Healthy ecosystems are fundamental to human survival and sustainable development. However, compared to ecosystem services (ES), the understanding of ecosystem disservices (EDS) significantly lags behind, resulting in a lack of comprehensiveness in ecosystem management and decision-making. In response to this gap, this study systematically reviews the latest advancements in EDS research and provides an empirical analysis using wildlife damage as a case study. The findings indicate that: (1) Since the EDS concept emerged in the 1990s, related phenomena have been reported in nearly 100 countries and regions. However, a unified consensus on the concept of EDS has not yet been reached within academia, and its formation mechanisms and effects have not been fully elucidated, lacking a systematic theoretical framework. (2) This study defines EDS as changes in the structure, processes, and functions of ecosystems driven by biological factors, where these changes have direct or indirect negative impacts on human well-being. Following the research paradigm of geography, a theoretical framework centered on "identification-mechanism-effect-regulation" is constructed, emphasizing that EDS exhibits significant scale dependency and spatiotemporal heterogeneity. (3) Typical cases reveal that wild boar damage in mountainous areas is a classic example of agricultural EDS, driven by socio-ecological factors such as ecological environment restoration and conservation policies. This has caused severe impacts on local residents' livelihoods. In the study area, 98.55% of surveyed farmers experienced wild boar damage, with damaged cultivated land reaching 30.62 hectares, accounting for 79.55% of the total surveyed cultivated land, and over 30% of this land was subsequently abandoned. (4) Globally, EDS issues are becoming increasingly prominent. Their impacts not only lead to direct economic losses but also threaten human life and health and trigger social conflicts, highlighting the urgency of EDS regulation. EDS research should focus on constructing multi-scale dynamic analysis frameworks to more accurately capture the evolutionary characteristics of EDS at different scales. Meanwhile, there is an urgent need to establish multi-dimensional evaluation indicator systems and standardized methodologies for the quantitative assessment and comparative study of EDS. Additionally, there is a need to establish regulation mechanisms involving multi-scenario simulation, multi-objective optimization, and multi-stakeholder collaborative governance to achieve integrated management that balances ES and EDS.

As an important lens for exploring human-environment interactions in geography, research on cultural ecosystem services (CES) has grown rapidly in recent years, particularly in the interpretation, identification, and quantification of CES. With the rise of internet technologies, scholars have increasingly turned to social media platforms to conduct or support CES studies. This review, based on English-language literature indexed in the Web of Science (WOS) from January 1, 2013 to June 30, 2025, systematically examines the evolving phases of CES research supported by social media data and summarizes the key limitation involved. Social media data offer advantages such as large data volume, easy accessibility, and the ability to capture users' spatial and temporal visitation patterns. However, several limitations remain. At the data level, persistent challenges include representation bias, since social media users are predominantly younger, the perspectives of older demographic groups may be underrepresented. Moreover, the use of web crawlers for data collection raises legitimate concerns regarding user privacy and data protection. At the technical level, unresolved challenges include removing noise and duplicate information, reliably recognizing linguistic and cultural diversity, and addressing the persistent opacity of algorithmic process. At the theoretical level, the research is constrained by the lack of robust frameworks for conceptualizing CES flows and the difficulty in systematically linking user characteristics to specific service demands. In response, this review proposes a developmental framework of "Data foundation-Technical optimization-Theoretical integration". At the data level, the framework suggests improving privacy protocols, conducting cross-platform validation, and encouraging public participation to enhance data robustness. At the technical level, it emphasizes building multilingual training datasets, automating de-noising processes, and improving spatio-temporal resolution to capture dynamic user engagement with green spaces. At the theoretical level, the integration of fine-grained population mobility data from big data science, spatio-temporal insights from geography, public demand analysis from sociology, and policy perspectives from management can help advance the conceptualization of CES flow. This framework connects empirical exploration with theoretical construction, and provides a new pathway for advancing CES research supported by social media data.

Ecological systems contain both positive flows, which contribute to stability, and negative flows, which amplify disturbances. However, the interaction between these flow types remains underexplored. Focusing on the period 2000-2022, this study aims to construct an integrated ecological-risk coupled network and derive a comprehensive ecological security pattern for the Weihe River Basin, a key catchment in Northwest China. To achieve this objective, we first identified ecological sources and risk sources based on evaluations of ecological regulation functions and ecological sensitivity using multi-source datasets. A unified resistance surface was then generated by integrating terrain, land use, hydrology, and human activity factors. Using the patch - corridor - matrix framework, least-cost path analysis, and circuit theory, we extracted ecological corridors, risk corridors, and bidirectional regulatory corridors linking ecological and risk sources.The results show that: (1) Ecological and risk networks exhibit strong spatial differentiation. A total of 18 ecological sources and 109 ecological corridors are concentrated in the Qinling Mountains and mountainous regions of central China, forming a longitudinal ecological barrier. In contrast, 17 risk sources and 116 risk corridors are clustered in the northern Loess Plateau and Guanzhong Plain, constituting a transverse risk transmission belt. (2) The two networks are tightly coupled through 50 ecological→risk regulatory corridors, 43 risk→ecological regulatory corridors, and 69 mutual regulatory nodes, highlighting the Guanzhong Plain as the key interaction zone that buffers ecological conflicts and enables cross-regional regulation. (3) The resulting ecological security pattern displays a "conservation in the south and control in the north" configuration, emphasizing strict ecological protection in the western upper reaches, targeted governance in the northern Loess Plateau, and coordinated regulation in the middle and lower basin. Overall, by integrating positive and negative ecological flows into a unified analytical framework, this study overcomes the traditional limitation of focusing on ecosystem services while neglecting ecological risks. The proposed zoning-grading-classification management strategy offers a scientific basis for basin-scale ecological governance and advances the theoretical and methodological development of ecological network research.

Risk assessment of soil erosion (SE) under specific conditions of soil water retention (SWR) is essential for effective resource management, particularly in water source regions. In this study, the Southern Gansu Plateau (SGP) was selected as the study area. Drawing on monitoring data from hydrological stations and field surveys, the InVEST and RUSLE models were rigorously validated through multiple approaches, enabling a comprehensive characterization of the spatiotemporal dynamics of SWR and SE. Innovative indices including soil water retention index (SRI) and soil erosion risk index (SEI) were developed to investigate the underlying patterns linking SWR and SE. Furthermore, future changes in SWR and SE were simulated under three emission scenarios derived from multiple GCMs, and targeted strategies were proposed to enhance soil and water conservation capacity in alpine-cold river source areas. The results of the study showed that, during the period 1990-2019, the SGP exhibited a multi-year average SWR of 21.19 mm, with a rising trend at a rate of 0.54 mm·a^-1. The average SE modulus was 26.55 t·hm^-2, showing a declining trend at a rate of 0.38 t·hm^-2·a^-1. Areas with high SWR were primarily located in regions with dense forestland and grassland cover, including the middle reaches of the Daxia River and Taohe River basins in the central plateau, Maqu in the southwest, and the Bailong River basin in the southeast. In contrast, high SE values were mainly concentrated in the loess region of the northern SGP. Notably, areas experiencing increases in both SWR and SE partially overlapped with their respective high-value regions. In addition, a conditional logarithmic relationship was identified between the SRI and the SEI. Except for unused land, increases in SWR across all other land use/cover (LUC) led to reductions in SE. In LUC with high vegetation cover, even slight gains in SWR significantly suppressed SE, whereas in low-coverage areas, SE reduction was more sensitive to changes in SWR. Over the next 30 years, changes in SWR and SE are expected to be most pronounced in forestland, grassland, and unused land. Under the low-emission scenario, most LUC are projected to experience declines in SWR and increases in SE, whereas the opposite trends are observed under the medium- and high-emission scenarios. Built-up land shows the most adverse outcomes under the medium-emission scenario, while wetland exhibits a consistent trend of declining SWR and increasing SE across all three scenarios. Notably, changes in SWR and SE in unused land show a high degree of synchronicity. This study provides a case-based reference for assessing precipitation-triggered SE under the context of regional SWR.