Localizing Indicators of SDG11 for an Integrated Assessment of Urban Sustainability—A Case Study of Hainan Province

: Rapid urbanization has brought many problems, including housing shortages, trafﬁc congestion, air pollution, and lack of public space. To solve these problems, the United Nations proposed “The 2030 Agenda for Sustainable Development”, which contains 17 Sustainable Development Goals covering three dimensions: economy, society, and environment. Among them, Sustainable Development Goal 11 (SDG11), “Make cities and human settlements inclusive, safe, resilient and sustainable”, can be measured at the city level. So far SDG11 still lacks three-quarters of the data required to accurately assess progress towards the goal. In this paper, we localized the indicators of SDG11 and collected Earth observation data, statistical data, and monitoring data at the city and county levels to build a better urban sustainable development assessment framework. Overall, we found that Haikou and Sanya were close to achieving sustainable development goals, while other cities were still some distance away. In Hainan Province, there was a spatial distribution pattern of high development levels in the north and south, but low levels in the middle and west. Through the Moran’s I Index of Hainan Province, we found that the sustainable development of Hainan Province did not yet form part of integrated development planning. The sustainable development assessment framework and localization methods proposed in this paper at the city and county levels provide references for the sustainable development of Hainan. At the same time, it also provides a reference for the evaluation of county-level sustainable development goals in cities in China and even the world.


Introduction
Over the past few decades, the world has been in the process of rapid urbanization. In 1950, only 30% of the world's population lived in cities. By 2018, this proportion had increased to 55% and is expected to increase to 68% by 2050 [1]. Although urban areas account for less than 1% of the global land cover, they contribute 75% of global GDP, consume 60-80% of energy, and produce 75% of the global waste [2]. Rapid urbanization has brought many challenges, including insufficient housing, traffic congestion, environmental pollution, and insufficient public space.
In order to measure, monitor, and report on the sustainable development of cities, the United Nations proposed 17 Sustainable Development Goals (SDGs) covering the three dimensions of economy, society, and environment. In the 2030 Agenda for Sustainable Development announced in 2015 [1], Sustainable Development Goal 11 (SDG11), "Building inclusive, safe, resilient, and sustainable cities and human settlements", is critical to achieving all the other SDGs [3]. The SDGs can be broken down into 169 indicators, and SDG11 includes seven technical targets and three policy targets, with a total of 15 measurable indicators of progress towards those targets. As of 29 March 2021, ten of those indicators presented problems to do with missing data in monitoring and evaluation, and one indicator was temporarily removed with no suitable substitute. In addition, SDG11 is linked to at least 11 other SDGs, and about one-third of the more than 230 SDG indicators can be measured at the city level [4][5][6][7]. China is the largest developing country in the world and has always insisted that development is the priority [8]. In 2016, China formulated the National Plan for the Implementation of the 2030 Agenda for Sustainable Development [9], and issued the Plan for the Construction of Innovation Demonstration Zones for the Implementation of the 2030 Agenda [10].
At present, the monitoring, comprehensive evaluation, and application of the United Nations' SDG indicators have become an international research frontier. The United Nations, national government departments, international organizations, and research institutes in China and elsewhere have carried out monitoring and comprehensive assessments of urban sustainable development indicators using statistical data. For example, experts from the Sustainable Development Solution Network (SDSN) carried out a comprehensive urban sustainability assessment for 45 European capital cities and 105 large American cities [11,12]. Based on statistical data, China's experts have also made sustainable development assessments of China's national, provincial, and municipal levels [13][14][15]. The recent studies did not use geospatial data, so it was difficult for them to effectively reflect the nation's spatial pattern and its variation in the SDG indicators. With the rapid development of satellite earth observation, some researchers have attempted to conduct evaluation and monitoring of SDGs combining statistics and geographic data [16][17][18][19][20]. However, in general, they are still at the stage of conceptual design, method discussion, and single indicator evaluation of small-scale pilot projects, and there has not yet been a comprehensive evaluation report of a complete administrative region [21][22][23].
Hainan Province is the newest province in China, and its urbanization development started relatively late, so it is easier to discover the impact of urbanization on sustainable development. In 2018, the Chinese government proposed the construction of the Hainan Free Trade Port, and released an overall plan for construction in 2020. Therefore, it is of great research significance to comprehensively assess urban sustainable development within the context of SDG11 in Hainan Province, and it is also a national strategic demand. The increasing urban population in Hainan brings many difficulties to the planning of urban public space, especially in the replanning of the old city. Therefore, in the context of provincial urbanization, unplanned urban expansion makes the urban ecological environment relatively fragile, and the ecological security situation is very grim [24][25][26]. Hainan Province has four cities and fifteen counties. Among them, 15 cities/counties have scarce statistical data, which cannot satisfy the sustainable development assessment of cities in Hainan Province.
Therefore, this paper takes Hainan Province as the study area, using Earth observation, statistics, and other data sources to carry out: (1) the construction of a localization system of SDG11 indicators in Hainan Province; and (2) quantitative and systematic monitoring and comprehensive assessment of SDG11 indicators at the municipal and county level of Hainan Province from 2010 to 2018. Our research can provide approaches and solutions for other provinces and regions to implement the SDGs.

Study Area
Hainan Province is located in the southernmost part of China, in a Special Economic Zone and Pilot Free Trade Zone. This research area includes three prefecture-level cities, five county-level cities, four counties, and six ethnic minority autonomous counties in Hainan Province, but excludes Sansha City, for a total of 18 cities and counties ( Figure 1).   [3]. Considering the actual situation of Hainan Province, however, the quantity of relevant data and the evaluation method cannot match these indicators completely. Therefore, it is necessary to adopt reliable, high-quality data applicable to the local situation of Hainan Province to localize the SDG11 indicators. This study mainly adopts statistics, Earth observation, and ground observation data, and the time scale is from 2010 to 2018 (Table 1).  [27] Note: + (−) means the higher (lower) the indicator, the better the level for sustainability.

Remote Sensing Data Preprocessing
In this study, the remote sensing data used were obtained from Google Earth Engine and the Atmospheric Composition Analysis Group [30,31]. From these sources, the urban built-up area was obtained based on the same definition of urban agglomeration area [33].
The standardized built-up area was converted from the impervious surface (ISA) data, which were obtained from the time series of Sentinel-1A/2 data in ascending/descending orbits [34]. First, the areas of urban patches were linked to form a continuous main urban area. After the small polygons and holes were removed, the built-up area was obtained through visual interpretation of the images on Google Earth Engine following the definition of the United Nations. For more details, please refer to the study by Jiang et al. [35].

Urban Sustainable Development Index System Based on SDG11
Given that SDGs are monitored based on country-level assessments, it is not possible to fully reflect the specific situation of different regions and local-scale administrative areas. Therefore, SDG indicators cannot be directly applied to a designated research area [36,37]. According to the regional characteristics of Hainan Province, the actual significance and use of each indicator of SDGs should be analyzed and improved to build a localized SDG indicator system [38]. Due to different administrative levels in the study area, some data cannot be used directly. Therefore, data need to be collated and processed to be applicable to all cities and counties and match SDG indicators. The data processing method is shown in Table 2. LCRPGR represents the ratio of land use rate to population growth rate; Urb t represents the area of the built-up area in the past as the initial value; Urb represents the area after n years of urban expansion, as the final value; Pop t represents the number of the city's population in the past as the initial value; Pop t+n represents the number after n years of urban population growth as the final value.

Methods
In order to better identify the development degree of a single indicator of a city, a dashboard like a traffic light is introduced using green, yellow, orange, and red to give the scores of the indicators [41]. Its purpose is to highlight the need of each city to pay special attention to the indicators with poor performance, remind relevant departments to prioritize their optimization, and improvement measures for the indicators. The comprehensive assessment of urban sustainable development mainly includes three steps: (I) screen data extremes and obtain a single index score through normalization; (II) build a dashboard for each indicator with its trend; and (III) calculate the city's comprehensive sustainable development score.

Single Indicator Score
Due to different city levels, some data are different in orders of magnitude. To ensure the comparability of these data, the largest or smallest outliers of each indicator are excluded. The maximum and minimum thresholds are obtained from the three largest and smallest data scores, respectively. After normalizing the data, each data score is 0-100, where 0 represents the lowest level of sustainable development, and 100 represents the best level of sustainable development.
The single index scoring method, namely normalization method is described below: where x is the original data, x max /x min represent the upper and lower bounds of the data, respectively, and x represents the normalized value after scaling the single index score. If the larger indicator means low sustainability, the normalized Equation (2) is used to get the score: where x is the original data, x max /x min represent the upper and lower bounds of the data, respectively, and x represents the normalized value after scaling the single index score.

Single Indicator Trend
In order to better show the long-term trend of a single indicator, we introduced the same method as used to derive the score of a single indicator to calculate the trend of a single indicator.
Specifically, we calculated the growth rate of each index. If the large index means high sustainability, then the growth rate formula (3) can be used to obtain the growth rate: where X GR is the growth rate, x is the data value of this year, and x n is the data value of n years ago. If the high index means low sustainability, then Equation (4) is used to obtain the growth rate: If the denominator is 0 in the calculation of the trend index, the trend index value is determined in three cases: if the numerator is positive, the trend index value is 3; if the numerator is 0, the trend index value is 0; and if the numerator is negative the trend index is −1.
In consultation with experts, the indicators were divided into four levels, as shown in Table 3 [42,43].

Comprehensive Evaluation Method
The sustainable development level of a city cannot be shown by the evaluation of a single index, so it is necessary to construct a comprehensive urban score to evaluate the relative sustainable development level of a city. We averaged all the indicators for each city in each year with equal weights to get a comprehensive score for the city: where n is the total number of indicators and x i is the score of a single indicator.

Single Index Ranking and Change Trend Analysis
Sanya had the largest number of "green indicators", with household garbage and per capita green area scores leading those of most cities and counties every year. Haikou had a stable ranking in each index, with rates of subsistence allowance, passenger volume, and solid waste leading those of most cities and counties. Changjiang and Ledong counties had the highest number of "red indicators" in 2010, Qiongzhong County had the highest number in 2015, and Ding'an and Qiongzhong counties had the highest number in 2018 ( Figure 2).

Analysis of Temporal Evolution and Spatial Patterns
The cities' highest score increased from 72.   Figure 5). Due to the rapid urbanization of Changjiang County, Wenchang County, and Ledong County in 2015, they neglected the impact of public space and the environment, resulting in a low total score. Wuzhishan City was the city with the lowest score for sustainable development.
As shown in Figure 6, Hainan's development was unevenly distributed, mainly high in the north and south while low in the central and western regions. Haikou, the administrative center of Hainan Province, and Sanya, an important tourist city in China, had the highest level of sustainable development, with annual comprehensive scores of more than 70 points. Ledong County and Wuzhishan City in the southwest performed poorly in the sustainable development indicators. After 2015, influenced by Sanya City, the sustainable development level of Ledong County rose rapidly to third place in the province in 2018. In Danzhou City, Dongfang City, Chengmai County and Changjiang County, the main heavy industry bases in Hainan Province, the overall sustainable development index performance was poor, although the urban sustainable development index performed well in some years. The central inland region had a low urbanization rate, poor urban modernization, and the lowest level of sustainable urban development. In 2018, the city with the best sustainable development level in the province was Sanya City (81.34 points), followed by Haikou City (73.60 points). In the same year, Wuzhishan City only scored 31.93 points, while other cities scored between 45 and 60 points.

Spatiotemporal Clustering Analysis
By the local spatial autocorrelation analysis of Moran's I from 2010 to 2018 (Figure 7, Table 4), it was found that the overall Moran's I of Hainan Province was low and the sustainable development efforts did not have a comprehensive effect. Most cities developed independently along their own lines, not directly benefiting neighboring cities. From 2011 to 2013, the urban agglomerations in Hainan Province, led by Ledong County, Baisha County, and Wuzhishan City, showed low-low clusters and high-low clusters, which indicated that these cities had a negative impact on the surrounding areas. High-high clusters appeared in Sanya City in 2018. The development of Sanya City drove the development of surrounding counties and cities, but an integrated development did not form. Obviously, the sustainable development of cities was influenced by the combined effects of nature, economics, and culture and was not determined by any single aspect.

Discussion
In contrast with previous studies, Ma et al. used four SDG11 indicators to obtain the conclusion that the level of sustainable urbanization in Jilin Province fluctuated and the development trend was increasing year by year [15]. We have also reached similar conclusions that the level of sustainable urbanization in Hainan Province has been increasing year by year, based on twelve SDG11 indicators. Research by Xu et al. found that cities in the Yangtze River Delta have not formed an integrated development model [14]. Similarly, our research found that cities and counties in Hainan Province have not formed an integrated development model yet, indicating that large cities in Hainan, such as Haikou and Sanya, have not prompted the development of surrounding cities. This is possibly due to the fact that the cities in Hainan basically focus on a single tourism economy structure originating from their own environment and cultural characteristics, and lack of high-tech industries [44]. It will be necessary to adjust the urban development model gradually to promote integrated development through cooperation between cities to improve sustainability [45,46].
Judging from our comprehensive assessment, the cities and counties of Hainan Province initially achieved sustainable development, but the degree of development of each city is different. High-scoring indicators should be maintained in the current development path and sustained; low-scoring indicators, such as disaster indicators, should be viewed as warnings. From 2010 to 2018, Ding'an's per capita green area indicator scores were low and still in a downward trend. The government should introduce relevant policies to increase the per capita green area. The air quality indicators of Hainan Province are in a good and stable state. The municipal and county governments in Hainan should continue to adhere to green development and maintain this state.
Similar studies mainly focused on a larger spatial scale (for example at city level) and most of them used statistical data [11,12,15,16]. Our method can be applied at finer countylevel scale due to the introduction of earth observation data, and provide a demonstration of evaluating sustainable development at the county level. Compared with the study of Ma et al., which used city-related indicators, our study focused on SDG11, which may have some limitations regarding indicator numbers [15]. Some spatial-related indicators still came from statistics. In the future, we will attempt to use more earth observation data instead of traditional statistics with more city-related indicators.

Conclusions
The conclusions of this paper are as follows. (1)  Combining earth observation and statistical data is proposed to solve the problem of insufficient data for assessing the sustainable development status of cities/counties in Hainan Province, China. This evaluation system and method can also be used in other administrative regions with various scales.
Author Contributions: Conceptualization and methodology, Z.S., Q.X. and C.Z.; validation, J.S., Z.S. and Q.X.; investigation, T.X. and H.Y.; resources, data curation, and writing-original draft preparation, C.Z.; writing-review and editing, Q.X.; visualization, C.Z.; supervision and project administration, Z.S. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
The data used to support the findings of this study will be available from the corresponding authors upon request.