Boundary Dataset of Built-up Areas in Chengdu-Chongqing Urban Economic Circle Based on POI&ISA Composite Index (2010, 2020)

Zhu, Y. L.^1,2  Zhang, Y.^1,2,3*  Yang, R. Z.¹  AShuo, A. Y.¹  NaiGuMe, E. W.¹

1. College of Geography and Planing, Chengdu University of Technology, Chengdu 610059, China;

2. Key Laboratory of Digital Drafting and Land Information Application, Ministry of Natural Resources, Wuhan 430079, China;

3. College of Architecture, Southeast University, Nanjing 210096, China

Abstract: Urban built-up area is an important basic information for urban planning and constr­uction management, which can reflect the urbanization process and urban spatial pattern. Acc­ura­te extraction of urban built-up areas’ boundaries is of great significance for promoting sus­tainable urban development. We using Landsat TM/OLI images in 2010/2020 and electronic map points of interest as data source to construct a comprehensive index (POI&ISA) reflecting the build-up level of cities, then through the best threshold selection developing a proper boundary dataset of built-up areas in Chengdu-Chongqing urban economic circle. This dataset includes boundaries data and areas data of the built-up area in the Chengdu-Chongqing economic circle which combining the landscape and functional features. Dataset is archived in .shp and .xlsx data formats, consisting of 257 data files with data size of 974 KB (compressed into 1 file, 611 KB).

Keywords: built-up area; POI&ISA index; Chengdu-Chongqing economic circle; electronic map point of interest; impervious surface

DOI: https://doi.org/10.3974/geodp.2023.04.06

CSTR: https://cstr.escience.org.cn/CSTR:20146.14.2023.04.06

Dataset Availability Statement:

The dataset supporting this paper was published and is accessible through the Digital Journal of Global Change Data Repository at: https://doi.org/10.3974/geodb.2023.09.02.V1 or https://cstr.escience.org.cn/CSTR:20146.11.2023.09.02.V1.

1 Introduction

Urban built-up area refers to that^[1] in the urban administrative region that has actually been developed and constructed, with basically available municipal and public facilities. The boundary of the area serves as an effective reflection of the urbanization process and urban spatial pattern and provides an basic information for urban planning, construction and management^[2,3]. It is widely used in the demarcation of urban development boundaries and the layout of public service facilities. In the context of the new urbanization strategy, precise demarcation of urban built-up area boundaries proves helpful in the quality assessment of urbanization quality and the optimization of urban spatial patterns, which holds great significance for sustainable urban development.

The urban built-up area includes the characteristics of both landscape and function. In terms of landscape, it is mainly reflected in the physical development and construction, with artificial surfaces dominating the area; in terms of function, it requires basic municipal and public facilities. Existing extraction methods have mainly focused on the landscape characteristics of urban built-up areas, wherein contiguous constructed land is identified as part of the urban built-up area. This includes techniques such as the neighborhood expansion method^[4] and normalized and differential construction index methods^[5]. With the advent of various forms of geographic big data in the information age, an increasing number of studies have shifted their attention towards the functional characteristics of urban built-up areas. They classify areas with relatively higher levels of social and economic development into built-up areas and have successively proposed a variety of extraction methods^[6–8] based on electronic map points of interest (Point of interest, POI) and night light images. In 2021, the Ministry of Natural Resources issued the industry standard of “Regulations for Urban Scope determination”, which places stringent demands on the basic data extracted from the built-up areas and involves a relatively complex process^[9,10]. Meanwhile, it suggests the use of big data and other means for auxiliary judgment. Zhang et al.^[11] comprehensively considered the actual surface coverage and the configuration of municipal public facilities and public facilities. They developed the built-up area comprehensive index (POI&ISA) based on POI and impervious surface index, which achieved high accuracy in the extraction of the built-up area of Wuhan.

The Chengdu-Chongqing region has always held an essential position in the national regional strategic layout. In 2011, there was an initial integration of the spatial layout of the entire urban system. In 2014, the sixth meeting of the Chengdu-Chongqing economic circle was established as a national strategy. In recent years, the central cities of the Chengdu- Chongqing economic circle have seen continuous enhancement of their radiating and driving influence, which accelerated the development of small and medium-sized cities coupled with gradual improvements in infrastructure. However, there is still a big gap between these cities and the more developed urban centers in the eastern regions. Accurate extraction of the urban built-up areas of the Chengdu-Chongqing twin cities economic circle is not only beneficial for understanding the characteristics of the urban development stage in this region but also serves as an essential step to support the goal proposed in the Planning Outline of Chengdu-Chongqing twin cities economic circle: “seamless matching between the social and economic development goals and the spatial support system”.

2 Metadata of the Dataset

The metadata of Boundary dataset of built-up areas in Chengdu-Chongqing urban economic circle based on POI&ISA composite index (2010, 2020)^[12] is summarized in Table 1.

3 Data Source and Methods

3.1 Data Source

The basic data sources for the development of this dataset are shown in Table 2.

Table 1  Metadata summary of the boundary dataset of built-up areas

3.2 Methods

The built-up area extraction method proposed by Zhang et al^[11] is centered on the construction of the POI&ISA index. This index is calculated using the following formula:    

                        (1)

where, POI represents the POI kernel density estimate, and ISA stands for the impervious surface index.

The POI kernel density estimation is mainly based on the regular region with a certain bandwidth around any given point in the region as the calculation range for density analysis. Weights are assigned based on the distance from the center point, with closer distances carrying higher weights. The estimated density for each point is the weighted density of all points in the region^[16]. The calculation equation is as follows:

                                                                   (2)

Table 2  Introduction of data sources

where, P_i  is the kernel density of any point i in the region; K_j  is the weight of the study subject j; D_ij is the Euclidean distance between point i and the study subject j; R is the bandwidth (D_ij<R); n is the number of study subjects j within the bandwidth range.

On the basis of Ridd’s^[17] VI-S (Vegetation-Impervious surface-Soil) model, where the urban land cover is a combination of vegetation, impervious surface and soil. The proportion of impermeable surface, that is, the impervious surface index, can be obtained by adding the abundance of high and low illumination obtained after the decomposition of the linear spectral mixing model^[18]. The equation for this is:

                                                                                 (3)

where, R_j is the spectral reflectance of band j; N is the number of elements; f_i is the proportional weight of end i in the image element; R_i,j is the reflectivity of element i in band j; e_j is the unmodelled residual error values.

3.3 Technical Route

The data development process in this study (Figure 1) mainly includes four steps: impervious surface index extraction, POI kernel density estimation, POI&ISA index calculation and optimal threshold selection, and post-processing.

3.3.1 Index Extraction of Impervious Surface

First, Landsat TM/OLI images underwent preprocessing, including radiation calibration and atmospheric calibration. Image registration was conducted based on the 2020 OLI images. Then, through the calculation of the improved normalized differential water index^[19], water masks were generated from the images. Finally, the influence after the water mask was separated by minimum noise, and four pure end elements (high illumination, low illumination, vegetation and bare soil) were extracted based on the scatter plot. After the linear spectral mixing decomposition, the abundance of high and low illumination was added to obtain the impermeable surface index.

3.3.2 POI Kernel Density Estimation

The full-category POI data was filtered based on the definition of built-up areas, retaining only the POI categories related to municipal utilities and public utility coverage. Since the service radius of such utilities and facilities typically fall between 500 m and 1,000 m, a bandwidth of 800 m was selected for the kernel density estimation in this dataset.

3.3.3 Calculation of POI&ISA Index and the Best Threshold

The POI core density and impervious surface index were geometrically averaged to calculate the POI&ISA index. The Densi-graph method^[6] was then applied to select the optimal threshold for the identification of built-up areas and the extraction of the initial range of the built-up area.

3.3.4 Data Processing

Referring to the post-processing method outlined in the Regulations for Urban Area Scope, the independent map spots with a distance from the built-up area were removed, and the adjacent map spot areas were considered for removal if their area was less than 0.2 km². Independent map spots with a distance of less than 100 m from the built-up area were also excluded. To ensure continuity in built-up areas, any inner pores (mainly water, green space, etc.) in the initial boundaries of the built-up area were filled in.

Figure 1  Technical route of the dataset develpoment

4 Data Results and Validation

4.1 Data Composition

This dataset consists of two sets of data items:

(1) The boundaries of the built-up area of 16 cities (2010, 2020) in the Chengdu- Chongqing economic circle. This includes 32 data files in the format of shape-file, with the coordinate system being the Albers equal cone projection.

(2) A statistical table of the area of 16 cities (2010, 2020) in the format of table.

4.2 Data Products

As shown in Figure 2, Chengdu and Chongqing dominate in terms of the area and scale of the experimental built-up area, with a considerable gap compared to the other 14 prefecture-level cities, among which Mianyang, Deyang and Nanchong are the leading built-up areas, while Ya’an has the smallest built-up area. From 2010 to 2020, the absolute increment of the built-up area of Chengdu and Chongqing is significantly higher than that of the other 14 prefecture-level cities, reaching 617.97 km² and 359.7 km², respectively. Following closely is Mianyang city, with an increase in area of 120.18 km²; the built-up area of other cities is under 100 km². In terms of the growth rate of urban built-up area, Ziyang exhibits the highest annual growth rate of 49%; only Chengdu, Deyang, Nanchong, and Leshan have growth rates below 10%; for the rest of the cities, the annual growth rate of built-up area ranges from 10% to 20%.

Due to the different spatial structure types and development conditions of the 16 cities in the Chengdu-Chongqing economic circle, their urban expansion types from 2010 to 2020 also show significant differences (Figure 3). According to the urban expansion classification rules proposed by Wilson and others^[20], the 16 cities in the Chengdu-Chongqing economic circle cover five expansion types: marginal expansion, linear expansion, filling expansion, enclave expansion and mixed expansion mode. Chengdu and Deyang, located in the vast Chengdu Plain, exhibit a typical marginal expansion mode for their built-up areas; the built-up

Figure 3  Maps of built-up area of Chengdu-Chongqing economic circle

areas of Yibin, Guang’an and Ya’an show a linear expansion mode along main roads and rivers; in most other cities, various expansion modes are mixed, with Chongqing city and Meishan city demonstrating enclaves into built-up areas based on marginal expansion.

4.3 Data Validation

The built-up areas of 16 cities in the Chengdu Chongqing economic circle for 2010 and 2020 were verified and compared with the respective reference built-up areas. It was observed that there was consistency in the center and form of the built-up area. The correlation coefficient between the urban built-up area and the reference built-up area in 2010 and 2020 reached 0.96 and 0.98, respectively.

The image verification of the experimental results of the built-up area was found between the size and continuity of the dataset and the reference dataset. First, the reference dataset uses the impervious surface as the indicator of urban built-up area, whereas the POI&ISA index also considers the facility POI core density index, which leads to the classification of some areas with low impervious surface index but high POI core density as the built-up area; according to the requirements of the urban built-up area, the green space, water body, etc. should also be included within the contiguous construction land, which is often surrounded by the POI&ISA index and the urban built-up area.

Figure 4  Verification of extraction results in built-up area (Chongqing)

5 Discussion and Conclusion

This dataset extracted the POI&ISA index to obtain built-up areas in both 2010 and 2020, which surpassed the validation dataset in terms of accuracy. The study verifies the applicability of the extraction method based on POI and impervious surface index for the extraction of built-up areas in cities of varying sizes. This contribution extends to the broader application of the POI&ISA index. This method considers the contiguity of urban construction land and the completeness of municipal public and public facilities, which provides a comprehensive representation of the landscape and functional characteristics of urban built-up areas. Furthermore, the operational steps are simpler compared to those outlined in the Regulations for Determination of Urban Scope. The data sources used are readily accessible and undergo dynamic updates. It is worth noting that the POI&ISA index not only enables the extraction of built-up areas through threshold selection to study the urban size and form but also reflects the construction level of these areas through its absolute value, which proves helpful for related research on the expansion of built-up areas from the perspective of scale and efficiency.

Author Contributions

Zhang, Y. designed the algorithms of dataset. Zhu, Y. L., Yang, R. Z. and AShuo, A. Y. contributed to the data processing and analysis. NaiGuMe, E. W. and Yang, R. Z. did the data verification. Zhu, Y. L. wrote the data paper.

Conflicts of Interest

The authors declare no conflicts of interest.

References

[1]       Ministry of Construction of P. R. China. Standard for basic terminology of urban planning (GB/T 50280—98) [S]. Beijing: China State Construction Industry Press, 1998.

[2]       Hu, T., Huang, X., Li, D. R., et al. Comprehensive evaluation of the urban built-up areas mapping ability from Luojia 1-01 nighttime light imagery over China [J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(3): 432‒442.

[3]       Ao, Y., Wu, B. L., Bai, Z. D., et al. Temporalspatial changes of urban built- up area expansion in Guangdong-Hong Kong-Macao Greater Bay Area, China based on NPP VIIRS like night light data [J]. Journal of Earth Sciences and Environment, 2022, 44(3): 513‒523.

[4]       Tan, X. Y., Chen, Y. G. Urban boundary identification based on neighborhood dilation [J]. Progress in Geography, 2015, 34 (10): 1259‒1265.

[5]       Bhatti, T. Built-up area extraction using Landsat 8 OLI imagery [J]. GIScience & Remote Sensing, 2014, 51(4): 445‒467.

[6]       Xu, Z. N., Gao, X. L. A novel method for identifying the boundary of urban built-up areas with POI data [J]. Acta Geographica Sinica, 2016, 71(6): 928‒939.

[7]       Lin, Z. L., Xu, H. Q., Huang, S. L. Monitoring of the urban expansion dynamics in China’s east coast using DMSP/OLS night-time light imagery [J]. Journal of Geo-information Science, 2019, 21(7): 1074‒1085.

[8]       Li, F., Yan, Q. W., Zou, Y. J., et al. Extraction accuracy of urban built-up area based on nighttime light data and POI: a case study of Luojia 1-01 and NPP/VIIRS nightime light images [J]. Geomatics and Information Science of Wuhan University, 2021, 46(6): 825‒835.

[9]       National Land and Space Planning Bureau of the Ministry of Natural Resources, Tongji University, Peking University, et al. Code of practice for standard urban built-up area delineation[S]. Ministry of Natural Resources of the People’s Republic of China, 2021.

[10]    Huang, M., Zhang, M., Zhang, B., et al. Exploration on the concept of “Urban built-up area” and method for “Urban built-up area Delineation”: taking 115 cities as the practice object [J]. City Planning Review, 2022, 46(5): 17‒26.

[11]    Zhang, Y., Zheng, F. J., Liu, Y. F., et al. Extracting urban built-up area based on impervious surface area and POI data [J]. Scientia Geographica Sinica, 2022, 42(3): 506–514.

[12]    Zhu, Y. L., Zhang, Y., Yang, R. Z., et al. Boundary dataset of built-up areas in Chengdu-Chongqing urban economic circle based on POI&ISA composite index (2010, 2020) [J/DB/OL]. Digital Journal of Global Change Data Repository, 2023. https://doi.org/10.3974/geodb.2023.09.02.V1. https://cstr.escience.org.cn/

CSTR:20146.11.2023.09.02.V1.

[13]    GCdataPR Editorial Office. GCdataPR data sharing policy [OL]. https://doi.org/10.3974/dp.policy.2014.05 (Updated 2017).

[14]    Jiang, H. P., Sun, Z. C., Guo, H. D., et al. A standardized dataset of built-up areas of China’s cities with populations over 300,000 for the period 1990‒2015 [DS/OL]. Science Data Bank, 2021.

[15]    Sun, J., Sun, Z. C., Guo, H. D., et al. Dataset of Chinese urban 2020 [J]. Chinese Scientific Data, 2022, 7 (1): 190‒204.

[16]    Cao, F. J, Zou, Y., Zou, Y. A fast extraction method of built-up area based on H/T breaks method and POI data [J]. Geography and Geo-information Science, 2020, 36(6): 48‒54.

[17]    Ridd, M. K. Exploring a V-I-S (Vegetation-Impervious surface-Soil) model for urban ecosystem analysis through remote sensing comparative anatomy for cities [J]. International Journal of Remote Sensing, 1995, 16(12): 2165‒2185.

[18]    Zhang, Y., Liu, Y. F., Liu, Y. Spatial and temporal patterns analysis of impervious surface in Wuhan city [J]. Scientia Geographica Sinica, 2017, 37(12): 1917‒1924.

[19]    Xu, H. Q. A study on information extraction of water body with the Modified Normalized Difference Water Index (MNDWI) [J]. National Remote Sensing Bulletin, 2005(5): 589‒595.

[20]    Wilson, E. H., Hurd, J. D., Civco, D. L., et al. Development of a geospatial model to quantify, describe and map urban growth [J]. Remote Sensing of Environment, 2003, 86(3): 275‒285.