1 Introduction

Land use/land cover (LULC) change is essential to global environmental change and sustainable development research. Urbanization, afforestation, reforestation programs and land development have accelerated land use and land cover change in recent decades [1]. Therefore, the study of land use change, predicting land use patterns, and identifying the dynamic characteristics of land use areas are crucial to sustainable urban development [2].

In land use planning, various land types are allotted based on their quantity and spatial location to achieve the best outcome. This research has been conducted at various scales, including global [3], national, regional, city, and landscape dimensions. Scholars have studied urban land use structure and urban sprawl from a variety of scales and perspectives, including assessing the impact of land use dynamics on the value of total ecological services under multiple scenarios [4], analyzing the simulation of multi-scenario land use change and landscape ecological risk [5], analyzing evolution characteristics and simulation prediction of landscape fragmentation [6], exploring urban growth boundaries [7, 8], and investigating urbanization drivers and mechanisms [1]. Changes in the spatial and temporal distribution of LULC substantially affect applications such as ecological security, ecological services, hydrological response, and carbon emissions, among others [9, 10]. In general, little research has been conducted on urban environment characteristics unique to urban areas.

Currently, geographic information systems (GIS) are utilized extensively in studying land use change at various spatial scales [11]. The simulation of spatial and temporal land use dynamics consists of two components: the historical characterization of land use change and the simulation of future land use. First, the land use transfer matrix allows for historical land cover change analysis [12]. Intensity analysis [13] models can be used to analyze the transfer level of each type of land, including the interval, category, and transition levels [3, 14]. Future land use demand is computed utilizing the Markov chain model [15, 16], a stochastic method that describes land use demand using a transfer probability matrix. The next step involves simulating and predicting land use. The most frequently employed models for future land use scenarios are the Cellular Automata (CA) [17], the Markov chain model [16], the CLUE-S model [4], the Future Land Use Simulation (FLUS) model [18], and the Patch Generation Land Use Simulation (PLUS) model [19]. Based on their existing studies, the PLUS model has a higher simulation accuracy [20] than other traditional models such as CLUE-S, ANN-CA, and Logistic-CA. In this paper, the Markov and PLUS models are combined and used to analyze the development and optimization of land use simulation.

Foshan, a city in the Pearl River Delta, is regarded as the entrance to China's reform and opening up and the city of the Guangdong-Hong Kong-Macao Greater Bay Area. The development of urban land use patterns affects the quality of life of its inhabitants. Foshan's urban economic development has influenced the evolution of the dike-pond, so it is essential to examine the driving mechanisms underlying the region's historical land use changes. Previous studies emphasized constructing simulation models of land use rather than analyzing or comparing the characteristics of past and future land use [2]. According to the outline of the development plan of Guangdong-Hong Kong-Macao Greater Bay Area, Foshan City will be an essential factor in the region's development. However, Foshan still faces the issue of relatively sloppy development, a heavy burden of pollution control in the air, water, and soil, and the pressing need to improve its ecological civilization system. In response, the Foshan Municipal People's Government issued the "Foshan National Ecological Civilization Construction Demonstration City Plan (2016–2025),” which details the development of a national ecological civilization construction demonstration city based on the theory of natural-economic-social composite ecosystems.

While previous studies have focused on predictive models of future land use modeling, our paper combines past and future land use patterns to explore the characteristics of land use change in prefecture-level cities, emphasizing the characteristics of "dike-pond" type transitions. Multiple scenarios are used in conjunction with the economic and policy characteristics of prefecture-level cities to simulate the future land use characteristics of cities: (i) modeling and forecasting land use changes using Markov and PLUS models to comprehend land use patterns better; (ii) revealing the stability of land use systems, identifying major land use types, and examining the characteristics of urban development; and (iii) conducting simulations of multiple scenarios of urban land use to forecast the structure and function of urban space in support of sustainable development initiatives. A comprehensive evaluation and simulation of the study can provide regional spatial planning with scientific references. It enables the establishment of a reasonable city boundary, a basic red line for the protection of farmland, and an ecological red line. In addition, the region's future development will better safeguard the territory's resources and result in coordinated and sustainable development.

2 Materials and methods

2.1 Study area

The city of Foshan is located in the south-central part of Guangdong Province, China (22°38′-23°34′ N,112°23′-113°23′ E), and is one of the most important cities in the Guangdong-Hong Kong-Macao Greater Bay Area (Fig. 1). There are five districts within Foshan, including Chancheng, Nanhai, Shunde, Gaoming, and Sanshui. Foshan is located in the humid subtropical monsoon climate zone, which has a mild climate and abundant rainfall. The average annual temperature is 22.5 °C. The coldest temperature is recorded in January with an average of 13.9 °C, and the hottest temperature is recorded in July with an average of 29.2 °C. The city has a total area of 3,792.67 square kilometers, a resident population of 9.1 million in 2020, and a level of urbanization of 98.7%. The Foshan City Master Plan (2011–2020) focuses on preserving a regionally sustainable ecological environment. The Foshan City Territorial Spatial Master Plan (2020–2035) safeguards the city's characteristic landscape environment by emphasizing the protection of the dike-pond agricultural areas and river systems characteristic of the Lingnan Water Village and the strict maintenance of the dike-pond system following reclamation and production regulations.

Fig.1
figure 1

Study area (location map)

Full size image

2.2 Data sources

This study primarily utilizes geospatial and socioeconomic data. The geospatial data includes land use, DEM, road, and river data for 2010, 2015, and 2020 (Figs. 2, 3). The land use data were categorized as cropland, forest land, grassland, water, dike-ponds, construction land, and undeveloped land. The dike-pond is listed separately from the water in this document.

Fig.2
figure 2

Foshan City’s land use in 2010, 2015, and 2020

Full size image
Fig.3
figure 3

Research framework

Full size image

We chose DEM, slope, road datasets, and various land datasets as the urban driver construction factors (Table 1). The open street map was utilized to download the road data, which included the vector data of the road network, which included six different types of data: highways, railways, national, provincial, county, and rural roads. The socioeconomic data are derived from three categories: the gross domestic product, the population, and the government.

Table 1 Data sources and descriptions
Full size table

2.3 Methods

The study is divided into two major sections to examine Foshan's past and future characteristics of land use. In the first section, characteristics of land use change from 2010 to 2020 are analyzed along three dimensions: interval, category, and transition level. Land type and area changes are analyzed based on the land use change matrix and intensity analysis, emphasizing dike-pond land. Second, the simulation of future land use requires the selection of driving factors, the modification of the simulation mechanism, and the simulation of multiple scenarios. We have proposed the land use change of the city of Foshan in 2020, based on land use in 2015 and three driving factors, namely socioeconomic conditions, transportation, and the natural environment, by combining the Markov model and PLUS model. It will be compared to 2020 Foshan land use data to verify the model's accuracy and make necessary adjustments. Three scenarios regarding land use types in Foshan in 2030 have been simulated and analyzed using a corrected PLUS model and a Markov model: Natural development (ND), Crop protection (CP), and Economic development (ED).

2.3.1 Intensity analysis

Intensity analysis is a technique for measuring the change of land categories over time [21]. The Land Use Transfer Matrix contains three levels of analysis: interval, category, and transition.

  1. 1.

    Land use transfer matrix

The Markov model can express the transfer of system states over different time intervals [22]. This model is utilized to investigate the internal flow of the land use process within the study area. We construct a land use transfer matrix to analyze the structural transformation of land use functions and reflect the quantitative changes over time. ‘i’ and ‘j’ represent the land use types at the beginning and end of the study, respectively; ‘n’ represents the number of land types; Sij represents the total area of land types that have been transformed from ‘i’ to ‘j’.

$${\mathbf{S}}_{{\varvec{i}}{\varvec{j}}}=\left|\begin{array}{ccc}{S}_{11}& \dots & {S}_{1n}\\ \dots .& \dots & \dots \\ {S}_{n1}& \dots & {S}_{nn}\end{array}\right|$$
  1. 2.

    Interval level

Interval levels characterize the intensity of change in land use over time as "fast" or "slow." St represents the average annual variation intensity over time, whereas U represents the equilibrium variation intensity over time. If St > U, the change is rapid.

$${S}_{t}=\frac{\left\{\sum_{j=1}^{J}\left[\left(\sum_{i=1}^{J}{C}_{tij}\right)-{C}_{tij}\right]\right\}/\left[\sum_{j=1}^{j}\left(\sum_{i=1}^{J}{C}_{tij}\right)\right]}{{Y}_{t+1}-{Y}_{t}}\times 100\%$$
$$U=\frac{\left\{\sum_{t=1}^{T-1}\left[\left(\sum_{j=1}^{J}{C}_{tij}\right)-{C}_{tij}\right]\right\}/\left[\sum_{j=1}^{J}\left(\sum_{i=1}^{J}{C}_{tij}\right)\right]}{{Y}_{t+1}-{Y}_{t}}\times 100\%$$
  1. 3.

    Category level

The transition level is classified as "active" or "dormant" based on the intensity of change in the area's land types. Lti indicates the average intensity of loss for land category i over time; Gti indicates the average annual intensity of increase for land category j over time. If Lti > St or Gti > St, the change is active; otherwise, it is dormant.

$${L}_{ti}=\frac{\left[\left(\sum_{j=1}^{i}{C}_{tij}\right)-{C}_{tij}\right]/\left({Y}_{t+1}-{Y}_{t}\right)}{\sum_{i=1}^{J}{C}_{tij}}$$
$${G}_{ti}=\frac{\left[\left(\sum_{j=1}^{J}{C}_{tij}\right)-{C}_{tij}\right]/\left({Y}_{t+1}-{Y}_{t}\right)}{\sum_{i=1}^{J}{C}_{tij}}$$
  1. 4.

    Transition level

The transition level measures the intensity of transfer into and out of land types, and the tendency to change land types is expressed as "tend" and "avoid." Rtin represents the average annual transfer intensity from class i to class n in the time interval; Wtn represents the average rate of change from other classes to class n; Qtmj represents the average transfer intensity from class m to class j; Vtm represents the mean rate of change from class m to the other classes.

$${R}_{tin}=\frac{{C}_{tin}/\left({Y}_{t+1}-{Y}_{t}\right)}{\sum_{i=1}^{J}{C}_{tij}}\times 100\%$$
$${W}_{tn}=\frac{\left[\left(\sum_{j=1}^{J}{C}_{tin}\right)-{C}_{tnn}\right]/\left({Y}_{t+1}-{Y}_{t}\right)}{\left({Y}_{t+1}-{Y}_{t}\right)}\times 100\%$$
$${Q}_{tmj}=\frac{{C}_{tmj}/\left({Y}_{t+1}-{Y}_{t}\right)}{\sum_{i=1}^{J}{C}_{tij}}\times 100\%$$
$${V}_{tm}=\frac{\left[\left(\sum_{j=1}^{J}{C}_{tmj}\right)-{C}_{tmm}\right]/\left({Y}_{t+1}-{Y}_{t}\right)}{\sum_{i=1}^{J}\left[\left(\sum_{j=1}^{J}{C}_{tij}\right)-{C}_{tim}\right]}\times 100\%$$

2.3.2 Land demand scale simulation

The PLUS model is an integrated model based on a traditional Cellular Automata (CA) model with modifications. The model has three main panels, including the extract land expansion, the land expansion analysis strategy (LEAS), and the CA based on multiple random seeds (CARS). This model considers the complexity of macro-level drivers of land use systems and the characteristics of micro-level pattern evolution laws to improve simulation and emulation capabilities for real-world land use conditions. The following is the formula.

$${S}_{t+1}={P}_{ij}+{S}_{i}$$

St and St+1 denote the state of the land at moment ‘t’ and ‘t + 1’, respectively, with t as the year; Pij is the state transfer probability matrix, denoting the probability of transferring from land type ‘i’ to land type ‘j’.

  1. 1.

    Driving factors

Taking Foshan's unique characteristics into account, three socioeconomic factors, six traffic conditions, and three environmental factors were chosen as driving conditions for a total of twelve (Fig. 4). Transportation factors, which account for 48.6% of construction land expansion, have the greatest impact on the growth of construction land. Urban transportation provides the primary conditions for the layout of urban functions. Natural environmental factors account for 28.7% of the total influence on construction land expansion. The land's topography, geomorphology, and slope significantly affect the site selection construction. Socioeconomic factors account for 22.4% of the expansion of construction land as the third factor. The further development of land for construction can result in a level in land utilization and output; therefore, the GDP has the most significant influence of all the factors.

  1. 2.

    PLUS-based simulation process

Fig. 4
figure 4

Driving factors

Full size image

Foshan is a highly urbanized city with intensive and frequent flows of various factors of production. In this paper, a five-year interval is used to simulate the demand scale for each land type and to select the operation parameters of the land data from 2015 and 2020 to test the accuracy of the model and predict the land status and ecological spatial development trend in the interval years (Fig. 5). Furthermore, the CA based on Multiple Random Seeds (CARS) module is utilized to simulate the model and validate its precision. As a result, the kappa coefficient for the 2020 simulation was 83.9%, the overall accuracy was 88.1%, and the FOM coefficient was 0.091, as determined by the accuracy analysis. Consequently, the PLUS model performs well in simulating land use in the city of Foshan, and the simulation accuracy is high, allowing it to be used to simulate land use data in the city of Foshan.

  1. 3.

    Multi-scenario simulation

Fig. 5
figure 5

Actual and simulated land use in Foshan City in 2020. a Actual land use in Foshan City in 2020; b Simulated land use in Foshan City in 2020; c Difference between actual and simulated area in 2020

Full size image

The PLUS model and Markov model were used to simulate the expansion scenarios of land use in 2030 within the context of natural development based on the 2020 land use simulation results. To simulate urban development realistically, the ecological control line, the basic cultivated land, the water protection area, and others were combined with random seeds in the model to determine the conversion probability (Fig. 6). To promote the coordinated development of ecological protection, cropland protection, and economical construction, it is essential to analyze the response characteristics of the ecological environment to changes in land use under various scenarios.

Fig. 6
figure 6

Policy restrictions for three scenarios

Full size image

Consequently, three land use scenarios are established, with 2030 as the target. (i) Scenario I: Natural Development scenario (ND). The natural expansion of land according to the previous land use transition method. Following the Foshan City Master Plan (2011–2020) [23], which is divided into primary and secondary ecological control areas, the ecological control line has also been added. (ii) Scenario II: Cropland Protection scenario (CP). During the development process, the objective is to maximize the benefits of cropland protection, increase cropland protection, and minimize the adverse effects of urban expansion on the ecosystem. According to the Protection of Basic Farmland [24] regulations and other pertinent laws and regulations, the concentrated areas of basic cropland are rigorously protected. In this scenario, changes in land use are influenced more by policy factors. Basic cropland is chosen as a constraint to prevent encroachment on regional cropland and improve regional ecological environments' quality. (iii) Scenario III: Economic Development scenario (ED). In the context of rapid urbanization, the urban master plan of Foshan City (2020–2035) [25] takes urban–rural integration as the development goal. In this scenario, the developable area is constrained by construction land. As a result of this scenario, policies and market interventions have less impact and tend to adhere more closely to urbanization laws. Priority is given to economic benefits to achieve rapid economic growth in the region.

3 Current characteristics of land Use

3.1 Temporal changes in land use

Figure 7 depicts the percentage of land area that changed in each time interval, the annual intensity of change (St) of land types in each time interval, and the uniform intensity of change (U) for each time interval, as indicated by the dashed line on the right. From the left, the area of interval change for both time intervals is the same, 16.2%. The intensity of change in Foshan city varies over time on the right side. Over the period 2010 to 2015, the area (St) has changed annually by 4.27%, which is greater than U (3.24%), indicating that the land type has changed rapidly. During this period, Foshan experienced rapid urbanization and industrialization due to China's reform and opening. The annual changed area (St) for 2015–2020 is 2.21%, less than U (3.24%); therefore, the intensity of land use change is relatively low. From 2010 to 2020, the overall trend in land use intensity is first rapid, then gradual.

Fig. 7
figure 7

Interval level intensity analysis

Full size image

3.2 Land use area and type change

The figure depicts the intensity analysis for the level of land category. The area on the left represents the average annual increase and decrease of area for each category, the area on the right represents the average annual gain intensity (Gtj) and loss intensity (Lti) for each category, and the red dashed line represents the uniform intensity (U) for each category during that period.

At the category level, cropland gain intensity (Gt = 7.34%) is greater than the average intensity of change (U = 4.27%), indicating that the gain intensity of cropland was active from 2010 to 2015. The loss intensity (Lt) of dike-pond (13.28%), grassland (7.26%), and unused land (4.89%) were all greater than uniform intensity (U), indicating that the loss intensity was active for these three types of land. From 2015 to 2020, the gain intensity (Gt) values of grassland, water, dike-pond, and construction land were 13.07%, 3.73%, 2.99%, and 2.76%, respectively, all exceeding the uniform intensity (U = 2.21%). The loss intensity (Lt) values for cropland, grassland, water, dike-pond, and unused land were 2.41%, 4.25%, 2.51%, 3.69%, and 2.31%, respectively, all of which exceeded the average change intensity and are currently active. Significantly, the intensity of dike-pond reduction was active in both periods, reaching its peak in both periods from 2010 to 2015 and from 2015 to 2020, with annual reduction values of 1323.39 ha and 585 ha, respectively (Fig. 8). Therefore, this study develops the following research on dike-pond transition levels.

Fig. 8
figure 8

The intensity changes of the category level in Foshan from 2010 to 2020

Full size image

3.3 Transition level of the dike-pond

The dike-pond is a traditional agricultural landscape in the Pearl River Delta region, and a typical ecological circular agriculture model is proposed. This study examined the transition levels of "avoid" and "target" conversion data of the dike-pond in the background of Foshan City. During the two intensity periods, 2010–2015 and 2015–2020, cropland, grassland, and unused land were converted to dike-ponds at an intensity that exceeded the uniform rate (0.3 and 0.24). "Cropland—Dike-pond" transitions had an annual area of 278 ha and 295 ha, respectively (Fig. 9). There are numerous reasons for converting cropland to dike-pond farming, including the higher economic benefits compared to traditional agricultural farming. Common in the Foshan districts of Nanhai and Shunde is the transformation of waste fields into dike-ponds. Despite this, traditional "dike-pond" agriculture, which primarily cultivates mulberries, sugar cane, and fruits, is gradually losing its competitiveness as the market economy expands. In addition, some dike-ponds are being converted into water farms as the cyclic pond agriculture system evolved into a freshwater fish pond farming agriculture system.

Fig. 9
figure 9

Transition level of the dike-pond from 2010 to 2020

Full size image

There is an additional significant direction in the conversion of dike-ponds. In the target area, dike-ponds are being converted into construction land. The average conversion of “dike-pond—construction land” during these two periods was 3,435 ha and 375 ha, respectively. As a result of urbanization, dike-ponds were converted into construction lands. The secondary sector of industry and the tertiary sector of services gradually replaced the primary sector of agriculture. These two periods are officially considered "rapid urbanization," with uncontrolled growth of construction land and incremental urbanization.

3.4 Transfer matrix of intensity change

A land use transfer matrix was utilized to compute the intensity changes for two time periods in Foshan, including quantity, exchange, and shift for seven different types of land (Table 2). From 2010 to 2015, the intensity change in the number of dike-ponds was the highest of all land use types, at 82.69%, followed by cropland (48.82%) and construction land (41.50%). The greatest exchange occurs between forest land and water area (86.12% and 87.15%, respectively). During 2015–2020, cropland, water area, and construction land increased by 336,213 ha, 2,705 ha, and 156,164 ha, respectively, with grassland experiencing the greatest change intensity at 74.96%. Generally, the exchange intensity is high across all land types, indicating that land use has changed substantially during this period. The grassland, water, and construction land area increased by 13,096 ha, 13,096 ha and 70,452 ha, respectively. It is more evident that forest land and dike-ponds continued to decline during both periods due to the destruction of ecological land caused by urban economic development and the gradual decline of fishing and farming. On the other hand, the demand for construction land and water grew due to urbanization and flooding.

Table 2 Intensity of change across categories in two periods in Foshan
Full size table

As a result of urban expansion, the spatial pattern of land use is evolving into a "core–edge" differentiation pattern. The districts of Nanhai and Chancheng, which are in the corridor between Guangzhou and Foshan City, are well-connected and conveniently located. As a result, they are the central areas where land use is changing the most rapidly. Generally, peripheral areas with slower land use change rates are relatively remote. They have few economic connections to the surrounding areas, such as the southwest of the Gaoming district, the east of the Shunde district, and the south of the Nanhai district (Fig. 10).

Fig.10
figure 10

Foshan City’s land use exchange in 2010–2015 and 2015–2020

Full size image

4 Characteristics of future land use simulation

4.1 Scenario comparison analysis

The three spatial planning scenarios for Foshan in 2030 differ significantly from one another (Table 3, Fig.11). According to the natural development scenario, the principal growth lands from 2020 to 2030 are construction land and forest land, with change rates of "exchange" of 64.46% and 53.98%, respectively. The principal decreased lands are cropland, water, and dike-pond, with change rates of "shift" of 8.85%, 22.2%, and 7.89%, respectively. The ecological land in the urban core grows and develops naturally, whereas the ecological land on the periphery shrinks. In this scenario, the construction land exhibits the characteristics of intensive use, which is consistent with Foshan's current policy development. Following the implementation of the land-spatial planning system, the urban boundary has been constrained and cannot be expanded unabatedly, resulting in a tightening development trend within the city. As a result of this arrangement, the erosion of the construction land on the ecological land in the city's central area can be compensated by the ecological land located on the city's periphery.

Table 3 Intensity of change of various land types for three simulated scenarios in Foshan
Full size table

In cropland protection, the cropland has continuously developed and been effectively protected. The construction of an ecological civilization has been accomplished through cropland restoration and compensation measures. As a result, the cropland is at the highest exchange level (82.35%), while the forest land is at the second transition level (63.94%) under the quantity label of “gain”. The protection of basic farmland primarily inhibits construction land growth. In this scenario, the city shifts from incremental to decremental development.

In the economic growth scenario, most occupied land consists of cropland and forest land, with 5,446.27 ha and 3,582.42 ha, respectively. In this scenario, urban land use is marked by the disorderly expansion of construction land and the fragmented development of ecological lands. On the other hand, the forest land has been preserved due to the urban development boundary, resulting in a "shift" ratio of zero.

4.2 Time interval level of multiple scenarios

In the 2030 simulation scenario (Fig. 12), the annual change area of the CP scenarios is less than the uniform intensity, which has a slower rate of development than the 2020–2030 intensity. According to the data, the annual change percent of the interval's domain is 1.16%, 1.13%, and 1.17% for ND, CP, and ED scenarios, respectively, arranged as ED > ND > CP. According to the city's development pattern, the city transitions from a period of rapid growth to a period of stock development and then to a period of reduced development. The results of the three scenario simulations are, therefore, plausible.

Fig.11
figure 11

Simulated land use in Foshan City in 2030 for multiple scenarios. a Land use map for the Natural Development Scenario (ND) 2030; b Land use map for the Cropland Conservation Scenario (CP) 2030; c Land use map for the 2030 Economic Development Scenario (ED); d Change area under ND scenario in 2020-2030; e Change area under CP scenario in 2020-2030; f Change area under ED scenario in 2020-2030.

Full size image
Fig.12
figure 12

Foshan City’s land use in 2030 for multiple scenarios

Full size image

4.3 Category level of simulated scenarios

In the natural development scenario that continues the transformation mechanism from 2015 to 2020, grassland transformation intensity is the greatest and most dynamic of all land types. Among the increasing land types, grassland, water, dike-pond, and construction land accounted for 1.75%, 1.58%, 1.60%, and 1.26% of the gain intensity percent, respectively, with construction land having the highest uniform annual gain of 1,459.66 ha. Grassland was the most active in the decrease in intensity at 7.01%. Even though forest land is generally transitioning to a dormant intensity, the total amount of forest land is growing. In 2030, Foshan City will continue to prioritize the conservation of ecological land (forest land) and the moderate expansion of construction land based on the historical development scenario.

Under the cropland protection scenario, the increased intensity of both cropland and dike-pond is active, at 1.2% and 1.51%, respectively, while the uniform annual gain numbers of cropland and construction land are the highest, at 1,405 ha and 1,393 ha, respectively. The decreased intensity (0.92 and 0.45 for cropland and forest land, respectively) is less than the uniform intensity. In this scenario, cropland and land are both generally strictly protected. Due to land-class transfer restrictions, construction land cannot encroach on cropland, reducing cropland development. Therefore, when the government creates regulations to manage the conversion between the two types of land, it must develop policies to precisely manage and balance the development of these two types of land.

In the economic development scenario, the construction land annual gain rate of 1,505 ha is the highest of all land types. Its gain intensity is also relatively active, at 1.42%, compared to 1.18% for the uniform intensity. In this scenario, cropland is inactive during gain intensity and active during loss intensity, indicating that cropland is continuously decreasing. Due to population growth and economic development, construction land encroaches on rural cropland.

According to the analysis of simulation results, the natural development scenario is more compatible with urban development: during this period, the urbanization process will slow, urban construction land will be developed intensively, and ecological land, including forest lands, dike-ponds, and water areas, will also be protected. However, the expansion of land for construction has affected the integrity and stability of ecological land. Therefore, the government should consider a sustainable development model with multiple objectives in the urbanization process.

Notable is that in the three simulation scenarios, the intensity of the transformation of the gain and loss of the dike-pond is active in each land type, launching a comprehensive analysis of the dike-pond land type.

4.4 Transition level of the dike-pond at three simulated scenarios

In the three simulation scenarios, the stable land target system conversion patterns are “cropland – dike-pond”, “water – dike-pond.” The stable land avoid system conversion patterns are “forest land – dike-pond” and “construction Land – dike-pond.” Under the influence of the regional culture of Foshan, the conversion of cropland and water into dike-ponds, as well as the operating systems of mulberry-based and fruit-based fish ponds, are likely to provide substantial benefits to farmers, farm owners, and dike-pond owners. In addition, tourists will be more familiar with a model of tourism other than the traditional one, especially as suburban tourism continues to proliferate in the future. In the future, however, the dike-pond will inevitably be encroached upon by rapid urban development. Figures 13, 14 illustrates that the dike-pond–construction Land transition is stable under the economic development and cropland protection scenarios. In both scenarios, the transition intensity percentage is greater than the uniform intensity percentage, at 22.1%.

Fig.13
figure 13

Category level of three simulation scenarios from 2020 to 2030

Full size image

5 Discussion

5.1 Urban development and ecological environment

The results of simulations indicate that urbanization has significantly altered land use patterns. For rapid urban sprawl to continue, three significant obstacles must be addressed: the loss of high-quality farmland, the development of urban hollowing out, and the ecological consequences. However, the stability of land use change from 2015 to 2020 has significantly improved compared to the stability of land use change from 2010 to 2015. Compared to the stability of land use change from 2015 to 2020, the stability of land use change from 2020 to 2030 demonstrates a downward trend. Therefore, while pursuing urban development, the government should continue to protect the ecological environment. Although the urban area is limited, urban planners are encouraged to consider re-using abandoned urban land and rebuilding old cities to increase land use efficiency instead of acquiring cropland to develop urban boundaries. Reduced cropland is likely to have a significant impact on food security. Therefore, the government must consider sustainable development with multiple objectives as part of the urbanization process.

5.2 Cropland conservation and efficient use

Throughout three periods, cropland has decreased steadily due primarily to urbanization. Between 2010–2015 and 2015–2020, cropland increased at an accelerated rate. According to the findings, cropland plays a significant role in the change in land use. Consequently, cropland must be conserved and utilized efficiently. The government's implementation of policies to conserve cropland and improve land use efficiency has contributed to cropland's prominence in the land use system. As a result of agricultural intensification, the cropland area has decreased, whereas cropland production has increased. It does not mean, however, that we should occupy cropland indefinitely. Notably, this does not imply that we should continue to occupy agricultural land. Additionally, it is necessary to strengthen cropland protection. It is a large-scale project requiring considerable time and substantial government investment to restore the cultivated layer and crop production capacity. To maximize the benefits of agricultural scale, unused and low-utility land should be evaluated and incorporated into the existing cropland area as much as possible. In the general land use plan, cropland should be safeguarded, and its area should increase gradually.

5.3 Policy trade-offs between conservation and development

Future scenario simulations indicate a decline in dike-pond, water, and grassland, as well as fragmentation of the ecological landscape, because of human activity. As a local government, it is necessary to implement new ecological regulations to safeguard existing ecological resources continuously. Specifically, the dike-pond, a symbol of Foshan City's local culture, is a unique landscape (Fig. 15). According to the Foshan Historical and Cultural City Protection Plan (2020–2035) [26], the historical landscape features of "mountain-water-dike-pond-city-village" and the agricultural and cultural heritage of Foshan dike-pond are protected.

Fig.14
figure 14

Transition level of the dike-pond at three simulated scenarios from 2020 to 2030

Full size image
Fig.15
figure 15

Dike-pond landscape. (Photo by the authors)

Full size image

At the local level, there has always been a conflict between environmental protection and economic development, as well as the need to implement cropland protection and scientifically control the rapid expansion of cities. In the future simulation scenario, it is anticipated that urban expansion will continue to increase while farmland will continue to decrease. The loss of farmland due to urbanization has harmed the original agricultural ecosystem. Nonetheless, cropland is the primary assurance of food security. Therefore, urgent protection of cropland is required. Countering the rapid growth of cities is one of the most effective methods for protecting cropland. The process of urban economic development must prioritize quality over quantity. In the future, local governments should intensify land use and reduce unscientific urban sprawl. Decision-makers must pay close attention to the inefficient utilization of urban and abandoned areas.

6 Conclusions

This study analyzes and predicts the characteristics of LULC change patterns in the area using land use data from the preceding ten years. Using a land use transfer matrix and an intensity analysis model, we analyze the characteristics of LULC change from 2010 to 2020 in Foshan City. To analyze land use change in 2030, the Markov and PLUS models simulate the future scenario of land use in Foshan. Assessing the stability and intensity of land use in cities at the prefecture level in comparison to current and future scenarios, particularly for dike-pond land, is part of innovation research. In addition, we investigate the characteristics of the transition mechanism between construction land and cropland to provide references for sustainable development and scientific and reasonable land use planning in Foshan City.

  1. 1.

    According to interval level intensity analysis, the overall trend of land use intensity between 2010 and 2020 is first rapid and then gradual. At the category level, the construction land is consistently active with increasing intensity, whereas the dike-ponds is consistently active with decreasing intensity. In the transition level, cropland and unused land are transferred to the dike-pond, while construction land is transferred.

  2. 2.

    According to the results of future multi-scenario simulations, the intensity of site conversion between 2020 and 2030 will exhibit a slow trend at the time interval level. At the category level, there are disparities between the three scenarios. In the natural development scenario, ecological lands, such as forest land and water areas, will continue to be developed following historical urban patterns and will be protected to the greatest extent possible. Except in cropland protection, urban land use will further compress cropland space. In the economic development scenario, construction land is in a state of increasing intensity. The natural development scenario aligns more closely with the city's development.

  3. 3.

    The dike-pond demonstrates a continuous decreasing trend in all future urban scenario simulations. The target systems of land use conversion at the future transition level will be "forestland—dike-pond", "water—dike-pond", and "dike-pond—construction land". Government, the market, and farmers benefit from the conversion of dike-pond, construction land, and cropland. It is unacceptable for the government to encroach on cropland in a disorderly manner to advance economic interests and destroy the dike-pond landscape and the city's historical and cultural characteristics.