Land Use Dynamics and Ecosystem Service Trade-offs in China’s Terraced Landscapes
A
DieChen1
WeiWei2,3✉Emailweiwei@rcees.ac.cn
LidingChen2,3
1School of Ecology and Nature ConservationBeijing Forestry University100091BeijingChina
2
A
State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-environmental SciencesChinese Academy of Sciences100085BeijingChina3University of Chinese Academy of Sciences100049BeijingChina
Die Chen1, Wei Wei2,3,*, and Liding Chen2,3
1 School of Ecology and Nature Conservation, Beijing Forestry University, Beijing 100091, China
2 State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China
3 University of Chinese Academy of Sciences, Beijing 100049, China
* Correspondence: weiwei@rcees.ac.cn
Abstract
China’s terraced landscapes represent a critical integration of agricultural heritage and ecological functionality, providing vital ecosystem services (ESs) such as water yield, soil conservation, and carbon storage. Despite their ecological importance, the spatiotemporal dynamics of ES provision in terraced systems and their trade-off/synergy relationships remain inadequately quantified. Here we address this knowledge gap by (1) quantifying nationwide changes in terraced land use from 2018 to 2023, (2) evaluating the corresponding variations in key ESs, and (3) elucidating the trade-off and synergy relationships among these services to inform sustainable land management. Utilizing the InVEST model integrated with Landsat imagery, climate, and soil data, we reveal significant transformations in China’s terraces. Rice terraced area declined dramatically by 39.2% (91400 to 55600 km2), while dryland terraces expanded by 10.1% and woodland terraces surged by 104.2%.These land-use shifts led to only a marginal increase in overall water yield (211.6 to 218.1 mm/year), but a substantial rise in soil conservation (+ 18.5%, from1.57×109 to 1.86×109 t), and a moderate increase in carbon storage (+ 1.7%, from 2.96×109 to 3.01×109 t). Trade-offs among services overwhelmingly outweighed synergies, particularly between water yield and carbon storage, though strong regional disparities were evident. Notably, synergies prevailed across 72.5% of the semi-arid Loess Plateau, whereas trade-offs dominated in humid southern regions (e.g., 78.9% of the Sichuan Basin). These findings reveal how climate-mediated land-use dynamics shape multi-service interactions and highlight the need for region-specific governance to balance ecological restoration with water security in China’s terraced landscapes.
Keywords:
Terraces
Ecosystem services
Land use change
Trade-offs
Synergies
InVEST model
China
A
A
1 Introduction
Terraced landscapes—sculpted by centuries of agricultural activity on steep terrains—not only symbolize human ingenuity but are also crucial for rural livelihoods and ecological stability1–4. They epitomize a symbiotic relationship between human society and natural ecosystems, delivering multiple ecosystem services fundamental to sustainable development5. With the shift toward an ecological civilization in agriculture, the value of terraces has transcended their original food production role to become an integral component of modern ecosystem management6. In China’s vast and topographically diverse environment, terraced fields are prominent features of mountainous regions2,7. Whether it is the breathtaking Hani Rice Terraces in Yunnan, the majestic Longji terraces in Guangxi, or the arid Loess Plateau terraces in the northwest, the cultural and ecological impact of these systems extends far beyond their initial agricultural function2. Terraces optimize land resource utilization, enhance agricultural productivity, and play an irreplaceable role in water regulation, soil preservation, carbon sequestration, ecological restoration, and biodiversity protection8–10.
Ecosystem services (ESs) are the benefits that humans obtain directly or indirectly from ecosystems11,12. They link ecosystem processes with human well-being, supporting societies from basic needs to advanced development goals13,14. As methods to quantify ES across spatial and temporal scales have matured, ES metrics have emerged as pivotal benchmarks for sustainable ecosystem stewardship15. The multifunctionality of terrace systems extends beyond food production to a broad range of ES that underpin environmental health and social welfare9,16. These services—including hydrological regulation, soil conservation, carbon storage, microclimate modulation, biodiversity habitat provision, and even cultural/aesthetic values—are crucial for sustaining local ecosystems and the communities that depend on them. In enhancing the ecological performance of terraced agriculture, these ES also provide significant economic and social benefits to surrounding communities17,18. However, despite their seemingly enduring presence, the provision of ES in terraced landscapes is dynamic and sensitive to human activities and broader environmental changes19.
A key challenge in managing terraced ecosystems is understanding the complex interactions among multiple ES. Enhancing one service often incurs trade-offs with others, while in some cases synergistic improvements are possible20, 21. Such trade-offs and synergies have profound implications for the overall health of terrace ecosystems and human well-being. Researchers have long emphasized that quantifying and managing these ES relationships is critical for sustainable ecosystem management and regional development22,23. Under escalating pressures from global climate change and population growth, the interdependence of ES—manifesting as both co-benefits and conflicts—has become a central research focus24. In terraced landscapes, multiple ES form intricate networks of mutual influence where co-benefits and competitive compromises dynamically coexist25,26. For example, afforestation can significantly enhance carbon storage but may reduce water yield via increased evapotranspiration, creating a carbon–water trade-off27,28. Similarly, agricultural intensification aimed at maximizing crop yields can undermine hydrological regulation and biodiversity if it reduces vegetative cover or landscape heterogeneity29. Conversely, well-designed ecological restoration efforts have shown synergistic outcomes: strategic afforestation not only sequesters carbon and stabilizes soil via root networks (reducing erosion), but also can create habitat heterogeneity that boosts biodiversity and improves water retention30,31. Vegetation acts as a “green barrier”, optimizing local hydrological cycles by enhancing soil water storage, thus achieving dual benefits for water yield and soil preservation, especially in arid regions32.
Despite their importance for soil conservation in sloping environments, many terrace systems have experienced accelerated degradation and land-use change due to combined natural forces (e.g. more intense rainfall events) and anthropogenic pressures (e.g. rural depopulation or expansion of intensive agriculture)33–35. Terrace degradation or abandonment can directly diminish local ES delivery, notably by reducing water retention capacity and carbon sequestration potential9. Globally, land-use transitions driven by population growth, urbanization, and economic globalization are reconfiguring resource exploitation paradigms36,37. These shifts often undermine landscape integrity through two primary pathways. First is biogeochemical cycle disruption: for instance, equivalent land disturbances can cause ~ 42% greater soil carbon loss in arid regions compared to humid regions, due to lower resilience of dryland soils38. Second is habitat fragmentation: conversion and abandonment of traditional agro-ecosystems can break continuous habitats—e.g., in North China’s Beijing-Tianjin-Hebei region, landscape fragmentation (1980–2020) led to a 68% reduction in average habitat patch size and a 3.7-fold increase in patch isolation, severing wildlife dispersal corridors39. These examples illustrate how unchecked land-use change can erode the ecological functions of terraced and non-terraced landscapes alike.
Terraces in China are increasingly recognized as critical green infrastructure for simultaneously boosting agriculture and safeguarding the environment. Their widespread implementation across hillslopes aims to mitigate soil erosion, enhance water regulation, and expand arable land in otherwise marginal terrains. Nonetheless, significant knowledge gaps remain regarding the current ES performance of China’s terraced landscapes. Comprehensive, up-to-date evaluations of their ES provision—particularly in water yield, soil conservation, and carbon storage—are lacking, as is an understanding of how these services co-vary and trade off across different regions and scales. In addition, the spatial heterogeneity of ES in terrace systems (i.e., how trade-offs and synergies might differ between arid vs. humid regions or across different land-use contexts) is insufficiently understood. Addressing these gaps is essential for establishing a robust baseline of terrace-derived ES, refining ecosystem modeling approaches for complex agricultural landscapes, and guiding sustainable terrace management and policy.
This study systematically assesses the spatiotemporal dynamics of terraced land use across China (2018–2023), quantifying the temporal evolution of three critical ESs—water yield, soil conservation, and carbon storage. We further analyze the complex synergies and trade-offs governing ES interactions and quantify their regional heterogeneity to identify spatially explicit land management priorities. Our findings offer a robust scientific foundation to inform evidence-based policy formulation, ensuring the long-term ecological and cultural sustainability of these vital landscapes.
2 Data and methods
2.1 Data sources
This study integrates diverse data from multiple sources. A nationwide terraced field spatial distribution dataset (30 m resolution) was obtained from Cao et al. 7, which served as the baseline raster dataset of terrace locations. Land use/land cover data for the years 2018 and 2023 (30 m resolution, interpreted from Landsat imagery and validated by field surveys) were acquired from the Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn). Climatic inputs were derived from daily precipitation records (2018–2023) at 235 meteorological stations, sourced from the National Climate Centre of China (http://data.cma.cn). These station records were interpolated into continuous precipitation surfaces (1 km resolution) using ANUSPLIN v4.4 spline interpolation with elevation as a covariate40, then downscaled to 30 m resolution via elevation-dependent regression to match the fine spatial scale of terrace mapping. We obtained Moderate Resolution Imaging Spectroradiometer (MODIS) NDVI time-series data (MOD13A, 250 m resolution) for 2018–2023 from NASA’s EOS Data Gateway, to support vegetation and land cover assessments. Topographic data came from the Shuttle Radar Topography Mission digital elevation model (SRTM DEM, 90 m resolution), and a 1:1,000,000 scale soil type map was obtained from the RESDC platform. Soil property measurements (e.g. texture, organic matter) for each soil type were determined through field surveys to parameterize soil-related model inputs (such as erodibility factors). All spatial datasets (land use, DEM, soil maps) were projected and resampled to a common 30 m grid using nearest-neighbor interpolation, ensuring consistent alignment and resolution for subsequent spatial analyses and modeling.
2.2 Ecosystem service quantification
We focused on three key ecosystem services in terraced landscapes—water yield, soil conservation, and carbon storage—selected based on: (i) relevance under the Common International Classification of Ecosystem Services (CICES) framework (covering provisioning, regulating, and supporting categories), (ii) importance to local stakeholders (especially rural communities relying on terraces), and (iii) data availability for robust modeling. Each ES was quantified using modules from the InVEST (Integrated Valuation of Ecosystem Services and Tradeoffs) model suite41, a widely used tool for spatially explicit ES assessment.
Water yield service was quantified using the water yield module in InVEST model (https://naturalcapitalproject.stanford.edu/invest/). In the InVEST model, the water yield module is based on water balance model and determines the annual water yield of different pixels by using the Budyko curve and annual total precipitation data. The specific formula is:
, where
is the annual water yield (mm) of land use type ‘j’ in pixel ‘i’,
is the actual annual evapotranspiration (mm) of land use type ‘j’ in pixel ‘i’, and
is the annual total precipitation (mm) of pixel ‘i’. Hence, the water yield module of InVEST model requires annual total precipitation, reference evapotranspiration, root depth, plant available water content, land use type, watersheds delineation, biophysical properties and Z parameter values as input data. Further methodological details can be found in the InVEST User’s Guide41.
Soil conservation service was quantified using the Sediment Delivery Ratio (SDR) module in the InVEST model (https://naturalcapitalproject.stanford.edu/invest/). In the InVEST model, the SDR module estimates the capacity of ecosystems to prevent soil erosion by calculating the difference between potential soil loss (without vegetation cover) and actual soil loss (under current land cover). The model integrates the Revised Universal Soil Loss Equation (RUSLE) and sediment routing processes. The specific formula for soil conservation is:
. where
is the soil conservation amount (t·ha−1·yr−1),
is the rainfall erosivity factor (MJ·mm·ha⁻¹·h⁻¹·yr⁻¹), K is the soil erodibility factor (t·ha·h·ha−1·MJ−1·mm−1), LS is the slope length and steepness factor (dimensionless), C is the cover-management factor (dimensionless), and P is the support practice factor (dimensionless). Actual sediment export is further adjusted by sediment delivery ratios based on hydrological connectivity. The SDR module of the InVEST model requires annual rainfall erosivity (R), soil erodibility (K), digital elevation data (for LS), land use/land cover data (for C and P), watershed boundaries, biophysical property tables, and sub-catchment flow thresholds as input data. Further methodological details can be found in the InVEST User’s Guide41.
Carbon storage service was quantified using the carbon storage module in InVEST model (https://naturalcapitalproject.stanford.edu/invest/). For each land use type, carbon storage module estimates the amount of carbon in each of four major carbon pools (aboveground, underground, soil, and cadaver organic matter). The specific formula is:
. Where
is total carbon storage (t),
is aboveground carbon storage (t),
is underground carbon storage (t),
is soil carbon storage (t),
is cadaver organic matter carbon storage (t). The carbon storage module of the InVEST model requires land use data and the property table of carbon amount across the basic four pools for each land use pixel as input data41.
2.3 Trade-offs and synergies analysis
To evaluate interactions among the three ES, we conducted spatial correlation analyses at two scales: national and regional. Trade-offs are defined as situations where an increase in one ES is associated with a decrease in another, whereas synergies occur when two ES either increase or decrease together (indicating co-benefits or co-losses). For each pair of services (water yield vs. soil conservation, water yield vs. carbon storage, and soil conservation vs. carbon storage), we calculated the Pearson correlation coefficient (r) between their values across all terraced pixels. A significantly positive r indicates a synergy (both services tend to be high or low in tandem), while a significantly negative r indicates a trade-off (one service is high where the other is low). We categorized correlation strength as: strong synergy (r > 0.7), moderate synergy (0.3 < r ≤ 0.7), weak or no correlation (-0.3 ≤ r ≤ 0.3), moderate trade-off (-0.7 ≤ r<-0.3), and strong trade-off (r<-0.7). This classification provides a spatially explicit picture of where and how strongly each type of ES relationship prevails.
1
Where rxy represents the spatial correlation coefficient, ranging between [-1,1]. When rxy>0 (indicating a positive correlation), it denotes a synergistic relationship between the two services. When rxy<0 (indicating a negative correlation), it signifies a trade-off relationship. xij and yij represent the respective grid cell values in the spatial datasets of different ESs.
In addition to correlation analyses, we standardized the ES values (z-scores) and applied a grid-based overlay to visualize trade-off and synergy hotspots. Maps were created to show areas of particularly high or low values for combinations of services. We also disaggregated the analysis by region to explore regional heterogeneity, specifically, we examined four major terraced regions of China: the Loess Plateau (semi-arid north), the Yunnan-Guizhou Plateau (southwest), the Sichuan Basin (south-central), and the Middle-Lower Yangtze River basin (humid east-central), reflecting distinct climatic and socio-ecological contexts. Within each region, we computed the same set of service correlations and categorized the outcomes, allowing comparison of trade-off/synergy patterns across regions. These analyses enable us to identify where trade-offs are most acute and where win-win synergies might be harnessed, thereby informing region-specific management recommendations.
3 Results
3.1 Spatiotemporal changes in terraced land use
Land use and land cover change in terraced areas reflects direct human-environment interactions and has major implications for the stability and functionality of ecosystem services. Across China’s terraced landscapes, we classified seven primary terraced land-use types: rice terraces, dryland (rainfed) terraces, woodland terraces, sparse woodland terraces, shrubland terraces, grassland terraces, and bareland terraces (terraced land temporarily fallow or without significant vegetation). Comparative analysis of terraced land use between 2018 and 2023 reveals significant spatial transformations. Notably, extensive conversions from dryland terraces to grassland occurred across northern Yunnan, eastern Shandong, and throughout Shanxi Province. Concurrently, key agricultural provinces such as Sichuan, Chongqing, and Hunan experienced pronounced shifts from rice paddies to dryland farming on terraces (Fig. 1). These transitions indicate dynamic alterations in terrace land-use structure, with potential impacts on regional ES provision.
Fig. 1
The spatial distribution of terraced land use in China, comparing (a) 2018 and (b) 2023. Major shifts are visible in several regions: blue areas (rice terraces) shrink notably in Sichuan, Chongqing, and Hunan, replaced by green areas (dryland terraces), while orange areas (Grassland terraces) expand in many locales, especially the Loess Plateau and southwest China.
From 2018 to 2023, China’s terraced landscapes underwent dramatic land-use transitions. The most striking change was the drastic 39.2% contraction of rice terraced area, plummeting from 91400 km2 to 55600 km2. The vast majority of lost rice terrace area (42,600 km2) was converted to dryland (rainfed) terraces, reflecting a shift away from paddy cultivation in many regions. Conversely, dryland terraces expanded by 10.1% (from 217,000 km2 to 239,000 km2), fueled mainly by conversions from other land covers into terraced arable land - notably grassland (37,000 km2), sparse woodland (17,000 km2), and shrubland (10,400 km2), though they also lost substantial area to woodland (18,400 km2) and grassland (56,200 km2). Most remarkably, woodland terraces surged by 104.2% (31,000 km2 to 63,300 km2), primarily sourced from dryland conversions (18,400 km2) and shrubland (7,000 km2). This expansion occurred despite 10100 km2 reverting to dryland. Meanwhile, shrubland terraces collapsed by 67.4% (33,100 km2 to 10,800 km2) and sparse woodland terraces halved (42,100 km2 to 21,100 km2), largely transitioning to dryland and denser woodland. Grassland terraces saw moderate growth (22.5%, 109,600 km2 to 134,300 km2), while bareland terraces changed minimally (11,000 km2 to 11,300 km2). These complex shifts highlight intense land-use dynamics and conversions among terrace types over a short five-year period, as visualized in the summary diagrams (Fig. 2&3).
Fig. 2
Changes in terraced land use area (2018–2023). This bar chart compares the total area of each terraced land-use category between 2018 and 2023. Rice terraces show a sharp decrease, dryland and woodland terraces an increase, while other categories like shrubland and sparse woodland decline.
Fig. 3
Land-use transitions in China’s terraced landscapes (2018–2023) – Sankey diagram. Flows illustrate the magnitude of conversion from each 2018 terraced land-use type to 2023 types. Thick flow lines from rice to dryland indicate massive rice-to-dryland conversions, while flows into woodland signify afforestation of various terrace types.
3.2 Water yield services of terraced fields
Applying the InVEST water yield model to China’s terraced landscapes, we estimated the total water yield and its spatial distribution for 2018 and 2023. Nationally, the aggregated water yield from all terraced lands was approximately 1.33×1011 m3 in 2018, increasing slightly to 1.37×1011 m3 in 2023. In terms of area-weighted average, the mean water yield across terraced fields rose from 211.6 mm in 2018 to 218.1 mm in 2023, a modest gain of about 3.1%. This marginal increase at the national scale masks considerable spatial variability. Figures 4a and 4b map the water yield per unit area for terraced fields in 2018 and 2023, respectively, highlighting both geographic patterns and temporal changes.
Overall, terraced regions in southern and southeastern China (with higher rainfall, such as parts of Guangxi, Guangdong, and Fujian) show the highest water yields per unit area, often exceeding 400–500 mm/year. In contrast, terraces in the semi-arid north and northwest (e.g., Loess Plateau regions) have much lower water yields, often below 150 mm/year, due to limited precipitation and higher evapotranspiration ratios. From 2018 to 2023, some areas exhibited increases in water yield: for instance, parts of Jiangxi, Zhejiang, and Guangdong provinces saw local water yield rises, potentially attributable to slight precipitation increases or land cover changes that reduced water consumption. Meanwhile, other areas, such as Yunnan and Sichuan, showed stable or even declining water yield in terraces, possibly reflecting reduced rainfall or increased evapotranspiration from growing tree cover (Fig. 4).
Fig. 4
Water yield from Chinese terraced fields, (a) 2018 and (b) 2023. These maps illustrate annual water yield (mm) for terraced pixels. Higher values are notable in regions with high rainfall (e.g., parts of southeast China), whereas lower values appear in drier terraced areas.
At the provincial scale, Jiangxi achieved the highest mean water yield per terraced hectare (531.65 mm in 2018 rising to 557.61 mm in 2023), followed by humid provinces like Guangdong and Zhejiang, which also showed upward trends over the five years. These provinces benefit from abundant rainfall and dense vegetation cover that, in moderation, can enhance infiltration and baseflow. On the other hand, traditionally drier provinces with extensive terraces, such as Gansu or Shanxi, have much lower water yield values (often < 100 mm) with little change over time. Such spatial disparities underscore the influence of climate on water service provision from terraces.
A
Fig. 5
Water yield capacity by terraced fields in different provinces, (a) 2018 and (b) 2023.
3.3 Soil conservation services of terraced fields
Terraced agriculture is historically renowned for its soil conservation benefits on steep slopes. Our analysis with the InVEST SDR model confirms that China’s terraced landscapes provide substantial erosion control, with notable spatial variability and recent temporal improvements. We calculated total soil conservation as the difference between potential erosion (no terraces or vegetation) and actual erosion on terraced land. In 2018, Chinese terraces conserved an estimated 1.57×109 t of soil (that would have been lost without these systems). By 2023, this annual soil retention increased to 1.86×109 t, an 18.5% rise, aligning with the observed expansion of vegetative cover (grasslands and woodlands) on terraces. The mean soil conserved per unit area also climbed from approximately 25.0 t/hm2·yr in 2018 to 29.60 t/hm2·yr in 2023, indicating improved erosion control efficacy on the average terraced hectare (Fig. 6).
Fig. 6
Soil conservation service provided by terraced fields, (a) 2018 and (b) 2023. Maps show the annual soil retention (tons per hectare per year) on terraces. Warmer colors (red/brown) indicate areas of high soil conservation value (usually steep, high-erosion-risk slopes where terraces prevent large soil losses), whereas green/blue areas indicate lower soil conservation demand or impact.
Soil conservation service levels exhibit a high degree of spatial heterogeneity, influenced by factors such as rainfall erosivity, slope steepness, soil erodibility, and land cover management. Generally, terraces in regions like the Loess Plateau and southwestern China achieve very high soil retention (often > 50 t/hm2·yr) because these areas have steep slopes and erosive rains, meaning terraces there prevent massive soil losses that would occur on un-terraced land. In 2018, provinces with the greatest total soil conservation by terraces included Yunnan (4.75×108 t/yr) and Sichuan (3.98×108 t/yr), reflecting their large terraced extents and high erosion potential terrain. By 2023, Yunnan’s soil conservation further increased to 5.27×108 t, and Sichuan’s to 4.70×108 t, thanks to ongoing re-vegetation and conservation efforts. Other provinces like Guizhou, Guangxi, Hubei, and Fujian also saw notable gains in soil retention. For example, Fujian’s terraced fields increased soil conservation from 1.02×108 t to 1.11×108 t over the five years (Fig. 7).
Fig. 7
Soil conservation by terraced fields in different provinces, (a) 2018 and (b) 2023.
3.4 Carbon storage in terraced fields
Carbon storage is a critical regulating service, tied to climate mitigation efforts. Our results show that China’s terraced ecosystems hold a considerable and growing stock of carbon. Using the carbon stock approach, we estimate that in 2018 the total carbon stored in vegetation and soils of China’s terraced lands was approximately 2.96×109 t. By 2023, this had increased to 3.01×109 t, an overall rise of about 1.7% in five years (Fig. 8). While modest in percentage, this increase corresponds to roughly 50 million tons of additional carbon sequestered, which is notable given the short period.
Fig. 8
Carbon storage in terraced fields across China, (a) 2018 and (b) 2023.
The spatial distribution of carbon storage on terraces is highly uneven (Fig. 9). In general, woodland terraces and grassland terraces have much higher carbon densities than cultivated terraces, due to greater biomass and soil organic matter accumulation under perennial vegetation. Thus, regions that saw afforestation or natural revegetation of terraces show significant carbon gains. For instance, the Yunnan-Guizhou Plateau and parts of the Loess Plateau, where many terraces were converted to forest or grass, now exhibit some of the highest per-hectare carbon stocks. In contrast, areas with active cropland terraces (especially annual crops) hold less carbon. In 2018, the provinces with the highest average carbon stock per terraced hectare were Zhejiang (84.1 t C/hm2) and Hunan (69.4 t C/hm2), both heavily forested provinces with many orchard or woodland terraces. By 2023, Fujian’s terrace carbon climbed to 82.3 t C/hm2, nearly catching up to Zhejiang (84.1 t C/hm2). Other provinces with notable carbon densities on terraces include Guizhou (63 t C/hm2 in 2023), Yunnan (61.5 t C/hm2), and Anhui (56.6 t C/hm2), reflecting substantial tree or shrub cover on many of their terraces.
Fig. 9
Carbon storage of terraced fields in different provinces, (a) 2018; (b) 2023
3.5 The trade-off and synergy relationship of ESs in terraced fields
3.5.1 The trade-offs and synergies among ESs of terraced fields at national scale
Ecosystems deliver a diverse array of services that are crucial to human welfare, yet these services often come with complex interdependencies, including both trade-offs and synergies. This study conducted a national-scale spatial correlation analysis which revealed that for terraced fields, the trade-offs between any two out of three examined ESs were more pronounced than the synergies (Fig. 10). These trade-offs occur due to the interactions within ecosystems and their common driving forces.
A notable pattern observed is the synergy between water yield and carbon storage services in the northern Loess Plateau, in contrast to the southern regions such as Sichuan and Chongqing where trade-offs are more evident (Fig. 10). These findings underscore the necessity of considering the delicate balance between different ESs when managing and planning land use to ensure sustainable outcomes.
Fig. 10
Trade-offs and synergies among water yield, soil conservation, and carbon storage on China’s terraces. (a) shows Water yield vs. Carbon storage, (b) shows Water yield vs. Soil conservation, (c) shows Soil conservation vs. Carbon storage.
In this study, the trade-offs and synergies based on the correlation coefficients are categorized into five types: strong synergy (0.7 to 1), synergy (0.3 to 0.7), no correlation (-0.3 to 0.3), strong trade-off (-1 to -0.7), and trade-off (-0.7 to -0.3). At the national scale, for terraced ecosystems, the relationship between water yield and soil conservation services showed that 43.1% of areas displayed no significant trade-off or synergy, 27.2% exhibited strong trade-offs, and 13.3% showed strong synergies. About 9.5% presented with moderate trade-offs, while 6.9% had moderate synergies, suggesting a slight prevalence of trade-offs over synergies. For soil conservation and carbon storage services, 36.4% of areas had no clear trade-offs or synergies, 29.6% indicated strong trade-offs, and 27.4% demonstrated strong synergies. Areas with moderate trade-offs made up 2.3%, and those with moderate synergies constituted 4.3%, indicating that trade-offs and synergies were roughly balanced. Different from the other two pairings, for water yield and carbon storage services, only 3.3% of areas had no significant trade-offs or synergies, 55.4% showed strong trade-offs, and 27.4% exhibited strong synergies, with the latter mainly distributed in the northern mountainous regions (Fig. 11). These findings point to complex interactions within terraced ecosystems that require balanced management approaches to optimize ESs.
Fig. 11
Trade-offs and synergies between ESs in terraced fields at national scales
3.5.2 The trade-offs and synergies among ESs of terraced fields at regional scale
The aggregate national picture can hide important regional differences. We therefore parsed the analysis into four macro-regions with distinct climatic and land-use contexts: (1) Loess Plateau (Northern semi-arid), (2) Yunnan-Guizhou Plateau (Southwestern highlands), (3) Sichuan Basin and Upper Yangtze (humid subtropical southwest), and (4) Middle-Lower Yangtze River Basin (humid subtropical east). Within each region, we evaluated the dominant ES interactions on terraced lands (Fig. 12).
In the Loess Plateau, strong synergy was predominant, accounting for 72.5%, whereas the Yunnan-Guizhou Plateau, Sichuan Basin, and the middle and lower reaches of the Yangtze River region were primarily characterized by strong trade-offs, accounting for 58.0%, 78.9%, and 78.2%, respectively (Fig. 12). Between soil conservation and carbon storage services in the Loess Plateau, a high proportion of strong trade-offs was observed, at 45.6%, while strong synergy accounted for only 14.8%. In contrast, synergies exceeded trade-offs in the Yunnan-Guizhou Plateau, Sichuan Basin, and the middle and lower reaches of the Yangtze River. In these regions, strong synergies comprised 34.8%, 35.3%, and 39.9%, respectively, while strong trade-offs were 20.7%, 18.3%, and 14.4% (Fig. 12).
The relationship between water yield and soil conservation services showed little variation across the four regions, with the largest proportion indicating no significant trade-offs or synergies. This was followed by strong trade-offs, strong synergies, moderate trade-offs, and moderate synergies (Fig. 12). In all regions, trade-offs were more prevalent than synergies. The proportions of strong trade-offs in the Loess Plateau, Yunnan-Guizhou Plateau, Sichuan Basin, and the middle and lower reaches of the Yangtze River were 27.5%, 29.9%, 24.8%, and 28.9%, respectively, while strong synergies accounted for 10.6%, 14.3%, 15.7%, and 14.4%.
Fig. 12
Trade-offs and synergies between ESs in terraced fields in different regions
4 Discussion
4.1 Land use changes and their impact on ESs
A
Our results show that the period 2018–2023 brought rapid land-use transitions in China’s terraced landscapes, with significant consequences for ecosystem services. The dominant shift from rice terraces to dryland terraces (over 42,600 km2 converted) is primarily an adaptation to water stress. This trend was especially pronounced in provinces like Sichuan, Chongqing, and Hunan, areas that traditionally relied on paddy rice but faced decreasing water availability due to recent extreme droughts (e.g., the unprecedented Yangtze Basin drought in 2022) and a broader climate drying trend42. The widespread loss of paddy terraces raises concerns beyond water supply: rice terraces often function as artificial wetlands supporting unique biodiversity (e.g., frogs, fish, aquatic plants)43. Their conversion has likely led to habitat loss and fragmentation, contributing to observed declines of amphibious and aquatic species by an estimated 18–25% in affected areas44. In other words, the resilience of rural biodiversity might be compromised as terrace agriculture shifts away from its wetland-like form.Conversely, the large-scale expansion of woodland terraces (+ 104.2%) and grassland terraces (+ 22.5%) reflects China’s aggressive ecological restoration policies, including afforestation campaigns such as the Grain for Green Program. These programs have clear benefits: newly established woodland terraces contribute to higher carbon sequestration and better soil stability. Our findings confirm that afforestation on terraces increased carbon storage density significantly (by tens of tons per hectare in many places) and bolstered soil conservation (roots and litter improved soil retention by 82.7 t/hm2·y in the Loess Plateau example)45. However, these land-use changes also trigger trade-offs in water services, evapotranspiration in new woodlands increases by 30–40%, reducing downstream water yield by 15% in Yunnan’s watersheds46,47. This finding mirrors global observations in other reforested drylands where water yield declines as canopy cover increases (the so-called “forest-water paradox”). Thus, while the shifts toward forest cover align with climate mitigation and erosion control goals, they must be managed carefully to avoid unintended water scarcity for irrigation and communities48.
Land-use transitions also varied in their regional impacts on bundles of ES. For instance, on the Loess Plateau, replacing some dryland terraces with grassland has proven beneficial for soil conservation by 82.7 t/(hm2·yr) and even helped carbon (via improved soil organic carbon) while not seriously harming water yield because precipitation is low to begin with. Yet there was a trade-off in terms of agricultural output: the reduction in cultivated terraced area meant an estimated ¥225 million annual loss in grain production value from the Plateau45. This highlights a socio-economic dimension of ES trade-offs: improving regulating services (soil, carbon) sometimes comes at the cost of provisioning services like food production, necessitating compensation or livelihood alternatives for farmers.
ES interactions are further modulated by localized geophysical and anthropogenic drivers. In the Yunnan-Guizhou Plateau, the conversion to woodland terraces significantly elevates carbon storage density (reaching 143.6 t/hm2) but concurrently exacerbates water consumption, reducing streamflow by 15% in vulnerable watersheds 44,49. Similarly, European farming systems show that shifts from irrigated to rainfed agriculture alter the balance between water yield and carbon storage in ways consistent with our observations of paddy-to-dryland transitions in southern China50. Furthermore, the abandonment of terraces on steep slopes (> 25°) degrades their structural integrity, increasing landslide susceptibility by an estimated 23% despite potential gains in carbon stocks34,35.Conversely, in the semi-arid Loess Plateau, grassland expansion promotes synergistic gains in soil conservation (+ 82.7 t/hm2·yr) and carbon sequestration (+ 1.8 t C/hm2·yr) through root biomass accumulation and improved soil aggregate stability45. This stark regional contrast underscores a climate-mediated ecological threshold: in semi-arid regions (rainfall < 550 mm/yr), water limitation constrains plant growth, allowing restoration efforts to co-enhance multiple ESs synergistically. In contrast, humid zones (> 800 mm/yr) are primarily energy-limited; here, carbon-centric afforestation maximizes biomass production at the direct expense of water yield, creating an inherent and often unavoidable trade-off51,52.
Elsewhere, converting shrubland terraces to dryland cropland has boosted short-term agricultural output but at an ecological cost. We found evidence that such conversions can decrease pollinator habitat connectivity, with a modeled ~ 40% drop in pollinator abundance in those areas53, echoing global findings that landscape homogenization (even if productive) undermines pollination services54. Thus, certain land-use changes entail more nuanced trade-offs: while agricultural yields may increase, supporting ecosystem services such as pollination are often compromised, potentially leading to long-term feedback effects on crop productivity.
These multifaceted outcomes highlight the importance of policy integration and regional tailoring. No single land-use policy will optimize all services everywhere. In water-rich, biodiversity-rich high-rainfall zones (e.g., Fujian, Jiangxi), our results suggest value in maintaining traditional rice terraces where possible. These systems provide water regulation (flood control by storing water), cultural services (tourism and heritage), and habitat for wetland species, complementing the upland forests. Conversely, in semi-arid regions (e.g., Gansu, Inner Mongolia terraces), there is a need to introduce drought-adapted agroforestry mosaics. Our discussion points to strategies like planting low-water native shrub-grass mixes that can sequester carbon and hold soil without heavy water use (potentially even leveraging phytolith carbon in grasses for sequestration). By designing land-use practices that fit local climate regimes, one can minimize trade-offs – for instance, a Gansu case might use phreatophyte shrubs that capture carbon while their deep roots stabilize soil and require minimal water, thus preserving runoff for downstream users.
4.2 Decoupling trade-offs through spatially explicit governance
Our findings of strong national-scale trade-offs (notably between water yield and carbon sequestration) underscore the need for targeted governance that can decouple or mitigate these conflicts. One significant finding is the delineation of a climate-driven ecological threshold demarcating semi-arid and humid terraced landscapes50. In areas below ~ 550 mm annual rainfall, water scarcity naturally limits vegetation growth; restoration efforts there (grass/low-density shrubs) tend to produce synergistic ES gains – improving soil and carbon without much water penalty. In contrast, in areas above ~ 800 mm rainfall, water is plentiful enough that vegetation growth is energy-limited, meaning afforestation leads to very high biomass (carbon) at the direct expense of water yield55. Evidence from Italy further underscores the strong synergy between soil conservation and carbon sequestration56, while studies in European rural landscapes confirm the widespread nature of the forest–water paradox observed in humid terraced areas of China55. These parallels suggest that climate-mediated ecological thresholds between semi-arid and humid systems are not unique to China but part of a broader global phenomenon57,58. This suggests governance should delineate zones by climatic criteria to apply different land management rules.
In water-stressed valleys (annual rainfall < 500 mm), policies should restrict further afforestation projects or the establishment of high water-demand plantations, as such measures could aggravate existing water shortages. In these zones, water-efficient agroforestry practices—for example, the cultivation of drought-tolerant shrubs or the implementation of rainwater-harvesting terraces—provide more sustainable alternatives, enhancing soil stability and carbon sequestration while minimizing water use. It is therefore recommended that geographically explicit regulations delineate areas where additional forest cover would drive evapotranspiration beyond thresholds that compromise downstream flows, and limit such land-use conversions accordingly. By contrast, in humid upland regions where flood control is a priority, a stronger emphasis on forest cover may be acceptable despite water-related trade-offs, as downstream water availability is ample but sediment regulation remains a critical management concern.
Our study also shows that localized geophysical and anthropogenic drivers modulate ES relationships. For instance, in Yunnan-Guizhou, turning terraces into woodlands raises carbon but exacerbates water consumption (streamflow down 15% in vulnerable catchments)57, and steep slope abandonment raises landslide risks34. In the Sichuan Basin, rapid urbanization around terraced areas intensifies the water-carbon trade-off, a 10% increase in impervious area (cities) can increase trade-off intensity as runoff patterns change59. Recognizing these, governance must not only be climate-specific but also account for landscape connectivity (e.g., urban encroachment) and topography (e.g., slope thresholds for safe abandonment).
A
Technical solutions can aid in mitigating trade-offs. We highlight the example of optimized ridge-furrow terrace designs in the Yangtze Plain, which experimentally showed improved water retention (+ 25%) and reduced soil erosion (-42 t/hm2·yr) simultaneously60. Such engineered terraces alter micro-topography to trap more rain while maintaining productivity. Likewise, our findings on management interventions illustrate that well-designed practices can convert trade-offs into synergies. In Fujian, an innovative intercropping of tea with nitrogen-fixing shrubs was reported to boost carbon sequestration by 2.3 t C/hm2·yr while also improving soil moisture by 18%, yielding a strong synergy (correlation r = 0.68) between carbon and water61. In Gansu, a combined “agro-geoengineering” approach (terraces with check dams plus drought-tolerant grasses) raised water use efficiency by 33% and sediment retention by 41%, outperforming monoculture terraces62. These examples from different regions show that integrating ecological engineering with traditional terrace farming can alleviate inherent trade-offs. Essentially, by tailoring cropping systems or physical structures, one can find approaches that allow multiple ES to improve together.Another dimension is incorporating socio-economic innovations to support ecological outcomes. For instance, the Hani terraces case in Yunnan demonstrates that blending traditional water management with modern technology (sensors for irrigation scheduling) improved water distribution efficiency by 30%, and reinvesting eco-tourism revenue into terrace maintenance significantly cut landslide risk by 73%63,64. This shows that economic incentives and community engagement can amplify the benefits of technical measures. Therefore, decoupling trade-offs isn’t purely a biophysical challenge—it also requires aligning incentives and knowledge systems.
Our analysis underscores the importance of spatially explicit governance of terraced landscapes, emphasizing that policies and management strategies must be precisely calibrated to local climatic regimes, topographic conditions, and socio-economic contexts. Strategies should include setting region-specific ES targets, such as maintaining minimum water yield indices in each watershed while pursuing carbon sequestration objectives, and employing instruments such as payment schemes or land-use zoning to enforce these targets. Recognizing critical thresholds—for example, evapotranspiration tipping points in humid zones—and drawing on successful models of synergistic farming, policymakers can guide terraces towards a more balanced multifunctionality. Prioritizing regionally adapted interventions, ranging from ridge–furrow engineering in flat, wet areas to shrub–grass mosaics in arid hills, is essential to sustain terraces as resilient, multi-benefit landscapes under the dual pressures of climate change and socio-economic development.
4.3 Policy recommendations and management implications
Effectively addressing the complex ES trade-offs and synergies identified in this study requires a multifaceted policy framework that is spatially differentiated and aligned with regional ecological thresholds and socio-economic conditions65. Our results highlight the necessity of tailoring governance to distinct hydro-climatic zones, slope gradients, and socio-economic realities.
Regulations must be geographically explicit. In water-stressed valleys (< 500 mm/yr rainfall), strict limits on further afforestation are critical to avoid exacerbating water scarcity. Instead, drought-resilient agroforestry mosaics (e.g., native shrub–grass systems) should be promoted on gentle slopes (< 15°), which have been shown to synergize soil conservation and carbon sequestration without heavy water demand66. On steep slopes (> 25°), mandatory programs to restore terrace walls and vegetation are essential, as evidence shows such measures can reduce landslide risks by up to 73% while protecting existing carbon stocks67. Precision zoning should also map “synergy hotspots” and “risk hotspots”, guiding targeted interventions and preventing maladaptive land-use conversions.
Economic incentive mechanisms require innovation to translate ecological gains into local livelihoods. Establishing a Payment for Synergy Services (PSS) system, which directly compensates farmers ¥1200/hm2/year for adopting practices that provenly enhance multiple ESs simultaneously, such as the tea-shrub intercropping system in Fujian, which has been shown to boost farmer incomes by 15–20%45,68. Furthermore, cross-regional ecological compensation should be institutionalized. Downstream beneficiaries (e.g., urban basins like Chengdu in Sichuan) should provide financial support to upstream communities for maintaining terraces that provide vital flood mitigation and water purification services 69,70. This model can draw inspiration from the eco-socio-economic zoning approach of the Guangdong-Hong Kong-Macao Greater Bay Area. Additionally, cultural-ecological financing should be leveraged by actively pursuing UNESCO GIAHS (Globally Important Agricultural Heritage Systems) designation, as seen with the Hani terraces. This status secures restoration funds and boosts tourism revenue, creating a sustainable financing loop where heritage-based income is reinvested into terrace maintenance, thus preserving indigenous knowledge and reducing physical risks63,64. Such heritage-centered models both preserve indigenous knowledge and reduce biophysical risks.
Effective management requires harmonizing top-down policy with bottom-up community agency. Lessons from forest landscape management71,72 and European high-nature-value farming systems43 indicate that targeted policies, compensation schemes, and community-driven governance can help reconcile trade-offs and maximize synergies. Terrace management policies should formally incorporate Local Ecological Knowledge (LEK) into restoration design and conflict-resolution frameworks. Additionally, fostering “1 + N” co-management systems—which pair technology firms (providing IoT sensors and data platforms) with local farmers (contributing labor and traditional practices)—can democratize benefits and enhance terrace resilience, as evidenced by the successful case of Ziquejie terraces in Hunan73.
The Chinese case mirrors global dynamics: Mediterranean terraces show that co-production of ES via community participation improves resilience16, while Italian and European cases illustrate both soil–carbon synergies and the forest–water paradox56,74. Recognizing these parallels highlights China’s potential to pioneer a globally relevant model of terrace governance, grounded in precision zoning, incentive-based financing, and participatory institutions. Optimizing ES provision in China’s terraced landscapes demands regionally adapted, socially inclusive, and economically viable policies. By integrating climatic thresholds, incentive systems, and community-led governance, terraced systems can evolve into resilient socio-ecological infrastructures that simultaneously support food security, water regulation, cultural heritage, and climate mitigation.
5 Conclusions and suggestions
This study systematically quantified the dramatic transformations in ’China’s terraced landscapes between 2018 and 2023. We observed a stark decline in rice terraces (-39.2%) alongside a remarkable surge in woodland terraces (+ 104.2%), shifts primarily driven by accelerating climate aridification and the implementation of large-scale afforestation policies. These land-use changes have generated complex trade-offs in ecosystem services (ES): while significantly enhancing soil conservation (+ 18.5%) and carbon storage (+ 1.7%), they have concurrently constrained gains in water yield (+ 3.1%). Critically, our analysis reveals that trade-offs, particularly between water yield and carbon sequestration (affecting 55.4% of terraced areas), dominate at the national scale. This pattern exhibits sharp regional disparities: synergies prevail on the semi-arid Loess Plateau (72.5% of areas), whereas trade-offs are most pronounced in humid southern basins (e.g., 78.9% in the Sichuan Basin)
Resolving this sustainability trilemma necessitates a paradigm shift towards spatially explicit and precision stewardship: (1) Geospatially targeted management should be applied, including strict afforestation restrictions in water-scarce valleys (annual rainfall < 500 mm), promotion of slope-tiered agroforestry systems on gentle slopes (< 15°), and structural stabilization combined with ecological restoration of high-risk terraces on steep slopes (> 25°) to mitigate soil erosion and landslide hazards; (2) Innovative economic incentive mechanisms must be established, such as scaling up Payments for Synergy Services (PSS), institutionalizing cross-regional ecological compensation to support upstream conservation of water and soil services by downstream beneficiaries, and leveraging the Globally Important Agricultural Heritage Systems (GIAHS) designation to foster cultural-ecological financing that channels ecotourism revenue into maintenance practices; (3) A co-evolutionary knowledge and adaptive governance framework should be developed by integrating traditional ecological wisdom with digital monitoring technologies and establishing multi-stakeholder co-management platforms to enhance community resilience. Through the comprehensive implementation of these strategies, terraced systems can transition into carbon-smart, water-secure, and high-efficiency ecological landscapes, offering a Chinese model for sustainable mountain agriculture worldwide and contributing directly to the achievement of Sustainable Development Goals (SDGs)—particularly Zero Hunger (SDG 2), Clean Water and Sanitation (SDG 6), and Climate Action (SDG 13).
A
Declaration of competing interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
A
A
Acknowledgements
This research was supported by the Fundamental Research Funds for the Central Universities (Grant XJJSKYQD202548), and the National Natural Science Foundation of China (Grants No. 42201100, U21A2011, 41991233).
References
1.
Wei W et al (2016) Global synthesis of the classifications, distributions, benefits and issues of terracing. Earth Sci Rev 159:388–403
2.
Tarolli P, Preti F, Romano N (2014) Terraced landscapes: From an old best practice to a potential hazard for soil degradation due to land abandonment. Anthropocene 6:10–25
3.
Yu M, Li Y, Luo G, Yu L, Chen M (2022) Agroecosystem composition and landscape ecological risk evolution of rice terraces in the southern mountains, China. Ecol Ind 145:109625
4.
D'Angelo J, Palmate SS, Descroix L (2023) Mitigating risks of hybrid rice use in terrace agriculture. Geogr Sustain 4:1–5
5.
Wei W, Pan D, Yang Y (2021) Effects of terracing measures on water retention of pinus Tabulaeformis forest in the dryland loess hilly region of China. Agric For Meteorol 308–309:108544
6.
Brown AG et al (2021) Ending the Cinderella status of terraces and lynchets in Europe: The geomorphology of agricultural terraces and implications for ecosystem services and climate adaptation. Geomorphology 379:107579
7.
Cao BW et al (2021) A 30 m terrace mapping in China using Landsat 8 imagery and digital elevation model based on the Google Earth Engine. Earth Syst Sci Data 13:2437–2456
8.
Chen D, Wei W, Chen L (2017) Effects of terracing practices on water erosion control in China: A meta-analysis. Earth Sci Rev 173:109–121
9.
Li W et al (2024) Impacts of rice terrace abandonment on soil ecosystem multifunctionality: Insights into reduced microbial diversity and exacerbated nitrogen limitation. J Clean Prod 437:140731
10.
Shi Y, Wei W, Wang W, Tarolli P, Chen L (2025) Food provision responses to changes in mountainous terraced and sloping cropland: Implications for land management based on land dynamics and terrain gradient. Geogr Sustain 6:100302
11.
Costanza R et al (1997) The value of the world's ecosystem services and natural capital. Nature 387:253–260
12.
Felipe-Lucia MR et al (2018) Multiple forest attributes underpin the supply of multiple ecosystem services. Nat Commun 9
13.
Ciftcioglu GC (2017) Social preference-based valuation of the links between home gardens, ecosystem services, and human well-being in Lefke Region of North Cyprus. Ecosyst Serv 25:227–236
14.
Cruz-Garcia GS, Sachet E, Blundo-Canto G, Vanegas M, Quintero M (2017) To what extent have the links between ecosystem services and human well-being been researched in Africa, Asia, and Latin. America? Ecosyst Serv 25:201–212
15.
Steur G, Verburg RW, Wassen MJ, Verweij PA (2020) Shedding light on relationships between plant diversity and tropical forest ecosystem services across spatial scales and plot sizes. Ecosyst Serv 43:101107
16.
Zoumides C, Bruggeman A, Djuma H (2025) Revitalising terraced landscapes: Co-production of ecosystem services for sustainable futures. Geogr Sustain 6:100306
17.
Yang L, Li Y, Wei X (2025) Validating social media data reliability for cultural ecosystem services mapping in data-scarce GIAHS: A mixed-methods approach in hani rice Terraces. Ecol Ind 175:113530
18.
Zheng H et al (2024) Long-term evolution of soil quality in citrus orchard terrace system: A whole life cycle perspective. Sci Total Environ 951:175754
19.
Liu SS et al (2025) Multiple targeted grassland restoration interventions enhance ecosystem service multifunctionality. Nat Commun 16
20.
Kassun BW, Kallio M, Trømborg E, Rannestad MM (2025) Land use and land cover change, trade-offs, and synergies between ecosystem services in a dry Afromontane Forest. J Nat Conserv 85:126874
21.
Simeon M, Wana D (2025) Synergies and trade-offs among key ecosystem services in Maze National Park and its environs, southwestern Ethiopia. Glob Ecol Conserv 57:e03398
22.
Bennett EM, Peterson GD, Gordon LJ (2009) Understanding relationships among multiple ecosystem services. Ecol Lett 12:1394–1404
23.
Burgos-Ayala A et al (2024) Mapping ecosystem services in Colombia: Analysis of synergies, trade-offs and bundles in environmental management. Ecosyst Serv 66:101608
24.
Rodriguez JP et al (2006) Trade-offs across space, time, and ecosystem services. Ecol Soc 11
25.
Chen L et al (2024) Long-term terrace change and ecosystem service response in an inland mountain province of China. CATENA 234
26.
Deng CX et al (2021) Advantages and disadvantages of terracing: A comprehensive review. Int Soil Water Conserv Res 9:344–359
27.
Cao SX, Suo XH, Xia CQ (2020) Payoff from afforestation under the Three-North Shelter Forest Program. J Clean Prod 256
28.
Wang YC, Li BY (2022) Dynamics arising from the impact of large-scale afforestation on ecosystem services. Land Degrad Dev 33:3186–3198
29.
Giller KE, Beare MH, Lavelle P, Izac AMN, Swift MJ (1997) Agricultural intensification, soil biodiversity and agroecosystem function. Appl Soil Ecol 6:3–16
30.
Zheng H et al (2016) Using ecosystem service trade-offs to inform water conservation policies and management practices. Front Ecol Environ 14:527–532
31.
Zhang Y, Zheng H, Chen X (2024) Effects of ecological restoration projects on ecosystem services flows. Ecosyst Serv 70:101681
32.
Kovacs KF, Haight RG, Moore K, Popp M (2021) Afforestation for carbon sequestration in the Lower Mississippi River Basin of Arkansas, USA: Does modeling of land use at fine spatial resolution reveal lower carbon cost? For Policy Econ 130
33.
Pijl A, Wang WD, Straffelini E, Tarolli P (2022) Soil and water conservation in terraced and non-terraced cultivations: an extensive comparison of 50 vineyards. Land Degrad Dev 33:596–610
34.
Long M, Tan R, Zhao Y, Li X, Zhou C, Zhang D (2025) The rice terrace abandonment and its challenge for food security in the rural villages of southern China. Habitat Int 160:103398
35.
Lu D et al (2025) Patterns and drivers of terrace abandonment in China: Monitoring based on multi-source remote sensing data. Land Use Policy 148:107388
36.
Gao Y-H, Han B, Liu H-W, Bai Y-N, Li Z (2025) Spatial–Temporal Restructuring of Regional Landscape Patterns and Associated Carbon Effects: Evidence from Xiong’an New Area. 17:6224
37.
Niu B et al (2021) The Evolution of the Interactive Relationship between Urbanization and Land-Use Transition: A Case Study of the Yangtze River Delta. Land-Basel 10
38.
Xu FJ, Zhao SQ, Li SC (2024) Dynamic Responses of Soil Organic Carbon to Urbanization: A Global Perspective. Glob Change Biol 30
39.
Fang GJ, Sun X, Liao C, Xiao Y, Yang P, Liu QH (2023) How do ecosystem services evolve across urban-rural transitional landscapes of Beijing-Tianjin-Hebei region in China: patterns, trade-offs, and drivers. Landscape Ecol 38:1125–1145
40.
Hutchinson MF, Xu T, Houlder D (2011) ANUSPLIN Version 4.4 User Guide. Australian National University
41.
Sharp RTHT, Ricketts T et al (2015) InVEST 3. 2. 0 User's Guide. The Natural Capital Project, Stanford
42.
UNCCD. The Global Threat of Drying Lands: Regional and Global Aridity Trends and Future Projections (2024)
43.
Plieninger T, Torralba M, Hartel T, Fagerholm N (2019) Perceived ecosystem services synergies, trade-offs, and bundles in European high nature value farming landscapes. Landscape Ecol 34:1565–1581
44.
Xie S, Yin GY, Wei W, Sun QZ, Zhang Z (2022) Spatial-Temporal Change in Paddy Field and Dryland in Different Topographic Gradients: A Case Study of China during 1990–2020. Land-Basel 11
45.
Qin ZX, Liu XL, Lu XY, Li MF, Li F (2022) Grain Production Space Reconstruction and Its Influencing Factors in the Loess Plateau. Int J Env Res Pub He 19
46.
Yang H et al (2015) Ecosystem Evapotranspiration as a Response to Climate and Vegetation Coverage Changes in Northwest Yunnan, China. PLoS ONE 10:e0134795
47.
Xiao Y, Xiao Q (2018) Impact of large-scale tree planting in Yunnan Province, China, on the water supply balance in Southeast Asia. Environ Monit Assess 191:20
48.
Wu B, Smith WK, Zeng H (2024) Dryland Dynamics and Driving Forces. In: Fu B, Stafford-Smith M (eds) Dryland Social-Ecological Systems in Changing Environments. Springer Nature Singapore
49.
Li Y, Peng YL, Peng HN, Cheng WY (2025) Assessment and multi-scenario prediction of ecosystem services in the Yunnan-Guizhou Plateau based on machine learning and the PLUS model. Front Ecol Evol 13
50.
Silva JF, Santos JL, Ribeiro PF, Marta-Pedroso C, Magalhaes MR, Moreira F (2024) A farming systems approach to assess synergies and trade-offs among ecosystem services. Ecosyst Serv 65
51.
Yan Y, Wang R, Chen S, Zhang Y, Sun Q (2023) Three-dimensional agricultural water scarcity assessment based on water footprint: A study from a humid agricultural area in China. Sci Total Environ 857:159407
52.
Wei CW et al (2023) Increase in precipitation and fractional vegetation cover promote synergy of ecosystem services in China's arid regions-Northern sand-stabilization belt. Front Ecol Evol 11
53.
Rahimi E, Barghjelveh S, Dong P (2021) Using the Lonsdorf model for estimating habitat loss and fragmentation effects on pollination service. Ecol Processes 10:22
54.
Garibaldi LA et al (2011) Stability of pollination services decreases with isolation from natural areas despite honey bee visits. Ecol Lett 14:1062–1072
55.
Bartolini F, Vergamini D (2023) Trade-Offs and Synergies between Ecosystem Services Provided by Different Rural Landscape. Agronomy-Basel 13
56.
Medina-Roldán E, Lorenzetti R, Calzolari C, Ungaro F (2024) Disentangling soil-based ecosystem services synergies, trade-offs, multifunctionality, and bundles: A case study at regional scale (NE Italy) to support environmental planning. Geoderma 448
57.
Zhang JX et al (2023) Trade-offs and synergies of ecosystem services and their threshold effects in the largest tableland of the Loess Plateau. Glob Ecol Conserv 48
58.
Zuo LY, Gao JB (2022) Dynamic analysis of the determinants of trade-off and synergy between karst soil loss and water yield with integration of geomorphological differentiation. Ecol Ind 137
59.
Chen Y, Li J, Lu H, Xia J (2020) Tradeoffs in water and carbon footprints of shale gas, natural gas, and coal in China. Fuel 263:116778
60.
Zhao J, Li C (2020) Investigating spatiotemporal dynamics and trade-off/synergy of multiple ecosystem services in response to land cover change: a case study of Nanjing city, China. Environ Monit Assess 192
61.
Chen L-L, Pozsgai G, Li X-Y, Li L, Reddy GVP, You M-S (2021) Effects of cover crops on beetle assemblages in tea plantations. Crop Prot 149:105783
62.
Liu C et al (2018) Soil carbon and nitrogen sources and redistribution as affected by erosion and deposition processes: A case study in a loess hilly-gully catchment, China. Agric Ecosyst Environ 253:11–22
63.
Yuan Y et al (2024) Aquaculture PPP development in China-case study from Hani Terrace. Mar Policy 163:106075
64.
José Gabriel V, Wang R, Mahato D, Wei C (2025) Slope stability and disaster mechanisms in the Honghe Hani Terraces: a systematic review. Egypt J Remote Sens Space Sci 28:411–425
65.
Qu Y, Su D, Wei C, Zhang Q, Jiang G (2023) How to prevent landscape ecological risk with a land use optimal allocation system: An empirical study of the Yellow River Delta in China. Ecol Ind 154:110888
66.
Xu Y, Chen C, Deng W, Dai L, Yang T (2025) Spatial eco-socio-economic trade-offs inform differentiated management strategies in mega-urban agglomerations. npj Urban Sustain 5:43
67.
Wang Y, Yang A, Shen W, Yang J, Wang Y, Li X (2024) Spatial patterns, determinants, future trends, and implications for the sustainable use of terraces abandonment in China. J Clean Prod 467:142860
A
68.Han Z, Song W (2019) Spatiotemporal variations in cropland abandonment in the Guizhou-Guangxi karst mountain area, China. J Clean Prod 238
69.
Su K et al (2020) The establishment of a cross-regional differentiated ecological compensation scheme based on the benefit areas and benefit levels of sand-stabilization ecosystem service. J Clean Prod 270:122490
70.
Ding Z, Yao S (2022) Theory and valuation of cross-regional ecological compensation for cultivated land: A case study of Shanxi province, China. Ecol Ind 136:108609
71.
Njoya HM, Hounkpati K, Adjonou K, Kokou K, Sieber S, Löhr K (2025) Institutional Design of Forest Landscape Restoration in Central Togo: Informing Policy-making through Q Methodology Analysis. Environ Manage 75:1239–1256
72.
Jafarzadeh AA, Mahdavi A, Shamsi SRF, Yousefpour R (2021) Assessing synergies and trade-offs between ecosystem services in forest landscape management. Land Use Policy 111
73.
Xiao PY et al (2025) Mechanisms of land use change effects on soil quality in ancient terraces based on the minimum data set approach. CATENA 254:108990
74.
Bartolini F, Vergamini D (2023) Trade-Offs and Synergies between Ecosystem Services Provided by Different Rural Landscape. Agronomy-Basel 13