Ecosystem Services Changes on Farmland in Response to Urbanization in the Guangdong–Hong Kong–Macao Greater Bay Area of China

: Extensive urbanization around the world has caused a great loss of farmland, which signiﬁcantly impacts the ecosystem services provided by farmland. This study investigated the farmland loss due to urbanization in the Guangdong–Hong Kong–Macao Greater Bay Area (GBA) of China from 1980 to 2018 based on multiperiod datasets from the Land Use and Land Cover of China databases. Then, we calculated ecosystem service values (ESVs) of farmland using valuation methods to estimate the ecosystem service variations caused by urbanization in the study area. The results showed that 3711.3 km 2 of farmland disappeared because of urbanization, and paddy ﬁelds suffered much higher losses than dry farmland. Most of the farmland was converted to urban residential land from 1980 to 2018. In the past 38 years, the ESV of farmland decreased by 5036.7 million yuan due to urbanization, with the highest loss of 2177.5 million yuan from 2000–2010. The hydrological regulation, food production and gas regulation of farmland decreased the most due to urbanization. The top ﬁve cities that had the largest total ESV loss of farmland caused by urbanization were Guangzhou, Dongguan, Foshan, Shenzhen and Huizhou. This study revealed that urbanization has increasingly become the dominant reason for farmland loss in the GBA. Our study suggests that governments should increase the construction of ecological cities and attractive countryside to protect farmland and improve the regional ESV.


Introduction
Unprecedented urbanization has been occurring worldwide, which has caused the urban population to currently account for more than half of the world s population; this proportion is expected to increase to nearly three-quarters by 2050 [1]. Considerable urban expansion has mostly taken place in developed countries in the past, but it will mainly occur in developing countries in the coming decades. It is estimated that the urban area in developing countries will increase to 1.2 million km 2 by 2050, which is four times the urban area in 2000 [2]. As the largest developing country in the world, China has been experiencing unprecedented urban expansion after the implementation of the reform and opening up policy in 1978, with the increase of the urban population from 17.92% in 1978 to 58.5% in 2017, and a projected rise to 70% by 2030 [3,4]. Urban agglomerations in China

Study Area
The GBA is located in the south of China, covering a total area of approximately 55,000 km 2 , with a population of approximately 45 million people in 2018 ( Figure 1). The study area consists of 11 cities; specifically, Guangzhou, Shenzhen, Zhuhai, Foshan, Dongguan, Zhongshan, Jiangmen, Huizhou, Zhaoqing, Hong Kong and Macao. It has several typical geomorphic types, including mountains, hills, terraces and plains. The study area is located in a subtropical climate zone with an average annual temperature of 22.3 • C and an average annual rainfall of 1832 mm, and 83% of the rainfall occurs in the rainy season [33]. The study area has transformed from a predominantly agricultural region into one of the most developed urban agglomerations in the world since China's reform and opening-up policy [30,34]. As one of the regions in China with the highest degree of openness and a notable economic vitality, the GBA occupies an important strategic position in the overall national development situation.
Land 2021, 10, x FOR PEER REVIEW 3 of 18 farmland changes due to urbanization during the past 38 years; and (3) to evaluate the ESV of farmland loss under the effects of urbanization over the past four decades.

Study Area
The GBA is located in the south of China, covering a total area of approximately 55,000 km 2 , with a population of approximately 45 million people in 2018 ( Figure 1). The study area consists of 11 cities; specifically, Guangzhou, Shenzhen, Zhuhai, Foshan, Dongguan, Zhongshan, Jiangmen, Huizhou, Zhaoqing, Hong Kong and Macao. It has several typical geomorphic types, including mountains, hills, terraces and plains. The study area is located in a subtropical climate zone with an average annual temperature of 22.3 °C and an average annual rainfall of 1832 mm, and 83% of the rainfall occurs in the rainy season [33]. The study area has transformed from a predominantly agricultural region into one of the most developed urban agglomerations in the world since China's reform and opening-up policy [30,34]. As one of the regions in China with the highest degree of openness and a notable economic vitality, the GBA occupies an important strategic position in the overall national development situation.

Data Source
The land use datasets of the farmland and construction land in this study related to 1980,1990,2000,2010 and 2018 were extracted from the Land Use and Land Cover of China (CNLUCC) database [35]. The CNLUCC includes the land use and land cover (LULC) datasets for 1980,1990,1995,2000,2005,2010, 2015 and 2018. These datasets were mainly produced by visual interpretation on the basis of a series of multitemporal Landsat images, including Thematic Mapper (TM), Enhanced Thematic Mapper (ETM) and Operational Land Imager (OLI) images. As one of the most accurate remote sensing monitoring

Data Source
The land use datasets of the farmland and construction land in this study related to 1980,1990,2000,2010 and 2018 were extracted from the Land Use and Land Cover of China (CNLUCC) database [35]. The CNLUCC includes the land use and land cover (LULC) datasets for 1980,1990,1995,2000,2005,2010, 2015 and 2018. These datasets were mainly produced by visual interpretation on the basis of a series of multitemporal Landsat images, including Thematic Mapper (TM), Enhanced Thematic Mapper (ETM) and Operational Land Imager (OLI) images. As one of the most accurate remote sensing monitoring data products in China, CNLUCC has played an important role in the investigation and research of national land resources, hydrology and ecology. In this study, the classification system included two primary classes and five subclasses, and the detailed information is listed in  Table 1. The two primary classes are farmland and construction land, and both were divided into several subclasses. Farmland is classified into two subclasses, namely paddy fields and dry farmland. Construction land includes three subclasses: urban residential land, rural residential land and other construction land. In addition, specific social statistical data, including the yield and area of the main grains in Guangdong Province and China, were obtained from the Statistical Yearbook of Guangdong Province and China of 1980,1990,2000,2010 and 2018 and were downloaded from the China National Knowledge Infrastructure (CNKI) database (http://data.cnki.net/Yearbook/Navi?type=type&code=A (accessed on 20 May 2020)). Table 1. Classification of farmland and construction land in this study.

Primary Classes Subclasses Implications
Farmland This refers to land for planting crops, including cultivated land, newly opened wasteland, leisure land, rotation rest land, grassland rotation cropland, land for fruits such as mulberry, agriculture and forest areas mainly for planting crops, and beach and sea land areas cultivated for longer than three years.

Paddy fields
This refers to cultivated land with guaranteed water sources and irrigation facilities, which are commonly irrigated in normal years and used to grow aquatic crops such as rice and lotus root, including cultivated land under rotation of rice and dry farmland crops.

Dry farmland
This refers to cultivated land for dry crops without irrigation water sources and facilities, cultivated land for dry crops that contain water sources and irrigation facilities, which are commonly irrigated in normal years, the cultivated land used mainly for planting vegetables, and the leisure land and rotation land under normal rotation.
Construction land This refers to urban and rural residential areas and industrial, mining, transportation and other land areas.
Urban residential land This refers to land for large, medium and small cities and built-up areas larger than counties and towns. Rural residential land This refers to rural residential areas independent from construction land.
Other construction land This refers to land for factories, mines, large industrial areas, oil fields, salt fields, quarries, etc., and roads, airports and special land.

Analysis of the Changes in Land Use Types
To analyze the degree of changes in land use types, the annual change area (ACA, km 2 /year) was calculated as follows: where A 1 and A 2 are the area of farmland or construction land at the start and end dates, respectively, and t is the time span of the study period.

Assessment of the Ecosystem Service Value of Farmland
The ESV was divided into four classes and eleven subclasses based on previous research [24,26,36]. The four classes were provision services, regulation services, support services and cultural services. Provision services include food production, primary production and water supply. Regulation services comprise gas regulation, climate regulation, environmental purification and hydrological regulation. Support services consist of soil conservation, nutrient cycling and biodiversity conservation. Cultural services include recreational and aesthetic values. Based on the research of Costanza et al. and Xie et al. [24,26,37], the ESV of farmland in the GBA from 1980 to 2018 was quantitatively estimated with the ecosystem service value coefficient method as follows: where ESV denotes the total annual value of the ESs, and A k and V ck are the area and value coefficient, respectively, for land use type k. In this study, the ESV of paddy fields and dry farmland were calculated first, then they were summed up to get the ESV of farmland.
To better reflect the ESV changes in the different study periods, the main grain yield in the GBA compared to that in all of China was used to revise the ecosystem services equivalent value [27]. The correction coefficient β t was calculated as follows: where y t and Y t are the grain yields in Guangdong Province and China at time t, respectively. Because the data of grain yield in the GBA in the early study periods was not easy to acquire, the grain yield in the GBA was replaced by that of Guangdong Province in this study. Thus, β 1980 , β 1990 , β 2000 , β 2010 and β 2018 were 1.34, 1.24, 1.38, 1.05 and 0.99, respectively. The economic value of the ecosystem services equivalent value per unit area (Ea) was calculated as follows [8,28]: where Y and A are the total yield and area of the main grains in Guangdong Province in 2018, respectively; P is the average price of the main grains. In order to eliminate the impacts of changing values of the currency (yuan), P was given the value of 3.77 yuan/kg, which is the average price of the main grains in 2018, obtained from the Grain Net of South China (https://gdgrain.com/#!/list?params=%7B%22type%22:8%7D (accessed on 15 May 2020)).

Urbanization in the GBA from 1980 to 2018
From 1980 to 2018, the area of construction land in the GBA showed a discernible growth trend, increasing from 2607.4 km 2 to 8243.5 km 2 , a more than twofold increase ( Figure 2a). A similar trend was observed in urban residential land and other construction land, which increased nearly sixfold (from 694.7 km 2 to 4778.5 km 2 and from 285.7 km 2 to 1991.2 km 2 , respectively) ( Figure 2b). The area of rural residential land peaked in 2000 with 1855.9 km 2 and then declined. By 2018, urban residential land occupied a dominant position among construction land (58.0%), followed by other construction land (24.2%) and rural residential land (17.9%). Regionally, construction land expanded rapidly in the 11 cities of the GBA in the past 40 years (Figure 2c). Among the 11 cities, only one had an area of construction land over 500 km 2 Figure 2f). Lastly, although the urbanization and development degree of Hong Kong and Macao were much higher than that of the other cities in the GBA, their construction land expansion area and rate seemed much smaller than those of the other cities because their own administrative area was relatively small.

Changes in Farmland in the GBA from 1980 to 2018
Over the past 38 years, the total area of the farmland in the GBA varied from 16,640.1 km 2 to 12,417.4 km 2 , and more than a quarter of the total farmland area of 1980 had been lost ( Figure 3a). An overall decreasing trend was also observed in the different categories of farmland and 11 cities (Figure 3b,c). The area of paddy fields was far larger than the area of dry farmland in the GBA; correspondingly, the paddy field loss (3086.3 km 2 ) was much greater than the dry farmland loss (1136.

Changes in Farmland in the GBA from 1980 to 2018
Over the past 38 years, the total area of the farmland in the GBA varied from 16,640.1 km 2 to 12,417.4 km 2 , and more than a quarter of the total farmland area of 1980 had been lost (Figure 3a). An overall decreasing trend was also observed in the different categories of farmland and 11 cities (Figure 3b,c). The area of paddy fields was far larger than the area of dry farmland in the GBA; correspondingly, the paddy field loss (3086.3 km 2 ) was much greater than the dry farmland loss (1136.   From 1980 to 2018, the conversion from farmland to construction land was mainly distributed at the center of the GBA; e.g., Guangzhou, Dongguan, Shenzhen, Foshan and Zhongshan ( Figure 4). Over the past 38 years, 3711.3 km 2 of farmland was converted into construction land. Among the four adjacent periods, the farmland loss due to urbanization increased drastically from the first study period (479.5 km 2 ) to the third study period (1772.4 km 2 ) but then dropped sharply in the last study period (439.2 km 2 ) (Figure 5a). However, the percentage of farmland loss due to urbanization in the total farmland loss exhibited a growth trend, increasing from 39.3% to 96.6% from the first study period to the last study period (Figure 5b), which demonstrates that urbanization increasingly became the dominant reason for farmland loss. From 1980 to 2018, the urbanization-triggered losses of paddy fields and dry farmland were 2599.4 km 2 and 1038.7 km 2 , respectively ( Figure 5c). The paddy field area lost due to urban sprawl was much more considerable than that of dry farmland over the past four decades. From 2000-2010, the losses of paddy fields and dry farmland due to construction land expansion were the highest, with areas of 1183.1 km 2 and 589.3 km 2 , respectively. Over the past 38 years, 2086.7 km 2 , 708.0 km 2 and 916.5 km 2 of farmland disappeared because of the expansion of urban residential land, rural residential land and other construction land, respectively (Figure 5d). The largest farmland loss due to urban residential land expansion was 1000.5 km 2 from 2000-2010, From 1980 to 2018, the conversion from farmland to construction land was mainly distributed at the center of the GBA; e.g., Guangzhou, Dongguan, Shenzhen, Foshan and Zhongshan ( Figure 4). Over the past 38 years, 3711.3 km 2 of farmland was converted into construction land. Among the four adjacent periods, the farmland loss due to urbanization increased drastically from the first study period (479.5 km 2 ) to the third study period (1772.4 km 2 ) but then dropped sharply in the last study period (439.2 km 2 ) (Figure 5a). However, the percentage of farmland loss due to urbanization in the total farmland loss exhibited a growth trend, increasing from 39.3% to 96.6% from the first study period to the last study period (Figure 5b), which demonstrates that urbanization increasingly became the dominant reason for farmland loss. From 1980 to 2018, the urbanization-triggered losses of paddy fields and dry farmland were 2599.4 km 2 and 1038.7 km 2 , respectively ( Figure 5c). The paddy field area lost due to urban sprawl was much more considerable than that of dry farmland over the past four decades.

The Impact of Urbanization on the Ecosystem Service Value of Farmland
An unprecedented expansion of construction land has led to a great loss of farmland, which has further impacted the ESs and functions of farmland in the GBA. In the past four decades, the ESV of farmland decreased by 5036.7 million yuan due to construction land encroachment, accounting for 43.1% of the total ESV loss of farmland (Table S1). The highest ESV loss of farmland because of urbanization was 2177.5 million yuan from 2000-2010, followed by 1655.1 million yuan from 1990-2000 (Figure 6a). For paddy fields and dry farmland, the ESV losses caused by urbanization were 3442.1 million yuan and 1594.7 million yuan, respectively (Figure 6b). The largest ESV losses of both paddy fields and dry farmland appeared from 2000-2010 (1438.8 million yuan and 738.7 million yuan, respectively), and the second-largest ESV losses were from 1990-2000 (1040.0 million yuan and 615.1 million yuan, respectively). Because of urbanization, the most affected ecosystem service function was hydrological regulation with a loss of 2514.2 million yuan, followed by food production and gas regulation with losses of 1541.4 million yuan and 1248.6 million yuan, respectively (Figure 6c). Specifically, the ESV of the water supply of farmland was positive under the effects of urbanization, because paddy fields need a large amount water for the growth of aquatic plants. To some degree, the loss of paddy fields is a benefit for the water supply. Regionally, the total ESV loss of farmland caused by urbanization in Guangzhou was the largest (1110.85 million yuan), followed by that in

The Impact of Urbanization on the Ecosystem Service Value of Farmland
An unprecedented expansion of construction land has led to a great loss of farmland, which has further impacted the ESs and functions of farmland in the GBA. In the past four decades, the ESV of farmland decreased by 5036.7 million yuan due to construction land encroachment, accounting for 43.1% of the total ESV loss of farmland (Table S1). The highest ESV loss of farmland because of urbanization was 2177.5 million yuan from 2000-2010,

Farmland Changes Caused by Urbanization
The results of this study demonstrate that construction land in the GBA experienced dramatic changes in the overall extent, different types and regional extent, especially after 1990. Farmland is the land use type that contributed most to the expansion of construction land over the past four decades ( Figure S1). Because of urbanization, 3711.3 km 2 of farmland disappeared in the GBA from 1980 to 2018. The highest farmland loss caused by urbanization was observed from 2000-2010 (1772.4 km 2 ), followed by that in the period from 1990-2000 (1020.1 km 2 ), but it was small from 1980-1990 (479.5 km 2 ) and from 2010-2018 (439.2 km 2 ). Our results are different from others showing the global conversion of farmland to other land use types; globally, farmland was mainly converted to grasslands and woodlands, which accounted for 57% and 36%, respectively [38]. These changes were strongly influenced by the social development background in China and local governments' behavior [39,40]. From 1980-1990, the first ten years after the implementation of the reform and opening-up policy in 1978, social and economic development was in the initial stage, production technologies lagged behind and infrastructure was relatively imperfect. All of these reasons contributed to a slower process of urbanization, which resulted in less farmland being converted to construction land. However, after ten years of development, the economy expanded to a certain extent, and the process of urbanization and industrialization also gradually accelerated, which led to farmland being intensively encroached on by construction land. Subsequently, China entered the World Trade Organization in 2001, which indicated that China's reform and opening up had entered a new stage in history. Rapid economic and scientific technology development impelled extensive urbanization, which caused the area of farmland loss to construction land to reach a peak from 2000-2010. As construction land expansion produced a great loss of farmland, a series of environmental problems appeared which threatened farmland preservation, food security, production capacity and social stability [41,42]. In addition,

Farmland Changes Caused by Urbanization
The results of this study demonstrate that construction land in the GBA experienced dramatic changes in the overall extent, different types and regional extent, especially after 1990. Farmland is the land use type that contributed most to the expansion of construction land over the past four decades ( Figure S1). Because of urbanization, 3711.3 km 2 of farmland disappeared in the GBA from 1980 to 2018. The highest farmland loss caused by urbanization was observed from 2000-2010 (1772.4 km 2 ), followed by that in the period from 1990-2000 (1020.1 km 2 ), but it was small from 1980-1990 (479.5 km 2 ) and from 2010-2018 (439.2 km 2 ). Our results are different from others showing the global conversion of farmland to other land use types; globally, farmland was mainly converted to grasslands and woodlands, which accounted for 57% and 36%, respectively [38]. These changes were strongly influenced by the social development background in China and local governments' behavior [39,40]. From 1980-1990, the first ten years after the implementation of the reform and opening-up policy in 1978, social and economic development was in the initial stage, production technologies lagged behind and infrastructure was relatively imperfect. All of these reasons contributed to a slower process of urbanization, which resulted in less farmland being converted to construction land. However, after ten years of development, the economy expanded to a certain extent, and the process of urbanization and industrialization also gradually accelerated, which led to farmland being intensively encroached on by construction land. Subsequently, China entered the World Trade Organization in 2001, which indicated that China's reform and opening up had entered a new stage in history. Rapid economic and scientific technology development impelled extensive urbanization, which caused the area of farmland loss to construction land to reach a peak from 2000-2010. As construction land expansion produced a great loss of farmland, a series of environmental problems appeared which threatened farmland preservation, food security, production capacity and social stability [41,42]. In addition, farmland conversion to construction land could also lead to intensification of farming, and abandonment and degradation of farmland [43,44]. Governments try their best to maintain a balance between urbanization development and farmland preservation by implementing much stricter policies for farmland conversion and ecological urbanization and construction [10,14]. As early as 2005, the central land policy for the preservation of 1.8 billion mu (1.2 million km 2 ) farmland was established, which is a key environmental policy to cope with China's land transformation crisis since the 1990s [45]. In addition, to improve the urbanization quality and to strengthen the protection of farmland and natural ecosystems, the National Plan on New Urbanization and the Opinions on Accelerating the Construction of Ecological Civilization were promulgated in 2014 and 2015, respectively [39]. Under these policies, farmland loss due to urbanization sharply declined after 2010. In addition, local governments' behavior has vital impacts on the conversion of farmland to construction land [14,46,47]. Driven by particular interests, local governments converted farmland to urban land at a low compensation and leased to developers at a much higher price [48], which significantly affected farmland change under rapid urbanization [14]. To preserve agricultural land and food security, a centralized fiscal reform in 1994 induced local governments' land financing behavior to significantly influence farmland conversion [14]. Although the central government limits the total construction land quotas that local governments can lease to developers, local governments still have sufficient autonomy to determine which parcel of land to lease out [14], and they resorted to land leasing to gain extra revenue to finance urban construction and balance fiscal expenditures [49]. This extra local revenue from land leasing was approximately 33.7% of the local revenues from 2007 to 2012, which mainly came from farmland conversion [50,51]. Lastly, the urbanization in Guangzhou, Dongguan, Foshan, Shenzhen and Huizhou was faster than in other cities over the past four decades ( Figure 2); correspondingly, the area of farmland conversion to construction land was greater in these cities than in the other cities ( Figure 5). The same is observed in other developing counties, like India, where farmland conversion to construction land predominantly occurred in the districts with high rates of economic growth and higher agricultural land suitability [22].

Ecosystem Service Value Changes due to Urbanization
Extensive urbanization has resulted in the degradation of farmland, which poses a great threat to food provision security, ecological environmental protection, ESs and regional sustainability [52][53][54][55]. In the past four decades, the ESV loss of farmland caused by urbanization was 5036.7 million yuan in the GBA, and the highest loss was 2177.5 million yuan from 2000-2010, followed by 1655.1 million yuan from 1990-2000. The change trend of the ESV loss of farmland is similar to the change trend of farmland lost because of urbanization, which is because the calculation of the ESV is based on area. In addition, our results indicate that ESs such as hydrological regulation, food production and gas regulation are particularly vulnerable to urbanization impacts ( Figure 6c). As construction land expands, a large area of farmland is converted to impermeable surfaces, which causes rain and runoff to not penetrate the ground in time and participate in the natural water cycle, which further heavily affects hydrological regulation [56]. For food production, urbanization has multiple effects. Rapid urbanization in the GBA has led to an increase in the agricultural land use intensity, cropping frequency and chemical fertilizer use, a decline in the per capita availability of food grains, and soil and water pollution [10,22,57,58]. In this complex context, both food production and security are seriously influenced. The process of urbanization and industrialization inevitably causes environmental problems such as air pollution [59], which predominantly originates from industry, transport, power generation and construction [60]. Air pollution could lead to air quality deterioration and thus influence regional gas regulation [59,60].

Farmland Conservation and Ecosystem Services Protection
Governments have been striving to maintain a balance between urbanization development and farmland preservation [10,14], and the results of this study show that farmland loss due to urbanization sharply declined after 2010, but the results also indicate that construction land expansion has increasingly become the dominant reason for the loss of farmland in the GBA over the past 40 years. Therefore, it is still vital for governments to take more effective steps to regulate construction land expansion and to develop trade-offs and synergies among urban development, agricultural production and ecosystem preservation [14,39]. Based on the results of this study, we provide several recommendations for farmland preservation. First, to protect farmland and improve regional ESs, central and local governments should strictly control the area of new construction land and strengthen the construction of ecological cities and beautify countryside. In urban areas, the government could protect, plan and establish natural parks or green belts, including forests, grasses and wetlands. In rural areas, the government should promote reasonable planning for farmland use or exploitation. Governments can also 'green' abandoned industrial and mining land with forests, grasses or water bodies to improve ESs. Second, governments must strictly protect the red line of high-quality farmland and prohibit exploitation without rational reasons. Third, local governments should maintain the requisition-compensation balance of farmland, including the quantity, quality and ecological balance. For example, if the quality of compensatory farmland is lower than the quality of requisitioned farmland, then governments should invest more money to improve the quality of medium-and low-yield farmland or exploit new farmland rather than invest in urban construction [41]. Fourth, local governments should take measures to prevent environmental pollution, such as educating farmers to strictly control the use of agrochemicals and other soil additives to prevent soil pollution, and strictly prohibit the direct discharge of domestic sewage and industrial wastewater into rivers without treatment. Finally, the public should respond to government policies and protect the environment. For example, to protect the water and soil environment of farmland, enterprises and factories near farmland should strictly treat sewage according to regulations, and discharge the sewage only after it reaches the required standard. Farmers could reduce the use of pesticides and other chemicals to reduce the pollution of the farmland environment.

Limitations and Future Works
Estimates of ESV regionally and globally in monetary units play a critical role in heightening awareness and estimating the overall level of importance of ESs relative to and in combination with other contributors to sustainable human well-being [37]. Therefore, it is better for decision-makers and the public to consider the ESs as public goods or natural resources, and take these values into account when scenarios and policies are changed. However, this valuation method could not be used to examine the spatial changes in ESV. In addition, the ESV of construction land in this study was considered as zero, which is not appropriate in some degree and probably led to errors in the results. In our future work, we would like to combine the valuation methods and other evaluation methods to acquire more accurate results.
Farmland conservation and ecosystem services protection require cooperation in many aspects, including from governments, experts and the public. Therefore, it is necessary to build a system integrating geographic information system (GIS), spatial multicriteria evaluation (SME) and participatory GIS (PGIS) approaches, where decision-makers, experts and the public can participate and identify a range of ecosystem services [61][62][63][64]. Decision maker can easily collect and manage the results, which is useful for land use planning. That would be conducted in our future work.

Conclusions
This study explored the impacts of urbanization on farmland in the GBA from 1980 to 2018 based on multiple temporal land use datasets of the CNLUCC database. The ESV of farmland was also estimated by the valuation methods to study the ecosystem service changes caused by urbanization. Our results showed that the total area of farmland loss caused by urbanization was 3711.3 km 2 over the past 38 years, leading to a direct decline in total ESVs by 5036.7 million yuan. Paddy fields suffered much more losses than dry farmland because of urbanization. The expansion of construction land increasingly became the dominant reason for farmland loss in the GBA with the influence increasing from 39.3% to 96.6%. The value of hydrological regulation, food production and gas regulation showed much more decline. Guangzhou, Dongguan, Foshan, Shenzhen and Huizhou had the greatest total ESV loss of farmland caused by urbanization. The social development background in China and local governments' behavior played a vital role in the farmland conversion to construction land. To protect farmland and improve regional ESs, the central and local governments should strengthen the construction of ecological cities and beautify countryside by increasing capital investment, strengthening supervision, and raising public awareness of environmental protection, and the public should respond to land use policies and protect the environment.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/land10050501/s1, Figure S1: Conversion percentages of the different land use types into construction land in the GBA from 1980 to 2018, Table S1: ESV changes of the farmland in the GBA from 1980 to 2018 (million yuan).