This study investigated the interactive effects of warming and inundation on methane (CH₄) fluxes and soil physicochemical mechanisms in the littoral wetland of Qinghai Lake. Soil samples were collected from the Bird Island littoral wetland. Eight treatments were established: natural control (CK), different inundation depths (S0, S10, S20), warming alone (ZWCK), and warming combined with inundation (ZW0, ZW10, ZW20). CH₄ fluxes were measured over one year using an ABB LGR analyzer. Principal component analysis (PCA) and Mantel tests were used to identify environmental drivers. The main findings are as follows: (1) Under different water level treatments, CH₄ fluxes showed a unimodal seasonal pattern, peaking in autumn. Warming and the interactive treatments shifted the emission pattern to bimodal or multimodal and significantly increased emission intensity. The warming-alone group had the highest annual emission, with anomalously high winter emission (47.683 μg·m⁻²·h⁻¹). Under the ZW20 treatment, emissions were synergistically enhanced in summer and autumn but turned to suppression in winter. (2) PCA showed that the carbon nitrogen pool (70.5%) and the salinity pH gradient (14.9%) were the main drivers of soil variation. The interactive effects on carbon-nitrogen dynamics shifted with season: warming promoted accumulation in spring; warming with shallow inundation retained carbon-nitrogen in summer, but deep inundation caused loss; warming with deep inundation formed a nutrient center in autumn; inundation dominated accumulation in winter, while warming increased loss. (3) Mantel tests showed that carbon-nitrogen components were highly correlated across seasons, but were strongly and positively correlated with CH₄ flux only in autumn (Mantel’s r ≥ 0.4, p < 0.05), indicating autumn as the key window. These findings provide important insights into carbon cycling processes and regulatory mechanisms of alpine wetlands under future climate change scenarios.

Lake-level rise driven by changing precipitation is increasing inundation in littoral wetlands, but responses of cbbL-harboring carbon-fixing microorganisms remain uncertain. A controlled soil-column experiment included four hydrological regimes: unflooded soil (CK), water level aligned with the soil surface (S0), and standing-water depths of 10 cm (S10) and 20 cm (S20). Amplicon sequencing of cbbL showed that the Shannon index was significantly higher in S20 than in CK (p = 0.0394), whereas Simpson, Chao1, and ACE did not differ significantly among treatments. Overall amplicon composition also lacked a significant treatment effect (PERMANOVA: R² = 0.344, p = 0.112). Pseudomonadota represented 96.9–97.7% of taxonomically assigned sequences and remained dominant across treatments, although several genera varied in relative sequence proportion. Exploratory LEfSe screening identified Rhodoblastus-, Hydrogenophaga-, and Azospirillum-related lineages as candidate indicators of CK, S10, and S20, respectively. Multivariate soil profiles did not differ significantly among treatments (PERMANOVA: R² = 0.374, p = 0.220). After collinearity filtering, EC had the highest normalized mean absolute out-of-fold SHAP contribution among EC, pH, and TC (43.8–59.1%). However, outer leave-one-out cross-validation produced negative predictive R² values (−0.43 to −0.11), indicating limited model generalizability. Overall, S20 was associated with higher Shannon diversity than CK, while Pseudomonadota remained dominant and several genera varied in relative sequence proportion. EC should therefore be regarded as an exploratory candidate correlate of cbbL alpha diversity rather than a validated environmental driver, and this association requires further evaluation under natural hydrological conditions.

To clarify the effects of precipitation changes on greenhouse gas (CO₂, CH₄, N₂O) emissions from the lakeshore wetland ecosystem of Qinghai Lake, this study establishes five precipitation treatments (D25%, D75%, CK, I25%, I75%) using a rainfall manipulation device. During the growing season, greenhouse gas fluxes were measured via the static chamber–gas chromatography method, while soil temperature, moisture, total carbon (TC), total nitrogen (TN), pH, electrical conductivity (EC), and aboveground and belowground biomass were simultaneously monitored. The results showed that at 11:00, both increased and decreased precipitation inhibited CO₂ and CH₄ emissions (with D25 showing the strongest inhibition for CO₂ and I75 for CH₄); however, increased precipitation promoted N₂O emissions (I75 strongest), while decreased precipitation suppressed them (D25 strongest). At 15:00, only D25 inhibited CO₂ emissions, whereas all other treatments promoted them (I75 strongest). CH₄ fluxes under all treatments were higher than those in CK, indicating a promoting effect (I25 strongest). N₂O emissions responded to precipitation changes as follows: wetting promoted emissions (I25 optimal), while drying suppressed them (D75 stronger). Correlation analyses revealed treatment- and time-specific patterns. At 11:00, significant negative correlations between CO₂ and soil temperature and significant positive correlations between CH₄ and soil temperature were observed only under I75; soil moisture was significantly positively correlated with CH₄ under D25; TN significantly inhibited all three gases under wetting treatments; and TC showed a highly significant negative correlation with N₂O under I75 (p < 0.01). At 15:00, soil temperature generally promoted CO₂ and CH₄ emissions (except for CO₂ under D75); soil moisture promoted CH₄; TC was negatively correlated with both CO₂ and N₂O; pH was positively correlated with CO₂ only under CK; the correlation between EC and CO₂ shifted from positive (CK) to negative (I75); aboveground biomass suppressed CO₂ under I25 and I75 but promoted N₂O under D25. In summary, precipitation changes significantly modulate greenhouse gas emissions and their relationships with soil factors, with strong temporal and treatment dependencies. Future studies should integrate long-term observations and process-based models to further quantify the comprehensive effects of precipitation changes on the greenhouse gas budget of this wetland ecosystem.