多年冻土区是陆地生态系统最大的碳库。随着全球变暖，多年冻土不断升温融化，大量的冻土碳被激活参与到生物地球化学循环中，而当前对冻土碳的观测和模拟存在很大不确定性，使得冻土碳排放与气候变化之间的反馈作用更为复杂。热融湖塘是多年冻土退化最直观和最广泛的形式，也是多年冻土碳参与全球碳循环的重要媒介，使得热融湖温室气体排放的时空差异及其过程机制研究成为地球科学领域的热点与前沿问题。环北极多年冻土区热融湖是温室气体排放的热点区域，是全球湖泊温室气体排放的重要组成部分，但是青藏高原热融湖塘温室气体研究仍十分缺乏。

本研究以青藏高原典型区热融湖塘为研究对象，利用区域调查、定位观测和融冰期定点观测等多种手段，探究青藏高原典型区106个热融湖塘甲烷（CH₄）和二氧化碳（CO₂）排放时空差异。揭示青藏高原热融湖塘CH₄和CO₂源/汇特征、区域和季节差异及控制因素；通过漂浮箱布放（累积布放690个）明确热融湖塘CH₄不同排放途径（气泡和扩散）通量及其对总通量贡献；量化热融湖塘CH₄和CO₂冬春季累积和排放特征及其对全年通量的重要性；基于2020-2021年野外观测数据，结合湖冰物候和统计模型，整体估计了青藏高原热融湖塘CH₄和CO₂排放年总量。主要得出以下结论：

1）青藏高原热融湖塘CO₂扩散通量具有较大时空差异。发育在高寒沼泽草甸的热融湖塘具有最高的CO₂扩散通量，融冰期末期和秋季翻转期（水体快速垂直混合）通常具有较高的CO₂扩散通量，多种因素共同塑造了热融湖塘CO₂扩散通量时空格局。

青藏高原热融湖塘（N=106）CO₂扩散通量为19.03 ± 65.38（mmol m^?2 d^?1），热融湖塘CO₂源汇特征在区域和季节之间存在较大变化，整体上为CO₂排放的源，7月和8月存在较多热融湖塘为CO₂吸收通量；热融湖塘CO₂扩散通量区域差异主要受植被变化调控，高寒沼泽草甸热融湖塘CO₂浓度和扩散通量（181.40 ± 180.56 mmol m^?2 d^?1）显著高于高寒草甸、高寒草原和高寒荒漠（p<0.01）。热融湖塘CO₂吸收通量主要发生在植被发育较好的区域，2020?2021年7月和8月具有最多的CO₂吸收通量。热融湖塘在融冰期末期和秋季翻转期具有高的CO₂浓度和有利的扩散条件，表现为高的CO₂排放。热融湖塘沉积物有机碳、水体营养物质、电导率、气体传输系数等因素均对热融湖塘水体CO₂产生和传输有显著影响，是调控青藏高原热融湖塘CO₂通量变化的重要影响因素。

2）青藏高原热融湖塘CH₄扩散和气泡通量表现出明显的时空差异。气泡是热融湖塘CH₄排放的主要途径，植被类型和有机质含量是影响热融湖塘CH₄排放的重要影响因素。

青藏高原热融湖塘均为大气CH₄的源，2020?2021年所有湖塘（106个）CH₄扩散通量平均值为5.11 ± 16.73 mmol m^?2 d^?1，湖塘CH₄气泡通量贡献约为63.46 %。CH₄扩散和气泡通量均受植被类型变化控制，整体表现为高寒沼泽草甸>高寒草甸>高寒草原>高寒荒漠，发育在高寒沼泽草甸的湖塘CH₄扩散通量显著高于高寒草原（p<0.001）和高寒荒漠（p<0.001），但是与高寒草甸之间差异不显著（p=0.12）；而CH₄气泡通量变化按照植被变化梯度变化，均存在显著差异（p<0.05）。热融湖塘CH₄扩散通量主要受沉积物有机碳密度、水体营养物质和电导率影响；气泡通量主要受沉积物有机碳密度影响。

3）青藏高原热融湖塘CH₄和CO₂具有明显的冰封期累积和融冰期排放特征，但累积量级有明显差异，水深、水文连通性及降雪事件是调控热融湖塘冬季累积和春季释放的重要控制因素。融冰期CH₄和CO₂扩散通量对全年贡献不可忽视。

热融湖塘水体CH₄和CO₂浓度累积量级明显不同，相比较于夏季CH₄和CO₂浓度，热融湖塘水体冰封期CH₄浓度高出夏季无冰期2?3数量级，累积效应明显，而冰封期CO₂浓度仅高出夏季无冰期数倍；随热融湖塘深度增加，冰封期和融冰期CH₄浓度逐渐增加，但是CO₂浓度并没有随着湖泊深度显示出增加的规律。水文连通性和降雪事件可以解释部分湖塘CH₄和CO₂浓度随湖冰消融变化规律。热融湖塘融冰期CH₄扩散年通量平均贡献为42.2%，融冰期CO₂扩散年通量平均贡献为19.2%；部分热融湖塘无冰期整体表现为CO₂吸收通量，融冰期CO₂排放贡献了100%，这可能导致部分热融湖塘从碳汇转换为碳源，外源输入是调控CO₂浓度和通量变化的重要因素。

4）青藏高原热融湖塘CH₄和CO₂年（2020?2021）总通量分别为0.086 ± 0.032（0.02?0.18）Tg CH₄ yr^?1和0.60±0.14（0.27?1.13）Tg CO₂ yr^?1。热融湖塘CH₄气泡通量控制着CH₄年总通量，CO₂仍是热融湖塘碳排放的主要形式。

The permafrost zone is the largest carbon reservoir of terrestrial ecosystems. As global warming triggers the continuous warming and melting of permafrost, a large amount of permafrost carbon have been activated to participate in biogeochemical cycles. Current observations and modelling of permafrost carbon, on the other hand, are characterised by large uncertainties, limiting the understanding of the feedback between permafrost carbon emissions and climate change. Thermokarst lakes and ponds (hereafter referred to as thermokarst lakes) are the most intuitive and widespread form of permafrost degradation. Meanwhile thermokarst lakes are an important medium for permafrost carbon to participate in the global carbon cycle. The study of the spatial and temporal variation of greenhouse gas emissions from thermokarst lakes and their process mechanisms has become a hot and advanced issue in the field of earth sciences. Thermokarst lakes in the circumpolar permafrost region are hotspots for greenhouse gas emissions and are an important component of global lake carbon emissions, but greenhouse gas research on thermokarst lakes in the Tibetan Plateau (TP) is still lacking.

In this study, we investigated the spatial and temporal differences in CH₄ and CO₂ emissions from thermokarst lakes in typical areas of the TP (a total of 106 thermokarst lakes) by using various means, including regional surveys, localized observations and positioning observation during ice-melting. The main objectives of this study are: 1) to reveal the characteristics, regional and seasonal differences of CH₄ and CO₂ source and sink in thermokarst lakes on the TP and their controlling factors, 2) to clarify the fluxes of different emission pathways of CH₄ from thermokarst lakes and their contributions to the total fluxes through floating box deployment (690 in total), 3) to quantify the accumulation and emission characteristics of CH₄ and CO₂ in thermokarst lakes in winter and spring and their importance to the annual fluxes. Finally, estimation of total annual CH₄ and CO₂ emissions from thermokarst lakes on the TP based on 2 years of field observations (2020-2021), combined with lake ice phenology and statistical models. The following main conclusions were drawn.

(1) Thermokarst lakes on the TP showed large spatial and temporal differences in CO₂ diffusive fluxes. Thermokarst lakes developed in alpine swamp meadows (ASM) have the highest CO₂ diffusive fluxes, while at the end of the melt-ice period and the autumn overturning season usually higher CO₂ diffusive fluxes were recorded, and several factors together shape the spatial and temporal patterns of CO₂ diffusive fluxes in thermokarst lakes. The CO₂ diffusive flux from thermokarst lakes (N=106) on the TP was 19.03 ± 65.38 (mmol m^?2 d^?1), and the CO₂ source or sink characteristics of thermokarst lakes varied considerably between regions and seasons, with an overall source of CO₂ emissions and more lakes showing CO₂ uptake fluxes in July and August compared to other seasons. Regional differences in CO₂ diffusive fluxes from thermokarst lakes were mainly regulated by vegetation changes, and CO₂ concentrations and diffusive fluxes from thermokarst lakes in ASM (181.40 ± 180.56 mmol m^?2 d^?1) were significantly higher than those in alpine meadows, alpine steppes and alpine deserts (p<0.01). CO₂ uptake fluxes in thermokarst lakes on the TP mainly occur in areas with better vegetation development, with the highest CO₂ uptake fluxes in July and August from 2020?2021. At the end of the melt ice period and the autumn overturning period high CO₂ emissions were encountered due to high CO₂ concentrations and favorable diffusive conditions. Factors such as organic carbon of thermokarst lake sediments, water nutrients, electrical conductivity, and gas transport coefficients all influence CO₂ production and transport in thermokarst lake waters, and are important influencing factors regulating CO₂ flux changes in thermokarst lakes on the TP.

(2) CH₄ diffusive and ebullitive flux in thermokarst lakes on the TP showed obvious spatial and temporal differences. Bubbles are the main pathway of CH₄ emission from thermokarst lakes, and vegetation type and organic matter content are important influencing factors on CH₄ emission from thermokarst lakes. All the thermokarst lakes in the TP are sources of atmospheric CH₄ release, and the mean CH₄ diffusive flux for all lakes (106) is 5.11 ± 16.73 mmol m^?2 d^?1, and the percentage of ebullitive flux for all lakes and ponds is 63.46 % in 2020?2021. Both CH₄ diffusive and ebullitive fluxes were controlled by vegetation type changes, and the overall performance was alpine swamp meadow > alpine meadow > alpine steppe > alpine desert. The CH₄ diffusive fluxes of thermokarst lakes developed in alpine swamp meadow were significantly higher than thermokarst lakes development in alpine steppe (p<0.001) and alpine desert (p<0.001), but not significantly different from those in alpine meadow (p