Multi-Scale Analysis of Spatial and Temporal Evolution of Ecosystem Health in the Harbin–Changchun Urban Agglomeration, China

Abstract

Urban agglomerations represent the pinnacle of spatial organization for fully developed cities. Gaining insight into the ecosystem health status of urban agglomerations in various geographical and temporal settings is essential for the long-term sustainability of both humans and the environment. Nevertheless, current research overlooks the impacts of human activities on the well-being of ecosystems, along with the effects of scaling and their implications for ecological management policies and future urban growth plans. This study enhances and refines the classic model and establishes the Vigor Organization Resilience Services Human activities (VOR-SH) evaluation model to assess the ecosystem health of the Harbin–Changchun urban agglomeration on three scales. The results reveal that the changes in the five indicators of ecosystem health within the Harbin–Changchun urban agglomeration differed across three unique periods from 2000 to 2020. In particular, energy, organization and human activities increased, whereas resilience and ecosystem services decreased. On all three scales, the overall ratings for ecosystem health showed improvement. Multi-scale spatial autocorrelation showed strong positive spatial correlations between ecosystem health clusters in the studied area. Multi-scale review results help locate key regions on a grid scale, coordinate regional management at the district-county scale and monitor huge ecosystems at the city scale. This study improves the ecosystem health model and expands multi-scale regulatory theory. This study’s findings help guide urban expansion and environmental management.

1. Introduction

Over the past few years, the growth of cities and the concentration of human activities have resulted in various ecological and environmental issues. These include the urban heat island effect, the fragmentation of landscapes and a significant decline in biodiversity. These problems are not limited to individual cities but also have negative impacts on nearby regions and cities [1], posing a significant threat and challenge to the overall health of ecosystems [2]. Promoting the healthy development of ecosystems is a widely accepted worldwide objective; consequently, conducting an exhaustive assessment of ecological health is crucial for regional progress [3]. Hence, evaluating ecosystem health can facilitate the harmonious advancement of ecosystems and socio-economics. Assessing ecosystems from a multi-scale perspective allows for the analysis of cascade effects, providing a scientific foundation for coordinating regional ecological security and economic growth [4]. Moreover, it serves as a theoretical basis for implementing the construction of an ecological civilization and promoting green sustainable development in China’s new era.

Ecosystem health assessments provide a comprehensive evaluation of the stability and potential for the sustainable development of regional ecosystems. They assess the ecosystems’ ability to maintain internal structural stability, self-regulation mechanisms, resilience to adversity and service functions necessary for sustainable human development [5]. Currently, ecosystem health assessments are extensively employed in the initial evaluation and impact assessment of ecological restoration, urban planning, urban renewal and other related fields [2,6]. The primary objective of ecosystem health assessment research is to develop a comprehensive, precise and local framework and research system for assessing ecosystem health. There are three primary types of evaluation models. The first type is the Vigor Organization Resilience (VOR) model, which can measure ecological integrity and ecosystem quality [7]. The second type is the Pressure State Response (PSR) model [8], which focuses on the impact of human social activities on the natural environment. Additionally, there are the Driver State Response (DSR) model [9] and the Driver Pressure State Impact Response (DPSIR) model [10], which consider factors such as the driving force. The third category classifies and evaluates the health of ecosystems based on ecological, social and economic factors [11]. The VOR model has become frequently used in evaluating regional natural ecosystems due to its ability to capture the fundamental essence of regional ecosystem health and effectively measure ecosystem integrity and vigor [12]. Over time, this model has evolved to encompass larger territorial spaces, such as administrative districts [13], economic zones [14], national parks [15] and urban agglomerations [16], enabling comprehensive evaluations and dynamic analyses. Existing evaluation frameworks offer multiple perspectives on the health of ecosystems. However, current research primarily concentrates on the essential significance and unity of ecosystems or examines their drivers and impact mechanisms [17]. Few studies have explored the interplay between human activities and ecosystems to develop a comprehensive framework that incorporates the assessment of ecosystem integrity, stability and the capacity of natural ecosystems to support human activities. Simultaneously, larger study areas exhibit a deficiency in the internal spatial heterogeneity of regional ecosystems, necessitating an analysis of interactions and spatial limitations between regional ecosystems inside urban areas [18].

Situated strategically within China’s “two horizontal and three vertical” urbanization strategic pattern, the Harbin–Changchun urban agglomeration plays a crucial role in advancing the development of new-type urbanization and ensuring the balanced spatial growth of China’s national territory [19]. The Harbin–Changchun urban agglomeration is a regional urban agglomeration characterized by a complex internal pattern. The rapid development of the regions within this urban agglomeration has disrupted the balanced state of human–land relations, leading to increased pressure on the ecological environment and posing a significant threat to the ecosystem health of the cities [20]. The majority of previous studies on assessing the health of larger-scale territorial spaces have primarily focused on economic changes over time [21], the value of the ecosystem [22], ecological vulnerability [23] and other individual aspects. However, there is a lack of comprehensive assessment analyses that can effectively support the sustainable development of the territorial space. Additionally, the complex relationship between the urbanization processes of different cities on a regional scale has not been adequately considered [24]. Consequently, to fully analyze the state of ecosystem health in the Harbin–Changchun urban agglomeration, this study focuses on two novel research areas: enhancing the evaluation framework for ecological model assessments and conducting a multi-scale assessment of the urban agglomerations and their spatial self-assessment. First, the Vigor Organization Resilience Services Human activities (VOR-SH) assessment model is proposed, which considers the influence of human production and lifestyle on the overall health of the ecosystem. It also improves the comprehensiveness of the ecosystem health evaluation by incorporating factors related to ecosystem services and environmental characteristics that are closely linked to human activities. Furthermore, an evaluation and analysis of alterations in the internal composition of ecosystems and their underlying factors are conducted across three levels of analysis (grid scale, district-county scale and city scale) inside the Harbin–Changchun urban agglomeration between the years 2000 and 2020. The study explores the spatial self-assessment and coupling effects, and the findings reflect strategies for ecological protection and rapid urban development across all levels within urban agglomerations over 20 years.

This project aims to address the following issues: (1) It improves the reliability of the ecosystem health assessment model by incorporating both natural and human influences. (2) This study focuses on the variations in ecosystem health across urban agglomerations, considering both spatial and temporal factors. It explores the management of human activities that disrupt ecosystems and the restoration of natural elements. Additionally, it provides recommendations for improving and optimizing ecological health. This study’s findings alleviate the conflict between ecological preservation and economic growth in urban agglomerations. Additionally, they can identify crucial places for ecological protection and guide future ecological restoration efforts. Simultaneously, it serves as a crucial foundation and point of comparison for investigating the quantitative evaluation and accurate improvement of the health of regional ecosystems.

2. Materials and Methods

2.1. Study Area

The Harbin–Changchun urban agglomeration (122°24′–131°18′ E, 42°00′–48°55′ N) is situated in the Heilongjiang and Jilin Provinces in Northeast China. It is located at the northern extremity of the longitudinal axis of the Beijing–Harbin–Guangzhou corridor, which is part of the national “Two Horizontal and Three Vertical” urbanization strategy pattern. It serves as a crucial center for the rejuvenation and advancement of the former industrial hub in northeastern China, a significant entry point for the expansion of the northern region, a pioneering area for the innovation of the old industrial base’s system and mechanism, and a cluster of environmentally friendly cities. The Harbin–Changchun urban agglomeration encompasses the cities of Harbin, Daqing, Qiqihar, Suihua and Mudanjiang in Heilongjiang province, as well as Changchun, Jilin, Siping, Liaoyuan, Songyuan and Yanbian Korean Autonomous Prefecture in Jilin province. It spans a total area of approximately 263,640.92 km² (Figure 1). The research area is characterized by a combination of flat plains and elevated mountains, with a significant disparity between the eastern and western landscapes. Additionally, there is a noticeable pattern of higher elevation in the southeast and lower elevation in the northwest. The eastern portion of the study area is primarily occupied by the Changbai Mountains, which are known for their forested mountain terrain in the east and low hills in the east-central region. In contrast, the western portion mainly comprises the Nenjiang Plain, which features agricultural land in the central area and grassland in the western region. The Harbin–Changchun urban agglomeration, being the first among China’s second-class urban agglomerations, has experienced significant economic and social growth in recent years, contributing significantly to its overall development. Therefore, it is vital to grasp the degree of ecosystem health and the course of ecosystem health in different times and spaces.

2.2. Data Sources

The accuracy and sources of administrative boundaries, Net Primary Productivity (NPP), Digital Elevation Model (DEM), annual evapotranspiration, annual precipitation, soil moisture content, plant root depth and road distribution data for the years 2000, 2010 and 2020 used in this study are shown in

Table 1. Study data sources.

Data	Data Type/Resolution	Source
Administrative boundary	Vector data	The Resource and Data Sharing Center of the Chinese Academy of Sciences (https://www.resdc.cn, accessed on 10 April 2023)
Land use data	Raster data/30 m
Net Primary Productivity (NPP)	Raster data/1000 m
Digital Elevation Model (DEM)	Raster data/90 m	Geospatial data cloud (https://www.gscloud.cn, accessed on 28 April 2023)
Annual precipitation	Raster data/1000 m	China Meteorological Data Network (http://data.cma.cn, accessed on 5 March 2023)
Annual evapotranspiration	Raster data/500 m
Soil moisture content	Raster data/1000 m	National Earth System Science Data Center (http://www.geodata.cn, accessed on 10 May 2023)
Plant root depth	Raster data/1000 m
Road distribution	Vector data	Open Street Map (https://www.openstreetmap.org, accessed on 10 May 2023)

. The data were preprocessed using ArcGIS10.8 and fragstats4.2 software. ArcGIS10.8 was applied to uniformly project Krasovsky_1940_Albers coordinates, and the raster data were uniformly resampled to a 30 m resolution using the Nearest Neighbor Allocation Method (NEAREST) in this software [25,26].

2.3. Ecosystem Health Assessment System Construction

2.3.1. Indicators for Ecosystem Health Assessment

Through an analysis of the current conditions of the study area and previous studies on ecosystem health evaluations [27,28], we enhanced the traditional VOR model [29] by incorporating two additional evaluation metrics: ecosystem services and human activities. This new evaluation system, called VOR-SH, considers the influence of human production and lifestyle choices on the overall health of the ecosystem. This study aimed to enhance the comprehensiveness of the ecosystem health evaluation system by incorporating 14 assessment factors across five indicators: ecosystem vigor, ecosystem organization, ecosystem resilience, ecosystem services and human activities (Figure 2).

1.

Ecosystem vigor. Ecosystem vigor refers to the overall health and vitality of an ecosystem, which is determined by factors such as its metabolism, capacity for net primary production and richness of biological resources [30,31]. For the purpose of assessing ecosystem vigor, this study specifically chose NPP and biological abundance as two essential evaluation variables. NPP data provide information on the quantity of organic matter produced by biological populations, biomes and ecosystems over a specific period and area. On the other hand, the bioconcentration indicator reflects the level of biodiversity and resource abundance within the ecosystem. The determination of the biological abundance indicator was based on the guidelines outlined in the “Technical Specification for Evaluation of Ecological Environment Condition (Trial) HJ/T192-2006” [32].

$$\begin{gathered} {\mathrm{A}}_{\mathrm{biological}\mathrm{abundance}}={\mathrm{A}}_{\mathrm{b}\mathrm{i}\mathrm{o}}\times (0.35\times \mathrm{F}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{s}+0.21\times \mathrm{G}\mathrm{r}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{d}+ \\ 0.2\times \mathrm{Water}\mathrm{bodies}+0.11\times \mathrm{Cultivated}\mathrm{land}+0.04\times \mathrm{Construction}\mathrm{land}+ \\ 0.01\times \mathrm{Unused}\mathrm{land})/\mathrm{Total}\mathrm{area} \end{gathered}$$

(1)

where ${\mathrm{A}}_{\mathrm{b}\mathrm{i}\mathrm{o}}$ represents the normalized index of biological abundance and ${\mathrm{A}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$ represents the maximum value of the biological abundance index before normalization treatment.

2.

Ecosystem organization. Ecosystem organization pertains to the stability, complexity and structure of ecosystems [33,34,35]. It is mostly analyzed statistically using Landscape Pattern Index (LPI) indicators, which assess landscape heterogeneity and connectedness [17,36]. The representation of landscape heterogeneity primarily involves the use of two metrics: Area-Weighted Patch Fractal Dimension (AWMPFD) and Shannon diversity (SHDI) [37]. Landscape connectivity can be categorized into two components. The first component involves assessing the overall landscape connectivity using the selected landscape fragmentation index (PD) and landscape spreading index (CONTAGE) [38]. The second component focuses on the connectivity of important ecological patches, such as woodland and grassland, which play significant roles in carbon sequestration, air purification, soil retention, water containment and biodiversity preservation. Consequently, the assessment index incorporates the landscape fragmentation indices (PD2, PD3) and patch aggregation indices (CONHESION2, CONHESION3) of woodland and grassland as crucial components. AWMPFD is employed to assess the intricacy of landscape structure, particularly about the shape and arrangement of patches. SHDI measures the variation in the landscape and is especially responsive to the uneven distribution of different types of patches across the terrain. It highlights the significance of rare patch types in providing valuable information. PD primarily indicates the concentration of a specific area within the landscape, serving as an indicator of the landscape’s overall diversity and fragmentation, as well as the level of fragmentation within a particular patch type. CONTAG is a term that quantifies the level of resemblance between adjacent patches or habitats within a landscape. CONHESION is a metric that quantifies the level of cohesiveness or connectedness among patches within a landscape. Citing prior research findings [39,40], the weights for landscape heterogeneity, overall landscape connectedness and the connectivity of significant ecological patches were allocated as 0.35, 0.35 and 0.3, respectively. The calculation of ecological organizing force was determined using the following formula:

EO = 0.35LH + 0.35LC + 0.3IPC = (0.25SHDI + 0.1 AWMPFD) + (0.25PD + 0.1CONTAGE) + (0.1PD₂ + 0.1PD₃ + 0.05CONHESION₂ + 0.05CONHESION₃)

(2)

where LH is landscape heterogeneity; LC is overall landscape connectivity; and IPC is ecologically important patch connectivity.

3.

Ecosystem resilience. Ecosystem resilience is the capacity of an ecosystem to withstand and maintain its structure and function in the face of external disturbances. It also includes the ability to recover to its original state after being disrupted. This study evaluated ecosystem resilience by measuring resilience and resistance [36,41,42], which are typically quantified using the resistance coefficient (${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}}$) and the resilience coefficient (${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}}$). The resistance coefficients and resilience coefficients of various land types were assigned values based on the contribution degree of different land use types, as indicated by the associated literature [36,39] (

Table 2. Table of ecosystem resistance and resilience assignment for different land use types.

Land Use Type	Cultivated Land	Forests	Grassland	Water Bodies	Construction Land	Unused Land
${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}}$	0.5	1	0.7	0.8	0.3	0.2
${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}}$	0.4	0.5	0.8	0.7	0.2	0.1

Where ${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}}$ represents the resistance coefficient, and ${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}}$ represents the resilience coefficient.

). The exact weights in this study were chosen based on an assessment of whether the external disturbance surpassed the ecosystem’s capacity for recovery. The ecosystem resilience weights were prioritized over the resistance weights due to the significant need for production and development in the Harbin–Changchun urban agglomeration, severe industrial impacts and the intensity of human intervention [43]. The specific formula for ecosystem resilience composition is as follows [44]:

$$\mathrm{E}\mathrm{R}=0.4\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{A}}_{\mathrm{i}}\times {\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}}{\mathrm{C}}_{\mathrm{i}}+0.6\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{A}}_{\mathrm{i}}\times {\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}}{\mathrm{C}}_{\mathrm{i}}$$

(3)

where ER represents ecosystem resilience; ${\mathrm{A}}_{\mathrm{i}}$ represents the area ratio of land use type i; ${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}}{\mathrm{C}}_{\mathrm{i}}$ represents the resistance coefficient of land use type i; ${\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}}{\mathrm{C}}_{\mathrm{i}}$ represents the resilience coefficient of land use type i; and n is the number of land use types.

4.

Ecosystem services. Ecosystem services refer to the capacity of natural ecosystems to offer ecological services and benefits to humans, which fulfill human requirements for ecosystems [14,24]. This study selected five exemplary services, namely soil conservation functions, carbon sequestration functions, water supply functions, biodiversity functions and food supply functions. The selection was based on an analysis of the current ecological environment and the production and living patterns of the residents in the study area [22,45]. The quantification of these five services was carried out using the InVEST 3.13.0 model [46]. The precise computations are displayed in

Table 3. Methods for quantifying ecosystem services.

Ecosystem Services	Meaning	Formula
Water supply service	Water supply services play a crucial role in influencing ecological functions, such as biomass and carbon cycling, within ecosystems [46,47].	${\mathrm{W}\mathrm{Y}}_{\left(\mathrm{x}\right)}=\left(1-\frac{{\mathrm{A}\mathrm{E}\mathrm{T}}_{\left(\mathrm{x}\right)}}{{\mathrm{P}}_{\left(\mathrm{x}\right)}}\right)\times {\mathrm{P}}_{\left(\mathrm{x}\right)}$ where ${\mathrm{W}\mathrm{Y}}_{\left(\mathrm{x}\right)}$ is the average annual water production; ${\mathrm{A}\mathrm{E}\mathrm{T}}_{\left(\mathrm{x}\right)}$ is the actual evapotranspiration on image x; ${\mathrm{P}}_{\left(\mathrm{x}\right)}$ is the average annual precipitation on image x.
Soil conservation service	Soil conservation services (SDR) refer to the capacity to preserve and control soil, which is crucial for mitigating regional erosion and minimizing the likelihood of floods [48,49].	$\mathrm{S}\mathrm{C}=\mathrm{R}\mathrm{K}\mathrm{L}\mathrm{S}-\mathrm{R}\mathrm{U}\mathrm{L}\mathrm{S}\mathrm{E}$                   $\mathrm{R}\mathrm{K}\mathrm{L}\mathrm{S}=\mathrm{R}\times \mathrm{K}\times \mathrm{L}\mathrm{S}$                 $\mathrm{R}\mathrm{U}\mathrm{L}\mathrm{S}\mathrm{E}=\mathrm{R}\times \mathrm{K}\times \mathrm{L}\mathrm{S}\times \mathrm{C}\times \mathrm{P}$ where SC is the actual soil conservation; RKLS is the potential soil erosion; and RULSE is the actual soil erosion. r is the rainfall erodibility factor; K is the soil erodibility factor; and LS is the slope length factor. c is the vegetation management factor, and P is the factor of soil and water conservation measures.
Biodiversity service	Habitat quality is a crucial factor in determining the capacity to sustain regional ecological diversity [50,51].	${\mathrm{Q}}_{\mathrm{x}\mathrm{j}}={\mathrm{H}}_{\mathrm{x}\mathrm{j}}(1-\frac{{\mathrm{D}}_{\mathrm{x}\mathrm{j}}^{\mathrm{z}}}{{\mathrm{D}}_{\mathrm{x}\mathrm{j}}^{\mathrm{z}}+{\mathrm{K}}^{\mathrm{z}}})$ ${\mathrm{Q}}_{\mathrm{x}\mathrm{j}}$ is the habitat quality of image element x in the jth land use type; ${\mathrm{H}}_{\mathrm{x}\mathrm{j}}$ is the habitat suitability of image element x in the jth land use type; ${\mathrm{D}}_{\mathrm{x}\mathrm{j}}^{\mathrm{z}}$ is the habitat degradation of raster x in the jth land use type; z is a normalization constant, which is usually taken to be the default value of 2.5; and k is a half-saturation constant [52].
Carbon sequestration service	Carbon sequestration, also known as carbon reserves, often refers to the quantity of carbon stored in various carbon reservoirs such as forests, oceans and land. The concept has garnered more interest following China’s introduction of the “dual carbon targets” program in 2020 [53].	$\mathrm{C}\mathrm{S}={\mathrm{C}\mathrm{S}}_{\mathrm{a}\mathrm{b}\mathrm{o}\mathrm{v}\mathrm{e}}+{\mathrm{C}\mathrm{S}}_{\mathrm{b}\mathrm{e}\mathrm{l}\mathrm{o}\mathrm{w}}+{\mathrm{C}\mathrm{S}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}+{\mathrm{C}\mathrm{S}}_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}$ where CS is the total carbon stock; ${\mathrm{C}\mathrm{S}}_{\mathrm{a}\mathrm{b}\mathrm{o}\mathrm{v}\mathrm{e}}$ is the aboveground biogenic carbon; ${\mathrm{C}\mathrm{S}}_{\mathrm{b}\mathrm{e}\mathrm{l}\mathrm{o}\mathrm{w}}$ is the belowground biogenic carbon; ${\mathrm{C}\mathrm{S}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}$ is soil carbon; ${\mathrm{C}\mathrm{S}}_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}$ is soil carbon; and CS_dead is dead organic carbon [54].
Food supply service	Food supply is essential for human survival and ensures the sustainability of regional development.	$$\begin{gathered} {\mathrm{Y}}_{\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{G}\mathrm{C}}=\mathrm{m}\mathrm{i}\mathrm{n}({\mathrm{Y}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\left(1-{\mathrm{b}}_{\mathrm{N}\mathrm{P}}\exp \left(-{\mathrm{c}}_{\mathrm{N}}{\mathrm{N}}_{\mathrm{G}\mathrm{C}}\right)\right), \\ {\mathrm{Y}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\left(1-{\mathrm{b}}_{\mathrm{N}\mathrm{P}}\exp \left(-{\mathrm{c}}_{\mathrm{P}}{\mathrm{P}}_{\mathrm{G}\mathrm{C}}\right)\right), \\ {\mathrm{Y}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\left(1-{\mathrm{b}}_{\mathrm{N}\mathrm{P}}\exp \left(-{\mathrm{c}}_{\mathrm{N}}{\mathrm{K}}_{\mathrm{G}\mathrm{C}}\right)\right) \end{gathered}$$ where ${\mathrm{N}}_{\mathrm{G}\mathrm{C}}$ represents the rate of nitrogen application provided; ${\mathrm{P}}_{\mathrm{G}\mathrm{C}}$ represents the rate of phosphorus application provided; and ${\mathrm{K}}_{\mathrm{G}\mathrm{C}}$ represents the rate of potassium application provided [55].

.

5.

Human activities. An urban agglomeration represents the most advanced level of spatial organization in the latter phase of urban development. Its progress is closely intertwined with human production activities, leading to increasingly evident conflicts between humans and the environment [56]. Evaluating human activities is a crucial component of assessing ecological health. The utilization of the human disturbance index allows for the quantification of human activities in a given area, providing a direct measure of the extent to which human presence affects the natural environment. Moreover, changes in land use within the human disturbance index serve as a direct and effective indicator of the kind and intensity of human activities per unit area. Hence, this study investigated the precise effects of human activities on the study area, focusing on alterations in land use [57,58]. The formula is as stated:

$$\mathrm{H}\mathrm{A}=\frac{{\mathrm{A}}_1+{\mathrm{A}}_2}{\mathrm{C}\mathrm{A}}$$

(4)

where HA is human activities; ${\mathrm{A}}_1$ is the total area of cultivated land in the study area; ${\mathrm{A}}_2$ is the total area of construction land in the study area; and CA is the total area of the study area.

2.3.2. Ecosystem Health Evaluation Index Weights and Evaluation System Construction

Analytic Hierarchy Process (AHP) analysis was used to identify the weights and create the indicator discriminant matrix for a consistency test. This helped define the weights of the evaluation criteria in the system, as shown in

Table 4. Ecosystem health evaluation factor weights.

Estimation Target	Evaluation Indicators	Indicator Weights	Evaluation Factors	Positive or Negative	Factor Weights
Ecosystem health evaluation factors	Ecosystem vigor (EV)	0.23	NPP (Net Primary Productivity)	+	0.115
			Biological Abundance	+	0.115
	Ecosystem organization (EO)	0.35	Area-Weighted Patch Fractal Dimension (AWMPD)	+	0.035
			Shannon’s Diversity Index (SHDI)	+	0.0875
			Patch Density (PD)	-	0.0875
			Forests Patch Density (${\mathrm{P}\mathrm{D}}_2$)	-	0.035
			Grassland Patch Density (${\mathrm{P}\mathrm{D}}_3$)	-	0.035
			Contagion Index (CONTAGE)	+	0.035
			Forests Patch Cohesion Index (${\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{H}\mathrm{E}\mathrm{S}\mathrm{I}\mathrm{O}\mathrm{N}}_2$)	+	0.0175
			Grassland Patch Cohesion Index (${\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{H}\mathrm{E}\mathrm{S}\mathrm{I}\mathrm{O}\mathrm{N}}_3$)	+	0.0175
	Ecosystem resilience (ER)	0.18	Resistance $({\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}}$)	+	0.072
			Resilience $({\mathrm{R}}_{\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}}$)	+	0.108
	Ecosystem services (ES)	0.14	Water supply service	-	0.0308
			Soil conservation service	+	0.021
			Biodiversity service	+	0.035
			Carbon sequestration service	+	0.035
			Food supply service	+	0.0182
	Human activities (HA)	0.1	Human activities	-	0.1

.

To address the issue of varying scales for each indicator, the 14 assessment factors within the five indicators of vigor, organization, resilience, ecosystem services and human activities were initially standardized in the calculation. Subsequently, the comprehensive evaluation of ecosystem health was determined by overlaying these factors based on their calculated weights.

2.4. Indicators for Ecosystem Health Assessment

Empirical research has demonstrated that spatial correlation is present among entities as a result of spatial interactions and diffusion. Furthermore, entities in closer proximity exhibit a higher degree of correlation [59]. An analysis of spatial autocorrelation was conducted to examine the geographical distribution characteristics of ecosystem health, ranging from a global to a local scale [60]. This study utilized the global Moran’s I index in Arc GIS10.8 Spatial Statistics Tools to assess the geographical correlation and variance within the study area. The objective was to determine if the area demonstrates a spatial inclination toward clustering or dispersal. The ${\mathrm{G}\mathrm{e}\mathrm{t}\mathrm{i}\mathrm{s}-\mathrm{O}\mathrm{r}\mathrm{d}\mathrm{G}}_{\mathrm{i}}^{\mathrm{*}}$ index in Arc GIS10.8 Spatial Statistics Tools was utilized to detect the local spatial grouping of major hot spots and cold spots [2]. The hot spot clusters represent regions with higher concentrations of ecosystem health, whereas the cold spot clusters represent regions with lower concentrations of ecosystem health. The calculating formula is as follows:

$$\text{Moran’s} \mathrm{I}=\frac{\sum _{\mathrm{i}=1}^{\mathrm{n}}\sum _{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{i}\mathrm{j}}({\mathrm{x}}_{\mathrm{i}}-\overline{\mathrm{x}})({\mathrm{x}}_{\mathrm{j}}-\overline{\mathrm{x}})}{S^2\sum _{\mathrm{i}=1}^{\mathrm{n}}\sum _{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{i}\mathrm{j}}}$$

(5)

where n is the total number of study units in space; ${\mathrm{x}}_{\mathrm{i}}$ and ${\mathrm{x}}_{\mathrm{j}}$ represent the ecosystem health values of region i and region j, respectively; ${\mathrm{W}}_{\mathrm{i}\mathrm{j}}$ denotes the spatial matrix of regions i and j; $S^2$ the ensemble of the factor weights; and $\overline{\mathrm{x}}$ denotes the average value of ecosystem health.

$${\mathrm{G}}_{\mathrm{i}}^{\mathrm{*}}\left(\mathrm{d}\right)=\frac{\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{i}\mathrm{j}}\left(\mathrm{d}\right) {\mathrm{W}}_{\mathrm{i}}}{\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{x}}_{\mathrm{i}}}$$

(6)

where ${\mathrm{x}}_{\mathrm{i}}$ is the observed value of study unit i and ${\mathrm{W}}_{\mathrm{i}\mathrm{j}}\left(\mathrm{d}\right)$ is the spatial weight matrix.

3. Analysis of Results

3.1. Analysis of Land Use Transfer in the Harbin–Changchun Urban Agglomeration

The primary land use type in the Harbin–Changchun urban agglomeration is cultivated land, with forests being the second most prevalent. Throughout the study period, the cultivated land area in the Harbin–Changchun urban agglomeration was relatively constant and generally stable. However, the forest area experienced a continuous decline, resulting in a total area of 1076.296 km². The forest area experienced a consistent decline, with a total reduction of 1076.296 km². The grassland area showed an initial increase followed by a decrease, but the overall trend was a decrease, with a total reduction of 558.711 km². Water bodies and unused land areas underwent significant changes, with a continuous decline in water bodies, resulting in a total reduction of 3,894.830 km². Conversely, the area of unused land areas exhibited a continuous increase, with a total reduction of 3,414.040 km². As a result of ongoing industrialization and urbanization, the percentage of land dedicated to construction is growing. Regarding land transfer, arable land has contributed the largest proportion of its area to forest land, construction land and grassland, accounting for 44.21%, 30.66% and 10.53% respectively. Forest land, on the other hand, has transferred 70.62% of its area to arable land, 15% to cultivated land and 15% to grassland. The amount of arable land that was converted into construction land reached a substantial 107.674 km² (

Table 5. Harbin–Changchun urban agglomeration land use transfer matrix, 2000–2020 (km²).

	2020 (km2)	Cultivated Land	Forests	Grassland	Water Bodies	Construction Land	Unused Land	2000 Total Area	Transfer Out Area
2000 (km2)
Cultivated land		140,181.760	4392.746	1046.398	630.586	3046.380	819.431	150,117.301	9935.541
Forests		4979.202	108,783.802	1057.538	256.829	168.773	588.403	115,834.547	7050.745
Grassland		1566.886	939.013	13,170.586	76.353	107.674	1023.592	16,884.104	3713.518
Water bodies		866.623	202.489	187.919	6516.404	45.977	3841.305	11,660.718	5144.314
Construction land		1755.119	128.356	41.268	35.455	9350.238	27.509	11,337.944	1987.706
Unused land		1403.615	311.845	821.685	250.260	98.795	13,695.919	16,582.118	2886.199
2020 Total area		150,753.204	114,758.251	16,325.393	7765.888	12,817.837	19,996.159	322,416.731
Transfer in area		10,571.444	5974.449	3154.807	1249.484	3467.599	6300.239

and Figure 3).

3.2. Assessment of Multi-Scale Ecosystem Health Indicators in the Harbin–Changchun Urban Agglomeration

3.2.1. Multi-Scale Assessment of the Spatial and Temporal Evolution of Ecosystem Vigor

This study categorized ecosystem vigor into five levels using the natural breakpoint method: Level I (>0.741), Level II (0.611~0.74), Level III (0.466~0.611), Level IV (0.333~0.46) and Level V (0~0.333).

Table 6. Mean values of ecosystem vigor of the Harbin–Changchun urban agglomeration at different scales, 2000–2020.

Year	2000	2010	2020	2000–2020 Mean Change
Grid scale	0.455769	0.491964	0.496651	0.040882
District-county scale	0.455935	0.491713	0.496393	0.040458
City scale	0.455859	0.491684	0.496363	0.040504

displays the average values of ecosystem vigor in the Harbin–Changchun urban agglomeration. In 2000, the Harbin–Changchun urban agglomeration had a mean level of IV for ecosystem vigor across the grid, district-county and city scales. However, in 2010 and 2020, the mean values of ecosystem vigor decreased to Level III across the same scales. This suggests that the overall level of ecosystem vigor in the Harbin–Changchun urban agglomeration has been consistently increasing over the years. Nevertheless, the augmentation is particularly evident at the grid level.

The analysis of the Harbin–Changchun urban agglomeration from 2000 to 2020 revealed the proportion and arrangement of various levels of ecological vitality grading within the area (Figure 4 and Figure 5). At the grid scale, Level V had the greatest area, whereas Level III had the smallest area. The percentage of Level I regions experienced a notable increase annually, whereas those of Level II and Level III regions consistently declined. The Level IV area saw an initial decline followed by a subsequent increase, displaying an overall rising trajectory. The Level V area initially had a decline, followed by a rise, but the general tendency was a fall. The Level IV region saw an initial decrease followed by an increase, displaying an overall rising trajectory. The Level V area initially had a decline, followed by a rise, but the overall trend indicated a reduction. The rise in level I primarily occurred in the eastern forest zone and was attributed to the successful enforcement of ecological regulations, such as the conversion of farms back into forests. Level V in the research region was predominantly found in the eastern grassland landscape and the central agriculture landscape. At the district-county scale, the region classified as Level V had the greatest extent, whereas the region classified as Level I had the smallest extent. The Level I region had slight growth, but the Level V area underwent a decrease. The areas of Levels II, III and IV had fluctuations, initially increasing and subsequently decreasing, but with an overall upward tendency. At the scale of cities, between the years 2000 and 2010, the levels of vigor were mostly characterized by Levels V and II, with the absence of Level I. However, in 2020, Level III became the dominant level, with the disappearance of Levels I and II. Within this group, the size of Level II reduced, the size of Level III increased greatly, the size of Level IV initially increased and then fell, and the overall size decreased. Additionally, the size of Level V increased and then decreased, with no change in the overall size. The process of urbanization has resulted in a reduction in green areas in the study region, causing a significant decline in the overall vitality at the municipal level. The enlargement of the observable range conceals the regions with the highest and lowest levels of intensity, resulting in a standardization of their degrees of intensity.

An analysis was conducted to examine the variations in ecosystem vigor levels between 2020 and 2000. This analysis was performed at three different scales, and it was found that the most noticeable changes occurred at the grid scale. The vigor value of the Little Xing’an Mountains Nature Reserve in the northeastern region of the country increased at the grid scale. Conversely, the vigor values of the Songhua River Basin and Nenjiang River Basin declined. The locations exhibiting a decline in vigor value indices at the district-county scale were primarily concentrated in the metropolitan regions of major cities, including Daoli District in Harbin City and Nanguan District in Changchun City, among others. The levels of ecological vigor at the municipal level were primarily concentrated in the metropolitan regions of big cities. Overall, the ecosystem vigor values increased at the city scale, with a greater increase observed in the vitality index of the cities around Harbin City.

3.2.2. Multi-Scale Assessment of the Spatial and Temporal Evolution of Ecosystem Organization

The assessments of ecological organization evolution in this study were categorized into five levels using the natural breakpoint method. These levels are as follows: Level I (>0.732), Level II (0.684~0.732), Level III (0.628~0.684), Level IV (0.521~0.628) and Level V (0~0.521). The

Table 7. Mean values of ecosystem organization of the Harbin–Changchun urban agglomeration at different scales, 2000–2020.

Year	2000	2010	2020	2000–2020 Mean Change
Grid scale	0.666987	0.668347	0.673743	0.006756
District-county scale	0.668996	0.670291	0.67568	0.006684
City scale	0.668996	0.670291	0.67568	0.006684

displays the average values of ecosystem capacity in the Harbin–Changchun urban agglomeration. From 2000 to 2020, the mean value of the ecological organizational force index climbed progressively across all three dimensions, consistently reaching the highest level of ecological organizational force. The ecological organization capacity index showed a faster increase at both the district-county and city scales.

The proportions and spatial distribution of ecological organization hierarchy at various scales in the Harbin–Changchun urban agglomeration from 2000 to 2020 are illustrated in Figure 6 and Figure 7. The spatial distribution of the organization level within the region was more evident at the grid scale. The spatial extent of Level I, Level II and Level V organization expanded, but the spatial extent of the other levels contracted, with a notable expansion in Level I. Level I ecosystem organization was primarily found in the western grassland landscape area and the central and eastern forests area. Through ecological management practices such as converting farmland to grassland and converting farmland to forests, the connectivity of ecological patches in the region was enhanced. In the central region, the shift in land use to construction land, driven by the need for regional development, resulted in a reduction in organization within this area. The district-county scale was primarily characterized by a prevalence of Level III organization, and Level V was becoming obsolete. Level I ecosystem organization had the lowest area, but it progressively expanded each year. Level II organizational power expanded, whereas Level III ecosystem organization shrank. Level IV organization, on the other hand, had an initial increase followed by a fall. The enhancement of organizational prowess at this scale exhibited a focused pattern, facilitating adjacent districts and counties to further elevate their organizational prowess. For instance, the primary urban area of Daqing City and its surrounding regions, as well as the area of Yanshou County and its surrounding regions in Harbin City, signify that the ecosystem structure exhibited spatial spillovers. The changes in organizational levels at the city scale were not readily apparent. However, it is noteworthy that Level III organization was predominant, with only Daqing and Songyuan cities possessing Level II ecosystem organizational levels.

We conducted an analysis of the fluctuations in the amount of ecosystem organization power between the years 2000 and 2020. The most prominent alterations were found at the grid scale when comparing the differences in changes. The ecological organization values of Little Xing’an Mountains Nature Reserve and Zhangguancailin Nature Reserve at the grid scale showed a considerable rise in the study region. At the district-county scale, the organization experienced growth in most places; however, the organization’s strength values notably declined in Dongfeng County, Yushu City, Yuhua City and Liaoyuan City. Only Liaoyuan City had a decline in organization at the city level, whereas the other regions demonstrated a consistent upward trend.

3.2.3. Multi-Scale Assessment of the Spatial and Temporal Evolution of Ecosystem Resilience

The ecosystem resilience values were categorized into five levels using the natural breakpoint method: Level I (>0.658), Level II (0.596~0.658), Level III (0.530~0.596), Level IV (0.468~0.530) and Level V (0~0.468). The table displays the mean values of ecological elasticity in the Harbin–Changchun urban agglomeration. From 2000 to 2020, the mean values of ecological elasticity in the research area declined each year across all three scales, consistently reaching Level III elasticity. As the study scale increased, the average elasticity value at the grid scale reached its peak in the same year. The difference in the average values at the district-county and city scales was more noticeable in comparison to the grid scale (

Table 8. Mean values of ecosystem resilience of the Harbin–Changchun urban agglomeration at different scales, 2000–2020.

Year	2000	2010	2020	2000–2020 Mean Change
Grid scale	0.566235	0.565171	0.56297	−0.003265
District-county scale	0.565904	0.564839	0.562464	−0.00344
City scale	0.565853	0.564789	0.562413	−0.00344

).

Between 2000 and 2020, the Harbin–Changchun urban agglomeration saw changes in the quantity and spatial distribution of ecosystem resilience classed areas at various scales, as shown in Figure 8 and Figure 9. At the grid scale, Level I exhibited the highest level of resilience in a particular area. The percentage of Level I and Level II areas continued to decline, whereas the percentage of Level IV and Level V land continued to rise. The Level III region exhibited a fluctuating pattern, initially expanding and subsequently contracting while displaying an overall increasing trajectory. Level I was mostly found in the southeastern forested landscape and the northwestern grassland environment, characterized by dense vegetation and robust ecosystem resistance and resilience. Conversely, situated in the southern region of the research area inside the Mudan River basin, the watershed area experienced a consistent reduction, resulting in deterioration in the resilience of its ecosystem. At the district-county scale, the distribution of ecological resilience was mostly influenced by Level II ecological resilience. The fraction of the Level I resilient area showed an initial increase followed by a decrease, and the overall situation remained rather stable. The Level II and Level III regions exhibited a declining pattern, but the Level IV and Level V regions demonstrated an ascending trend. Level V resilience mostly lay in the region between the central urban areas of Harbin City and Changchun City. This was due to the execution of the “Harbin–Changchun Economic Belt” development policy, which significantly amplified the effects of urbanization in this area. The assessment of the city’s scale was mostly influenced by the distribution of Level III ecological resilience. Jilin City initially progressed from Level III resilience to Level II and subsequently reverted to Level III resilience. Yanbian Korean Autonomous Prefecture’s flexibility was reduced from Level I to Level II. As the scale decreased, the measured elasticity level became more detailed, and increasing the size hid the larger areas with high values in the region.

By analyzing the changes in ecosystem resilience values from 2000 to 2020, it can be seen that the changes were more obvious at the raster scale, in which the resilience indices around the Songhua River Basin and the Mudan River Basin decreased significantly due to the decrease in the watershed area, and the resilience index of the forests increased significantly. At the district-county scale, the elasticity index decreased in most areas, with the most significant decrease in the main urban area of Changchun City. At the city scale, only Siping City and Mudanjiang City showed an increase in the resilience index.

3.2.4. Multi-Scale Assessment of the Spatial and Temporal Evolution of Ecosystem Service

To ensure the accuracy of the InVEST model’s output, this study conducted repeated experimental adjustments and referred to relevant data from the Heilongjiang Water Resources Bulletin, Jilin Water Resources Bulletin, China River Sediment Bulletin and previous studies [61,62]. As a result, the validity of each ecosystem service’s results was determined. The natural breakpoint method was employed to classify ecosystem services into five levels: Level I (>0.593), Level II (0.500~0.593), Level III (0.398~0.500), Level IV (0.268~0.398) and Level V (0~0.268).

Table 9. Mean values of ecosystem service of the Harbin–Changchun urban agglomeration at different scales, 2000–2020.

Year	2000	2010	2020	2000–2020 Mean Change
Grid scale	0.482491	0.486793	0.464452	−0.01804
District-county scale	0.485762	0.490022	0.467542	−0.01822
City scale	0.485718	0.490003	0.467515	−0.0182

displays the mean values of ecosystem services in the Harbin–Changchun urban agglomeration. The mean number of ecosystem services over all three scales was categorized as Level III. This value exhibited a pattern of initially increasing and then dropping, ultimately showing an overall declining tendency. The mean value of ecosystem services experienced the most substantial decline at the district-county levels.

Within the Harbin–Changchun urban agglomeration, the proportion and arrangement of ecosystem services between 2000 and 2020 at various scales can be observed in the area distribution (Figure 10 and Figure 11). At the grid scale, the percentage of Level II services was the highest and continued to grow annually. The proportion of Level I services initially rose and then fell, showing an overall decline. Conversely, the proportion of Level III services initially declined and then rose, indicating an overall gain. The regions of Level IV and Level V services consistently grew each year. The distribution of Level I ecosystem services was mostly concentrated in the eastern forests region, which exhibited superior biodiversity, carbon sequestration and water supply services. The area of Level I experienced a substantial decline between 2010 and 2020 due to a notable rise in economic construction activity in the city. Level V and Level IV services were primarily concentrated in highly developed metropolitan regions. However, due to the ongoing growth of urban construction land, the level of ecosystem services declined. Level V services were not available at the district-county scale. The largest area was covered by Level III services, which initially decreased and then grew, resulting in an overall increase. The provision of Level I and Level IV services steadily declined, and it was projected that Level I services would cease to exist by 2020. Level II services exhibited an initial increase followed by a subsequent fall, resulting in a notable overall decline. The distribution of Level I and II service levels was primarily concentrated in the eastern region of the research area, where the Little Xing’an Mountains Nature Reserve and Zhangguancailing Nature Reserve are located, exhibiting rich biodiversity. The expanded coverage of Level IV services primarily encompassed the central metropolitan areas of Daqing City and Changchun City, as well as the neighboring districts and counties. At the city scale, there were no service levels categorized as Level I, Level II or Level V. The largest area was covered by Level III services. The service level in Harbin City was downgraded from Level II to Level III, and in Qiqihar City and Suihua City, the service level was initially upgraded from Level III to Level II and then downgraded from Level II to Level III.

This study examined the alterations in the evolution of ecosystem service functions over a span of 20 years, considering three different scales. Ecosystem services declined in the northwest, northeast and southeast regions of the study area, whereas they increased in the southwest. The changes observed at the grid scale were even more pronounced. Additionally, the value of ecosystem services in densely populated areas, such as Zhang Guangcailings Nature Reserve, decreased significantly, whereas eco-land increased. At the district-county level, there was a significant drop in the value of ecosystem services in the center and northern regions over a vast area. Conversely, there was a minor gain in the southern regions, specifically in Yongji County, Panshi City and adjacent districts and counties. Ecosystem services exhibited an increase at the urban level in Songyuan City, Changchun City and Jilin City, all situated in the southeastern region.

3.2.5. Multi-Scale Assessment of the Spatial and Temporal Evolution of Human Activities

Human activities exert a detrimental influence on the health of ecosystems, and the magnitude of this influence is directly proportional to the level of damage posed to the environment. This study employed the natural breakpoint approach to categorize human activities into five levels: Level I (>0.841), Level II (0.623~0.841), Level III (0.392~0.623), Level IV (0.623~0.841) and Level V (0~0.164). The mean values of human activities in the Harbin–Changchun urban agglomeration are presented in

Table 10. Mean values of human activities of the Harbin–Changchun urban agglomeration at different scales, 2000–2020.

Year	2000	2010	2020	2000–2020 Mean Change
Grid scale	0.50069	0.502225	0.507232	0.006542
District-county scale	0.500601	0.502143	0.507173	0.006572
City scale	0.500686	0.502204	0.507246	0.00656

. The average human activity rating across all three scales was Level III, and it consistently increased each year. The average value of human activities experienced the most notable increase at the district-county levels. From 2000 to 2010, and then from 2010 to 2020, there was a fluctuation in the mean value of human activity as the research scale increased. However, the overall fluctuation was minimal.

The proportion and arrangement of human activities in the Harbin–Changchun urban agglomeration from 2000 to 2020 at various levels can be observed in Figure 12 and Figure 13. From 2000 to 2020, the extent of Level I human activities consistently expanded at the grid level. Moreover, the fraction of Level V human activities was slightly more favorable but showed a progressive decline over time. The west had a high degree of human activity, primarily due to its cultivation of land and grasslands. In contrast, the east was predominantly characterized by forests and mountains, resulting in a lower frequency of human activities. At the district-county scale, Level IV human activities encompassed a progressively bigger area, experiencing slight annual growth. The Level I area experienced growth, whereas Levels II, III and V exhibited a decline. Level I human activities were predominately concentrated in the Harbin–Changchun Economic Belt, a region characterized by significant economic advancement and frequent human engagement. At the city scale, the variation in human activity level was not substantial. Cities with significant industrial growth and a strong regional economy, such as Daqing and Changchun, tended to have higher levels of human activity. On the other hand, cities like Yanji, which contain more nature reserves and ecological forests, tended to have lower levels of human activity.

We conducted an analysis of the fluctuations in human activities levels between the years 2000 and 2020. At the grid scale, the human activity value increased significantly due to the density of cities in the west, which was heavily urbanized. In the central part, the land use type was mostly cultivated land, and due to the implementation of ecological policies such as returning farmland to forests, the human activity index decreased in the central part. At the district-county scale, most of the areas with increased human activity indexes were in the east, and some of the eastern forested landscapes’ human activity indexes decreased more significantly. At the city scale, the human activity index increased significantly in the west, such as Daqing City, and decreased in the center and east, most notably in Mudanjiang City and Siping City.

3.3. Spatially and Temporally Evolving Multi-Scale Ecosystem Health Assessment of the Harbin–Changchun Urban Agglomeration

This study employed the natural breakpoint approach to categorize ecosystem health into five levels: well(>0.695), relatively relatively well (0.613~0.695), ordinary (0.530~0.613), relatively relatively weak (0.453~0.530) and weak (0~0.453). The average ecosystem health value in the Harbin–Changchun urban agglomeration is presented in

Table 11. Mean values of ecosystem health of the Harbin–Changchun urban agglomeration at different scales, 2000–2020.

Year	2000	2010	2020	2000–2020 Mean Change
Grid scale	0.566062	0.574553	0.575084	0.009022
District-county scale	0.566085	0.5746	0.575147	0.009062
City scale	0.56602	0.574564	0.575099	0.009079

. From 2000 to 2020, the health mean value remained at Ordinary across all three scales. Additionally, the mean health value climbed steadily each year. The mean level of ecosystem health experienced the greatest increase at the city scale and the smallest increase at the grid scale.

This study examined the urban agglomeration of Harbin–Changchun from 2000 to 2020, specifically focusing on the proportion and spatial arrangement of ecosystem health grading at various levels. The findings are presented in Figure 14 and Figure 15. The regions classified as good were primarily situated within the eastern forests, and the remaining portions were found in the northern region of the Little Xing’an Mountains and its surrounding vicinity. Over the past few years, the adoption of ecological policies, such as converting farmland to grassland, protecting water ecology and preserving forest land, has enhanced the connection between different patches of land, increased carbon storage, improved soil conservation and safeguarded biodiversity. As a result, there have been substantial ecological advantages, leading to enhanced ecosystem health at the regional level. At the district-county level, the proportion of areas in relatively weak ecological health initially rose and then declined, with an overall decrease in the affected area. The area experiencing a relatively weak health status initially increased and then decreased, and the overall area remained relatively stable. Conversely, the area with an ordinary status initially decreased and then increased, leading to an overall increase. The regions with a good status were primarily located in the eastern districts and counties, whereas the regions with a weak status were predominantly found in the Harbin–Changchun Economic Belt. This resulted in haphazard expansion of construction land and frequent human production activities, which disrupted the structural stability of ecosystems and led to a decline in the quality of urban habitats and ecological benefits. At the city scale, Jilin City and Suihua City were the only ones that exhibited changes in their health status. Jilin City’s health status improved from average to relatively good between 2000 and 2020. On the other hand, Suihua City’s health status progressed from relatively poor to average and then declined from average to relatively weak. This demonstrates that the grid scale more accurately represented the attributes of the health condition of smaller geographical areas. The district-county scale illustrated the variations in and patterns of ecosystem health among districts and counties with varying levels of development. At the city scale, the expansion in size obscured places with favorable health conditions within the overall context, while simultaneously intensifying the concentration of these same health conditions in the geographical distribution. Changes in the three scales from 2000 to 2020 were analyzed, with the most pronounced difference in change at the grid scale. At the grid scale, the distribution of ecosystem health change areas was relatively scattered, with significant value added near the Little Xing’an Mountain in the north and the forested area in the center, and ecosystem health was significantly reduced in the dense urban areas in the south and northwest of the study area. At the district-county scale, the ecosystem health of the main urban areas of Qiqihar City, Harbin City and Changchun City and its surroundings decreased significantly. At the city scale, only Qiqihar City and Liaoyuan City showed decreases, and the ecosystem health of Qiqihar City decreased significantly. Ecosystem health increased in the rest of the cities, with the largest increase in Jilin Province.

An examination of the changes in the three scales over a span of 20 years revealed a notable increase in the human activity value at the grid scale. This increase can be attributed to the high concentration of metropolitan areas in the western region. In the central region, the predominant land use was agriculture, and the execution of ecological measures, such as converting cropland into forests, resulted in a decline in human activity in this area. At the district-county scale, the majority of places experiencing higher levels of human activity were located in the eastern part, whereas certain forests in the east showed a more pronounced decline in their human activity indexes. At the city scale, there was a substantial increase in the human activity index in the western region, particularly in Daqing City. Conversely, there was a fall in the index in the central and eastern regions, with notable declines observed in Mudanjiang City and Siping City.

3.4. Spatial Self-Assessment of the Harbin–Changchun Urban Agglomeration

This study examined the spatial clustering of ecosystem health at various scales by conducting a global autocorrelation analysis and cluster distribution analysis of the study area. This study chose various grid scales for experimental comparison [63], considering the relevant literature [4], the number of patches in the grid and data accuracy. Based on the experimental results, a 10 km grid scale was determined to be the most appropriate research scale for characterizing spatial variations in ecosystem health. In this study, we acknowledge the distinctiveness of administrative scales at the district, county and city levels. However, due to the insufficient number of cities at the city scale for spatial autocorrelation analysis, we chose to focus on the 10 km grid scale as well as the district and county scales for this analysis. The global Moran’s I index is displayed in

Table 12. Mean values of global Moran’s I index at different scales in the Harbin–Changchun urban agglomeration, 2000–2020.

Year	2000	2010	2020
10 km grid scale	0.673905	0.670782	0.680074
District-county scale	0.783365	0.77627	0.759915

. The Moran’s I index values for the whole Harbin–Changchun Urban Agglomeration all passed the 1% significance test and were positive. This suggests a substantial spatial positive correlation between ecosystem health values and the level of agglomeration. This suggests that regions with greater levels of ecosystem health tended to group in terms of their geographic location. Moran’s I index showed a consistent decline at both the 10 km raster scale and district-county sizes from 2000 to 2020. The global Moran’s I index at the district and county scales exceeded the global Moran’s I index at the 10 km grid scale in the same year, indicating a higher level of clustering in the ecosystem health of the study area at the district-county scale.

To examine the spatial distribution of ecosystem health in the Harbin–Changchun urban agglomeration, we conducted a local autocorrelation study using the hot spot analysis tool. This analysis allowed us to identify agglomeration patterns at different scales and determine the area share of ecosystem health from 2000 to 2020. The results of this research are presented in Figure 16 and Figure 17. In general, the spatial arrangement of the Harbin–Changchun urban agglomeration stayed rather consistent. Ecosystem health hot spots were primarily concentrated in the eastern region of the research area, characterized by densely wooded land, abundant biodiversity and robust ecosystem vigor and service capacity. Consequently, areas with high value clusters emerged. The areas with poor ecosystem health were primarily concentrated in the central region of the research area. These areas had clusters with low ecological value due to factors such as the significant need for urban growth, the imminent threat of urbanization and land expansion for construction and the frequent occurrence of human activities. Between 2000 and 2020, there was a minor increase in the area of hot spots at the 10 km grid scale. Conversely, the area of cold spots initially fell and subsequently climbed, resulting in an overall increase. The cold spot region situated in the northern section of the research area exhibited a noticeable decrease followed by a more pronounced increase. At the district-county scale, the spatial extent of hot spot areas expanded, and the spatial extent of cold spots exhibited a trend of initially expanding and then dropping. Overall, the cold spot areas decreased. By 2010, the high-value clusters remained unchanged, including Shuangyang District in Changchun City, Bayan County and Bin County in Harbin City and ten other districts and counties. These areas maintained their clustering features, whereas the low-value clusters did not undergo substantial changes. By 2020, eleven districts and counties, such as Bayquan County in Qiqihar City, Bayan County, and Bin County in Harbin City, transformed low-value agglomerations to having inconsequential agglomeration features. Yushu City transformed into a low-value cluster area, whereas Wuchang City transformed into a high-value cluster area.

From 2000 to 2020, the size of hot spot high-value clusters experienced more growth at the district-county level compared to the 10 km grid scale. The cold spot clusters exhibited a declining tendency in the area at the district-county scale but demonstrated an increasing trend at the 10 km grid scale. It is evident that individual patterns of cluster distribution at different scales differ, even though the overall geographical cluster distribution in the research area tends to be relatively consistent.

4. Discussion

4.1. Causes of Heterogeneity in Ecosystem Health

The VOR-SH model developed in this study was founded on a comprehensive analysis of several scales. It aims to assess and examine five variables that are closely linked to ecosystem health and general ecological well-being. The regional distribution of the overall ecosystem health and the five indicators contained in the Harbin–Changchun urban agglomeration exhibited clear variations. Based on our analysis at the raster scale, we found that the forested ecosystems in the eastern part of the study area exhibited greater vigor, organization, resilience, ecosystem services and overall ecosystem health compared to other areas. However, the assessment of anthropogenic factors is lacking. The presence of forests significantly contributes to environmental purification, water conservation, soil and water preservation and climate enhancement. Forested land generates ecological benefits that can mitigate the adverse effects of human activities, such as environmental pollution and urbanization. Within the research area, human activities were more pronounced in the construction land and cultivated land located in the center region. Conversely, other components and the general health of the ecosystem were notably poorer in comparison to the rest of the areas. Particularly detrimental to ecological development are the impermeable surfaces produced by human construction activities, which are predominantly found on cultivated and construction land, where human activities occur most frequently and cause substantial environmental harm [64]. Hunchun City, Antu County, Hualong City and the neighboring districts and counties exhibited superior ecosystem vigor, organization, resilience, ecosystem services and overall ecosystem health at the district-county scale. However, the human impact elements in these areas were more detrimental. This holds immense importance for the preservation of biodiversity and the promotion of ecological sustainability. In the central areas of Harbin City, Changchun City and the surrounding counties, there was a greater level of human activities. However, the other variables contributing to ecosystem health and the overall ecosystem health were lower. The Harbin–Changchun Urban Agglomeration Development Plan focuses on the area between the main cities of Harbin and Changchun. This region had a high concentration of districts and counties with intense human construction activities and a high level of urbanization. However, these factors negatively affected the ecological benefits of the area. The superior ecological vigor, structure, adaptability, provision of ecosystem services and overall ecosystem health, as well as the lower human activity values in Yanbian Korean Autonomous Prefecture at the municipal level, can be attributed to the presence of a superior ecological foundation consisting of nature reserves and its proximity to the Changbai Mountain Range. The overflow of ecological advantages from mountain ranges has a positive impact on biodiversity, soil and water preservation and the ecological environment within a city [65]. The increased intensity of human activities in Changchun City is responsible for the decreased vitality, organization, resilience, service functions and general health of the environment. Changchun, located in China, is a significant region for grain production and serves as an industrial hub. This characteristic has resulted in frequent human activities related to production and building, leading to increased exploitation of ecological resources and elevated levels of environmental degradation inside the city.

4.2. Multi-Scale Effects of Ecosystem Health

Due to the scale dependence of ecosystem health indicators, different scales exhibit distinct properties [66]. This study conducted a hierarchical assessment of the ecosystem health of the Harbin–Changchun urban agglomeration at different scales (grid, district-county and city) in the years 2000, 2010 and 2020. The objective was to offer diverse perspectives and solutions for managing the interaction between natural ecosystems and human activities. The analytical results indicate that grid-scale ecosystem health exhibited the most substantial changes, with more pronounced heterogeneity in the geographical distribution. This heterogeneity allowed for a detailed assessment of changes in tiny areas. The district-county scale provided a detailed representation of the shifting trends and spatial variability in local ecosystem health in the Harbin–Changchun urban agglomeration. It allowed for a more comprehensive assessment of the overall health of the area environment, while also possessing the qualities of an easily regulated administrative scale, facilitating the formulation of particular ecological construction recommendations. At the municipal level, the extreme values of ecosystem health within a small territory were concealed, resulting in a homogeneous environment that facilitated macro-control and strategic decision making for ecological purposes in a larger region. When considering the Harbin–Changchun urban agglomeration, the most accurate reflection of ecosystem health and distribution characteristics can be achieved by examining the district-county scale. By focusing on these smaller scales, it becomes possible to identify specific regional issues and propose targeted improvements. Additionally, managing and controlling at the district and county scales is generally easier than that at the larger city scale within the administrative framework.

4.3. Spatial Autocorrelation and Ecological Management Recommendations

To analyze spatial autocorrelation, we chose to evaluate and analyze at both the 10 km grid scale and the district-county scale. The findings indicate that the studied area exhibited pronounced clustering in global spatial autocorrelation. Furthermore, the local autocorrelation analysis reveals that the eastern half of the area was characterized by a high-value concentration, whereas the middle section exhibited a low-value concentration. The reason for this is the exceptional ecological environment and abundant biodiversity of the forested areas in the east, which were minimally impacted by human construction activities. In contrast, the central region experienced extensive construction and development, with a relatively flat terrain that was conducive to human production activities. Through our examination of spatial clustering analysis, we present and provide ideas for the management and restoration of ecological systems at several scales. In summary, it is crucial to protect and preserve the ecological environment in areas with high ecosystem health values and utilize their ecological benefits to improve the health of neighboring areas [67]. It is important to monitor and address ecological changes in moderately healthy areas and implement preventive measures and ecological regulations. Immediate attention should be given to areas with low ecosystem health, and prompt ecological restoration should be carried out to minimize disturbances and enhance the ecological health condition. In other areas, it is essential to identify and address key contradictions or problems to actively improve the ecological health situation. At the gride scale, the primary concern is the transfer of land use, and the fundamental indicator of changes in ecosystem health is the alteration of land use. This study offers a fundamental framework for investigating the factors that contribute to alterations in health conditions. The district-county scale specifically addresses the problems that arise between ecological development and production construction at a small administrative level and implements precise improvements to address these conflicts. At the municipal level, it is more appropriate for macro-control. When aligning with national strategic development decision making, it is important to consider the distinctiveness of the region to achieve harmonious integration of the two. It supports the creation of a comprehensive development plan for metropolitan areas that may successfully address the tensions between regional ecology and development while aligning with national strategic goals.

4.4. Limitations and Future Work

The swift growth of the Harbin–Changchun urban agglomeration presents a significant risk to the health of the ecosystem. Evaluating ecological health can effectively pinpoint crucial regions that require protection and restoration [68,69]. Existing studies were used as a basis to incorporate evaluation factors of human activities into the VOR-SH ecosystem health multi-scale assessment model. This addition enhanced the connection between natural and human factors and provided a more accurate and comprehensive reflection of the health of modern ecosystems. The VOR-SH model departs from the conventional assessment framework, which emphasizes the inherent characteristics of ecosystems and serves to bridge the theoretical divide between ecosystem features and the administration of intricate ecosystems. Given the vast scope of ecosystems and the intricate interplay of several elements, current research on ecosystem health is still in its nascent phase [70]. The assessment system implemented in this study is currently limited by temporal, spatial and regional constraints. To enhance the accuracy and credibility of the assessment results, it is necessary to reinforce the system by selecting region-specific indicators and effectively coordinating various factors in the future.

5. Conclusions

This study examines the spatial and temporal changes in ecosystem health in the Harbin–Changchun urban agglomeration from 2000 to 2020 using a multi-scale methodology. An analysis was conducted on the land use transfer of the Harbin–Changchun urban agglomeration between 2000 and 2020. Furthermore, the spatial and temporal evolution of five indicators—vigor, organization, resilience, ecosystem services and human activities—in the research region was evaluated hierarchically at three scales: grid, district-county and city. Subsequently, ecosystem health spatio-temporal variations and spatial clustering at several scales were examined. Last, the features of ecosystem health attributes at varying sizes were further examined. The research revealed the following:

• The predominant land types in the Harbin–Changchun urban agglomeration are cultivated land and woods. Between 2000 and 2020, the Harbin–Changchun urban agglomeration experienced substantial growth in construction land and unused land due to rapid urban development. This expansion led to a significant reduction in water bodies and grasslands, and the decrease in forested areas was relatively minor. Cultivated land, on the other hand, remained relatively unchanged.
• From 2000 to 2020, the Harbin–Changchun urban agglomeration had a decline in resilience and ecosystem service capacity, while witnessing an increase in vigor, organization and human activities. These changes can be attributed to variations in regional development levels and the variability in the ecological substrate. The geographical distribution of ecosystem vigor, ecosystem resilience and ecosystem services varied across different scales, with higher values observed in the eastern forests and lower values observed in the center plains area. Based on the assessment of the district-county scale, Hunchun City and Antu County in the eastern region of Yanbian Korean Autonomous Prefecture received higher evaluation ratings. Consequently, the evaluation ratings of Yanbian Korean Autonomous Prefecture at the city scale were notably superior to those of other cities. The ecological organizational worth was greater in the eastern forested region and the western grassland region, whereas it was lower in the middle plains area. Honggang District and Saltu District in the western part of Daqing City received the highest assessment ratings at the district-county scale. As a result, Daqing City’s overall evaluation ratings for organizational strength at the city scale surpassed those of other cities. On the other hand, human activities, which have a detrimental impact on ecosystem health, were less prevalent in the eastern forests area but more pronounced in the central urban agglomeration area. Human activities were most concentrated in the region between the large cities of Harbin and Changchun at the district-county level. As a result, the human activity index in Changchun was greater compared to other cities. Utilizing multi-scale analysis to elucidate the progression of each of the five indicators of ecosystem health offers specific avenues for reconciling the interplay between nature and humans.
• From 2000 to 2020, the ecosystem health of the Harbin–Changchun urban agglomeration improved at various scales. The extent of in-good-condition regions exhibited a progressive expansion at both the grid and district-county levels but ceased to exist at the city level. The regions exhibiting a generally favorable health status experienced a decline in area at both the grid and district-county levels while showing a modest gain at the city level. The extent of regions with average health exhibited varied degrees of decline. This decline persisted at the grid level, but at the district-county level, it initially decreased and then increased. Conversely, at the city level, it first increased and then decreased. Regions with relatively poor health conditions experienced negligible changes at the grid level, noteworthy declines at the district-county level and declines followed by increases at the city level, without any substantial overall changes. Regions exhibiting a low level of health demonstrated a consistent decline across all three measures. The distribution of ecosystem health levels at the three scales was distinct. The grid scale provided more detailed information, and the district-county scale considered both necessary details and administrative characteristics. The city scale complemented the macro perspective. This research offers a foundation and direction for boosting ecosystem sustainability and improving health by elucidating the overall evolution and distribution of ecosystem health on a multi-scale in the Harbin–Changchun urban agglomeration.
• An analysis was conducted to assess the clustering of ecosystem health in the Harbin–Changchun urban agglomeration. The findings from the local spatial autocorrelation analysis at both scales indicate that the areas with high values were primarily clustered in the eastern wooded land, whereas the areas with low values were primarily clustered in the center constructed land and cultivated land. The low-value clusters in the middle region of the research area had greater variation at the 10 km grid scale, and the high-value clusters in the east showed more significant variation at the district-county scale. In light of the spatial agglomeration analysis, multi-scale ecological regulation is suggested to preserve the ecological environment in the high-value areas of ecosystem health and to act as a spillover effect to improve the health of the surrounding areas. Additionally, real-time attention must be given to the low-value areas of ecosystem health to carry out ecological restoration promptly, thereby reducing interference and enhancing ecological health; attention must also be given to changes in the other areas’ ecological health to reinforce ecological regulation and prevention. The multi-scale hierarchical analysis and thorough assessment of ecosystem health provided by this study serve as a crucial foundation for planning the urban development and ecological management of agglomerations, offering a range of approaches to balance the coexistence of humans and natural ecosystems.