Urban expansion’s direct and historical ecological consequences are well recognized, but its indirect and future implications remains unclear. This study employs an integrated modeling approach, combining the InVEST, FLUS and linear regression models, to explore how urban expansion in the Chang-Zhu-Tan Urban Agglomeration (CZTUA) has driven and will continue to drive land use changes, resulting in the loss of ecosystem carbon storage both in the past and under future development pathways. From 1990 to 2020, the construction land increased significantly by 85,482.09 hm2, with newly expanded construction land mainly distributed around the periphery of existing built-up areas. Concurrently, the ecosystem carbon storage decreased significantly by 8.9 × 108 kg, and the expansion direction of low-carbon storage areas aligns with that of newly expanded construction land. Under the ecological protection scenario, construction land shows a contraction trend by 2030 and 2060, with a slow decrease in ecosystem carbon storage. Contrastly, under the natural and urban development scenarios, construction land demonstrates an expansion trend, accompanied by an accelerated reduction in ecosystem carbon storage. Construction land expansion primarily exerts significantly negative impacts on aboveground and belowground carbon storage, while its adverse effect on soil carbon storage remains relatively limited.
urban expansioncarbon storageInVEST modelFLUS modelspatial autocorrelation analysiscarbon densityCZTUAThe author(s) declare that financial support was received for the research and/or publication of this article. The Humanities and Social Sciences Planning Fund of the Ministry of Education of China (24YJA630052). A Project Supported by Scientific Research Fund of Hunan Provincial Education Department (23B0210).section-at-acceptanceLand, Livelihoods and Food SecurityIntroduction
Over the past five decades, China has experienced profound and extensive historical transformations, accompanied by rapid population expansion and unprecedented economic growth that has attracted global attention. The national urbanization rate increased significantly from 20.16% in 1981 to 63.89% in 2020—more than tripling. Concurrently, the urban built-up area expanded dramatically from 6,720 km2 to 58,355.3 km2 over the same period, representing an approximately ninefold increase. It is anticipated that these trends will continue, or even accelerate, during the first half of the 21st century (Kong et al., 2025). However, by converting vast expanses of natural and cultivated land into impervious surfaces, urban expansion—despite its socio-economic benefits—directly leads to a significant decline or even a complete loss of essential ecosystem services such as carbon storage (Liu et al., 2023). Simultaneously, China has implemented a national food security strategy alongside strict cropland protection policies, which intensify the multiple linkages between urban expansion and ecosystem carbon storage, making it an ideal case study for this topic (Zeng et al., 2025).
Urban expansion directly encroaches on ecosystem carbon storage while also triggering indirect and remote effects. In response, research has evolved along two key strands: one quantifies the historical loss of ecosystem carbon storage attributable to urbanization expansion (Wang et al., 2021; Niu et al., 2022; Li et al., 2023), and the other extensively employs land-use simulation models to project future ecosystem carbon storage dynamics under multiple scenarios. Among the most widely applied simulation frameworks are those based on cellular automata (CA) (Li and Lin, 2025Li and Lin, 2025; Ali et al., 2025, Li and Lin, 2025), such as CLUE-S (Movasaghi et al., 2025; Siabi et al., 2025), FLUS (Huang et al., 2024; Qin et al., 2025), and LANDSCAPE (Yang et al., 2023; Hou et al., 2024). The FLUS model, in particular, incorporates an innovative adaptive inertia coefficient mechanism and a roulette-wheel-based competition mechanism to simulate future land-use patterns (Zhang et al., 2024). It offers the advantages of high simulation accuracy, fast operational speed, user-friendly operation, and dynamic visualisation capabilities (Ma et al., 2025). Concurrently, the InVEST model has gained widespread application in ecosystem service assessments—particularly for quantifying ecosystem carbon storage—due to its operational simplicity and capacity for spatial visualisation (Chen et al., 2025). Coupling the FLUS model with the InVEST model enables in-depth investigation into the spatiotemporal effects and underlying mechanisms through which both historical and future urban expansion influences changes in ecosystem carbon stocks (Zhao et al., 2022).
Urban agglomeration, compact spatial arrangements of cities with strong economic ties formed through developed infrastructure networks, face significant challenges in achieving carbon neutrality (Jia et al., 2025). In China, for instance, urban agglomeration concentrate 75% of the population on just 25% of the land. This high density of population and industry leads to substantial energy consumption and high carbon emission intensity. Furthermore, carbon cycles within these urban agglomeration involve both natural and anthropogenic processes—such as construction, transportation, and industry—adding complexity to carbon storage dynamics (Meng et al., 2025). The Chang-Zhu-Tan urban agglomeration (CZTUA), serving as Hunan Province’s economic core, exemplifies these patterns. Its compact spatial arrangement facilitates efficient development of transport infrastructure and rapid circulation of productive factors. Moreover, the CZTUA is endowed with abundant natural resources and ecosystems, including the Xianjiang River and Yuelu Mountains. Despite its significance, few studies have examined the spatio-temporal effects of urban expansion on ecosystem carbon stocks in rapidly developing conurbations like CZTUA. Therefore, this study endeavors to answer the following overarching research question:
To what extent does urban expansion impact ecosystem carbon storage? How will urban expansion and ecosystem carbon storage change under different future development scenarios?
This paper therefore aims to elucidate the mechanisms by which urban expansion in CZTUA from 2000 to 2060 led to ecosystem carbon storage loss. The study deepens the understanding of the intricate relationship between urban expansion and ecosystem carbon storage loss, providing pioneering insights into carbon balance management, low-carbon urban planning, and climate adaptation strategies in other regions or conurbations. Specifically, the contributions of this research are twofold. Firstly, by integrating the land-use transition matrix, the FLUS model, and the InVEST model, the spatiotemporal effects and underlying mechanisms through which historical and future urban expansion have affected ecosystem carbon storage are revealed. Secondly, through the application of linear regression models, the differential effects of urban expansion on carbon pools—namely above-ground biomass, below-ground biomass, dead organic matter, and soil carbon density have been demonstrated.
Materials and methodsStudy area
The CZTUA, composed of the three cities of Changsha, Zhuzhou, and Xiangtan, is one of China’s most significant urban agglomeration and a demonstration zone of regional integrated development. Located in the east-central part of Hunan Province and along the middle-lower reaches of the Xiangjiang River, its covers a total area of 28,000 km2. The CZTUA exhibits a distinctive spatial composition, where basins, hills, river, towns, and villages interweave. With the successive implementation of national and regional strategies such as the Rise of Central China, the CZTUA, and the Yangtze River Economic Belt, the region has undergone rapid urbanization and industrialization. Urban expansion has extensively encroached on cultivated land and natural land, leading to a series of ecological crisis such as habitat loss and soil erosion, which significantly impact ecosystem carbon storage. Establishing a harmonious and stable relationship between territorial spatial evolution and ecosystem carbon storage has become an essential requirement for the region’s healthy development. This study selects the core area of the CZTUA as the research region, including the urban districts of Changsha, Zhuzhou, and Xiangtan, as well as Changsha County and Xiangtan County (Figure 1). This area represents the most dynamic zone in terms of territorial spatial transformation and carbon storage changes within the CZTUA.
CZTUA core area administrative divisions diagram.
Map showing Hunan Province on the left, highlighting the Chang-Zhu-Tan Urban Agglomeration. The area is outlined and enlarged on the right with a color-coded elevation map, indicating higher elevations in red and lower elevations in green. Scale bars and north arrows are included.MethodsLand use transition matrix
The land use transfer matrix serves as a fundamental tool to analyze the scale and identify the pathways of conversions among various land use types. The matrix for each interval was developed through an overlay analysis of multi-temporal land-use maps, followed by a quantitative calculation of the areal extent transferred between all land-use categories. The mathematical expression of this matrix is presented below:
Sij=∣S11S12…S1nS21S22…S2n…………Sn1Sn2…Snn∣
Where Sij denotes the area of land use type i converted to type j; n represents the land use types.
InVEST model
The InVEST model quantifies ecosystem carbon storage by integrating land use data with carbon density values specific to each land cover type. It replaces complex ecological processes with a computable carbon stock estimation framework based on area and carbon density, enabling objective and spatially explicit evaluation (Zhang et al., 2025). This approach generates spatially explicit outputs that visually depict the spatial patterns of ecosystem carbon storage, facilitate the identification of critical carbon sink areas, and allow for simulation analysis under different future development scenario (Dai et al., 2025). The assessment comprehensively accounts for four carbon pools: aboveground biomass, belowground biomass, dead organic matter, and soil organic carbon (Table 1). The specific calculation formula is as follows:
Ctotal=(Cabove+Cbelow+Csoil+Cdead)×S
Carbon density of different land use types in CZTUA kg/m2.
Land use type
Above-ground carbon density
Below-ground carbon density
Soil carbon density
Dead organic carbon density
Cultivated land
2.79
9.46
10.84
0
Forest land
5.05
15.18
21.32
0
Grass land
2.28
8.65
9.99
0
Water area
1.25
7.9
0
0
Unused land
0.51
2.43
0
0
Construction land
1.25
5.67
11.08
0
Where Ctotal represents total carbon storage; Cabove, Cbelow, Csoil, and Cdead represent the carbon density of aboveground biotic, belowground biotic, soil, and dead organic, respectively; S denotes the land use type.
FLUS model
The FLUS model integrates an Artificial Neural Network (ANN) module and a Cellular Automata (CA) module to simulate future land use patterns. The ANN module calculates the occurrence probability of each land use type for every pixel based on driving factors, while the CA module utilizes the probability outputs from ANN, combined with land use demand, neighborhood parameters, and a transition matrix, to simulate future land use patterns (Luan et al., 2024). This integrated framework enables a more realistic simulation of future land use patterns under the dynamic interplay of human activities and natural environment (Qin et al., 2025). Land use demand is projected using the Markov chain method, and neighborhood weight parameters are derived by calculating the ratio of the expansion area of each land use type to the total expansion area (Table 2). Three development scenarios—Natural Development, Urban Development, and Ecological Protection—were established for this study. Based on existing research (Lai et al., 2017), the corresponding land use transition matrices for these scenarios are specified in Table 3. The model’s accuracy was validated with a Kappa coefficient of 0.74 and an overall accuracy of 0.85, meeting the requirements for subsequent analysis.
Neighborhood parameters of land use types.
Land use type
Cultivated land
Forest land
Grass land
Water area
Unused land
Construction land
Domain factor parameters
0.319146
0.371444
0.001056
0.039663
0.000099
0.268592
Simulated land use transfer matrix under different scenarios.
Land use type
Natural development scenario
Urban development scenario
Ecological protection scenario
a
b
c
d
e
f
a
b
c
d
e
f
a
b
c
d
e
f
a
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
b
1
1
1
1
1
1
1
1
1
1
1
1
0
1
0
0
0
0
c
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
d
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
1
0
0
e
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
f
1
1
1
1
1
1
0
0
0
0
0
1
1
1
1
1
1
1
a, b, c, d, e, f represent cultivated land, forest, grassland, water area, unused land, and construction land, respectively; 0 indicates non-convertible, 1 indicates convertible.
Spatial autocorrelation analysis
Spatial autocorrelation analysis was employed to examine the spatial interdependence of geographical attributes (Dong et al., 2023). Global Moran’s I was applied to assess the spatial clustering pattern of carbon storage, identifying whether its distribution exhibited positive, negative, or random spatial autocorrelation. Subsequently, bivariate spatial autocorrelation analysis was conducted to quantify the spatial dependence between construction land area and carbon storage, with results classified into four relationship types: High-High, High-Low, Low-High, and Low-Low. The formula for Global Moran’s I is given as follows:
Where I denotes the bivariate global Moran’s I index; I′ represents the univariate global Moran’s I; n is the total number of fishnet grid cells; wij defines the spatial weight between cells i and j; xi and xj are the attribute values of cells i and j; x¯ is the mean value of the attribute; S2 is the variance of the attribute.
Hot spot analysis
Hot spot analysis was conducted using the Getis-Ord Gi* statistic to identify statistically significant spatial clusters of carbon storage. This method works by calculating the Gi* index for each location, which categorizes areas into high-value clusters (hot spots) and low-value clusters (cold spots) (Almeida et al., 2025). The statistical significance of these clusters was rigorously assessed at three confidence levels: 90, 95, and 99% (Ma et al., 2021). The specific calculation formula is as follows:
Gi∗=∑jnwijxj∑jnxj
where wij is the spatial weight between units i and j; xj represents the carbon storage value of spatial unit j; n denotes the total number of spatial units.
Data sources
The datasets used in this study include land use data and its associated driving factors. The land use data were obtained from the Resource and Environment Science and Data Center of the Chinese Academy of Sciences,1 comprising four temporal snapshots (1990, 2000, 2010, and 2020) at a spatial resolution of 30 meters. Land use change is influenced by a combination of natural and socioeconomic factors (Du et al., 2025). While natural factors such as climate and topography exert long-term control over land use patterns, socioeconomic factors often dominate short-term changes. Taking data availability into account, 12 driving factors were selected, the detailed sources of which are summarized in Table 4.
Driving factors and data sources of land use simulation.
Type
Data
Resolution (m)
Data source and processing
Natural factors
Annual mean temperature
1,000
National Qinghai-Tibet Plateau Scientific Data Center (https://data.tpdc.ac.cn/)
Annual mean precipitation
1,000
Elevation
30
Center for Resource and Environmental Science and Data, Chinese Academy of Sciences (resdc.cn)
Slope
30
Resource and Environment Science and Data Center of the Chinese Academy of Sciences (resdc.cn)
Normalized difference vegetation index
30
National Ecological Science Data Center (http://www.nesdc.org.cn/)
Distance to river
30
OpenStreetMap (OSM)
Socioeconomic factors
GDP
30
Center for Resource and Environmental Science and Data, Chinese Academy of Sciences (resdc.cn)
Population
30
World Pop (Open Spatial Demographic Data and Research–WorldPop)
Distance from railways
30
OpenStreetMap (OSM)l.
Distance from highways
30
Distance from major roads
30
Distance from local administrative centers
30
ResultsSpatio-temporal characteristics of construction land use changeTemporal dynamics of construction land use change
Between 1990 and 2000, land use change in the CZTUA primarily occurred among construction land, cultivated land, and forest land (Equation 1). The expansion of construction land largely took place at the expense of cultivated land and forest land, with a total increase of 85,482.09 hm2 in construction land area (Table 5). Across the three study periods—1990–2000, 2000–2010, and 2010–2020, the total construction land area increased by 17,819.55 hm2, 29,410.02 hm2, and 38,252.52 hm2, respectively. Cultivated land was the primary source of land converted to construction land, accounting for 15,716.16 hm2 (88.20%), 28,347.93 hm2 (81.11%), and 39,828.06 hm2 (93.62%) in each period. Forest land was the second most converted category, with 1,573.56 hm2 (8.69%), 5,092.47 hm2 (14.57%), and 1,942.11 hm2 (4.57%) converted, respectively. Over these three periods, construction land maintained a rapid growth trajectory, with expansion rates of 83.72, 75.21, and 55.83%, respectively.
Land use transfer matrix in CZTUA from 1990 to 2020 (hm2).
Type
Period
Cultivated land
Forest land
Grass land
Water
Unused land
Construction land
Total
Cultivated land
1990 → 2000
516,814.01
82,108.35
22.77
3,549.42
0.00
15,716.16
618,210.71
2000 → 2010
491,417.63
20,729.79
40.50
2,415.24
0.00
28,347.93
542,951.09
2010 → 2020
512,569.89
58,519.26
146.61
5,181.48
13.50
39,828.06
616,258.80
Forest land
1990 → 2000
25,131.51
401,775.65
0.54
10.98
0.00
1,573.56
428,492.24
2000 → 2010
113,308.38
365,404.68
23.40
260.91
0.00
5,092.47
484,089.84
2010 → 2020
44,572.41
340,031.06
15.21
585.27
0.54
1,942.11
387,146.60
Grassland
1990 → 2000
36.72
9.36
57.33
78.48
1.71
77.13
260.73
2000 → 2010
22.41
1.26
5.04
47.97
0.00
10.08
86.76
2010 → 2020
7.29
2.25
4.86
0.18
0.09
55.89
70.56
Water area
1990 → 2000
957.96
196.29
3.96
32,328.18
1.53
721.17
34,209.09
2000 → 2010
6,740.55
738.36
0.36
27,320.40
0.54
1,498.95
36,299.16
2010 → 2020
2,322.45
192.24
0.09
27,317.79
0.99
717.21
30,550.77
Unused land
1990 → 2000
0.99
0.00
2.16
52.29
7.20
21.42
84.06
2000 → 2010
0.09
0.09
0.00
9.36
0.00
0.90
10.44
2010 → 2020
0.00
0.00
0.00
0.00
0.45
0.09
0.54
Construction land
1990 → 2000
9.90
0.18
0.00
279.81
0.00
20,995.11
21,285.00
2000 → 2010
4,769.73
272.43
1.26
496.89
0.00
33,564.24
39,104.55
2010 → 2020
3,648.69
120.78
5.40
515.43
0.54
64,223.73
68,514.57
Total
1990 → 2000
542,951.09
484,089.83
86.76
36,299.16
10.45
39,104.55
1,102,541.84
2000 → 2010
616,258.79
387,146.61
70.56
30,550.77
0.54
68,514.57
1,102,541.84
2010 → 2020
563,120.73
398,865.59
172.17
33,600.15
16.11
106,767.09
1,102,541.84
To explore the future dynamics of land use in the CZTUA, the FLUS model was used to simulate the spatial patterns of land use under three scenarios—natural development, urban development, and ecological protection—for 2030 and 2060 (Table 6). From 2020 to 2030, the natural development scenario largely continued historical trends observed between 1990 and 2020, with construction land and forest land increasing by 29,405.97 hm2 and 7,544.97 hm2, respectively. Under the urban development scenario, the quantitative changes in land use were similar to those in the natural development scenario, with construction land expanding by 29,480.4 hm2. In the ecological protection scenario, the construction land decreased by 2,540.61 hm2 compared to 2020, while the increase in forest land was similar to that in the other two scenarios, and the reduction in cropland area was the smallest among the three scenarios. By 2060, the construction land areas under the three scenarios were 212,812.11 hm2, 213,302.07 hm2, and 100,389.33 hm2, respectively. Under the ecological protection scenario, construction land decreased by 6,375.87 hm2 compared to 2020. Overall, the ecological protection scenario significantly curbed the expansion of construction land, while the natural and urban development scenarios continued the trend of rapid expansion.
Land use data under different scenarios statistical table (hm2).
Type
2030
2060
Natural development scenario
Urban development scenario
Ecological protection scenario
Natural development scenario
Urban development scenario
Ecological protection scenario
Cultivated Land
525407.58
525407.31
555335.37
450720.45
450720.27
558592.38
Forest land
406410.56
406410.65
406410.74
401104.34
401105.42
401103.53
Grassland
170.91
170.91
170.91
153.90
153.90
153.90
Water area
34365.06
34290.63
36383.04
37737.27
37245.33
42287.94
Unused land
14.67
14.85
15.3
11.88
12.96
12.87
Construction land
136173.06
136247.49
104226.48
212814.00
213303.96
100391.22
Total
1102541.84
1102541.84
1102541.84
1102541.84
1102541.84
1102541.84
Spatial differentiation of construction land use change
Between 1990 and 2020, the expansion of construction land in the CZTUA demonstrated notable spatiotemporal heterogeneity (Figure 2). During the 1990–2000 period, construction land in Changsha expanded primarily along an east–west axis centered on Furong District, a process facilitated by the newly built Yinpenling Bridge and Yueliangdao Xiangjiang Bridge, which significantly improved transport connectivity across the Xiang River. Between 2000 and 2010, while Furong District remained the urban core, Changsha’s growth direction shifted markedly southward. Concurrently, construction activities in Xiangtan were concentrated in the southwestern part of Yuetang District, whereas newly developed land in Zhuzhou was predominantly located at the confluence of four districts. From 2010 to 2020, Tianxin and Yuhua Districts in Changsha continued to expand southward, while Changsha County extended eastward from existing built-up areas. Xiangtan exhibited a bidirectional north–south expansion pattern, and Zhuzhou’s urban growth aligned predominantly along a northeast–southwest axis, with the most pronounced development occurring in northern Tianyuan District. Overall, Changsha demonstrated the most rapid pace of urban expansion. Furong District, as the initial development nucleus, exerted a strong radiating influence on surrounding areas, consistently driving the outward spread of construction land in multiple directions.
The land use distribution of Changzhutan urban agglomeration from 1990 to 2020.
Land use maps of a region from 1990 to 2020 show changes in cultivated land, forest land, grassland, water areas, unused land, and construction land. The maps illustrate increased construction and reduced forest areas over time.
The land use patterns of the CZTUA under the three development scenarios for 2030 and 2060 are generally consistent (Figure 3). Construction land continues the spatial trend observed from 1990 to 2020, expanding outward from existing built-up areas. Forest land and grassland are predominantly distributed in the southern and northeastern parts of the CZTUA, while cultivated land is scattered between construction land and forest land. Under both the natural development and urban development scenarios, construction land dominates the central region, with most newly expanded areas converted from cultivated land, indicating substantial encroachment on agricultural areas. In contrast, under the ecological protection scenario, construction land exhibits a shrinking trend in both 2030 and 2060, effectively limiting the conversion of adjacent cultivated, forest land, and other land types.
The construction land change of CZTUA under different simulation scenarios from 2020 to 2030 and 2060.
Maps depicting land development scenarios within different timeframes for a region. The top row shows the natural, urban, and ecological scenarios for 2020-2030, highlighting existing construction in yellow and newly added areas in red. The bottom row illustrates scenarios for 2020-2060 with additional construction land in orange and decreased areas in red, alongside construction land in 2060 marked in light blue. A scale bar indicates 0 to 20 kilometers.Spatio-temporal characteristics of carbon storage changeTemporal dynamics of carbon storage
Between 1990 and 2020, the carbon storage in the CZTUA showed a net decrease of 8.9 × 108 kg (Equation 2; Table 7). This overall decline was the result of distinct decadal trends: an initial increase of 8.2 × 108 kg during 1990–2000 was followed by significant decreases of 14.9 × 108 kg and 2.0 × 108 kg in the subsequent periods of 2000–2010 and 2010–2020, respectively. Simulation results for 2030 project that under natural development, urban development, and ecological protection scenarios, the carbon storage of the CZTUA will reach 286.4 × 108 kg, 286.4 × 108 kg, and 287.6 × 108 kg, respectively. Relative to the 2020 baseline, both the natural and urban development scenarios indicate a slight decrease of 0.2 × 108 kg. In contrast, the ecological protection scenario shows an increase of 1.02 × 108 kg. By 2060, carbon storage is projected to be 281.6 × 108 kg, 281.6 × 108 kg, and 286.2 × 108 kg under the three scenarios. Compared to 2020, this represents a substantial decrease of 5.0 × 108 kg for both the natural and urban development scenarios, while the ecological protection scenario shows a considerably smaller reduction of only 0.4 × 108 kg.
Historical and contextual carbon stock data statistics table.
Year
1990
2000
2010
2020
2030
2060
Natural development scenario
Urban development scenario
Ecological protection scenario
Natural development scenario
Urban development scenario
Ecological protection scenario
Carbon storage (108kg)
295.3
303.5
288.6
286.6
286.4
286.64
287.76
281.6
281.6
286.2
Spatial differentiation of carbon storag<bold>e</bold>
Between 1990 and 2020, areas of carbon storage loss in the CZTUA were primarily distributed around Changsha, Xiangtan, and Zhuzhou, expanding continuously around the original low-value zones, largely aligned with the direction of construction land expansion (Figure 4). From 1990 to 2000, the high-value carbon storage areas in the northeastern part of the CZTUA increased, driven by the extensive conversion of cultivated land to forest land during this period, which resulted in a notable rise in carbon storage. Low-value carbon storage areas were mainly concentrated in the central part of the CZTUA, where the dominant land use types are construction land and water bodies. In contrast, high-value carbon storage areas were primarily located in the northeastern and southern parts of the CZTUA, where forest land and grassland are the main land use types. Under the three simulated scenarios for 2030 and 2060, the spatial distribution patterns of carbon storage were generally consistent. Under both the natural development and urban development scenarios, the low-value carbon storage areas expanded significantly. In contrast, under the ecological protection scenario, the expansion of low-value carbon storage areas was notably smaller than in the other two scenarios, indicating that ecological protection measures can, to some extent, curb the rate of expansion of low-value carbon storage areas.
Spatial distribution of ecosystem carbon storage in the CZTUA from1990 to 2060.
Maps show changes in carbon stock across different scenarios and years. The top row reflects years 1990, 2000, 2010, and 2020. The second row presents scenarios for 2030: natural, urban, and ecological protection. The third row covers 2060 with the same scenarios. A color scale from green to red indicates low to high carbon stock, ranging from 238.14 to 3365.55 kilograms. A scale bar shows a distance of 20 kilometers.Spatial clustering of carbon storage
The CZTUA was partitioned into 3 km × 3 km fishnet grid units, and spatial autocorrelation analysis was conducted on carbon storage. The global Moran’s I values for the four periods from 1990 to 2020 were 0.23, 0.20, 0.32, and 0.24, respectively (Equations 3, 4). All values are greater than zero, indicating a significant positive spatial autocorrelation in the distribution of carbon storage. Further analysis using Getis-Ord Gi* hotspot analysis (Equation 5) revealed that statistically significant cold spots (clusters of low carbon storage values) were primarily distributed in the central part of the CZTUA (Figure 5). Over the past three decades, these cold spots have generally shown an expanding trend, which aligns with the direction of construction land expansion. In contrast, hotspots (clusters of high carbon storage values) were mainly concentrated in the southern and northeastern parts of the CZTUA, corresponding to areas with concentrated forest land. These hotspots have exhibited a dynamically decreasing trend over the 30-year period.
Hot spot distribution of carbon storage in CZTUA from1990 to 2020. Note: Cold Spot indicates a cold spot, Hot Spot indicates a hot spot, Not Significant indicates not significant; *denotes p < 0.01, ** denotes p < 0.05, *** denotes p < 0.1.
Four maps show spatial distribution of cold and hot spots in an unspecified area across four years: 1990, 2000, 2010, and 2020. Symbols in blue represent cold spots with varying significance, while red symbols indicate hot spots. The maps indicate gradual changes in the distribution patterns over time, with both types of spots marked at increasing intensity levels, as seen by changes in the number of symbols. A scale of zero to twenty kilometers is provided for distance reference.Impact of urban expansion on ecosystem carbon storage
The bivariate spatial autocorrelation analysis between construction land area and carbon storage yielded results that were statistically significant at the p ≤ 0.001 level for all periods from 1990 to 2060. The global bivariate Moran’s I values were consistently negative, confirming a stable spatial negative correlation between construction land and carbon storage distribution. This indicates that areas with larger construction land areas tend to exhibit lower carbon storage values. Furthermore, the observed gradual decrease in Moran’s I values over time suggests an intensification of this negative spatial association. This trend points to a strengthening of dissimilar clustering, meaning that construction land and high carbon storage are becoming increasingly spatially segregated. Under the simulated scenarios for 2030 and 2060, the Moran’s I values were similar in both the natural development and urban development scenarios. In contrast, the ecological protection scenario demonstrated lower Moran’s I values compared to the other two scenarios. This indicates that the strength of the negative spatial correlation between construction land and carbon storage is most pronounced under the ecological protection scenario, highlighting its role in more effectively concentrating and preserving carbon stocks away from areas of urban expansion (Table 8).
Global bivariate autocorrelation coefficient between construction land and carbon storage (1990–2060).
Scenario
Global Moran’s I
Z-value
p-value
1990
−0.0653
−6.1661
0.001
2000
−0.1397
−12.8065
0.001
2010
−0.1859
−16.7441
0.001
2020
−0.2222
−20.1837
0.001
2030
Natural development
−0.2457
−22.4793
0.001
Urban development
−0.2474
−22.4657
0.001
Ecological protection
−0.2232
−20.2490
0.001
2060
Natural development
−0.2743
−25.0308
0.001
Urban development
−0.2746
−25.0714
0.001
Ecological protection
−0.2194
−19.8983
0.001
The results of the bivariate local spatial autocorrelation analysis reveal distinct spatial clustering patterns between construction land area and carbon storage from 1990 to 2020. Low-Low Clusters were primarily distributed in the peripheral areas of the CZTUA. Low-High Clusters were mainly located in the northeastern and southern parts of the CZTUA, where forest land serves as the dominant land use type with abundant carbon storage. However, the proportion of these Low-High clusters showed a declining trend over the study period. High-Low Clusters were predominantly concentrated in the central urban areas of Changsha, Xiangtan, and Zhuzhou, characterized by intensive construction land distribution. These clusters exhibited a notable expansion over the 30-year period, aligning closely with the direction of construction land expansion. High-High Clusters were relatively scarce and sporadically distributed within the CZTUA.
The spatial distribution patterns of construction land and carbon storage under different scenarios in 2030 and 2060 remain largely consistent with those observed during the 1990–2020 period (Figure 6). Among these patterns, the Low-Low, Low-High, and High-High clusters show no significant changes in their distribution or scale. In contrast, the High-Low clusters—primarily concentrated in Furong District, southern Yuetang District, and northeastern Tianyuan District—exhibit notable expansion. Under both the natural development and urban development scenarios, the spatial distribution patterns of construction land and carbon storage are nearly identical. However, under the ecological protection scenario, the extent of High-Low clusters is significantly smaller compared to the other two scenarios. This indicates that the ecological protection scenario can more effectively mitigate the spatial conflict between construction land and carbon storage.
Localized bivariate autocorrelation between construction land and carbon storage in CZTUA in 2030 and 2060 under different simulation scenarios.
Nine maps showing land change scenarios from 1990 to 2060. Each map uses colors to indicate significance levels: gray for not significant, dark blue for L-L, light blue for L-H, pink for H-L, and red for H-H. The maps are divided into three time periods (1990-2020, 2030, and 2060) with three scenarios each (natural development, urban development, and ecological protection). Each map includes a scale of 20 kilometers.
A linear regression model was further employed to analyze the impact of urban expansion on carbon storage in the CZTUA. The results show that for the period 1990–2020 (Figure 7), the regression between construction land and carbon storage yielded an R2 of 0.479, with a negative regression coefficient, indicating that the urban expansion led to a reduction in carbon storage. Specifically, the regression models between construction land and both above-ground and below-ground carbon storage showed R2 values greater than 0.3 with negative regression coefficients, confirming the adverse impact of urban expansion on vegetative carbon storage. Notably, the negative effect of urban expansion on below-ground carbon storage was consistently stronger than that on above-ground carbon storage across all study periods. However, the magnitude of this negative influence on both carbon compartments demonstrated a gradual weakening over time. Furthermore, the relationship between construction land and soil carbon storage showed an R2 consistently below 0.1, indicating minimal explanatory power and suggesting that urban expansion had limited adverse impact on soil carbon storage.
The liner regressionof urban expansion and ecosystem carbon storage from 1990 to 2000.
Twelve scatter plots display relationships between construction land use area and carbon storage (above-ground, below-ground, and soil) across different decades (1990-2020). Each plot shows regression lines with corresponding equations and R-squared values indicating the strength of correlation. All charts illustrate a general trend of decreased carbon storage with increased land use, with varying degrees of correlation evidenced by the R-squared values ranging from 0 to 0.763. The plots are labeled from a to d, corresponding to different time periods and types of carbon storage.
Regression analysis for the period 2030–2060 (Figure 8) showed that the R2 values for the relationship between changes in construction land and carbon storage were 0.919 under both natural and urban development scenarios, indicating a high degree of fit. However, under the ecological protection scenario, the fit was lower and the absolute value of the regression coefficient was highest, suggesting that carbon storage is more sensitive to changes in construction land in this scenario.
The liner regressionof urban expansion and ecosystem carbon storage during 2030–2060 under different simulation scenarios.
Three scatter plot graphs depict the relationship between construction land use area and carbon storage from 2030 to 2060 under three scenarios: natural development, urban development, and ecological protection. Each graph shows a downward trend with fitted lines. The natural and urban development scenarios exhibit a strong negative correlation, indicated by R-squared values of 0.919. The ecological protection scenario shows a weaker correlation with an R-squared value of 0.053. Axes are labeled with construction land use area in hectares and carbon storage in kilograms.DiscussionDiscussion of findings
The findings of this study align with existing research on the relationship between urban expansion and carbon storage. For instance, multiple studies have identified construction land expansion as a primary driver of carbon storage decline in regions like the Yangtze River Delta (Ding et al., 2021; Tao et al., 2022;) and Pearl River Delta (Wang et al., 2025; Qi et al., 2025; Chen et al., 2025), with a significant spatial overlap observed between areas of low carbon storage and high construction land density. This study corroborates these findings, demonstrating that low carbon storage areas in the CZTUA are predominantly located in the central urban zones characterized by concentrated construction land. This further validates the widespread negative impact of urban expansion on ecosystem carbon storage.
Regarding the methodology for carbon storage simulation, this study employed the InVEST model integrated with the FLUS model for multi-scenario prediction, an approach that has been increasingly applied in regional carbon storage assessments in recent years (Yang et al., 2020; Xiang et al., 2022; Li et al., 2023; Xu et al., 2025; Qin et al., 2025; Lai et al., 2025). Compared to studies utilizing models like CA-Markov or CLUE-S, the FLUS model demonstrates superior adaptability in simulating complex land-use conversions, enabling a more accurate representation of human activity-driven changes. Our simulations for 2030 and 2060 under three different scenarios reveal that the rate of carbon storage loss is significantly slowed under the ecological protection scenario. This conclusion aligns with findings from research conducted in the Changchun-Jilin-Tumen region (Wu et al., 2024), Yellow River Basin (Wang et al., 2024) and Hainan coastal zone (Gong et al., 2025), collectively highlighting the positive role of ecological protection policies in mitigating carbon loss.
Furthermore, this study employed bivariate spatial autocorrelation analysis and linear regression models to quantify the impacts of urban expansion on different carbon pools. The results indicate that urban expansion exerts a stronger negative effect on belowground carbon storage than on aboveground carbon storage, while its impact on soil carbon storage is not statistically significant. Discrepancies with some other studies, which might report more pronounced effects on soil carbon, likely stem from differences in regional land-use structures, soil types, and the intensity of human disturbance.
Limitations and implications
Despite efforts to maintain scientific rigor in methodology and data, this study has several limitations.
Firstly, relatively low resolution of land use data. The spatial resolution of the land use data used in this study is 30 meters. While sufficient for regional-scale analysis, it remains inadequate for detailing internal urban structures such as construction land patterns and green space distribution. Higher-resolution data (e.g., 10 meters or 5 meters) would allow more precise identification of small-scale land conversion processes, thereby improving the accuracy of carbon storage estimation.
Secondly, the limited representativeness of carbon density data introduces uncertainty in carbon stock estimation. The carbon density data applied in this study were primarily sourced from existing literature (Feng et al., 2025; Lei et al., 2025) and China’s terrestrial ecosystem carbon density dataset, lacking validation through field sampling. Given the substantial spatial heterogeneity in carbon density across regions, vegetation types, and soil conditions, the use of uniform parameters may lead to systematic over or under estimation of carbon storage. As carbon density serves as a core input to the InVEST model, such bias can propagate through the calculation, directly affecting the accuracy and reliability of the final carbon stock estimates. Future research should integrate field surveys and remote sensing inversion to develop a carbon density database that better reflects local characteristics.
Thirdly, uncertainties in model simulation: Although the FLUS model incorporated multiple scenarios for simulating future land use, it still cannot fully account for sudden policy changes, natural disasters, or other unforeseen events that influence land use. Moreover, the ecological protection scenario remains somewhat idealized, as its real-world implementation would likely face multiple socioeconomic development pressures. Future studies could introduce more complex system dynamics models or participatory scenario simulations to enhance predictive reliability.
Finally, spatial and temporal limitations. This study focused on the core area of the CZTUA and did not cover the entire region, potentially overlooking spatial heterogeneity in carbon storage across the broader area. Additionally, although the study spanned 70 years, it did not incorporate the effects of extreme climate events or major policy adjustments. Future work could consider longer time-series data or couple climate models for a more comprehensive analysis.
In summary, future research should further refine data accuracy, model mechanisms, carbon pool categorization, and multi-scale integration to more comprehensively and precisely assess the impact of urban expansion on ecosystem carbon storage. This will provide a more actionable scientific basis for regional low-carbon development and territorial spatial optimization.
Conclusion
(1) From 1990 to 2020, the construction land area in the CZTUA increased by 85,482.09 hm2, with newly expanded areas predominantly clustered around existing built-up zones. Simulation results indicate that under both the natural development and urban development scenarios in 2030 and 2060, the expansion of construction land is projected to follow the historical trend observed during 1990–2020. In contrast, under the ecological protection scenario, construction land exhibits a contraction trend.
(2) From 1990 to 2020, the ecosystem carbon storage in the region decreased by 8.7 × 108 kg. Spatially, areas of low carbon storage strongly overlapped with construction land expansion area. Simulations for 2030 and 2060 suggest that under both natural and urban development scenarios, carbon storage reduction will continue along the trajectory observed in previous decades. However, under the ecological protection scenario, the rate of carbon loss is significantly slowed, demonstrating that ecological conservation measures can effectively mitigate the decline in regional carbon storage.
(3) A persistent spatial negative correlation was observed between construction land and carbon storage from 1990 to 2020. High-Low clustering areas—where high construction land coincides with low carbon storage—were mainly concentrated in the central districts of Changsha, Xiangtan, and Zhuzhou, and showed an expanding trend over time. Although construction land expansion consistently contributed to carbon storage reduction, its negative effect has gradually weakened over the decades. Furthermore, the expansion exerted a stronger adverse impact on below-ground carbon storage than on above-ground carbon storage, while its effect on soil carbon storage was minimal.
In summary, our findings deliver a clear policy message: the ecological protection scenario plays a critical role in curbing construction land expansion, alleviating spatial conflicts, and preserving regional carbon sinks. Urban National Territorial Space Planning Strategy should prioritize ecological conservation policies to achieve a balance between urban development and carbon sequestration, ensuring sustainable regional development.
Data availability statement
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The authors declare that no Gen AI was used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
ReferencesAliM. A.JamalS.WahidN.AhmadW. S. (2025). Leveraging CA-ANN modelling for SDGs alignment: previse future land use patterns and their influence on Mirik Lake of sub-Himalayan region. World Dev. Sustain.6:100218. doi: 10.1016/j.wds.2025.100218AlmeidaB.MonteiroL.TiengoR.GilA.CabralP. (2025). Spatially explicit assessment of carbon storage and sequestration in forest ecosystems. Remote Sens. Appl. Soc. Environ.38:101544. doi: 10.1016/j.rsase.2025.101544ChenH. X.ChenB. Y.ChenW. H.ChenH.ChenB.ChenW.. (2023). Contribution of future urbanization to summer regional warming in the Pearl River Delta. Urban Clim.49:101476. doi: 10.1016/j.uclim.2023.101476ChenC.LiangJ. T.ZhangW. W. (2025). Quantifying dynamics of ecosystem carbon storage under influence of land use and land cover change in coastal zone from remote sensing perspective. Sustain. Horizons14:100146. doi: 10.1016/j.horiz.2025.100146DaiX. A.LiD. L.ZhengY. M.DaiX.LiD.ZhengY.. (2025). Analysis of spatiotemporal changes and driving factors of green and blue carbon stocks in Zhangzhou City based on PLUS and InVEST models. J. Clean. Prod.528:146743. doi: 10.1016/j.jclepro.2025.146743DingT. H.ChenJ. F.FangZ.DingT.ChenJ. (2021). Assessment of coordinative relationship between comprehensive ecosystem service and urbanization: a case study of Yangtze River Delta urban agglomerations, China. Ecol. Indic.133:108454. doi: 10.1016/j.ecolind.2021.108454DongY. H.PengF. L.LiH.MenY.-Q. (2023). Spatial autocorrelation and spatial heterogeneity of underground parking space development in Chinese megacities based on multisource open data. Appl. Geogr.153:102897. doi: 10.1016/j.apgeog.2023.102897DuA. T.TongS. Q.RenJ. Y.DuA.TongS.RenJ.. (2025). Spatio-temporal variation and influencing factors of carbon emissions from land use change in Xilingol region of Inner Mongolia, China. Ecol. Indic.176:113633. doi: 10.1016/j.ecolind.2025.113633FengX. Y.ZhaoX.TongL.FengX.WangS.DingR.. (2025). Impacts of land use and cover change on carbon storage: multi-scenario projections in the arid region of Northwest China. Reg. Sustain.6:100248. doi: 10.1016/j.regsus.2025.100248GongW. F.DuanX. Y.SunY. X.ZhangY.JiP.TongX.. (2025). Multi-scenario simulation of land use/cover change and carbon storage assessment in Hainan coastal zone from perspective of free trade port construction. J. Clean. Prod.385:135630. doi: 10.1016/j.jclepro.2022.135630HouY. L.WangL. X.LiZ. W.HouY.WangL.LiZ.. (2024). Landscape fragmentation and regularity lead to decreased carbon stocks in basins: evidence from century-scale research. J. Environ. Manag.367:121937. doi: 10.1016/j.jenvman.2024.121937, 39074435HuangM. L.MamitiminY. S.AbuliziA. B.HuangM.MamitiminY.AbuliziA.. (2024). Integrated assessment of land use and carbon storage changes in the Tulufan-Hami Basin under the background of urbanization and climate change. Int. J. Appl. Earth Obs. Geoinf.135:104261. doi: 10.1016/j.jag.2024.104261JiaB. Y.XieM. F.WuJ.JiaB.XieM.ZhaoJ. (2025). Towards low carbon urban agglomerations: spatiotemporal characteristics and influencing factors of carbon emission intensity and network linkages in China’s urban agglomerations. Ecol. Indic.177:113728. doi: 10.1016/j.ecolind.2025.113728KongW. L.ShenW. C.YuC. Y.KongW.ShenW.YuC.. (2025). The neglected cost: ecosystem services loss due to urban expansion in China from a triple-coupling perspective. Environ. Impact Assess. Rev.112:107827. doi: 10.1016/j.eiar.2025.107827LaiS.LeoneF.ZoppiC. (2017). Land cover changes and environmental protection: a study based on transition matrices concerning Sardinia (Italy). Land Use Policy67, 126–150. doi: 10.1016/j.landusepol.2017.05.030LaiJ. L.QiS.ChenJ. D.LaiJ.ChenJ.GuoJ.. (2025). Exploring the spatiotemporal variation of carbon storage on Hainan Island and its driving factors: insights from InVEST, FLUS models, and machine learning. Ecol. Indic.172:113236. doi: 10.1016/j.ecolind.2025.113236LeiX.ZhouY.HuoP. (2025). Impacts of multi-scenario land use change on carbon storage and its economic value in the Qilian Mountains region, China. J. Nat. Conserv.88:127042. doi: 10.1016/j.jnc.2025.127042LiG. Y.ChengG.LiuG. H.LiG.LiuG.ChenC.. (2023). Simulating the land use and carbon storage for nature-based solutions (NbS) under multi-scenarios in the three gorges reservoir area: integration of remote sensing data and the RF–Markov–CA–InVEST model. Remote Sens15:5100. doi: 10.3390/rs15215100LiL.HuangX. J.YangH. (2023). Optimizing land use patterns to improve the contribution of land use planning to carbon neutrality target. Land Use Policy135:106959. doi: 10.1016/j.landusepol.2023.106959LiS. H.LinW. (2025). A hybrid landscape metric-enhanced cellular automata model (LE-CA) for land use/land cover change simulation: an application to coastal wetlands. Ecol. Model.508:111209. doi: 10.1016/j.ecolmodel.2025.111209LiuC. L.LiuD. H.LiP.LiuC.LiuD.LiX.. (2023). Assessment of occupation of natural habitat by urban expansion and its impact on crucial ecosystem services in China’s coastal zone. Ecol. Indic.154:110682. doi: 10.1016/j.ecolind.2023.110682LuanC. X.LiuR. Z.LiY.LuanC.LiuR.ZhangQ. (2024). Comparison of various models for multi-scenario simulation of land use/land cover to predict ecosystem service value: a case study of Harbin-Changchun urban agglomeration, China. J. Clean. Prod.478:144012. doi: 10.1016/j.jclepro.2024.144012MaS. L.HuangJ. L.WangX. X.MaS.HuangJ.WangX.. (2025). Multi-scenario simulation of low-carbon land use based on the SD-FLUS model in Changsha, China. Land Use Policy148:107418. doi: 10.1016/j.landusepol.2024.107418MaJ.LiX. T.JiaB. Q.LiuX.LiT.ZhangW.. (2021). Spatial variation analysis of urban forest vegetation carbon storage and sequestration in built-up areas of Beijing based on i-tree eco and kriging. Urban For. Urban Green.66:127413. doi: 10.1016/j.ufug.2021.127413MengY.DuW. X.ChenL.DuW.ZhaoZ.DuH.. (2025). Ecosystem services in the northern Tianshan urban agglomeration: nonlinear responses to natural and human factors and threshold-based spatial optimization strategies. J. Hydrol. Reg. Stud.61:102682. doi: 10.1016/j.ejrh.2025.102682MovasaghiY.KhosravR.MohammadyM.PourghasemiH. R.GhoddousiA.KuemmerleT. (2025). Exploring spatial patterns of human–bear conflict in southwestern Iran due to future land-use change. Biol. Conserv.311:111459. doi: 10.1016/j.biocon.2025.111459NiuL.ZhangZ. F.LiangY. Z.ZhangZ.LiangY.HuangY. (2022). Assessing the impact of urbanization and eco-environmental quality on regional carbon storage: a multiscale spatio-temporal analysis framework. Remote Sens14:4007. doi: 10.3390/rs14164007QiL.HN.OskenbayevY.AlisherS.HairisK. (2025). Impact of rapid urban construction land expansion on spatial inequalities of ecosystem health in China: evidence from national, economic regional, and urban agglomeration perspectives. Ecol. Indic.172:113196. doi: 10.1016/j.ecolind.2025.113196QinY. W.YanY.DongJ. Q.ZhenyuZ.ShuangjiangL.JianshengC.. (2025). Scenario-driven modeling of mountain ecosystems: land use-carbon dynamics simulation based on the coupled SD-FLUS-InVEST framework. Ecol. Model.510:111293. doi: 10.1016/j.ecolmodel.2025.111293SiabiE. K.Kabo-bahA. T.AnornuG.AkpotiK.MorteyE. M.Manson IncoomA. B.. (2025). Future land use simulation modeling for sustainable urban development under the shared socioeconomic pathways in west African megacities: insights from Greater Accra region. J. Environ. Manag.376:124300. doi: 10.1016/j.jenvman.2025.124300, 39919577TaoY.TaoQ.SunX.QiuJ.PueppkeS. G.OuW.. (2022). Mapping ecosystem service supply and demand dynamics under rapid urban expansion: a case study in the Yangtze River Delta of China. Ecosyst. Serv.56:101448. doi: 10.1016/j.ecoser.2022.101448WangX. Y.SuF. Z.YanF. Q.WangX.SuF.YanF.. (2025). Balancing on carbon storage and ecological security in urban areas: the case of Dongguan city, China. Sustain. Futures9:100468. doi: 10.1016/j.sftr.2025.100468WangH. Y.WuL. H.YueY. S.WangH.WuL.YueY.. (2024). Impacts of climate and land use change on terrestrial carbon storage: a multi-scenario case study in the Yellow River Basin (199-2050). Sci. Total Environ.930:172557. doi: 10.1016/j.scitotenv.2024.172557, 38643873WangZ.XuL.ShiY.MaQ.WuY.LuZ.. (2021). Impact of land use change on vegetation carbon storage during rapid urbanization: a case study of Hangzhou, China. Chin. Geogr. Sci.31, 209–222. doi: 10.1007/s11769-021-1183-yWuT.AnM. Q.ZhangL. L.AnM.ZhangL.WuX.. (2024). Modeling urban expansion and its impacts on carbon storage through integrative scenario analysis for sustainable development in the Changchun-Jilin-Tumen region. Sustain. Cities Soc.117:105970. doi: 10.1016/j.scs.2024.105970XiangS. J.WangY.DengH.XiangS.YangC.WangZ.. (2022). Response and multi-scenario prediction of carbon storage to land use/cover change in the main urban area of Chongqing, China. Ecol. Indic.142:109205. doi: 10.1016/j.ecolind.2022.109205XuY. H.XiaoC. H.BaiJ. H.XuY.XiaoC.BaiJ.. (2025). Spatio-temporal evolution of coastal wetlands and carbon storage driven by natural and anthropogenic processes in Yancheng City, China. Estuar. Coast. Shelf Sci.325:109466. doi: 10.1016/j.ecss.2025.109466YangH.HuangJ. L.LiuD. F. (2020). Linking climate change and socioeconomic development to urban land use simulation: analysis of their concurrent effects on carbon storage. Appl. Geogr.115:102135. doi: 10.1016/j.apgeog.2019.102135YangG. T.SuC.ZhangH.YangG.ZhangX.LiuY. (2023). Tree-level landscape transitions and changes in carbon storage throughout the mine life cycle. Sci. Total Environ.905:166896. doi: 10.1016/j.scitotenv.2023.166896, 37717743YangH.WangY. W.TuP.WangY.ZhongY.HuangC.. (2024). Evaluating the effects of future urban expansion on ecosystem services in the Yangtze River Delta urban agglomeration under the shared socioeconomic pathways. Ecol. Indic.160:111831. doi: 10.1016/j.ecolind.2024.111831ZengS. Y.LiuJ. N.MaJ.ZengS.LiuJ.YangY.. (2025). Optimizing cropland expansion for minimizing ecosystem service loss in China. Geogr Sustainability6:100299. doi: 10.1016/j.geosus.2025.100299ZhangJ. P.CaoP. H.RoosliR. (2025). Assessing land use and carbon storage changes using PLUS and InVEST models: a multi-scenario simulation in Hohhot. Environ. Sustain. Indic.26:100655. doi: 10.1016/j.indic.2025.100655ZhangK. L.FangB.ZhangZ. C.ZhangK.ZhangZ.LiuT.. (2024). Exploring future ecosystem service changes and key contributing factors from a “past-future-action” perspective: a case study of the Yellow River Basin. Sci. Total Environ.926:171630. doi: 10.1016/j.scitotenv.2024.171630, 38508260ZhaoJ.ShaoZ.XiaC. Y.XiaC.FangK.ChenR.. (2022). Ecosystem services assessment based on land use simulation: a case study in the Heihe River basin, China. Ecol. Indic.143:402. doi: 10.1016/j.ecolind.2022.109402
Edited by: Liye Wang, Shandong University of Finance and Economics, China
Reviewed by: Hou Jiao, Wuhan Polytechnic University, China
Chenning Deng, Chinese Research Academy of Environmental Sciences, China