Land Use Change and Its Impact on Ecological Risk in the Huaihe River Eco-Economic Belt

Abstract

Exploring the landscape ecological security pattern and its driving mechanisms in key economic zones is of great significance for preventing and resolving landscape ecological risks and promoting regional sustainable development. This study quantitatively analyzed the land use change characteristics in the Huaihe River Eco-economic Belt from 1980 to 2020 using the land use transfer matrix and land use intensity index. Further, the evolution of ecological risks and their driving mechanisms were investigated using the landscape pattern index and hierarchical partitioning analysis. The results show that (1) in terms of absolute area, dryland, grassland, and paddy land decreased by 7075 km², 2708 km², and 1874 km², respectively, while urban–rural land and water area increased by 10,538 km² and 1336 km², respectively. In terms of change intensity, grassland, water area, urban–rural land, and unused land exhibited the most dramatic change, whereas forest land, paddy land, and dryland exhibited weaker change. (2) The conversions in the study area were primarily between dryland, paddy land, and urban–rural land. Paddy land and dryland tended to convert to urban–rural land, which is further likely to be transformed into dryland and unused land when converted. (3) The study area mainly presented medium to low ecological risk. Overall, the ecological risk remained stable throughout the study period. Nevertheless, Jining, Zaozhuang, and Bengbu show high ecological risks in the construction of the economic zone. (4) Forest land explained 40% of the variance in landscape risk, whereas urban–rural land and dryland each explained 20% of the variance. An increase in the proportion of urban–rural land and dryland will increase landscape ecological risk. However, after urban–rural land exceeds 15%, the ecological security risk does not increase significantly with increasing proportion of urban–rural land.

1. Introduction

Land use/cover change (LUCC) is an important component of the land system, involving various aspects, such as social, economic, and ecological environments, and directly reflects the interaction between human society and natural environment [1]. Human interference with the ecosystem leads to the shrinkage of vegetation cover, fragmentation of habitats, and a sharp decline in biodiversity [2,3], which have serious impacts on regional landscape ecological security [4,5,6,7]. Ecological security, as an important component of national security, is an important guarantee for the sustainable and healthy development of the economy and society. Effective prevention and resolution of environmental risks and achieving harmonious coexistence between humans and nature are crucial tasks in the construction of ecological civilization [8].

Landscape is a complex region composed of different land use/cover types [9]. Changes in landscape pattern refer to spatial variations in patch/landscape size, shape, aggregation, etc., which are a concrete reflection of spatial heterogeneity [10]. Land use change is closely related to the ecological environment and has direct or indirect impacts on the regional landscape pattern and quality of the ecological environment, which in turn have led to various ecological security risks [11]. Hasan et al. [12] examined the impacts of land use change on ecosystem services in a global context, and revealed the negative impact of land use change on ecosystem services. Foley et al. [13] studied the global impacts of land use on various aspects, such as climate change, biodiversity loss, water resources, and food production, and highlighted the need for sustainable land management practices to mitigate the negative impacts on ecosystems. Similarly, Opdam et al. [14] emphasized the importance of maintaining ecological functionality through land use planning and management to ensure the sustainable provision of ecosystem services. Furthermore, Liu et al. [15] investigated the socio-economic drivers of forest loss and fragmentation, highlighting that policy driver is the most influential one and that economic driver also has a strong effect on forest loss and fragmentation in the urban planning zone.

Regarding the evaluation of ecological security, the main methods include the pressure-state-response (PSR) and its derivative models [16], minimum cumulative resistance (MCR) model [17], and the landscape index evaluation method [18]. In the landscape index evaluation method, landscape ecological security is quantified by constructing evaluation models using certain landscape indices, such as fragmentation, isolation, dominance, and type of landscape vulnerability [19,20]. This method emphasizes the role of landscape spatial structure in influencing ecological risk processes, enabling a comprehensive understanding of the quality of the ecological environment and its potential risks [21,22]. In the evaluation of landscape ecological risk, the effects of landscape spatial structure on ecological risk processes are emphasized, highlighting the spatiotemporal heterogeneity and scale effect of the evaluation and comprehensively reflecting ecological environmental problems caused by natural and anthropogenic interactions [23,24]. Previous efforts have already overcome the limitation of characterizing regional risk status using a specific natural risk factor, making the evaluation of landscape ecological risk the most promising method for ecological risk evaluation [25], and it has achieved good results at different spatial scales of areas, such as river basins [26], administrative regions [27], cities [28], and economical belts [29].

Ecological risk analysis based on landscape pattern provides important information on the nature and possible consequences of the development and evolution of landscape ecological security [30]. Landscape pattern indices have been widely used in previous studies on landscape safety risks [31,32], but the driving mechanism from the perspective of land use has rarely been studied. To fill this gap, this study selects the Huaihe River Eco-economic Belt as a case study, constructing an ecological risk model based on landscape pattern indices and employing hierarchical partitioning analysis (HPA) to examine the contributions of different land use types to landscape risk. In doing so, this research not only builds upon existing studies but also brings novelty by highlighting the influence of land use on landscape pattern and risk. Moreover, the study’s focus on a specific geographical context provides transferable knowledge applicable to similar regions globally, contributing to the broader field of landscape ecology and supporting the sustainable management of landscapes worldwide.

2. Study Area, Data, and Methods

2.1. Study Area

The Huaihe Eco-economic Belt, as a national development strategy, covers an area of 243,000 km². It mainly includes seven prefecture-level cities, such as Hai’an and Yancheng in Jiangsu Province, eight prefecture-level cities, such as Bengbu and Huainan in Anhui Province, seven prefecture-level cities, such as Xinyang and Zhumadian in Henan Province, four prefecture-level cities, such as Zaozhuang and Jining in Shandong Province, and Suizhou, Xiaogan, and Guangshui in Hubei Province (Figure 1). The study area is located in a transitional zone between subtropical monsoon climate and temperate monsoon climate in China, with a warm temperate zone to the north of the Huaihe River and a subtropical zone to the south. The Huaihe Eco-economic Belt has an average annual precipitation of 992.88 mm and an average annual temperature of 13.7 °C. With its suitable water–heat combination, rich mineral resources, convenient water transportation, and superior geographical location, the study area is one of the most densely populated areas in terms of population distribution, urban distribution, and industrial and agricultural distribution in China. It is also one of China’s traditional food production bases, with abundant mineral resources, as well as a large number of fishery and agricultural resources, providing a unique development environment for industries, such as food and textiles. However, the urban construction, socio-economic development, and unbridled resource exploitation frequently cause ecological problems, attracting significant social attention.

2.2. Data

To obtain land use data for the study area, five time periods (1980, 1990, 2000, 2010, and 2020) were selected, and data with a spatial resolution of 1 km × 1 km were obtained from the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences (https://www.resdc.cn/, accessed on 20 May 2021). Based on the characteristics of the Huaihe River Eco-economic Belt, seven land use types were considered and classified as follows: forest land, grassland, water area, urban–rural land, unused land, paddy land, and dryland (Figure 1). On this basis, the economic belt was sampled using an equidistant grid with a square size of 15 km × 15 km, generating a total of 1320 risk zones via thr ArcGIS software.

2.3. Methods

2.3.1. Intensity of Land Use Change

The land use intensity index defined in Aldwaik et al. [33] was used to analyze the land use intensity in the Huaihe River Eco-economic Belt. The intensity analysis can quantitatively express the degree of change in various land use types during different time periods, including interval level, categorical level, and transition level. The specific calculation method can be found in the literature [34,35], and a brief description is provided in this paper.

Interval level analysis: It analyzes the overall rate of change in different land cover types during different time intervals. S_t defines the annual percentage change of each land cover type during each time interval, and U defines the overall change that occurs across all intervals from the first to the last time interval, defining the annual average change rate. If S_t = U, different land cover types change uniformly over time, i.e., the land cover change has absolute stability on a time scale. If S_t > U, the land cover type changes rapidly during that period. If S_t < U, the land cover type changes slowly.

$$S_t=\frac{[Y_t,Y_{t+1}]\mathrm{area} \mathrm{change}/\mathrm{area} \mathrm{of} \mathrm{study} \mathrm{area}}{duration of \mathrm{interval} [Y_t,Y_{t+1}]}\times 100$$

(1)

$$U=\frac{\mathrm{area} \mathrm{change} \mathrm{during} \mathrm{all} \mathrm{intervals}/\mathrm{area} \mathrm{of} \mathrm{study} \mathrm{region}}{Y_{t+1}-Y_t}\times 100$$

(2)

Categorical level analysis: It explores whether the intensity of change in a certain land use type during a certain period is relatively active or stable compared to the annual average change intensity. G_tj and L_ti represent the intensity of annual increase and annual loss for a category, respectively. S_t represents the uniform line of each time interval at the intensity level of the category. If G_tj/L_ti is less than S_t, then the change in that category is relatively stable during that time interval. If G_tj/L_ti is greater than S_t, the change in that category is relatively active.

$$G_{tj}=\frac{gross gain area of category j during [Y_t,Y_{t+1}]/duration [Y_t,Y_{t+1}] }{area of category j at Y_{t+1}}\times 100$$

(3)

$$L_{ti}=\frac{gross loss area of category i during [Y_t,Y_{t+1}]/duration [Y_t,Y_{t+1}] }{area of category i at Y_t}\times 100$$

(4)

Transition level analysis: The intensity analysis of the transition level reflects mutual conversion between different land cover types during a specific time interval. If R_tin > W_tn, land cover type n tends to expand through conversion from land cover type i, and vice versa. If Q_tmj > V_tm, land cover type m tends to convert to land cover type j, and vice versa.

$$R_{\mathit{tin}}=\frac{transition area from i to n during [Y_t,Y_{t+1}]/duration [Y_t,Y_{t+1}]}{area of category i at time Y_t}\times 100$$

(5)

$$W_{tn}=\frac{gross gain area of category n during [Y_t,Y_{t+1}]/duration [Y_t,Y_{t+1}]}{area that is not category n at time Y_t}\times 100$$

(6)

$$Q_{\mathit{tmj}}=\frac{transition area from m to j during [Y_t,Y_{t+1}]/duration [Y_t,Y_{t+1}]}{area of category j at time Y_{t+1}}\times 100$$

(7)

$$V_{tm}=\frac{gross loss area of category m during [Y_t,Y_{t+1}]/duration [Y_t,Y_{t+1}] }{area that is not category n at time Y_{t+1}}\times 100$$

(8)

2.3.2. Landscape Ecological Risk

The ecological risk in a regional ecosystem is influenced by both external and internal factors [36]. External factors encompass the disturbances imposed on the regional ecosystem, while internal factors denote the ecosystem’s resilience to withstand external disturbances. The impact of external factors is mediated by internal factors. From a landscape ecology perspective, the landscape disturbance index can be utilized to characterize external factors, with a higher index indicating a greater degree of external disturbances on the regional ecosystem. Conversely, the landscape fragility index can represent internal factors, where a higher index value signifies a lower capacity of the regional ecosystem to resist external disturbances [37,38]. Building upon existing research findings [39,40], the level of ecological risk can be expressed as a function of the landscape disturbance index and landscape fragility for different landscape types at the landscape level. The formula used for calculation is as follows:

$$ERI={\displaystyle {\sum }_{i=1}^N\frac{S_{ki}}{S_k}}\times L{L}_i$$

(9)

The comprehensive ecological risk index (ERI) is calculated by taking into account the total area (S_k) of the Kth risk area, the area of the ith landscape component (S_ki), and the degree of landscape loss (LL_i), which is determined by the landscape disturbance index (U_i) and the fragility index (S_i) of each landscape type. N represents the number of landscape types. The calculation formula is as follows:

$$L{L}_i=U_i\times S_i$$

(10)

The calculation of the landscape disturbance index is performed using the following formula:

$$U_i=a{C}_i+b{F}_i+c{D}_i$$

(11)

where, C_i, D_i, and F_i are the landscape fragmentation index, dominance index, and isolation index, respectively.

Table 1. Calculation of landscape index.

Index	Formula	Meaning
Fragmentation index (Ci)	$C_i=\frac{n_i}{A_i}$	The Ci fragmentation index measures the degree of fragmentation of a given landscape type under specific environmental conditions and over a defined period of time. Generally, a higher Ci value indicates a lower landscape stability.
Divisibility index (Fi)	$F_i=\frac{A}{2A_i}\sqrt{\frac{n_i}{A}}$	The divisibility index is an index used to describe the degree of separation between landscape types, and it is typically used to measure landscape diversity and spatial heterogeneity. A higher divisibility index indicates a greater degree of separation between landscape types, and correspondingly, higher landscape diversity and heterogeneity.
Dominance index (Di)	$D_i=\frac{M_i+L_i}{4}+\frac{P_i}{2}$	The dominance index is a measure of the proportion of a specific type in the landscape. In ecology and landscape ecology, this index is commonly used to describe the importance and influence of a species or vegetation type in an ecosystem. The larger the dominance index, the higher the proportion of that type in the landscape and the greater its impact on the ecosystem.
Fragility index (Si)	scored by experts and normalized	The landscape fragility index serves as a comprehensive measure of the ability of different landscape types to resist external disturbances and the ease with which they deviate from a stable state when exposed to external risks. A higher fragility index indicates a poorer capacity of a landscape type to resist external disturbances, resulting in a higher ecological risk. Drawing on previous research and taking into account the actual situation of the study area, the following values are assigned: 7 for unused land, 6 for water area, 5 for paddy land, 4 for dryland, 3 for grassland, 2 for forest land, and 1 for urban–rural land. These values are then normalized to obtain the fragility index.

Abbreviations used in the calculation of the fragmentation index are as follows: n_i for the number of patches of a landscape type, A_i for the total area of a landscape type, A for the total area of the landscape, M_i for the number of sample units in which a patch occurs divided by the total number of sample units, L_i for the number of patches divided by the total number of patches, and P_i for the area of a patch divided by the total area.

presents the formulas used for calculation and their meanings, with values of 0.5, 0.3, and 0.2 assigned to a, b, and c, respectively, based on both existing research results and the specific conditions of the Huaihe Eco-economic Belt.

2.3.3. Driving Mechanism for Landscape Ecological Risk

The relative importance of land use types on landscape ecological risk was analyzed using HPA, which is a statistical tool for analyzing the contribution of each variable on the model’s goodness of fit [41]. It determines the contribution variance explained independently and jointly by each variable by fitting a model with all possible combinations of n predictor variables (i.e., models (1), (2), …, (n), (1,2), …, (1, n), …, (1,2,3, …, n)) and computing the R² for each combination. The joint variance of these variables is then averaged, and the independent variance explained by each variable is obtained as the sum of the marginal variance and the joint variance [42]. HPA can help reduce multicollinearity by determining the independent contributions of each explanatory variable to the response variable, allowing for the sorting of the importance of covariates when explaining the response variable. HPA can be implemented in the R package “hier.part” [43], and more details are available in the literature [44].

3. Results

3.1. Changes in Land Use Transfer Matrix

The main land use types in the Huaihe Eco-economic Belt are dryland and paddy land, which accounted for approximately 67% of the entire study area in 2020, followed by urban–rural land and forest land, which accounted for 15% and 9%, respectively. From 1980 to 2020, dryland, grassland, and paddy land decreased by 7075 km², 2708 km², and 1874 km², respectively (

Table 2. Land use transfer matrix (in km² except for change rates) from 1980 to 2020.

1980/2020	Forest Land	Grassland	Water	Urban–Rural Land	Unused Land	Paddy Land	Dryland
Forest land	16,294	1849	315	263	7	1925	2904
Grassland	1539	3885	378	249	22	414	1589
Water area	414	639	7839	1093	73	2133	2758
Urban–rural land	596	589	995	7787	39	7896	22,336
Unused land	15	36	9	8	18	12	67
Paddy land	2006	429	1683	4698	42	34,883	6068
Dryland	2767	3357	2394	15,602	125	4420	95,968
Total in 1980	23,631	10,784	13,613	29,700	326	51,683	131,690
Total in 2020	23,557	8076	14,949	40,238	165	49,809	124,633
Absolute change rate	−74	−2708	1336	10,538	−161	−1874	−7057
Rate of change (%)	0	−34	9	26	−98	−4	−6

), and unused land and grassland decreased by 98% and 34%, respectively. Urbanization has led to a significant increase in urban–rural land, which has increased by 10,538 km² over the past 40 years. It is worth noting that under the background of global change, the water area has been significantly improved, increasing from 13,613 km² in 1980 to 14,949 km². The Sankey diagram of land use change (Figure 2) shows that land use change in the Huaihe River Eco-economic Belt is primarily characterized by conversions between dryland, paddy land, and urban–rural land, particularly the conversion from dryland to urban–rural land.

3.2. Intensity of Land Use Change

The relationship between S_t and u shows that S_t was greater than u during the periods of 1990–2000, 2000–2010, and 2010–2020, indicating a rapid change in land use since 1990 (Figure 3). In terms of administrative divisions, Linyi, Jining, and Zaozhuang in Shandong Province; Xuzhou, Suqian, and Huai’an in Jiangsu Province and Xinyang in Henan Province are experiencing rapid changes in land use. In contrast, land use change is relatively slow in Luohe, Zhoukou, and Zhumadian in Henan Province; Bozhou in Anhui Province and Yancheng in Jiangsu Province (Figure 3a). Overall, similar changes were observed across different time periods, and it is worth noting that Zaozhuang and Xuzhou have been in a state of rapid change throughout the period.

At the category level, regardless of the entire period or a specific period, the G_tj and L_ti of forest land, paddy land, and dryland were smaller than St, indicating that although the absolute area of these three land categories changed significantly, their relative changes were less active (Table S1). The G_tj and L_ti of grassland, water area, urban–rural land, and unused land were greater than S_t, indicating that their changes were more drastic. The G_tj and L_ti of urban–rural land and unused land were the largest, indicating the most significant changes. This is also confirmed by their changes in area. From 1980 to 2020, unused land and urban–rural land decreased and increased by 98% and 26%, respectively.

The above analysis revealed that urban–rural land is the most dramatically changing land use type in terms of both area and intensity of change. Therefore, further analysis of changes in urban–rural land at the transition level is necessary (Table S2). Regarding land conversion into urban–rural land, the R_tin of paddy land and dryland was greater than W_tn from 1980 to 2020, indicating that urban–rural land tended to expand through conversion from paddy land and dryland. The same trend was also observed during other periods, but from 2000 onwards, unused land also tended to convert into urban–rural land in addition to paddy land and dryland. Regarding land conversion from urban–rural land, only the Q_tmj of dryland was greater than V_tm from 1980 to 2020, indicating that urban–rural land tended to convert into dryland rather than other land use types. Similarly, from 2000 onwards, the Q_tmj of unused land was greater than V_tm, indicating that urban–rural land also tended to convert into unused land.

3.3. Ecological Landscape Index and Ecological Risk Evolution

The results of the ecological landscape index are shown in

Table 3. Landscape pattern index from 1980 to 2020 in the Huaihe Eco-economic Belt.

Land Use	Period	Fragmentation Index	Divisibility Index	Dominance Index	Disturbance Index	Fragility Index
Forest land	1980	0.0008	0.0467	0.0836	0.0311	0.0714
	1990	0.0008	0.0464	0.0836	0.0310	0.0714
	2000	0.0008	0.0468	0.0839	0.0312	0.0714
	2010	0.0007	0.0432	0.0785	0.0290	0.0714
	2020	0.0007	0.0435	0.0781	0.0290	0.0714
Grassland	1980	0.0015	0.0964	0.0586	0.0414	0.1071
	1990	0.0015	0.0944	0.0594	0.0410	0.1071
	2000	0.0017	0.1032	0.0582	0.0434	0.1071
	2010	0.0017	0.1175	0.0488	0.0459	0.1071
	2020	0.0018	0.1197	0.0501	0.0468	0.1071
Water area	1980	0.0022	0.1019	0.0957	0.0508	0.2143
	1990	0.0023	0.1088	0.0965	0.0531	0.2143
	2000	0.0022	0.1041	0.0976	0.0519	0.2143
	2010	0.0020	0.0927	0.1054	0.0499	0.2143
	2020	0.0022	0.0969	0.1096	0.0521	0.2143
Urban–rural land	1980	0.0046	0.1008	0.3941	0.1114	0.0357
	1990	0.0045	0.0976	0.3958	0.1107	0.0357
	2000	0.0042	0.0924	0.3965	0.1091	0.0357
	2010	0.0036	0.0795	0.4099	0.1076	0.0357
	2020	0.0033	0.0727	0.4058	0.1046	0.0357
Unused land	1980	0.0046	0.9572	0.0043	0.2903	0.2500
	1990	0.0018	0.3949	0.0045	0.1202	0.2500
	2000	0.0069	1.4975	0.0038	0.4534	0.2500
	2010	0.0046	1.1096	0.0033	0.3359	0.2500
	2020	0.0069	1.6553	0.0032	0.5007	0.2500
Paddy land	1980	0.0003	0.0203	0.1219	0.0306	0.1786
	1990	0.0003	0.0199	0.1201	0.0301	0.1786
	2000	0.0003	0.0207	0.1209	0.0306	0.1786
	2010	0.0003	0.0205	0.1211	0.0306	0.1786
	2020	0.0003	0.0209	0.1195	0.0303	0.1786
Dryland	1980	0.0001	0.0077	0.2418	0.0507	0.1429
	1990	0.0001	0.0077	0.2401	0.0504	0.1429
	2000	0.0001	0.0076	0.2391	0.0502	0.1429
	2010	0.0001	0.0079	0.2330	0.0490	0.1429
	2020	0.0001	0.0083	0.2336	0.0492	0.1429

. The disturbance index of forest land, urban–rural land, and dryland has generally decreased in recent years, while that of grassland and unused land have increased. The fragility index of unused land was found to be the highest among the seven land use types, while that of urban–rural land was the lowest. Using the natural breakpoint method, the ecological risk level of the Huaihe Eco-economic Belt was divided into five levels, and the spatial distribution map of ecological risk was generated for different periods (Figure 4). Relatively few very high-risk areas were observed in the study area, and the water area presented the highest ecological risk owing to its relatively high vulnerability and disturbance (Table 3). Very low-risk areas were mainly distributed in forest land due to the low value of disturbance index and fragility index. At the city level (Figure 5), very low-risk and low-risk areas were mainly distributed in areas with a large forest area, such as Suixian, Guangshui, Dawu, Xinyang, Lu’an, Tongbai, Chuzhou, and Pingdingshan. Middle-risk and high-risk areas were mainly distributed in the central region of the study area, with Jining, Zaozhuang, and Bengbu being high-risk areas throughout the studied time periods. It is worth noting that high-risk areas are decreasing (Figure 5e), which also indicates that the ecological environment in the study area is stable and there is no serious deterioration trend.

3.4. Driving Mechanisms of Ecological Risk

According to the analysis of landscape risk and the proportion of different land use types at the city level, landscape risk presented a significant negative correlation with forest proportion, with a R² of 0.88 (Figure 6). This indicates that increasing forest coverage can effectively reduce landscape ecological risk. It is worth noting that large areas of forests are concentrated in the southwest region of the Huaihe Eco-economic Belt, where ecological security risks have remained low for the past 40 years (Figure 4 and Figure 5). This can be attributed to the beneficial ecological functions provided by forests, including soil conservation, water regulation, and biodiversity preservation. Protecting and restoring forests play a crucial role in mitigating land degradation and environmental pollution, thereby reducing ecological risks.

Furthermore, our findings revealed positive correlations between landscape risk and the proportions of urban–rural land and dryland, with R² values of 0.59 and 0.36, respectively. The increase in urban–rural land and dryland significantly contributes to landscape ecological security risks. Positive correlations were observed between the proportions of urban–rural land and dryland and landscape risk. This could be due to the expansion of urban–rural land and the increase in dryland, which result in ecological fragmentation and degradation, thereby increasing the probability of ecological risks. Urbanization and population growth lead to excessive land development and improper land use, intensifying pressure on ecosystems and subsequently raising ecological risks. However, it is important to emphasize that once the urban–rural land exceeds 15%, the ecological security risk does not show a significant increase with further expansion (Figure 6d). This suggests that the linear relationship between landscape risk and urbanization is not universally applicable. This phenomenon can be attributed to the diminishing impact of further land expansion on the ecosystem once a certain threshold of urban–rural land has been reached. At this stage, effective land use management becomes crucial in maintaining landscape ecological security. Measures, such as protecting and restoring ecosystem functions and implementing ecological compensation can play a significant role in achieving this goal.

In the HPA variance decomposition (

Table 4. Contribution rate of different land use types to landscape ecological risk.

	All Year	1980	1990	2000	2010	2020
Forest land	41	38	40	38	39	38
Grassland	1	1	1	1	1	1
Water	6	5	5	6	6	6
Urban–rural land	22	24	25	24	25	26
Unused land	2	5	3	5	6	5
Paddy land	6	6	6	6	5	6
Dryland	21	20	20	20	17	18

), forest land explains approximately 40% of the ecological security risk. Therefore, forest land is the most important land use type for maintaining ecological security. Urban–rural land and dryland also have high explanatory power for ecological security risk, with R² values of 22% and 21%, respectively. Other land use types have a relatively small impact on landscape ecological security risk, with explanatory variance less than 7%. Overall, forest land, dryland, and urban–rural land are the main factors determining the landscape ecological risk in the Huaihe River Eco-economic Belt.

4. Discussion

Land use change in the Huaihe Eco-economic Belt is mainly reflected by conversions between dryland, water, and urban–rural land, particularly the conversion from dryland to urban land. Among all land use types, the absolute reduction in dryland, grassland, and paddy land is the largest. However, in terms of the intensity of land use change, the changes in grassland, water area, urban–rural land, and unused land are the most dramatic. This indicates that the analysis of the change intensity can reveal hidden information from land use transfer matrices and provide more useful information for decision-makers. Previous studies conducted on land use change in the Huaihe River Eco-economic Belt have also identified similar patterns and trends. For example, Yang et al. [45] found that the conversion of dryland to urban land is a dominant driver of land use change in the region. Similarly, Liu et al. [46] highlighted the significant reduction in dryland and grassland as key aspects of land use change. In addition to the conversion of dryland to urban land, other researchers have identified additional factors influencing land use change in the region. Liu and Long [47] examined the impact of economic development and population growth on land use change and found that these factors, along with urbanization, play crucial roles. Their findings underscore the importance of considering multiple drivers when analyzing land use change. Tang et al. [48] focused on the dynamic evolution and scenario simulation of habitat quality and highlighted the influence of government regulations and incentives in shaping land use patterns.

The study area has relatively low ecological risk, primarily corresponding to medium and low risks (Figure 4 and Figure 5). This finding is consistent with the research conducted by You et al. [49], who also identified decrease in fragmentation of landscape pattern, and increase in connectivity and landscape diversity in this region. During the period from 1980 to 2020, the ecological risk in the Huaihe River Eco-economic Belt did not increase significantly, indicating that urbanization is not the main cause of ecological deterioration. In other words, increasing urban–rural land does not necessarily lead to an increase in landscape risk.

The results of this study confirmed that ecological risk does not increase with urbanization beyond a 15% proportion of urban–rural land. In the entire Huaihe River Eco-economic Belt, forests explain 40% of the landscape ecological risk, indicating that afforestation is currently the most effective means for improving ecological security. The findings discussed are consistent with the research conducted by Sang et al. [50], which highlights the positive impact of afforestation on ecological restoration in the Huaihe River Basin. Promoting afforestation initiatives is essential for enhancing ecological security and mitigating landscape ecological risks. Furthermore, Zhang et al. [51] investigated the effects of land use change on ecosystem services in the region and emphasized the importance of implementing sustainable land management practices to mitigate ecological risks. These studies reinforce the observations made in our study and provide additional evidence of the importance of afforestation in enhancing ecological security.

The evolution mechanism of ecological risk is a complex process that involves the interaction of multiple factors. By exploring mechanisms related to forest conservation and restoration, urban–rural land planning and management, and ecosystem services, we can gain a better understanding and explanation of the evolution of ecological risks. These aspects play crucial roles in shaping the dynamics of ecological risks and provide valuable insights for effective risk assessment and mitigation strategies. It should be noted that this study analyzed only the impact of land use proportion on ecological risk, and other factors, such as topography, climate, economy, and population were not considered. Consequently, the research results may contain some uncertainties. The driving mechanisms of natural, social, and economic factors on ecological risk will be further analyzed in future research.

5. Conclusions

This study employed the landscape index method to quantitatively analyze ecological risks in the Huaihe River Eco-economic Belt and investigated the driving mechanism of ecological risks through the changes in land use. Based on the analysis, the following main conclusions can be drawn:

(1)

In terms of the intensity of land use change, grassland, water, urban–rural land, and unused land exhibited the most significant changes. In terms of absolute change area, dryland, grassland, and paddy land showed the largest decrease, whereas urban–rural land and water area showed the largest increase. Land use changes in the study area can be primarily characterized by conversions between dryland, water area, and urban–rural land. The expansion of urban–rural land often occurs through the conversion of both paddy land and dryland. Additionally, urban–rural land tends to undergo conversion, resulting in its transformation into dryland and unused land.

(2)

The ecological risk of the Huaihe Eco-economic Belt is at medium and low levels. Jining, Zaozhuang, and Bengbu have been in a state of medium to high ecological security risk throughout the studied time periods, and are therefore areas requiring attention in the construction of the economic belt. Despite the reduction in land area with low ecological risks, such as forests and grasslands, the ecological risks within the Huaihe Eco-economic Belt have not increased. This can be primarily attributed to the decrease in the disturbance index of forests, urban–rural land, dryland, and paddy fields, which occupy a significant proportion of the total area. This indicates that, in addition to absolute changes in land area, reducing external disturbances to the ecosystem and improving habitat quality are equally crucial for maintaining ecological security in the region.

(3)

Forest land, urban–rural land, and dryland play significant roles in landscape risk. Forest land exhibits a strong negative correlation with landscape risk, explaining 40% of the variation in landscape ecological risk. This finding highlights the effectiveness of afforestation as a means of ecological restoration. On the other hand, dryland shows a positive correlation with ecological risk, accounting for 20% of the variation in landscape ecological risk. Reducing the extent of dryland can help mitigate regional ecological risks. As the proportion of urban–rural land increases, landscape ecological security initially improves and then reaches a plateau. Importantly, the ecological security risk does not show a significant increase when the proportion of urban–rural land exceeds 15%. This suggests that implementing reasonable and effective urban planning and management practices are crucial for maintaining landscape ecological security.