The Bohai Rim region is an important industrial and agricultural base in China, with rich agricultural, mineral, oil and gas and tourism resources. This region, centered on Beijing and Tianjin, is an important economic zone in eastern China with a concentrated population and increasingly concentrated industries. With a variety of ecosystems, including estuarine delta wetlands, cropland, breeding ponds, salt flats, forests, coastal mudflats and towns, the Bohai Rim region is an important part of the ecosystems in northern China and provides a wide range of ESs (Ning et al., 2011). With the rapid development of urbanization and industrialization in the Bohai Rim region, coastal development activities have become increasingly frequent, and LULC patterns have changed significantly. In addition, the Bohai Sea is surrounded by land on three sides, and there are problems such as innate hydrodynamic insufficiency and weak exchange capacity of water bodies. Global climate change and human activities have exposed the ecological environment in the Bohai Rim region to multiple risks such as coastal wetland shrinkage, nearshore soil salinization, biodiversity reduction, and water environment deterioration. The Bohai Rim region studied in this article includes 15 coastal cities, including Tianjin, Hebei Province (Qinhuangdao, Tangshan, Cangzhou), Shandong Province (Yantai, Weihai, Dongying, Qingdao, Binzhou, Weifang), and Liaoning Province (Dalian, Yingkou, Panjin, Jinzhou, and Huludao) (Figure 1).

The main data of this study includes two categories: spatial data and statistical data. All spatial data were uniformly converted to the WGS_1984_UTM_50N coordinate system, and specific data information is shown in Table 1. Based on current situation of study area, we reclassified the snow land in the original LULC data into barren land and the wetland into water, which is divided into 7 types: cropland, forest, shrub, grassland, water, barren land, and impervious land. The MOD16A2GF product has a temporal resolution of 8 days, and the MOD13A3 product is monthly NDVI raster data. The two are batch projected by MRT (MODIS Reprojection Tools) software, and the Maximum Value Composites method was used to form the PET and NDVI datasets of each year. DEM data was filled in ArcGIS software. The final assessment results of each ES are at a resolution of 30 m.

The LULC types in the Bohai Rim region are mainly cropland land, impervious land and forest, among which cropland accounts for the largest proportion, exceeding 60% at all stages of the study period. Impervious land and forest account for more than 14% and 11%. Shrubs account for the smallest proportion, with an average area of only 1.1 km² over 20 years (Table 4).

From the perspective of the spatial distribution (Figure 2), cropland is widely distributed, covering most of the Bohai Rim region; forest is mainly distributed in the north and northwest, shrubs and grasslands are mainly distributed in the northwest of the study area and the northern area of Shandong Peninsula; water is mainly distributed on the edge of the Bohai Bay; barren land is mainly distributed in the edge of Laizhou Bay and Bohai Bay, as well as the Yellow River Delta; impervious land are mainly distributed in the marginal areas of the coast and the city centers.

From the perspective of the temporal distribution, the proportions of forest, water and impervious land have expanded from 11.16%, 3.45%, 14.08% in 2000 to 11.67%, 4.05%, 21.60% in 2020; while the area proportions of cropland, grassland and barren land decreased from 67.21%, 2.63%, 1.47%–60.71%, 1.77% and 0.21%. From 2000 to 2020, the area of forest and shrub showed a trend of first decreasing and then expanding. This study further calculated the LULC transfer matrix (Table 5) and drew the LULC transfer Sankey diagram (Figure 3). During the study period, 12,841.78 km² of cropland was transferred out and 2,957.05 km² was transferred in, mainly transforming into impervious land, forest and water; 761.37 km² of impervious land was transferred out and 12,184.74 km² was transferred in, mainly from cropland and water. 1,556.34 km² of forest was transferred out and 2,322.11 km² was transferred in, mainly from cropland and water; 2,111.99 km² of water was transferred out and 3,023.37 km² were transferred in, mainly from cropland and barren land. Overall, the LULC types in the Bohai Rim region changed drastically, mainly among cropland, impervious land, forest, and water.

At the landscape metrics (Figure 4), PD and SHDI both increased from 2000 to 2015 and decreased from 2015 to 2020, indicating that the degree of regional landscape fragmentation first increased and then decreased. ED and LSI increased rapidly from 2000 to 2015 and decreased from 2015 to 2020, indicating that the landscape shape first became complex and then tended to be simplified. AI and CONTAG both decreased significantly from 2000 to 2015 and increased from 2015 to 2020, indicating that the connectivity between landscape patches weakened and the separation increased in the early stage, which was alleviated in the later stage.

At the patch metrics (Figure 5), PD, ED and LSI of cropland increased, while AI decreased, indicating that cropland landscape has an obvious trend of fragmentation, complexity and dispersion. PD of forest decreased from 2000 to 2010 and increased from 2010 to 2020, its ED and LSI continued to decrease, and AI continued to increase, indicating that the degree of fragmentation in forest first decreased and then increased, with a process of simplified shape and aggregation of patches. Shrubs have fewer patches, are easily affected by external conditions, and change erratically. PD, ED, LSI and AI of grassland continued to decrease. PD, ED and LSI of water increased from 2000 to 2015, then decreased slightly from 2015 to 2020, but were still higher than the level of 2000; its AI decreased during 2000–2015 and increased significantly from 2015 to 2020. PD and LSI of barren land increased, AI decreased, ED increased during 2000–2015 and decreased to a level lower than that in 2000 during 2015–2020. PD and LSI of the impervious land continued to decrease, while ED and AI continued to increase.

The verification of the WY is based on the data from the Water Resources Bulletin. The actual water resources in the study area in 2000, 2005, 2010, 2015, and 2020 were 10.423 billion m³, 23.135 billion m³, 23.267 billion m³, 14.07 billion m³ and 23.54 billion m³ respectively. The error between the model simulation results and the actual water resources is controlled at around 10%, so the results are reliable.

The carbon density data of CS were considered in accordance with the following principles: the four basic carbon pool data were considered from published data obtained from field investigations. The missing data were used from neighboring regions. This approach and criteria ensured the credibility of the results to the greatest extent.

The assessment results obtained from the SC module were verified with reference to relevant research (Feng et al., 2022; Wang et al., 2022). Since the Bohai Rim region spans multiple watershed divisions, the results are also within the overall range. Therefore, it has a reasonable value.

The evaluation results of the HQ by province are basically consistent with the findings of Tianjin (Li et al., 2022b) and Dongying (Ding et al., 2021), with a difference of about 10%. In addition, the HQ change rate of Dongying from 2000 to 2020 is found to be consistent with the trends in the Yellow River Delta (Liu et al., 2022b), with a change rate of about 15.8%, this provides a good confidence.

The Bohai Rim region is vast, and due to differences in natural resource endowments and socio-economic factors among administrative, various ESs are highly heterogeneous and dynamic. For example, when comparing Qinhuangdao and Tianjin, the dominant ESs in Tianjin are provisioning services, while those in Qinhuangdao are regulating, supporting and cultural services (Table 6). The average CP in Qinhuangdao in the past 20 years is 80.28wt, accounting for 2.74% of the total area. While Tianjin has a greater advantage in CP, with an average of 166.95wt, accounting for 5.71%, about twice that of Qinhuangdao. However, the regulating service of Qinhuangdao is much higher than that of Tianjin, especially SC. The area of Tianjin is about 11,966.45 km² and the SC is 1632.15wt, while the area of Qinhuangdao is about 7750 km² and the SC is 7407.67wt, about 4.5 times those of Tianjin. In the past 20 years, the supporting and cultural service of Qinhuangdao have been much higher than the average level of the Bohai Rim region, while Tianjin is lower than it (Figure 9).

In the past 20 years, CP in Tianjin has continued to improve (Table 7). Compared with Qinhuangdao, the growth trend in Tianjin is stable, indicating that it has a stronger ability to resist natural risks and human interference, and is a key area for CP. WY in Tianjin has continued to decline since it increased to a maximum of 0.63 billion m³ in 2005, while the overall precipitation has not changed much from 2005 to 2020. It indicates that the WY in Tianjin is highly related to human activities. Although the amount of WY increased due to the dramatic expansion of impervious land, this high value has been offset by the reduction in cropland with similarly high WY capacity. At the same time, with the implementation of policy documents such as returning cropland to the lake, “Tianjin Land Use Master Plan (2006–2020),” “Tianjin Wetland Protection Ordinance,” “Tianjin Wetland Nature Reserve Plan (2017–2025),” Tianjin has established an eco-friendly land use pattern that coordinates rivers, beaches, bays, lakes and seas. The internal structure and spatial layout of ecological construction and wetland protection land have been optimized, and many artificial wetlands have been integrated to form a wetland ecological network with higher connectivity and better aggregation, making WY in Tianjin more balanced and reducing the risk of floods caused by urban construction.

Qinhuangdao has distinct regional characteristic and is a famous coastal touristic area, leisure and holiday resort in China. Tourism and port logistics are its traditional advantageous industries. Qinhuangdao has outstanding capabilities in regulating, supporting and cultural services, with CS, SC, HQ, and LA all at a leading level in the Bohai Rim region (Table 8). This is mainly due to the warm temperate semi-humid continental monsoon climate in this region, which is greatly affected by the ocean. The climate is mild, with little rain and dryness in spring, warm and no scorching heat in summer, cool and sunny in autumn, and long and no severe cold in winter. In addition, the northern part of Qinhuangdao is dominated by mountainous and low hills, with forests, shrubs and grasslands as its main LULC types, which have good CS, SC and HQ capabilities. However, in the past 20 years due to the development of large-scale construction projects such as Qinhuangdao Economic and Technological Development Zone, Qinhuangdao Seaport Economic Development Zone, and Qinhuangdao Port, the ESs have degraded to varying degrees, among which CS has declined by 243wt, HQ by 4.92% and LA index has increased by 4.18%. In the future, Qinhuangdao should continue to leverage its regional advantages. The southeastern coastal areas should promote the concentrated and contiguous protection of agricultural land, strictly control the occupation of cropland by new construction, and revitalize existing construction land. The northern areas should effectively protect water sources and ecologically fragile areas, conduct comprehensive management of soil erosion and land degradation to ensure the continuous supply of ESs.

The natural environmental conditions and human activities are important factors influencing the regional landscape patterns and ESs. No single development policy can improve all landscape patterns and ESs. How to resolve the contradiction between protection and development is the core issue of ecological civilization construction (Li et al., 2022a). Firstly, people should actively respond to the influence of natural factors. The Bohai Rim region has a temperate monsoon climate with uneven spatial distribution of rainfall, which is vulnerable to drought risks. Therefore, the provisioning service levels such as crop production and water yield in most areas fluctuate greatly. It is necessary to vigorously develop regional agricultural irrigation technologies and optimize agricultural irrigation strategies. The drip irrigation method will improve CP levels and save water resources (Wu, 2013). Water resources can be planned and managed to avoid waste through rainwater collection, wastewater recycling and freshwater protection. According to the characteristics of the ecosystem, sloping cropland and other cropland with low CP should be converted into forests and grasslands. In areas with heavy precipitation and strong soil erosion, forests should be considered priority, the areas with low precipitation and weak soil erosion, grasslands should be considered as priority basis. Selecting drought-resistant and water-saving species can also help alleviate water shortages. In addition, continuing to scientifically promote the implementation of water diversion projects is also one of the basic policies that local governments must adhere to, ensuring accurate and precise water diversion, refining water allocation plans, and strengthening precise scheduling from water sources to ecosystems.

Second, the irregular urban expansion should be restricted, and the urban landscape structure should be optimized. Correlativity study showed that rapid urbanization expansion and high-intensity land development are the main reasons for the ESs decline in the Yangtze River Delta urban agglomeration (Lu et al., 2024). Some scholars have compared the response of ESs in Shenzhen and Hong Kong to rural-urban transitions. And a persistent downward trend in the ESs was observed for both cities, although the decline in Shenzhen was much stronger than in Hong Kong. This was because as an island city, Hong Kong’s urban expansion was much slower compared to Shenzhen’s extremely rapid urbanization due to limitations in land availability (Xu et al., 2020). Therefore the results of ESs assessments should be incorporated into urban landscape planning. Adopting a green, low-carbon, and resource-saving urban development model. For example, when planning urban construction, attention should be paid to reducing the occupation of ecological land such as forests and grasslands, rationally constructing urban green spaces, integrating transportation networks, water networks, and ecological networks (Fu et al., 2022), configuring green infrastructure such as sponge cities to increase urban infiltration and reduce runoff (Yang et al., 2024), so as to improve WY levels, reduce land use intensity and ecological risks in urbanized areas. For urban parks and green spaces, it is necessary to arrange them in a hierarchical manner according to the spatial structure of population concentration, so as to form a complete park system, which can provide cultural services accurately while ensuring the fairness of green spaces, and ultimately promote the improvement of human wellbeing through ESs (Tu et al., 2019).

Third, people should strengthen the protection and restoration of key land types. Protect forests, shrubs, grasslands and wetlands with high regulating and supporting services, and maintain a sustainable supply of ESs by optimizing nature reserve management strategies and implementing key ecosystem protection projects (Ma et al., 2021). At the same time, strengthening the landscape connectivity of key land types, aggregating small plots into larger patches can improve the supply level of ESs. To protect cropland resources, improve cropland quality, county-level people’s governments should strengthen daily supervision of the conversion of cropland into construction land within the county area, and vigorously develop intensive cropland (Xing et al., 2020). When developing ecosystem optimization strategies, the resilience of different ecosystems should be considered (Zhao et al., 2019). For example, forest ecosystems tend to have stronger natural recovery capabilities and ecological resilience, while cropland, barren land, and wetland ecosystems are relatively fragile and less resilient (Liu et al., 2018; Seidl et al., 2014). Conducting refined design in this way will be more conducive to achieving sustainable management of regional ecosystems.

This study uses InVEST model to evaluate ESs, the quality of evaluation depends on the accuracy of the LULC data classification. However, the accuracy of the Wuhan University CLCD dataset used in this paper is only 79.31%, and no more detailed secondary classification was performed when using it due to the difficulty in obtaining other parameters. In the future research, more detailed LULC classification and higher resolution, more accurate LULC datasets should be used.

In addition, this study only evaluated six key ESs that dominate the Bohai Rim region. Due to the large scope of study area, it cannot fully represent the overall ecological status of the region. In the future, scholars can add ESs that reflect regional characteristics, such as aquaculture and water purification. Moreover, the use of LA to characterize cultural services is usually limited to specific locations (Dou et al., 2022), and it is difficult to accurately reflect their overall situation. In the future, when conducting cultural service assessments, it is necessary to comprehensively select assessment indicators based on research needs and data availability. Finally, although the assessment results have been validated by other studies, the lack of measured data may still lead to certain uncertainties in the results. Future studies can support the assessment results by obtaining observational data from field surveys.