The distribution of red spruce (Picea rubens) in the central Appalachian Mountains in Virginia and West Virginia, USA, was limited during post-Pleistocene warming followed by exploitative logging, wildfire, and other stressors in the early twentieth century. Moreover, future climate projections predict further fragmentation of this globally rare, montane boreal forest type. Ecotones provide a useful framework to identify changes in species distribution associated with environmental change. To evaluate long-term dynamics in unmanaged red spruce–northern hardwood forests, we remeasured 10 ecotonal transects across the central Appalachian Mountains that were established in 1992–1994 to assess post-exploitation recovery or lack thereof. In the tree strata, red spruce significantly increased its importance value, a metric of density and basal area, throughout the transects with gains in both the transitional ecotone and northern hardwood forest zones. Changes in red spruce sapling densities varied among transects, though decreases in sapling densities were correlated with increased trees per hectare. Over the 30-year study period, the red spruce–northern hardwood ecotone continued the compositional shift toward red spruce identified in the first decade of the study, while average annual temperature also increased.
Pollinators play a critical role in maintaining agricultural productivity and ecosystem functioning, yet evidence indicates that pollinator populations are under increasing pressure worldwide. Despite concern about pollinator decline, empirical evidence from smallholder farming systems in sub-Saharan Africa remains limited, particularly regarding how farmers perceive pollinator change and the socio-ecological conditions associated with these perceptions. This study examined smallholder farmers’ perceptions of pollinator change and associated socio-ecological factors in northern Ghana. Data were collected through a cross-sectional survey of 1,107 farming households across sixteen rural communities within the Black Volta and Western Wildlife corridors. Binary logistic regression models examined associations between socio-demographic, household and institutional, human-driven environmental, climate-related, and geographic factors and perceived changes in pollinator diversity, activity, and mortality. Most farmers reported decreased pollinator diversity (83.8
Although the characteristics of a region or site can have a large effect on the implications of climate change on wildfire risk, the associated causal relationships are poorly understood. This article reviews 424 studies to identify the environmental factors that mediate and moderate the relationship between climate change and wildfire risk, gathering the associated evidence from the perspective of regional and site-specific variability. Based on the review results, the study contributes a set of qualitative questions that provide a methodological basis for assessing the impact of climate change on wildfire risk at regional and site-specific scales, and a synthesis applying the questions to Sweden as an example region. This reveals feedback loops, in which fire activity alters the environmental and societal conditions that govern subsequent risk, that are difficult to capture in quantitative models but are essential for understanding real-world outcomes. For Sweden, this implies the need to further study the non-coastal parts of the predominantly rural counties in the northwest. For studies of regional environmental change, the questions and the synthesis reveal both the value and a method for allowing a diverse set of disciplinary experts to cooperate when conducting region or site-specific assessment of future wildfire risk, while considering both social and natural phenomena.
Cloudbursts are among the most destructive forms of extreme precipitation, characterized by short-duration, high-intensity rainfall that frequently triggers flash floods, landslides and debris flows, particularly in mountainous regions. Pakistan has experienced a significant increase in cloudburst-related disasters in recent years; however, national-scale susceptibility assessments of this catastrophic phenomenon remain limited. This study presents a comprehensive cloudburst susceptibility mapping of Pakistan using an innovative Expert Opinion (EO) and Large Language Model (LLM)-assisted weighted linear combination (WLC) framework. Twenty cloudburst-conditioning factors were compiled from multi-source datasets. Factor importance was determined using a hybrid ranking strategy that integrates EO and five LLMs (ChatGPT, DeepSeek, Gemini, Grok and Perplexity) through a standardized prompt-based ranking scheme. WLC-based susceptibility maps were produced for each weighting configuration and subsequently integrated to produce a composite cloudburst susceptibility map. The cloudburst maps reveal that Azad Kashmir, northern Punjab, Islamabad, central and northern Khyber Pakhtunkhwa, and some parts of Gilgit-Baltistan are very highly susceptible to cloudbursts. The integrated susceptibility map identifies northwestern Pakistan, particularly the Hindukush-Himalaya region, as highly prone to cloudbursts, reflecting the combined influence of complex topography, orographic lifting, atmospheric instability and high moisture availability. The results of the model performance reveal excellent predictive accuracy by all models, with the highest performance by Perplexity- and Grok-based models, closely followed by EO and ChatGPT. Overall, this study demonstrates that LLM-assisted multi-criteria decision analysis (MCDA) provides a strong, transparent and scalable framework for cloudburst susceptibility assessment, offering valuable insights into disaster risk reduction, early warning systems and climate-resilient planning in data-scarce regions.
Collaborative governance is an important pathway toward modernizing grassland ecological governance systems and governance capacity, and its effective implementation depends on the strategic interactions and dynamic adjustments among diverse stakeholders. Utilizing a benefit-cost driven analytical framework, this study constructs a four-party evolutionary game model involving government, enterprises, herders, and social organizations to analyze their participation in collaborative governance. The model examines strategy evolution trends, stability, and interactions among stakeholders, simulating evolutionary paths using empirical data from China’s northwest pastoral areas, and analyzes factors influencing stable strategy combinations. The study finds that: (1) the strategic choices of stakeholders are characterized by dynamic coupling and synergistic evolution, through which a multi-dimensional and multi-tiered collaborative governance network is formed via strategic gaming and interaction. The inflection points of the evolutionary trajectories of government and social organization strategies are mutually interdependent, whereas those of enterprise and herder strategies are both determined by the strategic decisions of social organizations. (2) The initial probability, as the starting parameter of the game system’s evolution, does not alter the final equilibrium direction of stakeholders’ strategic evolution, but it can influence the dynamic characteristics of the evolution process. At the initial phase of grassland ecological governance, rationally aligning the strategic propensities of multiple stakeholders via policy guidance, interest reconciliation, and the establishment of mutual trust mechanisms can significantly reduce the convergence time to the optimal strategy portfolio. (3) The strategic evolution of the four participating actors is asymmetric in nature, with their strategy choices being driven by heterogeneous factors. Specifically, government strategies are predominantly shaped by regional public benefits, central government incentives, and governance costs; enterprise behavioral adjustments hinge on returns from green transformation and technological innovation capabilities; herder strategies are highly sensitive to personal income expectations and the ease of participation; and the conduct of social organizations is co-determined by supervision costs and the level of financial support received. Accordingly, a combination of subsidy and reward policies, punitive regulations, interest alignment mechanisms, and supervision guarantee systems should be employed in an integrated manner, with differentiated regulatory measures applied according to the behavioral profiles of each stakeholder, in order to foster a stable and efficient collaborative governance framework. The findings provide insights for stakeholders in the collaborative governance mechanism of grassland ecosystems, particularly in terms of responsibility division and joint cooperation.
