A nationwide reclassification and simulation analysis of China’s land dynamics reveals that secondary vegetation serves as the pivotal structural bridge between contracting natural habitats and expanding human-dominated landscapes. Researchers Boxuan Zhao, Penghe Xu, and Yuanyuan Yang constructed an integrated spatial assessment combining remote sensing with human activity indicators to separate primary vegetation, secondary vegetation, managed plantations, and pasture across China’s terrestrial land base. Evaluating multi-decadal changes from 2000 to 2020 alongside projections through 2030 and 2050 under four Shared Socioeconomic Pathways and Representative Concentration Pathways, the team found that conventional land-cover schemes have obscured critical functional transitions. By decoupling surface greenness from ecological integrity, the study established that secondary ecosystems absorbed natural habitat degradation while continuously supplying converted land for agriculture, commercial forestry, and grazing systems.
Standard satellite-derived land-cover datasets have traditionally struggled to represent regional ecological conditions because broad categories such as forest and grassland frequently group ecologically distinct land uses together. High-resolution products typically rely on surface reflectance and structural canopy signatures, meaning that ecologically intact primary forests and single-species timber plantations can present nearly identical spectral appearances. Similarly, natural grasslands with minimal human intervention can look indistinguishable from heavily grazed livestock pastures in remote-sensing imagery. To overcome these limitations, the investigators established a rule-based reclassification framework that treats surface cover as distinct from management intensity and historical disturbance. Using the 30-meter China Land Cover Dataset as a baseline map aggregated to a standardized one-kilometer equal-area grid, the team separated the country into seven dynamic land-use classes: primary vegetation, secondary vegetation, plantation, pasture, cropland, urban, and bare land.
The methodological framework identified human-dominated types first before subdividing remaining natural vegetation into primary and secondary categories. Cropland, urban areas, and bare ground were extracted directly from the base dataset, while specialized spatial inventories were used to delineate plantations and pastures. Pasture was isolated by applying a conservative grazing intensity threshold of 0.5 stocking units per hectare from high-resolution livestock datasets, classifying higher-intensity grasslands as human-dominated grazing lands. Candidate cells containing remaining natural vegetation were subsequently evaluated against a dual disturbance detection scheme spanning rolling five-year observational windows across five historical intervals from 2000 to 2020. This screening captured active fires, satellite-recorded burned area, Hansen global forest loss events, and interannual Normalized Difference Vegetation Index anomalies derived from AVHRR and MODIS sensors to identify recent habitat degradation without conflating short-term fluctuations with permanent structural turnover.
To classify undisturbed natural lands conservatively, the authors paired the absence of recent disturbance with explicit spatial buffers reflecting conservation status, transport networks, population density, and topography. Undisturbed natural vegetation situated inside protected areas was categorized as primary vegetation. Outside established reserves, areas situated more than five kilometers away from major roads and railways were screened using demographic and terrain thresholds. Remote candidate zones with fewer than one person per square kilometer qualified as primary vegetation, while areas supporting between one and five persons per square kilometer were classified as primary vegetation only if local terrain slopes reached or exceeded 15 degrees, reflecting Chinese ecological policy benchmarks where steep topography restricts development. Natural vegetation failing to meet these strict preservation and accessibility criteria was classified as secondary vegetation, representing ecosystems regenerating after natural or anthropogenic disturbance.
Independent validation verified that the seven-class reclassification accurately separated land uses along clear ecological gradients. When evaluated against the 2020 Human Footprint Pressure dataset, median pressure scores increased systematically from bare land at 0.49 and primary vegetation at 1.87 up to pasture at 7.42, secondary vegetation at 8.74, plantation at 9.82, cropland at 20.66, and urban areas at 35.46. Within this distribution, primary vegetation exhibited significantly lower human pressure than secondary vegetation, with 70.37 percent of primary land classified under low pressure compared to just 15.42 percent for secondary cover. Forest age and biomass layers similarly corroborated the distinction: forest cells identified as primary vegetation averaged 66.0 years of age and 142.1 megagrams per hectare of aboveground biomass density, compared to an average age of 44.2 years and 127.1 megagrams per hectare for secondary forests, demonstrating structurally mature conditions.
Historical mapping revealed that China’s baseline landscape in 2020 was dominated by secondary vegetation, which occupied 24.15 percent of the national territory covering 2,322,003 square kilometers. Bare land formed the second largest category at 21.06 percent, followed closely by cropland at 20.20 percent, while primary vegetation spanned 13.80 percent or 1,327,000 square kilometers. Pasture and plantation accounted for 10.67 percent and 7.48 percent respectively, with urban surfaces covering 2.64 percent. Geographically, high-naturalness habitats showed strong regional clustering. Primary vegetation remained concentrated along the margins of the Qinghai-Tibet Plateau, the mountainous Southwest, and the forested tracts of Northeast China. By comparison, secondary vegetation was more widely distributed across Northeast China, the southern margin of North China, Central China, and the hilly regions of South China.
Between 2000 and 2020, cumulative land transitions affected 39.99 percent of China’s land area, or 3,843,826 square kilometers, with 26.06 percent of the country experiencing at least one land-use conversion. Across the study period, human-dominated lands grew by 15.78 percent while natural habitats contracted by 11.92 percent. Specifically, primary vegetation declined by 18.14 percent, representing a net reduction of 294,027 square kilometers, and secondary vegetation decreased by 7.92 percent or 199,799 square kilometers. Conversely, plantation areas expanded by 50.98 percent, pasture grew by 33.05 percent, and urban land expanded by 65.80 percent. The restructuring exhibited marked temporal shifts, characterized initially by rapid primary vegetation decline and pasture expansion between 2000 and 2005, followed by surging plantation establishment between 2005 and 2010, accelerated urban growth from 2010 to 2015, and continued loss of primary vegetation alongside renewed pasture expansion between 2015 and 2020.
A critical structural finding was that historical land turnover was driven by eight dominant pathways that accounted for 72.91 percent of all land conversion. Secondary vegetation was central to this dynamic, serving as both a primary sink for degraded habitats and a major source for economic land uses. Converting secondary vegetation to timber plantations constituted the single largest pathway nationally, representing 16.06 percent of all transitions. Conversion of secondary vegetation to pasture accounted for 11.94 percent, while the degradation of primary vegetation into secondary vegetation formed the third largest pathway at 10.74 percent. Plantations derived 88.83 percent of their new area from secondary stands, and newly created pastures drew 51.94 percent of their extent from secondary vegetation and 30.96 percent from primary vegetation, underscoring that economic expansion occurred largely at the expense of semi-natural and natural cover.
To model future dynamics, the authors integrated the 2020 reclassification baseline with subregional land-demand trajectories derived from the Land-Use Harmonization 2 project using the Patch-generating Land Use Simulation model. The model was calibrated across seven geographic subregions against historical 2010 to 2020 patterns, achieving an overall national agreement of 0.85, a Kappa coefficient of 0.81, and a Figure of Merit of 0.19. Simulations through 2030 and 2050 explored four development pathways: the sustainable SSP1-RCP2.6, the intermediate SSP2-RCP4.5, the regionally fragmented SSP3-RCP7.0, and the fossil-fueled development scenario SSP5-RCP8.5. Across all four scenarios, primary vegetation continued to shrink by 10.81 to 13.04 percent by 2050, while urban footprints expanded by 18.16 to 36.59 percent. However, the largest inter-scenario differences emerged not in urban expansion, which varied by only 46,729 square kilometers, but in secondary vegetation and cropland, which diverged by up to 649,799 and 437,475 square kilometers respectively.
The simulated pathways illustrated fundamentally different trade-offs in landscape management by 2050. Under the conservation-focused SSP1-RCP2.6 scenario, secondary vegetation expanded by 30.28 percent, absorbing extensive conversions from cropland, which declined by 11.84 percent, and pasture, which shrank by 17.08 percent. The SSP3-RCP7.0 projection generated the largest overall land-use changes, with secondary vegetation expanding by 35.42 percent while cropland contracted by 20.87 percent by 2050, representing the largest changes among all scenarios. In contrast, the development-heavy SSP5-RCP8.5 scenario saw the lowest gain in secondary vegetation at just 7.44 percent, the steepest urban increase at 36.59 percent, and net cropland expansion of 1.66 percent. Under SSP5-RCP8.5, cropland losses in urbanizing eastern provinces were offset by inward cropland conversion in southwestern mountain valleys, highlighting potential displacement of agriculture into ecologically fragile zones.
Spatial hot-spot analysis demonstrated that scenario divergences were concentrated in specific transitional geographies, including the northern agro-pastoral ecotone spanning Northeast and North China, marginal hilly margins bordering eastern plains, and the transition belt between southwestern mountains and plateaus. In these sensitive corridors, land competition between agricultural maintenance, secondary vegetation regeneration, and primary habitat preservation was acute. While primary vegetation contracted marginally along outer boundaries under all scenarios, SSP5-RCP8.5 caused direct conversions of primary lands into cropland, pasture, and plantations, whereas SSP1-RCP2.6 and SSP3-RCP7.0 transitions proceeded primarily via gradual degradation into secondary vegetation.
The authors emphasize several analytical boundaries and technical limitations inherent to the framework. Operating at a one-kilometer spatial resolution and relying on auxiliary datasets means the framework's ability to capture small fragmented patches and fine-scale transitions remains limited. Furthermore, because the reclassification pipeline builds upon the base China Land Cover Dataset, any underlying classification errors in initial grassland, forest, or shrubland extents may carry forward into derived secondary and primary classes. The researchers noted that while sensitivity testing showed the primary-secondary thresholds to be locally robust, future work could incorporate finer-resolution driving factors, more detailed protected-area information, and biodiversity response data to better assess the ecological consequences of land-use change.
Subject of Research: Land-use reclassification and scenario simulation of land system dynamics in China
Article Title: An integrated land-use reclassification and scenario-simulation analysis of land system dynamics in China
Article References: Zhao, B., Xu, P., & Yang, Y. (2026). An integrated land-use reclassification and scenario-simulation analysis of land system dynamics in China. Landscape Ecology, 41(11), Article 206. https://doi.org/10.1007/s10980-026-02444-6
Image Credits: AI Generated
DOI: 10.1007/s10980-026-02444-6
Keywords: land-use change, China land systems, secondary vegetation, primary vegetation, PLUS model, scenario simulation, SSP-RCP scenarios, China Land Cover Dataset, landscape ecology, ecological restoration, cropland dynamics, plantation forestry, grazing intensity, human footprint, spatial modeling, ecosystem disturbance, land management
News Source: Gavin Prescott. (November 19, 2026). Secondary Vegetation Acts as Central Intermediate Class in China Land-Use Dynamics. Scienmag.



