The decomposition of deadwood is a crucial process for the accumulation and sequestration of soil organic carbon (SOC) in forest ecosystems. However, the response of SOC to different decay classes of deadwood and the underlying mechanisms remain poorly understood. Here, we investigated the dynamics of SOC, soil properties, extracellular enzyme activities, and phospholipid fatty acid biomarkers across five decay classes (ranging from 1 to 5) of Masson pine (Pinus massoniana Lamb.) downed deadwood in a subtropical–temperate ecotone forest in Central China. Our results revealed a nonlinear response pattern of SOC along the deadwood decomposition gradient, with the maximum value at the decay class 4. Soil available nitrogen content, bacterial biomass, fungal biomass, the ratio of fungal-to-bacterial biomass, cellulase, activity and ligninase activity all increased with the intensification of deadwood decay, while soil pH decreased. The increase in SOC content was associated with a direct positive effect of bacteria and both direct and indirect positive effects of fungi by cellulose activity, but ligninase activity showed no significant relationship with SOC content. These findings suggest that cellulose and microbial biomass are key determinants of soil C formation and sequestration during deadwood decomposition. This study highlights the importance of the nonlinear response of SOC to deadwood decay, providing valuable insights for predicting future carbon-climate feedbacks.

Understanding the mechanisms influencing forest changes is crucial for developing effective adaptation strategies to mitigate the impacts of climate change and anthropogenic factors on ecosystems. In this study, we utilized the Theil–Sen median, Mann–Kendall (MK) test, contribution partition, partial least squares (PLS), geodetector, and residual analysis methods to explore the effects of human activities and climate change on forest cover and forest NDVI in the Altai Mountains. For forest cover, the results indicate that both forest management policies and climate change influence changes, with forest management policies being the primary factor. However, the mechanisms vary across different altitude belts. Natural mortality in alpine and subalpine forests, coupled with the promoting effect of climate change, has led to a dynamic equilibrium in forest cover, while human activities and climate change together promote forest cover expansion in the low mountain belt. For the forest NDVI, areas where climate change and human activities jointly increased NDVI growth account for the largest proportion (50.20%) across the six scenarios. Climate change has a wider influence on the NDVI than do human activities, mainly driving NDVI increases (30.28%). In contrast, forest degradation is caused primarily by human activities (19.39%), especially along forest edges. Among the climate factors, precipitation and snowmelt are the main controlling factors. Notably, snowmelt in March and April (SM_34) is the most important water source for forest growth before the growing season. This study offers essential scientific support for forest management and strategic planning in the Altai Mountains.

Phenological model is a useful tool to predict vegetation phenology and investigate the relationship between vegetation and climate. However, compared with temperate and boreal ecosystems, phenological modeling work in alpine regions has received much less attention. In this study, we established a semi-mechanistic phenological model, the AGSI model, considering different impacts of daily maximum and minimum air temperatures and constraints of precipitation and photoperiod for predicting foliar phenology of alpine grasslands on the Qinghai–Tibetan Plateau (QTP). The model was driven by daily minimum temperature (T_min), daily maximum temperature (T_max), precipitation averaged over the previous month (PA), and daily photoperiod (Photo). Based on the AGSI model, we further assessed impacts of T_min, T_max, PA, and Photo on modeling accuracy, and identified the predominant climatic controls over entire alpine grasslands on the QTP. Results showed that the AGSI model had higher accuracy than other GSI models. Total RMSE of predicted leaf onset and offset dates when evaluated using ground observations was 12.9 d, which decreased those of other models by 10.9%–54.1%. Total modeling accuracy of leaf onset and offset dates was improved by 20.2% after considering the effects of T_max and PA in the AGSI. Overall, PA and T_min showed more critical and extensive constraints on foliar phenology of alpine grasslands. The limiting effect of T_max was also considerable, primarily during July–November. This study provides a simple and effective tool to predict foliar phenology of alpine grasslands and evaluate the climatic effects on vegetation phenological development in alpine regions.