Dynamic Expansion of Urban Land in China’s Coastal Zone since 2000

: Although a major region with strong urbanization, there is not yet a systematic and comprehensive understanding of urban expansion during the last 20 years for China’s coastal zone. In this paper, based on remote sensing techniques, and using indicators such as new urban land proportion, annual urban increase, and annual growth rate, as well as a landscape expansion index reﬂecting the urban expansion type (e.g., edge-expansion, inﬁlling, and outlying), we measured the dynamic expansion of urban land in China’s coastal zone since 2000. The results indicated that: (1) China’s coastal zone experienced rapid urbanization from 2000 to 2020, with the new urban land and annual urban growth rate at 17,979.72 km 2 and 4.83%, respectively. The new urban land was mainly concentrated in economically advanced regions, such as Bohai Rim, Shandong Peninsula, the Yangtze River delta, and the Pearl River delta. (2) The urban growth rates of coastal cities in Liaoning, Hebei, Shandong, southeast Fujian, and Taiwan became slower over time, with a sharp decline during 2015–2020. In the mid and south of China’s coastal zone, such as coastal cities in Jiangsu, Guangxi, and Hainan, there was slow urbanization before 2015, and urban land expanded dramatically during 2015–2020. (3) The urban expansion of China’s coastal zone was dominated by edge-expansion after 2000, but it went through a low-speed and intensive development stage during 2010–2015, with an increase in urban land less than 50% of that in the other three ﬁve-year periods, and the most signiﬁcant ﬁlling of urban space compared with the other three ﬁve-year periods, which was probably caused by the global ﬁnancial crisis. (4) The spatial-temporal differences in the urbanization process in China’s coastal zone were largely consequent on national economic development strategies and regional development plans implemented in China’s coastal zone.


Introduction
Urbanization is the result of a combination of economics, politics, and technology [1,2]. With urbanization, more than half of the world's population live in urban areas, and the proportion will reach 68% by 2050, with the strongest increases in developing countries [3,4]. Rapid urbanization has led not only to a dramatic increase in urban populations, but has also brought a series of problems, including pollution [5][6][7], shortage of resources [8,9], decrease in biodiversity [10][11][12], and so on, which have seriously affected the local climate, ecological environment and global change [13]. It is well known that urban expansion and population expansion occur at different rates; it is expected that by 2030 the speed of global urban land expansion will be three times faster than that of population expansion [14,15]. The rapid expansion of urban land has led to the loss and abandonment of farmland, reduction in forest cover and other ecological damage. Analyzing the spatiotemporal expansion of 75 typical cities in China from the 1970s to 2020. This suggested that China's urbanization process was strongly correlated with the administrative rank and population size of cities. Sun and Zhao (2018) [32] found that the urban growth rate was inversely proportional to city size by analyzing 13 cities across the Jing-Jin-Ji urban agglomeration, which was consistent with the results of Zhao et al. (2015) [35]. However, urbanization is a dynamic evolutionary process with significant spatial differences [36], such that large cities with a continued increase in urban land might exhibit a downward trend in urban growth rate compared to small cities [26,32]. China has many cities of different scale and it is of value to research the dynamic changes in urban land as well as the relationship between city scale and urban growth rate to meet the needs of future urban planning and management.
As a region with advanced economic development, China's coastal zone spans a large area from north to south, with climatic and geographical differences leading to an extremely uneven distribution of the population along the coast [37,38]. The urbanization process in China's coastal zone is significantly affected by resource distribution and the impacts of national and regional development policies [39,40]. A number of studies have been carried out to monitor the spatiotemporal characters of urban land in China's coastal zone [24,39,[41][42][43][44]. Most of these have attached importance to large and developed cities, such as Guangzhou [24], and urban agglomerations [23,39]. However, little attention has been paid to small and medium-sized cities in China's coastal zone, although they constitute a higher proportion than large cities, and research studies focused on urban expansion of all the cities in China's coastal zone are even rarer [16]. Therefore, it is necessary to perform more research on urban expansion for the entire coastal zone in China. In this paper, we quantified the spatiotemporal characteristics of urban expansion in China's coastal zone over the past 20 years using multi-temporal land use data for 2000,2005,2010, 2015, and 2020 [45]. This research aimed to (1) quantify and compare urban extent, new urban increase, urban growth rates, and urban expansion types in cities of China's coastal zone during different periods; (2) analyze the spatiotemporal differences in urban expansion at city level and regional level; and (3) discuss the consistencies between social-economic development policies and urban expansion in the cities of China's coastal zone.

Study Area
China's coastal zone is located at the intersection of Eurasia and the Pacific Ocean, with a continental coastline of more than 18,000 km and an island coastline of more than 14,000 km, including Liaoning, Hebei, Tianjin, Shandong, Jiangsu, Shanghai, Zhejiang, Fujian, Guangdong, Guangxi, Hainan, Hongkong, Macao and Taiwan ( Figure 1). Over 50% of China's large cities, more than 40% of its population and 60% of its GDP, are concentrated in the coastal zone [46]. In mainland China, coastal prefecture-level cities and some prefecture-level cities that are closer, but not adjacent, to the continental coastline, are defined as the coastal zone in this paper; in particular, Taiwan island and Hainan island are both included in this study, with Taiwan as a provincial level and Hainan dividing into Haikou, Sanya and other cities.

Land Use Data
The land use data of 2000, 2005, 2010, and 2015 were obtained from the China Coastal Land Use Database [45,47], which is produced by a visual interpretation method based on multi-temporal Landsat images. This database divides land use types into 8 primary categories and 24 secondary categories ( Figure 2); the primary categories include farmland, forest land, grassland, built-up area, inland freshwater, coastal saltwater, human made wetlands, and unused land. Subsequently, we updated the land use database based on Landsat/OLI images of 2020 according to the method provided by Di et al. [45]. Thus, complete land use data covering 2000, 2005, 2010, 2015, and 2020 were accessed. The land use data for 2010, 2015, and 2020 were selected to evaluate mapping accuracy; in particular, validation samples were collected from high resolution images on Google Earth, and the overall accuracy and Kappa coefficient were calculated for the 8 primary categories of land use. The results (as shown in Figure 2) confirmed that high accuracy of land use mapping was achieved. In this paper, urban land information was derived from the secondary category denoted 'city' and numbered 41 in Figure 2.

Land Use Data
The land use data of 2000, 2005, 2010, and 2015 were obtained from the China Coas Land Use Database [45,47], which is produced by a visual interpretation method based multi-temporal Landsat images. This database divides land use types into 8 primary ca egories and 24 secondary categories ( Figure 2); the primary categories include farmlan forest land, grassland, built-up area, inland freshwater, coastal saltwater, human ma wetlands, and unused land. Subsequently, we updated the land use database based Landsat/OLI images of 2020 according to the method provided by Di et al. [45]. Thu complete land use data covering 2000, 2005, 2010, 2015, and 2020 were accessed. The lan use data for 2010, 2015, and 2020 were selected to evaluate mapping accuracy; in partic lar, validation samples were collected from high resolution images on Google Earth, an the overall accuracy and Kappa coefficient were calculated for the 8 primary categories land use. The results (as shown in Figure 2) confirmed that high accuracy of land u mapping was achieved. In this paper, urban land information was derived from the se ondary category denoted 'city' and numbered 41 in Figure 2.

Urban Expansion Quantification
Three indicators were calculated to quantify the magnitude of urban expansion: annual increase (AI, km 2 a −1 ), new urban land proportion (NP, %), and annual growth rate (AGR, %). AI and NP were used to measure the change in urban area for a single city over different periods, and AGR was used to compare the urban expansion of a number of cities for a certain period while discounting the impact of city size [22,23,31]. The calculation formulas are as follows: where U start and U end are the urban area extents at the start and end of a certain period, respectively, n is the number of years spanned, A t i is the area of new urban land in period t i , i is the number of periods.

Urban Expansion Types
According to the positional relationship between new urban land and the original urban land, urban expansion types are mainly divided into three types: infilling, edgeexpansion, and outlying [24,38]. In this paper, the landscape expansion index (LEI) was used to identify different types of urban expansion. The LEI is calculated as follows: where S refers to the area occupied by original land in the buffer zone for new urban land,  [26], the buffer distance was set to 1 m in this paper. Urban expansion type is defined as outlying, infilling, and edge-expansion when LEI = 0, LEI > 50 and 0 < LEI ≤ 50, respectively. For a particular city, the mean landscape expansion index can be used to define the degree of urban expansion aggregation [48]. The MLEI is calculated as follows: where LEI i is the LEI of new urban land i, n is the amount of new urban land. Generally, the larger the value of MLEI, the more concentrated the distribution of new urban land is.

Quantification Characteristics of Urban Expansion in China's Coastal Zone during 2000-2020
From 2000 to 2020, China's coastal zone experienced rapid urbanization, especially cities in the Bohai Rim, Shandong Peninsula, the Yangtze River delta, and the Pearl River delta (Figures 2 and 3). The urban extent of China's coastal zone increased from 11,473.88 km 2 in 2000 to 29,453.60 km 2 in 2020, with an annual increase (AI) and annual urban growth rate (AGR) of 898.99 km 2 a −1 and 4.83%, respectively.  At the regional level, for the coastal area in 14 provinces, municipalities and special administrative regions, the distribution of new urban land showed significantly spatial differences. For example, the area of new urban land in CCGd was 4532. 45   At the regional level, for the coastal area in 14 provinces, municipalities and special administrative regions, the distribution of new urban land showed significantly spatial differences. For example, the area of new urban land in CCGd was 4532.45 km 2 , which accounted for 25.21% of total new urban land in China's coastal zone, and approached the total increases of CCL, CCH, and CCS (27.62%). At prefecture level, cities with a large increase in urban land were Suzhou, Shanghai, and Dongguan; the areas of new urban land were 1312.24 km 2 , 1127.21 km 2 , and 1058.30 km 2 , with corresponding annual growth rates of 7.04%, 4.58%, and 10.91%, respectively. The cities with small increases in urban land were mostly concentrated in CCL, CCH, CCF, Mc, Hk, and Tw, with new urban land and annual growth rates under 200 km 2 and 4%, respectively. Some cities with small increases in urban land had high annual growth rates; for example, the areas of new urban land in Qinzhou and Sanya were 58.21 km 2 and 67.91 km 2 , with corresponding annual growth rates of 5.66% and 8.09%, respectively.  As shown in Figure 4, CCL, CCH, CCS, CCZ, and CCF experienced rapid urban expansion and had a significant increase in urban land area before 2015, while CCJ, CCGx, and Hn showed rapid urbanization after 2015. To highlight the differences in the magnitude of urban expansion among coastal cities, graded maps of AI and AGR for each city in different periods are shown in Figure 5.  During 2015-2020, the cities with higher NP, AI, and AGR values were concentrated in the Yangtze River Delta, northern CCGd, CCGx, and Hn, with a significant increasing trend compared to other periods, especially for CCGx and Hn which experienced rapid urbanization, with all the NP and AGR values greater than 60% and 8%, respectively. For example, the NP, AI, and AGR values of Sanya were 94.85%, 12.88 km 2 a −1 , and 31.78%, respectively. However, CCL, CCH, Tj, and CCS entered into slow urbanization, especially Yingkou, Tianjin, and Yantai, for which the values of NP were below 5%. In addition, Suzhou had the largest new urban land area, which even exceeded the total new urban lands of CCGx and Hn during the past 20 years, with NP, AI, and AGR values of 53.29%, 139.85 km 2 a −1 , and 10.62%, respectively.

Urban Expansion Types in Different Periods
Urban expansion was classified into three types using Equation (4), including edgeexpansion, outlying, and infilling, respectively. Figure 6 shows the area proportions of the three types in coastal cities for the four periods, and Figure 7 shows cities exhibiting different expansion characters. Generally, the edge-expansion type was the main type of new urban patches in China's coastal zone over the past 20 years, while there were still significant spatiotemporal differences. For most coastal cities, with urban expansion, the area proportions of the edge-expansion type tended to decrease in the first three periods and then to increase during 2015-2020. The infilling and outlying types showed an opposite trend and fluctuating changes, respectively.
During 2000-2005, the urban expansion of China's coastal zone was dominated by edge-expansion, especially CCJ, Sh, CCZ, and CCGd, for which the proportions of edgeexpansion were mostly over 70%. Cities with a relatively large proportion of the outlying type were concentrated in CCS, southern CCZ, and CCF, though some cities, such as Dongying and Weihai, were characterized by an outlying process. During 2010-2015, the dominant expansion type varied significantly. For cities in southern CCF, CCGd, CCGx, Hn, and Tw, the dominant type gradually shifted from edgeexpansion to infilling, except in Chaozhou which was dominated by the outlying type with a proportion at 69.52%. For most of the other cities, the urban expansion type still prevailed over edge-expansion, except for Dongying (as shown in Figures 6 and 7). In addition, in cities in which the edge-expansion type predominated, especially cities in   During 2010-2015, the dominant expansion type varied significantly. For cities in southern CCF, CCGd, CCGx, Hn, and Tw, the dominant type gradually shifted from edge-expansion to infilling, except in Chaozhou which was dominated by the outlying type with a proportion at 69.52%. For most of the other cities, the urban expansion type still prevailed over edge-expansion, except for Dongying (as shown in Figures 6 and 7). In addition, in cities in which the edge-expansion type predominated, especially cities in CCZ, the proportions of the outlying type increased to various degrees compared with the previous 5-year period.
During 2015-2020, the prevailing type of urban expansion returned to edge-expansion in most coastal cities of China, and the proportions of the infilling type decreased significantly. In addition, the outlying patches were dispersed all over the coastal zone in relatively small proportions.
As shown in Figure 8,

Spatiotemporal Dynamics of Urban Expansion in China's Coastal Zone
Time series land use data, and multi-temporal analysis methods combined with indictors representing urban expansion patterns and types can effectively reflect the dynamic process of urbanization in terms of urban land [17][18][19][20][21]49]. A large number of studies have considered the spatial and temporal characteristics of urban expansion in typical cities or urban agglomerations of China based on remote sensing techniques [16,20,22,23,25,[30][31][32]34,50], which have demonstrated that China's urban expansion has significant regional difference. However, studies on typical cities, such as provincial capitals [30] or urban agglomerations [50], might ignore the characteristics of small cities or cities not radiated by urban agglomerations; thus, it is necessary to conduct a comprehen-

Spatiotemporal Dynamics of Urban Expansion in China's Coastal Zone
Time series land use data, and multi-temporal analysis methods combined with indictors representing urban expansion patterns and types can effectively reflect the dynamic process of urbanization in terms of urban land [17][18][19][20][21]49]. A large number of studies have considered the spatial and temporal characteristics of urban expansion in typical cities or urban agglomerations of China  which have demonstrated that China's urban expansion has significant regional difference. However, studies on typical cities, such as provincial capitals [30] or urban agglomerations [50], might ignore the characteristics of small cities or cities not radiated by urban agglomerations; thus, it is necessary to conduct a comprehensive analysis for a wide range of regions with cities of different sizes. This study provided a new perspective on the whole coastal zone rather than a local area or a single city in a coastal zone. Overall, the urban extent of China's coastal zone has expanded significantly during the past 20 years, and the speed of urbanization in coastal cities has exhibited prominent spatial-temporal differences.
During 2000-2020, the magnitude and type of urban expansion showed obvious spatiotemporal differences at city and regional level, and there were significant temporal differences in the urbanization process. Spatially, new urban land was mainly concentrated in economically developed regions, with notable differences among cities. Temporally, the urbanization process was unstable. For CCL, CCH, CCS, Southeast CCF, and Tw, the urbanization rate decreased, with the magnitude of urban expansion sharply declining in the last 5-year period (2015-2020). Meanwhile, in the south of China's coastal zone, such as CCJ, CCGx, and Hn, there was slow urbanization before 2015, and new urban land increased dramatically in 2015-2020.
In addition, urban expansion in China's coastal zone was mainly dominated by edgeexpansion after 2000, and the area proportions of infilling and outlying varied significantly during the four five-year periods. In particular, during 2010-2015, the dominant expansion type shifted from edge-expansion to infilling in most cities of CCGd, CCGx, and Hn. Meanwhile, the increase in urban land in China' coastal zone was much lower than that during the other three 5-year periods-this was most probably related to the global financial crisis of 2008. Under the impact of the global financial crisis, economic activities reduced, and urban expansion slowed down markedly, even in developed cities, such as Guangzhou and Shenzhen.
Previous studies have found that urban growth rate is strongly related to city size [34,35], while other studies have claimed that urban growth rate is independent of city size, which is expressed in Gibrat's law [51][52][53]. However, this study found that urban growth rates of small cities could be higher than that of large cities in certain periods. For example, the urban growth rates of CCGx and Hn were higher than other cities in the last period (2015-2020) (Figure 4), but the original urban extents in 2015 of these cities were much smaller than others. Considering Sanya and Suzhou, in 2015-2020, the annual urban growth rate of Sanya was three times that of Suzhou, while the annual increase in urban land was less than one tenth of Suzhou's. This phenomenon might be related to the regional development strategies put forward in different periods.

Impacts of Policies on Urban Expansion
Major policies, such as economic development strategies and regional development plans, are significant drivers of the dynamic evolution of urban expansion [22,23,37]. In this paper, the results indicated that there were obvious spatial and temporal differences in urban expansion in China's coastal zone during 2000-2020, and such differences had significant consistencies with major policies, such as the Five-year Plan (Outline of the Five-Year Plan for Social and Economic Development), special economic zones, city clusters, revitalization of the northeast, and so on. Figure 9 shows the national and regional development plans implemented in China's coastal zone since 2000. After China's accession to the World Trade Organization (WTO), China's economic development began a new stage, with regions involving the Bohai Rim, Shandong Peninsula, Yangtze River Delta, Southeast Fujian and Pearl River Delta entering a rapid stage of urbanization, and urban extent increasing significantly (Figures 3-5). For example, as a special economic zone with resource and development advantages, Dongguan had the largest annual growth rate (AGR) in 2000-2005. However, with the subprime crisis spreading all over the world in 2008, causing the global stock market shock and economic recession [54,55], economic activities reduced in Dongguan, which led to new urban land in Dongguan declining dramatically, especially in 2010-2015 when the area of new urban land was less than one thirtieth that in 2000-2005. In order to promote the economic development of Guangdong, a series of national and regional development plans were proposed after 2008, such as the Pearl River Delta Industrial Layout Integration Plan implemented in 2009. With a series of policy developments, promoting economic recovery, the area of new urban land of CCGd increased significantly during 2015-2020 (Figures 3 and 4). As problems, such as environmental pollution, reduced ecological diversity, and land abuse, have intensified, governments have paid more attention to the sustainable utilization of resources, and more and more sustainable development strategies concerning land intensification and environmental protection have been proposed, such as the cultivated land protection policy. During the three early periods (2000-2005, 2005-2010, and 2010-2015), with the development concepts of intensive development and regional economic integration put forward, the values of MLEI increased significantly, and the proportions of urban land infilling increased. This indicated that the development strategies strongly promoted urban expansion change from extensive to intensive [30,40].
In addition, as the revitalization of the northeast strategy proposed in 2004, CCL experienced rapid urban expansion at an incredible rate after 2005, and the NP, AI, and AGR values increased significantly. This demonstrated that regional development strategy had impacts on the urban expansion in different periods.
Moreover, as a new territorial unit for national participation in global competition, urban agglomerations have gradually become the main form to achieve regional economic integration and to drive the development of neighboring cities by radiation effects [32,56,57]. The scale, urbanization level, and agglomeration of city clusters had huge variability and unevenness [38,[56][57][58], which manifested in significant spatial differences and stage differences in the urbanization process in China's coastal zone. Driven by the radiation effects of city clusters, CCJ joined the Yangtze River Delta city cluster [59] in 2016, As problems, such as environmental pollution, reduced ecological diversity, and land abuse, have intensified, governments have paid more attention to the sustainable utilization of resources, and more and more sustainable development strategies concerning land intensification and environmental protection have been proposed, such as the cultivated land protection policy. During the three early periods (2000-2005, 2005-2010, and 2010-2015), with the development concepts of intensive development and regional economic integration put forward, the values of MLEI increased significantly, and the proportions of urban land infilling increased. This indicated that the development strategies strongly promoted urban expansion change from extensive to intensive [30,40].
In addition, as the revitalization of the northeast strategy proposed in 2004, CCL experienced rapid urban expansion at an incredible rate after 2005, and the NP, AI, and AGR values increased significantly. This demonstrated that regional development strategy had impacts on the urban expansion in different periods.
Moreover, as a new territorial unit for national participation in global competition, urban agglomerations have gradually become the main form to achieve regional economic integration and to drive the development of neighboring cities by radiation effects [32,56,57]. The scale, urbanization level, and agglomeration of city clusters had huge variability and unevenness [38,[56][57][58], which manifested in significant spatial differences and stage differences in the urbanization process in China's coastal zone. Driven by the radiation effects of city clusters, CCJ joined the Yangtze River Delta city cluster [59] in 2016, while CCGx and Hn joined the Beibu Gulf city cluster [60] in 2017, which accelerated urban expansion in those areas. The results showed that there were significant increases in the NP, AI, and AGR values in the cities of Jiangsu, Guangxi and Hainan during the fourth five-year period (2015-2020), and new urban land of CCJ, CCGx, and Hn was 1.13, 1.80, and 4.86 times greater than the sum of that during the previous three five-year periods combined (2000-2015), respectively; in particular, the new urban land in Sanya increased by 17.41 times compared to the other three five-year periods. The results also showed that the proportion of the outlying type in cities of China's coastal zone increased to different degrees during 2015-2020. This was mainly due to the radiation-driven effects of urban agglomerations, which had led to the expansion of new urban land patches to non-core areas of urban agglomerations and the increase in outlying patches.

Current Limitations and the Way Forward
In this paper, the proportion of new urban land in different periods, the annual increase in urban land and the annual urban growth rate were calculated to reveal spatiotemporal differences at regional and city levels. Theoretically, urban expansion includes not only the spread of urban land, but also the concentration of population, and there is a clear inconsistency between urban expansion and population expansion [61][62][63][64]. Many studies have found that population and economy play very important roles in urban land expansion [26,65,66]; therefore, it is of value to research the relationship among migration movements, economic growth, and urban land expansion. In addition, since rapid urban expansion results in the loss of farmland, forests and water, destruction of the ecological environment, and so on, it is important to analyze the ecological effects of urban expansion in China's coastal zone [13,57,58,67]. In future, the relationship of population expansion, economic growth, and urban land expansion as well as the ecological effects of urban expansion, should be considered to further reflect the characteristics of urbanization, and to help policymakers balance the relationships among land, population, environment, and development.

Conclusions
In this study, the dynamic evolution of urban expansion in China's coastal zone during 2000-2020 was analyzed based on LUCC data. The results showed that China's coastal zone experienced drastic urbanization during 2000-2020, with an average urban growth rate (AGR) of 4.83%. Cities with large new urban land area were concentrated in the regions of the Bohai Rim, Shandong Peninsula, Yangtze River Delta and Pearl River Delta, and the variation and trends of NP, AI and AGR in different regions showed dramatic spatiotemporal differences. In addition, the values of MLEI showed incredible increasing trends before 2015, which indicated that the new urban patches gradually became more concentrated over time. The temporal differences in urban expansion confirmed that the global financial crisis of 2008 had severely impacted the process of urbanization in China's coastal zone. Furthermore, the high consistencies between development policies and urbanization showed that national and regional development strategies and plans had probably played a very important role with respect to the differences in urban expansion in different cities or regions. Furthermore, driven by economic development policies, cities originally with a small area of urban land would be likely to exhibit a higher speed of urban expansion than many large cities.
In brief, it is worth noting that even in similar geographic environments, there are still significant differences in the development of coastal cities, and such differences are strongly related to the economic status and size of the city. As a result of policy guidance, there are also significant differences with respect to the ability of cities to agglomerate population and to tap the potential for economic development, leading to unbalanced characteristics of urban development. In the future, it is important to study the relationship among population, policy, economy development, and urban expansion. This study will provide essential information to policymakers, governments and investors for formulating strategies or investing capital to improve the efficiency and quality of urbanization.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.