Abstract:

[Objective] Coastal blue carbon ecosystems, including mangroves, tidal flats, and seagrass meadows, constitute important natural carbon sinks and play a critical role in mitigating climate change. However, under the dual pressures of rapid urbanization and land reclamation, the ecological patterns and carbon stocks of coastal zones have undergone profound transformations. These impacts are particularly pronounced in the highly developed Pearl River Delta, where land-use dynamics, climatic variability, and human activities interact. Understanding how land-use and land-cover (LULC) change, climatic drivers, and anthropogenic factors jointly shape the spatiotemporal dynamics of coastal blue carbon has become an important topic in environmental research. In recent years, models such as InVEST have been widely applied to estimate blue carbon stocks, while statistical approaches, including geographical detectors and process-based decomposition, have been used to identify key drivers. These studies have revealed the dominant roles of land-use transitions and climatic conditions in determining carbon storage patterns, thereby providing an important basis for quantitative assessment of blue carbon. Nevertheless, existing research has predominantly focused on terrestrial carbon storage, while investigations of coastal blue carbon have often been restricted to short time periods, with limited attention paid to long-term trajectories. Methodologically, many studies rely on a single modelling framework and therefore provide only a partial understanding of the coupled natural-human mechanisms regulating blue carbon dynamics. Against this background, the present study aimed to address the knowledge gap concerning the long-term spatiotemporal evolution and combined driving mechanisms of blue carbon storage in the coastal zone of the Pearl River Delta and to provide quantitative evidence for regional blue carbon management and policy development. [Methods] Taking the coastal zone of the Pearl River Delta as the study area, this research integrated remote-sensing datasets, the InVEST carbon storage module, and the Optimal Multivariate-Stratification Geographical Detector (OMGD) to conduct the following analyses. First, land-use transition matrices were constructed using LULC maps from 1990 to 2020 to characterize the spatial and temporal evolution of coastal LULC patterns. Second, blue carbon storage from 1990 to 2020 was estimated using the InVEST model, a biophysical parameter table for blue carbon ecosystems, and LULC maps for 1990, 2000, 2010 and 2020, and its spatiotemporal variation was subsequently analyzed. Third, the initial-to-final transitions among coastal land-use types from 1990 to 2020 were classified into six process categories: accretion (A), natural succession (S), retrogressive succession (Rs), erosion (E), encroachment (Rc) and restoration (Re). A multistage process-decomposition framework was then used to quantify the contribution of each process to changes in blue carbon storage. Fourth, OMGD was applied to identify the dominant driving factors and their interactions in controlling the spatial heterogeneity of blue carbon storage. By optimising the spatial resolution and variable-discretization schemes, the explanatory power, expressed as the q-statistics, and statistical significance of 13 potential drivers were evaluated. The natural factors included mean annual temperature (T_mean), mean annual minimum temperature (T_min), mean annual maximum temperature (T_max ), air pressure (PRES), relative humidity (RHU), potential evapotranspiration (PetPM), annual precipitation (PREC), soil erosion (SE), the Normalized Difference Vegetation Index (NDVI), soil type (SOIL) and elevation (DEM), while the anthropogenic factors included population density (POP) and night-time light intensity (NL). The synergistic effects among multiple drivers were also quantified to clarify how natural and anthropogenic factors jointly shape the spatial heterogeneity of coastal blue carbon storage. [Results] The results showed that: 1) From 1990 to 2020, coastal LULC patterns in the Pearl River Delta changed substantially. Shallow offshore waters, tidal flats, cropland, grassland and bare land all exhibited decreasing trends, with the area of tidal flats declining by 84.7%. After an initial period of contraction, mangrove area recovered markedly from 2010 to 2020, resulting in a net increase of 18.6% compared with that in 1990. Over the same period, built-up land and mariculture ponds expanded continuously, with the area of built-up land increasing by 1.11×10⁵ hm², mainly at the expense of cropland and aquaculture areas. The dominant unidirectional conversion of shallow offshore waters to built-up land indicated that large-scale coastal reclamation projects had exerted persistent impacts on the regional coastal environment. 2) Blue carbon storage in the coastal zone of the Pearl River Delta exhibited nonlinear dynamics from 1990 to 2020, following a U-shaped trajectory. Total blue carbon storage decreased from 4.98×10⁷ Mg in 1990 to 4.92×10⁷ Mg in 2000 and further declined to a minimum of 4.91×10⁷ Mg in 2010 before recovering to 4.98×10⁷ Mg in 2020. These temporal fluctuations contrasted with the comparatively stable spatial distribution of blue carbon storage, with high-value clusters concentrated in the alluvial coastal plains surrounding the Pearl River Estuary. 3) Blue carbon storage responded sensitively to land-use transitions. Natural succession (S) enhanced wetland carbon sinks, whereas erosion (E) and reclamation-related encroachment (Rc) caused substantial carbon losses. These negative effects were further intensified when erosion occurred concurrently with ecological degradation. Artificial restoration (Re) generated only limited short-term gains in blue carbon storage, and neither land reclamation nor “occupy-and-compensate” strategies could offset the greater carbon storage capacity and ecological stability of mature natural wetlands. The process-based decomposition therefore highlighted the importance of protecting existing natural wetlands and regulating anthropogenic disturbances to stabilize and enhance coastal blue carbon storage. 4) The OMGD results indicated that the spatial differentiation in blue carbon storage was governed primarily by natural factors, particularly elevation (DEM, q=0.729) and NDVI (q=0.556), while spatial heterogeneity was further amplified by the nonlinear interactions involving anthropogenic drivers. Interactions between natural and anthropogenic factors contributed to the spatial clustering of blue carbon storage, underscoring the coupled influence of biophysical and socioeconomic controls on coastal carbon dynamics. These findings indicated that effective blue carbon management should consider not only geomorphological and ecological constraints but also the spatial impacts of human development. [Conclusion] Overall, this study provides a quantitative basis for coastal ecological compensation and blue carbon market mechanisms. By attributing gains and losses in blue carbon storage to specific land-use conversion processes, the study clarifies the carbon consequences of coastal land-use change and supports the formulation of differentiated compensation standards. Carbon losses caused by urban expansion and land reclamation could be internalized through ecological compensation funds or development fees, whereas additional carbon sinks generated by mangrove restoration and tidal-flat succession could potentially be converted into tradable blue carbon assets. Using long-term datasets covering the period from 1990 to 2020, this study establishes a relatively stable and internationally comparable baseline for blue carbon asset accounting, providing a reference for carbon-credit pricing, quota allocation and low-carbon transition strategies under China’s carbon-peaking and carbon-neutrality goals. The analytical framework and empirical findings may also be applicable to other highly urbanized delta regions, offering scientific support for improving blue carbon trading schemes, strengthening coastal ecological resilience, and promoting sustainable coastal development

Key words: pearl river delta, blue carbon stocks, InVEST model, spatio-temporal evolution, driver analysis

摘要：

【目的】沿海蓝碳生态系统是关键的天然碳汇，本研究旨在揭示快速城镇化背景下珠江三角洲海岸带蓝碳储量的长期演变规律，并量化土地利用转型及自然、人为因子的驱动作用。【方法】集成多源遥感数据与InVEST模型，结合最优多变量分层地理探测器，量化分析了珠江三角洲海岸带1990－2020年蓝碳储量的时空变化特征及其驱动机制。【结果】结果表明，1）1990－2020年，土地利用变化导致建设用地扩张了1.11×10⁵ hm²（主要由耕地、浅海水域及海水养殖区转化），红树林面积净增约18.6%。2）1990年，珠江三角洲海岸带的总蓝碳储量为4.98×10⁷ Mg，经过2000－2010年的衰减阶段后，在2020年恢复至与1990年相同的4.98×10⁷ Mg。蓝碳储量演变过程显示，滩涂自然演替过程增加了3.26×10⁵ Mg的蓝碳储量，而填海工程导致了4.00×10⁵ Mg的损失。3）蓝碳储量的空间异质性主要由地形高程（q=0.729）和NDVI（q=0.556）驱动，多因子交互分析显示，各因子组合整体呈协同增强效应，表明多因子综合影响蓝碳储量的空间分异特征。【结论】珠江三角洲海岸带蓝碳储量演变体现了快速城镇化背景下自然生态过程与人类活动共同塑造区域碳汇格局的基本规律，可为沿海地区生态保护修复、国土空间管控及“双碳”目标下蓝碳增汇政策制定提供科学依据。

关键词: 珠江三角洲, 蓝碳储量, InVEST模型, 时空演变, 驱动分析