization construction have grown, urban residents have continuously increased their demand for open spaces (Ma, 2020, Sikorska et al., 2020, Neutens, 2015), particularly those for which there is convenient, equal and fair access (Giuliani et al., 2021, Sun et al., 2022). UGS accessibility is an effective indicator for evaluating the rationality of UGS layout (Kabisch et al., 2016, Zhang et al., 2021); it can reveal the degree of spatial resistance associated with reaching a UGS, specifically in regard to whether visitors can freely enter the green space and whether it is welcome (Biernacka and Kronenberg, 2018, Biernacka and Kronenberg, 2019).

Spatial accessibility evaluation methods have received widespread attention around the world, and the methodological system is relatively rich and used extensively in the study of the reasonableness of UGS distribution and service fairness. The main methods for assessing accessibility are the minimum distance method (Apparicio et al., 2008), simple buffer analysis (Mallick and Routray, 2001), network analysis (Li and Liu, 2009), the gravity-based model (Hansen, 1959, Wu et al., 2017) and the two-step floating catchment area method (2SFCA) (Liu et al., 2021). The minimum distance method and the simple buffer analysis method are based on Euclidean distance and do not take into consideration the urban road network (Nicholls, 2001); thus, UGS accessibility is overestimated. The network analysis method, which is centered on utility points, estimates the service area along the road network based on a certain cost related to time or distance, thereby adding the consideration of the urban road network, but ignoring the relationship between supply and demand (Liu et al., 2010). The gravity-based model considers the magnitude and potential of the interaction between the UGS's ability to provide services and residents' demand. In this method, the choice of the distance decay function and its decay coefficient have a significant impact on the accessibility calculation results (Kong et al., 2010, Kwan, 1998); however, there are no strict criteria for choosing a function to evaluate the UGS accessibility (Hao et al., 2021, Lee et al., 2018), resulting in inconsistent evaluation criteria and a reduction in the comparability of evaluation results. Overall, the above accessibility evaluation methods have limitations that are difficult to overcome.

2SFCA is built on the gravity-based model and uses the road network to determine the accessibility between supply and demand by performing floating catchment searches twice; the public facility is the supply point and the residents represent the demand point. The advantages of this method are that it can consider the supply scale of urban public facilities, the demand scale, and the distance relationship between supply and demand, and the interpretation is more intuitive (Li et al., 2019, Luo and Wang, 2003). Although 2SFCA and its improved forms have yielded a number of useful results in the study of green space accessibility (Dony et al., 2015, Wei, 2017, Zhang et al., 2019), most studies use census data to estimate demand, which ignores the daily fluctuating demand of public transportation (Ryan, 2011), ignores residents' short-term location dynamics, lacks timeliness (Tong et al., 2021), and makes it difficult to estimate the actual population distribution. Furthermore, most of these studies estimate UGS accessibility for a single mode of transportation (e.g., walking or driving) and rarely consider and measure UGS accessibility for multiple modes of travel simultaneously (Yang et al., 2021). However, in practice, people choose different modes of transportation to UGSs depending on the distance (Chang et al., 2019, Ma and Lu, 2011), so it is necessary to explore the method of measuring the accessibility of UGSs in the context of actual population distribution and multimode transportation to provide a scientific basis for a more comprehensive and realistic study of the distribution logic and accessibility of UGSs.

To address the limitations of the current methods for assessing UGS accessibility in terms of data inaccuracy and lack of consideration of actual population activities, this study implements a dynamic two-step floating catchment area method (D2SFCA) by using real-time travel data from the Baidu direction application programming interface (API) and population activities from Baidu heatmaps. The D2SFCA has two advantages over traditional 2SFCA. To evaluate demand scale, the traditional 2SFCA uses population data based on the census (Lin et al., 2021), which have a large scale (generally the administrative region of the statistical unit) that makes it difficult to reflect more detailed differences in accessibility (Wu et al., 2018, Zhang et al., 2022); however, small scale (especially community scale) data are more likely to indicate the lack of accessibility (Tan and Samsudin, 2017). At the same time, the population census is conducted every 10 years in China, so population changes lag behind the actual situation. In contrast, short-term changing population heatmaps can capture the trajectory of population dynamic activities at a small scale, thereby reflecting the spatial distribution as the real-time population aggregation degree (Li et al., 2021, Wang et al., 2022). To measure the distance relationship between supply and demand, the road network data used for network analysis in the traditional 2SFCA is obtained from satellite image interpretation, which has insufficient data accuracy and relatively long update cycles. In contrast, the map platform can capture the latest status quo of the roads and calculate the real-time navigation time (Rong, 2020). Using the time distance, D2SFCA can reflect the changes in the distance relationship between population and UGS in real time. In summary, the improved D2SFCA evaluation results are more in line with the dynamic changes in the actual situation.

This article addresses the following objectives: (1) to use heatmaps to calculate the dynamic population distribution in the study area and compare it to the static population distribution; (2) to use dynamic population and multimode transportation data to implement the D2SFCA and study UGS accessibility in the study area; and (3) to propose a spatial optimization strategy for the UGS system.