Study on the Spatial and Temporal Evolution of the Ecological Environmental Quality in Linghekou Wetland

Abstract

Coastal wetlands are located in the overland area between land and sea and play an important ecological role, but social and economic development and the acceleration of urbanization have resulted in the degradation of the ecological functions of coastal wetlands and serious pollution within the wetlands. The study of the spatial and temporal changes in the ecological environmental quality of coastal wetlands can suggest feasible response strategies for the ecological construction of coastal wetlands. This study uses RS and GIS technology, based on the PSR model, AHP method and InVEST model, to study the spatial and temporal changes in the ecological environmental quality of the Linghekou wetland and to propose early warning on future ecological and environmental conditions. The results show the following: (1) The natural wetland area decreased, the landscape fragmentation index increased and the wetland landscape showed a degradation trend between 2005 and 2020. (2) The composite indices for the evaluation of the ecological environmental quality of the Linghekou wetland in 2005, 2010, 2015 and 2020 were 0.473, 0.380, 0.353 and 0.378, respectively, with the ecological environmental quality declining from a sub-healthy state in 2005 to a sub-sick state in 2020, with increasing interference from human activities, increasing differentiation of the internal organization of the wetlands and degradation of ecological services. (3) The habitat quality of the Linghekou wetland has improved since 2015 due to the implementation of local conservation measures, but the habitats are still under strong pressure from anthropogenic disturbance. (4) The predicted results for 2025 and 2030 show that the ecological environmental quality of the Linghekou wetland will continue to deteriorate in the future, especially in the northwestern and central-eastern parts of the study area, where anthropogenic disturbance will continue to intensify and habitat degradation will become more severe in the future. This study provides a scientific reference for coastal wetland management and ecological construction, and also enriches the research results on coastal wetlands in the field of ecological and environmental assessment.

1. Introduction

Ecological environmental quality is the result of the interaction between human activities and the ecological environment, and the quality of the natural environment is closely related to the development of human society. With the acceleration of economic development and urbanization, human interference with natural resources is becoming more frequent, leading to the overexploitation of natural resources and the deterioration of the quality of the ecological environment, resulting in a series of ecological and environmental problems, such as reductions in biodiversity, air pollution, land desertification and water pollution. People have realized that ecological and environmental problems have a profound impact on human and social development, and more and more scholars have conducted research on ecological and environmental quality assessment [1,2,3,4,5]. Based on ecological theory, the ecological environment’s quality reflects the suitability of the ecological environment for human survival and sustainable socioeconomic development from a spatial and temporal perspective, at the ecosystem level, and is an assessment of the nature of the ecological environment and the results of its state of change according to specific human requirements [6,7]. The core of ecological quality assessment is the construction of a reasonable indicator system, the selection of representative and easily accessible assessment indicators and the application of appropriate assessment methods to determine the degree of quality of the ecological environment [8,9,10]. The most widely used theoretical model in ecological assessment is the pressure-state-response (PSR) model, which is considered by many governments and organizations as one of the most effective frameworks [11,12]. The hierarchy process (AHP) and analytic network process (ANP) are methods for ecological quality assessment. Both are quantitative assessments of subjective views, with the AHP method being simple in structure, making it easier to construct judgement matrices and easier to calculate eigenvalues in a small number of dimensions, while the ANP method super-matrix is more complex and computationally intensive. When access to indicators is limited, the AHP method is usually used for evaluation. In addition to ecological quality assessment studies using PSR models, many scholars have conducted ecological degradation studies to address ecological quality issues from different perspectives and have produced rich research results, especially relating to the changes in habitat conditions caused by changes in landscape types, which has become a hot and cutting-edge research topic. The main research methods used are physiological indicators, ecological networks, water balance, mathematical optimization algorithms, fuzzy mathematics, stratigraphic analysis, etc. [13,14,15]. The models used include the HIS model, GUMBO model, MAXENT model, CLIMEX model, IDRIS model, SolVES model, BIB-LCJ model and InVEST model [16,17,18,19,20]. Compared with other models, the InVEST model has become a mainstream model for assessing habitat quality. Its advantages are that it does not require data such as biodiversity data and the number of species in the environment, it can quickly assess the impact of different threat sources and land use/cover on biodiversity, and it has easy access to data, fewer required parameters, high analytical power and simple data processing. There have been numerous empirical studies on habitat quality assessment based on the InVEST model in different regions and at different scales [21,22]. As can be seen, there are many methods and models for assessing ecosystem quality, and there is no agreed set of assessment criteria. Due to the complexity of each ecosystem and the biases of researchers’ perceptions, current research on ecological and environmental quality assessment suffers from high data requirements and poor visualization of research results. In addition, some methods are too subjective and lack spatial relationships to describe the links between evolutionary processes and ecological quality in the study area.

Wetlands are the kidneys of the Earth and play an important role in maintaining biodiversity, regulating floods and purifying water. However, in recent years, with the expansion of land for construction, wetland resources have been continuously eroded, resulting in the gradual reduction in natural wetlands, serious pollution within wetlands, declining ecological regulatory functions and increasing contradictions between economic development and wetland resource allocation, which seriously threaten the sustainable development of wetlands. Wetland ecological and environmental quality assessment can be carried out to better understand the causes of wetland development and degradation mechanisms, so that wetland resources can be better protected and managed. Located at the interface between land and sea, the Linghekou wetland has a unique ecological structure and environmental characteristics, and is rich in biodiversity, sensitive to external disturbances and has a relatively fragile ecosystem. In recent years, due to the development of aquaculture and accelerated urbanization, intensive human activities have caused greater disturbance to the ecological environment of the Linghekou wetland, and changes in the landscape pattern have led to changes in the spatial structure of ecosystem types in the area, resulting in varying degrees of damage to the integrity of ecosystem types and the stability of habitat quality. Long-term assessments of habitat quality are important for assessing the stability of regional ecosystem services and for maintaining ecological security and promoting sustainable development. Based on remote sensing image data from 2005, 2010, 2015 and 2020, this study obtained data on the spatial distribution of landscape classification within the study and used the landscape pattern index to analyze the characteristics of spatial and temporal pattern changes in landscape types of the Linghekou wetland over the past 15 years. The PSR model, the AHP method and the InVEST model were used to optimize and adjust their parameters in combination with the relevant literature and field research analysis to achieve the analysis of spatial and temporal changes in the ecological and environmental quality of the Linghekou wetland, and the ANN-CA-Markov model was used to propose an early warning study on the future ecological and environmental quality of the Linghekou wetland, providing a scientific basis for the future wetland management and restoration of the Linghekou wetlands. This study will provide a scientific basis for future management and restoration in the Linghekou wetland.

2. Data and Methods

2.1. Study Area

The Linghekou wetland is located between 121°00′ and 121°30′ and between 40°45′ and 41°00′ N. The total area is 838.66 km². The average total rainfall is 587.91 mm, the average total evaporation is 942.07 mm, the average temperature is between 8.0 and 8.7 °C and the average sunshine time is 2808.2 h. The Linghekou wetland is home to about 250 species of birds. Approximately 250 species of birds live and breed in Linghekou wetland, making it an important waterbird habitat in China. Figure 1 shows the geographical distribution of Linghekou wetland.

2.2. Data Processing

This study used the 2005, 2010 (Landsat7 ETM+), 2015 and 2020 (Landsat8 OLI) data with a resolution of 30 m as the base data. Referencing Current Land Use Condition Classification (GB/T 21010-2017) and previous scholars’ research results, combined with geomorphological characteristics [23,24,25], this study classified the landscape of Linghekou wetland into eight types: woodland, farming pond and lake, residential land, reed swamp, dry land, paddy field, mudflat, and river [26,27,28,29,30]. In this study, the overall accuracy and kappa coefficient were used as test indicators, and the results showed that the overall classification accuracy in 2005, 2010, 2015 and 2020 was 87.58%, 88.32%, 91.83% and 92.19%, respectively, and the decoding accuracy met the calculation requirements. The Landsat image series of Linghekou wetland from 2005 to 2020 is shown in

Table 1. Landsat image series data of Linghekou wetland from 2005 to 2020.

No	Acquisition Date	Sensors	Path/Row	Cloud (%)
1	2005-4-14	Landsat7 ETM+	120/31	0
2	2010-3-27	Landsat7 ETM+	120/31	0.05
3	2015-3-11	Landsat8 OLI-TIRS	120/31	0.03
4	2020-5-10	Landsat8 OLI-TIRS	120/31	0.04

. Figure 2 shows the landscape pattern distribution of Linghekou wetland.

2.3. PSR Model

PSR Model includes pressure, state and response indicators. The pressure indicator reflects the impact of the external environment, the state indicator reflects the current state and the response indicator reflects the impact after an external perturbation [31,32].

Combining the actual situation of Linghekou wetland, this study established an evaluation index system around the impact of human activities and ecological service functions of Linghekou wetland [33,34], as shown in

Table 2. Evaluation indicator system.

Pressure Indicator	Status Indicator	Response Indicator
Population density	Diversity index	Degradation index
Human interference index	Mean patch area	Fragmentation index
	Hydrological regulation index
	Shape index
	Uniformity index
	Maximum plaque index

.

Based on the actual situation of Linghekou wetland and combined with the quality classification of the existing literature, the ecological and environmental quality of the wetland is divided into five classes: Class I (healthy), Class II (healthier), Class III (sub-healthy), Class IV (sub-sick) and Class V (sick), corresponding to index intervals of (0.8–1.0), (0.6–0.8), (0.4–0.6), (0.2–0.4) and (0–0.2), respectively [35].

2.4. The Analytic Hierarchy Process (AHP)

The AHP method is a simple, flexible and practical multicriteria decision-making method for quantitative analysis of qualitative problems. It is characterized by organizing the various elements of a complex problem by dividing them into interrelated and ordered levels, directly and effectively combining expert opinion and the results of the analyst’s objective judgement according to the structure of subjective judgement about a certain objective reality (mainly two-by-two comparisons), and quantitatively describing the importance of two-by-two comparisons of elements at one level.

When using the hierarchical analysis method (AHP) to determine the influence weights of each indicator, in order to compare the indicators two by two to obtain a quantitative judgement matrix, a judgement matrix was constructed by the expert opinion method using a 1–9 level scale. The AHP method can decompose the more complex indicators affecting the ecological and environmental quality of the Linghekou wetland into several levels for step-by-step analysis and comparison, establishing a tree structure, obtaining the weight values of different nodes and performing the next evaluation based on the weight values.

2.5. Landscape Index

2.5.1. Diversity Index (H)

H refers to the change in the number of different landscape types and the proportion of area they occupy. The formula is as follows [28]:

$$H=-{\displaystyle \sum _{i=1}^m(P_i\times {\log }_2{P}_i)}$$

(1)

$P_i$ indicates the proportion of the area occupied by the ith landscape type, and m indicates that there are m landscape types in total.

2.5.2. Mean Patch Area (MPS)

MPS can be used to measure the degree of fragmentation of a wetland landscape. The formula is as follows [28,36]:

$$MPS=\frac{S_i}{N_i}$$

(2)

where $S_i$ denotes the area of the ith wetland landscape type and $N_i$ denotes the number of patches in the ith landscape.

2.5.3. Shape Index (LSI)

LSI indicates the complexity of the external shape of the landscape. The formula is as follows [28,37]:

$$LSI=\frac{0.25E}{\sqrt{A}}$$

(3)

where E denotes the sum of the lengths of all patch boundaries and A denotes the total area.

2.5.4. Evenness Index (SHEI)

SHEI indicates the evenness of the distribution of different landscapes with the following formula [28,36]:

$$SHEI=\frac{{\displaystyle \sum _{i=1}^m{P}_i\times {\log }_2{P}_i}}{{\log }_{2}m}$$

(4)

where $P_i$ is the percentage of the total area of the i type of landscape, and m is the number of landscape types.

2.5.5. Maximum Patch Index (LPI)

LPI is mainly used to reflect the dominant landscape, with the following formula [28,36]:

$$LPI=\frac{a_{\max }}{A}$$

(5)

where $a_{max}$ is the maximum patch area and A is the total area. Fragmentation and degradation indices are used as response indicators.

2.5.6. Fragmentation Index (PD)

PD is expressed as patch density. The higher the density, the higher the degree of fragmentation, and vice versa; this is expressed with the following formula [28]:

$$PD=\frac{{\displaystyle \sum N_i}}{{\displaystyle \sum A_i}}$$

(6)

where $\sum N_i$ is the total number of patches in the study area or the total number of patches of a given landscape type, and $\sum A_i$ is the total area of the study area or the total area of a given landscape type.

2.6. InVEST Model

2.6.1. Habitat Quality Index

The concept of habitat quality is actually the potential capacity of an ecosystem to provide the conditions for species survival and reproduction. Ecosystem types and disturbance sources are used as data and raster data as evaluation units to calculate the habitat quality index, which is calculated as follows [38,39]:

$$Q_{xj}=H_j(1-\raisebox{1ex}{$D_{xj}^Z$}\!\left/ \!\raisebox{-1ex}{$(D_{xj}^Z+k^z)$}\right.)$$

(7)

where Q_xj is the habitat quality of raster x in land use type j; H_j is the land use type j of habitat suitability; k is the half-saturation constant; and D_xj is the level of habitat stress for raster x in land use type j.

2.6.2. Habitat Degradation Index

The habitat degradation index is judged by the distance of influence of ecological threat factors, the high sensitivity of habitat type patches to threat factors and the number of threat factors in concert. The habitat degradation index is calculated using the following formula [40,41,42]:

$$D_{xj}={\displaystyle {\sum }_{r=1}^R{\displaystyle {\sum }_{y=1}^{Y_r}(\raisebox{1ex}{$W_r$}\!\left/ \!\raisebox{-1ex}{${\displaystyle {\sum }_{r=1}^R{W}_r}$}\right.)r_y}}{i}_{rxy}{\beta }_x{S}_{jr}$$

(8)

where D_xy refers to the magnitude of habitat degradation; R is the number of threat factors; Y_r is the number of rasters on the range of layers occupied by the stressor; W_r represents the weight of the threat factor, i.e., the relative magnitude of damage caused by a given stressor across all habitat type patches, and takes a value between 0 and 1; r_y is the number of threat factors on each grid in a given layer range; β_x is the accessibility level of the habitat type patch raster x and also takes a value between 0 and 1; S_jr is the sensitivity of the land use type j to the stressor r, again ranging from 0 to 1; and l_rxy indicates the stress factor value of r_y for grid y on the stress level of habitat grid x.

2.7. ANN-CA-Markov Model

The ANN-CA-Markov model must first determine the area of each land type in future years. Based on the historical land use of the two periods, a Markov chain can be used to obtain a land use transfer probability matrix for these two periods, from which the conversion threshold for each land type in future years can be calculated [40,41].

$$P=\left[\left.\begin{array}{cccc}{P}_{11}& P_{12}& \dots & P_{1n}\\ P_{21}& P_{21}& \dots & P_{2n}\\ ⋮& ⋮& \ddots & ⋮\\ P_{n1}& P_{n2}& \dots & P_{nn}\end{array}\right]\right.$$

(9)

where 0 ≤ P_ij ≤ 1, ${\displaystyle \sum _{j=1}^n{P}_{ij}=1(i,j=1,2,\dots ,n)}$, n is the number of land types and p_ij is the probability of converting land type i to land type j, the one-step transfer probability of land type j.

Based on the land use in year T, the formula for projecting future land use in T+M is as follows:

$$L{U}_{T+M}=L{U}_T\times P$$

(10)

ANN-CA-Markov models are trained using artificial neural networks with multiple input and output neurons to produce suitability probabilities for different land use types.

Neuron j receives from all input neurons on grid cell p at time t the signal expressed by the following equation:

$$ne{t}_j(p,t)={\displaystyle \sum _i{w}_{i,j}\cdot x_i(p,t)}$$

(11)

where x_i(p, t) is the training time t associated with the grid cell at an input neuron i; w_i, j is the i type variable associated with the input neuron i; and w_i, j are the adaptive weights between the input and hidden layers.

The probability of occurrence of land use type k at training time t with grid cell p can be expressed as

$$P(p,k,t)={\displaystyle \sum _j{w}_{j,k}\cdot \frac{1}{1+e^{-ne{t}_j(p,t)}}}$$

(12)

where w_j,k is the adaptive weight between the hidden layer and the output layer weights, similar to w_i,j.

The assessment results of 2005, 2010, 2015 and 2020 were used as the original data to complete the prediction of the habitat quality and habitat degradation indices of the study area for the next 5 years and 10 years.

The schematic framework of the ecological environmental quality assessment of Linghekou wetland is shown in Figure 3.

3. Results

3.1. Analysis of Landscape Area Change

From 2005 to 2020, the area of both natural and artificial wetlands decreased to some extent. The largest decrease was in paddy fields, with a total decrease of 40.56 km², followed by farmed lakes, with a decrease of 25.78 km². The main food crops in Liaoning Province are rice, corn and soybeans, and the scale of corn and soybean cultivation has increased year on year due to its economic benefits, resulting in an increase in the scale of dry land and a decrease in the scale of paddy fields as more residential land has become available. The Aquaculture Ecological Development Regulation enacted by Linghai City in 2017 limits the size of private aquaculture ponds, which to some extent protects wetland resources and improves the ecological quality of wetlands, resulting in a sharp decline in the area of farmed lakes from 2015 to 2020. In recent years, under the influence of relevant local government policies, the ecological environment of the Linghekou wetland has recovered, especially the natural wetland. Between 2005 and 2015, the natural wetland decreased in area sharply from 66.44 km² to 52.51 km², and with the implementation of conservation policies, the area of natural wetland recovered from 52.51 km² in 2015 to 62.41 km² in 2020. In terms of the changes in landscape area, this shows that the ecological environment of the Linghekou wetland is gradually improving, but the situation is still not optimistic. The landscape changes in the Linghekou wetland from 2005 to 2020 are shown in Figure 4.

3.2. Analysis of Evaluation Indicator Values

The landscape index data were obtained by processing with Fragstats 4.2 software. Population density was calculated by dividing the number of people by the area of the study area. Human disturbance refers to the impact of various human activities on ecosystems. The anthropogenic disturbance index uses the area of the most anthropogenically disturbed dry land and residential land divided by the area of the study area. The hydrological regulation index was calculated by dividing the area of rivers and mudflats that play an active role in hydrological regulation by the total area of the wetlands. From the above calculations, the values of the various assessment indicators for the Linghekou wetland were obtained, as shown in

Table 3. The evaluation indicator values of Linghekou wetland.

Guide Layer	Program Layer	Year
		2005	2010	2015	2020
Pressure Indicator	Population density	675.301	637.888	616.422	599.045
	Human interference index	0.489	0.516	0.551	0.561
Status Indicator	Diversity index	1.567	1.555	1.548	1.575
	Mean patch area	0.823	0.957	0.884	0.823
	Hydrological regulation index	0.039	0.036	0.033	0.038
	Shape index	13.974	13.68	14.823	15.773
	Evenness index	0.713	0.702	0.705	0.719
	Maximum plaque index	20.754	21.391	22.363	20.21
Response Indicator	Degradation index	419.61	397.16	368.64	355.43
	Fragmentation index	0.689	0.598	0.643	0.688

.

The pressure indicator reflects the impact of the wetland ecosystem on the external environment, in particular the impact of human activity. Once the external pressure exceeds the capacity of the wetland ecosystem, the internal organization of the wetland is damaged, leading to the degradation of the wetland. From the calculation results, it can be seen that from 2005 to 2020, the population density gradually decreased from 675.301 (units/km²) to 599.045 (units/km²), as the number of permanent residents living in the Linghekou wetland gradually decreased. As the permanent population decreases, the amount of residential land continues to increase, indicating an increase in population mobility in the area. In addition, the human disturbance index also gradually increased from 0.489 to 0.561 from 2005 to 2020, indicating that the wetlands are under increasing pressure from human disturbance.

The diversity of landscape types not only provides a variety of places for organisms to live and reproduce, but also makes the biochemical activity within the landscape more intense. The average patch area index gradually decreased from 0.713 in 2005 to 0.705 in 2015, then reverted to 0.719 in 2020. The main service function of the wetland is hydrological regulation, and strong hydrological regulation can maintain the wetland’s functions of water retention, water purification and being a migratory stopover for various rare bird species. The beach, river and reed swamp all play an important role in hydrological regulation. The beach area was 9.41, 9.72 and 11.28 km² in 2005, 2010 and 2015, respectively, with an increasing trend. Due to massive local reclamation and development of beaches, the beach area was only 2.63 km² in 2020 with a very high degree of degradation. The river area decreased by 7.27 km² from 2005 to 2015 and increased to 32.09 km² in 2020, with significant implications for hydrological regulation. The changes in the two landscape types directly affect the hydrological regulating capacity of the Linghekou wetland, which was calculated to have decreased from 2005 to 2015, with the overall capacity returning to the 2005 level by 2020 due to a significant increase in the river area. The shape index reflects the extent to which the landscape deviates from the standard shape, with a larger shape index indicating a more irregular landscape, and the evenness index can be used to determine several dominant landscape types in the wetland. From the calculation results, the landscape shape index and the evenness index gradually increased from 2010 to 2020, indicating a gradual development trend of irregularity and unevenness in each landscape type in the Linghekou wetland. The results of the maximum patch index calculation show that residential land and dry land are still the most dominant landscapes in the Linghekou wetland, and the dominance of rivers and reed swamps is also gradually increasing.

The response indicators visualize the evolutionary outcome of the wetlands in the study area and are indicator values that correspond to external pressures. The wetland degradation index is equal to the area of reduced wetlands. The wetland area decreased by 22.45 km² from 2005 to 2010, 28.52 km² from 2010 to 2015 and 13.21 km² from 2015 to 2020, a total of 64.18 km² from 2005 to 2020, according to the analysis of remote sensing images; most of the reduced wetlands have been converted into residential land, with a small proportion converted into dry land, and the trend of reduction has slowed down in recent years.

The fragmentation index shows the degree of fragmentation of each landscape in the study area. There is a decreasing trend in the number of rivers, paddy fields and reed swamp patches, indicating that fragmentation is decreasing and the situation is improving. In the study area as a whole, the fragmentation level improved from 2005 to 2010, when the patch fragmentation index decreased from 0.689 to 0.598, and then increased again to the 2005 level of 0.643 and to 0.688 in 2015 and 2020, respectively. This indicates that a series of human disturbance activities have had a greater impact on the Linghekou wetland and increased the fragmentation level of the landscape.

3.3. Analysis of Ecological Environmental Quality

In this study, AHP was used to determine the weights of each indicator and the calculations were performed under the SPSS26 platform. The target layer of this paper is the ecological quality of the Linghekou wetland, the criterion layer is the pressure indicator, status indicator and response indicator, and the scheme layer is the ten parameters, as shown in

Table 4. The parameters values of indicators.

Target Layer	Guide Layer	Program Layer	Indicator Weight
Eco-environmental quality of Linghekou wetland	Pressure 0.261	Population density	0.075
		Human interference index	0.186
	Status 0.504	Diversity index	0.046
		Mean patch area	0.027
		Hydrological regulation index	0.179
		Shape index	0.052
		Evenness index	0.131
		Maximum patch index	0.069
	Response 0.235	Degradation index	0.171
		Fragmentation index	0.064

.

The main external human pressures on the Linghekou wetland are the expansion of residential land, the reclamation of agricultural land and the development of engineering areas. Between 2005 and 2020, reclamation activities caused a rapid reduction in the area of the reed swamp. The emergence of private agricultural ponds has destroyed the original beach resources, and the quality of water resources in the Linghekou wetland has also been greatly reduced by the decline in hydrological regulation and anthropogenic wastewater discharges. The above-mentioned anthropogenic activities represent a major threat to the ecological environmental quality of the Linghekou wetland.

As can be seen from

Table 5. The ecological environmental quality index values of Linghekou wetland in 2020.

Individual Indicators	Indicator Measurements	Indicator Values	Indicator Weight	Overall Assessment Value
Population density	0.11	0.97	0.075	0.378
Human interference index	0.15	0.96	0.186
Diversity index	0.01	0.01	0.046
Average patch area index	0.01	0.01	0.027
Hydrological regulation index	0.03	0.01	0.179
Shape index	0.13	0.97	0.052
Evenness index	0.01	0.01	0.131
Maximum patch index	0.03	0.01	0.069
Degradation index	0.15	0.04	0.171
Fragmentation index	0.00	0.99	0.064

, the ecological environmental quality evaluation result of the Linghekou wetland is 0.378, and according to the classification grade, the wetland was in Class III (sub-sick) in 2020, which indicates that the Linghekou wetland suffers from high external pressure, there is a trend of deterioration of various functions, the vitality performance is average and the system is beginning to show a little abnormality.

The same method was used to calculate the overall evaluation indices of 0.473, 0.380 and 0.353 for 2005, 2010 and 2015, respectively. The results show that before 2015, the ecological and environmental quality of the Linghekou wetland was in a declining stage; the ecological environmental quality has improved to a certain extent since 2015. Because of the development of wetland protection awareness by the local government and the community and various protection measures, the ecological environmental quality has improved to a certain extent, but the situation is still not optimistic.

3.4. Habitat Quality Assessment of Linghekou Wetland

In this study, based on the data of the study area in 2005, 2010, 2015 and 2020, a comprehensive evaluation and analysis of the habitat quality of the Linghekou wetland in 2005, 2010, 2015 and 2020 was carried out using the habitat quality module of the InVEST model. The analysis results are shown in Figure 5 and Figure 6.

By collecting and processing data for each year and using statistical tools to calculate the mean habitat index and mean degradation index for the study area, the mean habitat index and mean degradation index were obtained as 0.247 and 0.041 in 2005, 0.246 and 0.041 in 2010, 0.264 and 0.042 in 2015, and 0.439 and 0.037 in 2020, respectively. The analysis results are shown in Figure 7.

As can be seen from Figure 5, Figure 6 and Figure 7, in terms of space, the habitat quality of residential land, dry land, farmed lakes and paddy fields is low, while the habitat quality of beaches, woodlands, rivers and reed swamps is high; in terms of the degree of habitat degradation, a small number of residential lands, reed swamps and woodlands are low, while dry land and beaches are high, and the degree of degradation of rivers has a distribution pattern, with rivers adjacent to farmed lakes in the south having a high degree of degradation and the rest having a low degree of degradation. In terms of time, the habitat degradation index was almost unchanged and flat from 2005 to 2015, with a significant decrease in 2020, indicating that the impact of human activities is gradually diminishing, while the habitat quality index showed a slight decrease from 2005 to 2010, an increase in 2015, and a significant increase in 2020, reflecting the treatment effect achieved after the government introduced corresponding treatment policies.

3.5. Early Warning on the Habitat Quality and Habitat Degradation

Based on the landscape classification data of 2010 and 2015, this study simulated the landscape types in 2020 using the GeoSOS-FLUS V2.4 software platform and analyzed them in comparison with the real landscape classification data of 2020. Seven influencing factors, namely the DEM, slope, slope direction, and distance from roads, railways, settlements and rivers, were selected as the driving factors of landscape type change(Billionnet, 2013). The DEM was downloaded from the Geospatial Data Cloud with a resolution of 30 m. The slope and slope direction were calculated from the DEM. Roads, railways, rivers and settlements were downloaded from the National Basic Geographic Information Centre website. In the model, the training sampling rate of the neural network was set to 2%, the number of hidden layers was set to 12 and the random sampling mode was selected to sample the training samples of each land type to accomplish the neural network training. Combined with the distribution of each driver after the standardization process, the final calculation of the suitability probability map of the landscape type was conducted on each image element. The predicted landscape classification maps of the Linghekou wetland in 2025 and 2030 based on the 2010, 2015 and 2020 data showed that the kappa coefficients were 0.831919 and 0.860895, respectively, with an overall accuracy of 0.892118 and 0.901253, which is a high simulation accuracy and meets the requirements of this study. The habitat quality and habitat degradation indices were assessed based on the predicted results to warn of future changes in ecological quality. The predicted results are presented in Figure 8 and Figure 9.

As can be seen from Figure 7, the habitat quality of the Linghekou wetland in 2025 and 2030 shows a continuing trend of deterioration, particularly in the northwestern and eastern parts of the study area. The landscape types predominantly distributed in the northwestern part of the study area are residential land and paddy fields, which are subject to perennial human disturbance. Combined with the outline of the 14th Five-Year Plan and the 2035 Vision Plan of Linghai City, the area will be developed into an emerging industrial zone in the future, forming new industries such as trade and logistics and high technology. There are some difficulties in improving the habitat quality of the area. The eastern part of the study area is dominated by residential land and shows a distribution pattern around the reed swamp. With the future expansion of the city, the habitat quality in this area has not improved and it is recommended that necessary protection measures are taken to avoid the destruction of the natural wetlands. As can be seen in Figure 8, the most degraded habitats in the Linghekou wetland in 2025 and 2030 are in the central-western and central-eastern parts of the study area, which are dominated by paddy fields and residential land. Combined with road and rail data, the central and western regions are served by major roads and railways, and the future relies on high-speed rail projects to create new industrial zones. In the central-eastern part of the region, where the dry land and paddy fields intersect, the focus will be on building a deep processing base for agricultural and sideline products. It can be seen that anthropogenic disturbances in the central and central-eastern parts of the region will continue to intensify in the future and habitat degradation will become more severe, suggesting a synergistic development of economic construction and ecological conservation in the region.

4. Discussion

According to existing research, there are currently many methods and models for studying ecological environmental quality. However, such studies lack spatial relationships and cannot reflect the relationship between the evolution of the situation within a region and the impact of ecological environmental quality. In addition, the spatial variability of the landscape within the region and future land use planning can have an important impact on ecological environmental quality. This study uses different methods and models to assess and analyze the ecological and environmental quality of the Linghekou wetland, improving the subjective nature of the previous application of the AHP method to coastal wetlands assessment, with the aim of combining expert scoring and spatial and temporal distribution to carry out a multifaceted assessment of the ecological and environmental quality of the Linghekou wetland, which helps to provide a more comprehensive understanding of the ecological and environmental conditions in the study area. In addition, the assessment process will be integrated with regional spatial planning to provide early warning of future changes in ecological quality, so that the assessment results are more in line with the actual situation of the study area and provide a more concrete scientific basis for the conservation and restoration of the Linghekou wetland.

According to the research results, with the rapid increase in population in recent years, a large amount of natural water and beach area is gradually being reclaimed by human activities for more economically beneficial farmed ponds, and the area of natural wetland is decreasing while the area of dry land and residential land is increasing. In addition, the construction of highways and the expansion of residential land continue to encroach on the natural wetland landscape, fragmenting it and damaging both the quality and quantity of natural wetland. It can be seen that anthropogenic disturbance is the main cause of changes in the landscape pattern and ecological environmental quality of the Linghekou wetland. From the perspective of the local land use vision, the anthropogenic pressure on the area has not abated and it is very difficult to improve the ecological environment. It is recommended that certain protective measures be taken to halt the continued deterioration of the ecological environment.

According to the limitations of the study, due to the influence of objective factors, the assessment results of this study still need to be improved. The low resolution of the remote sensing imagery, only 30 m, and the inability to extract small patches could affect the accuracy of the assessment results. In addition, the inability to visit and measure in the field limits the availability of much information. When using the PSR model for assessment, only 10 assessment indicators were selected, which may not fully reflect the real situation. All of the above factors could lead to bias in the assessment results. There are many factors that affect the ecological environments of wetlands, and the impact of social and economic factors on the ecological environments of wetlands should also be considered in the practical application process.

5. Conclusions

This study used the PSR model combined with the AHP method and the InVEST model to analyze the spatial and temporal changes in the ecological environmental quality of the Linghekou wetland. The results of the analysis show that the wetland area is decreasing during the period 2005–2020, with the most significant reductions in paddy fields and farmed lakes. The non-wetland area will continue to increase, especially the dry land area, due to a change in crop types as a result of trends in economic interests. In terms of ecological environmental quality, the ecological condition of the Linghekou wetland has gradually changed from sub-healthy in 2005 to sub-sick in 2020 due to increased human disturbance, and although the ecological environmental quality has improved in recent years, the situation is still not optimistic. According to the analysis of habitat quality and habitat degradation, the habitat quality of the Linghekou wetland has improved and habitat degradation has slowed down since 2015 with the introduction of local government conservation policies in the context of the 14th Five-Year Plan of Linghai City and the outline of the 2035 Vision. With the completion of the emerging industrial zone in the northwest of the study area and the deep processing base for agricultural and secondary products in the central-eastern part, anthropogenic disturbance within the study area will continue to intensify and habitat degradation will become more severe in the future, suggesting that economic construction should be accompanied by a focus on ecological protection of the region.