研究生知识库

本论文选取普定喀斯特生态系统观测研究站内的自然恢复样地，建立了以碳通量观测为核心的基于喀斯特关键带理念的综合观测系统，主要包括大气-冠层界面的涡度相关法观测系统、土壤-大气界面的多通道自动箱法观测系统和洞穴-大气界面的CO2浓度梯度观测系统。另外站内有标准气象站可提供相关辅助气象参数，洞穴内部也安装了温湿度、气压、风速、滴率等传感器来记录洞穴微气象变化，以及RAD7型氡气测量仪来记录洞穴氡气浓度变化，用来示踪洞穴通风过程。经过2015至2018年的连续观测，我们得到以下结论：（1）沙湾自然恢复样地在2010年退耕后，植被恢复明显，观测期间（2015年3月至2018年12月）样地基本被植被完全覆盖。观测期间，年际降雨变化幅度较大，经历了由极高（2015年降雨量为1637 mm）到极低（2016年降雨量为963 mm）到逐步升高的过程（2017和2018年降雨量分别为1166、1454 mm）；年均气温呈小幅下降趋势（16.5降至16.1℃）；全年光合有效辐射经历了先降低后升高的变化趋势，全年光合有效辐射与降雨量呈单峰曲线关系，在降雨量为1500 mm左右时全年光合有效辐射达到峰值。2016至2018年平均生态系统净交换（NEE）大小为544±77 g C m-2 a-1，其中雨季（4至9月）占全年NEE的68 %至82 %，NEE随全年光合有效辐射增加而增加，随土壤温度增加而减小。植被调查结果显示生物量增加所导致的碳汇大小约为95 g C m-2 a-1。植被生物量增加是NEE的主要组成部分，但其值远小于涡度相关法观测结果。可能与所选样方代表性不足以及根生物量无法准确计算造成植被调查法结果低估和涡度法所测碳汇包含岩溶有关，但两者相加在量级上仍不足以解释所测NEE较高的原因。本论文根据对洞穴通风过程的监测和洞穴碳排放的估算，发现土壤呼吸产生的CO2有一部分向下扩散至洞穴裂隙并集中以点源的方式排放到大气中，因为受涡度相关法的足迹与权重问题影响，该部分CO2无法被涡度相关法全部观测到，甚至一部分排放完全不在涡度相关法的足迹范围内，导致涡度相关法所测生态系统呼吸偏低，进而导致所测NEE的高估。（2）土壤碳通量异质性较大，以2号观测点为例，土壤碳排放大小平均为973±45 g C m-2 a-1，其中雨季约占70 %左右。土壤碳通量主要受土壤温度和土壤含水量的耦合变化控制。当5 cm处土壤温度低于30 ℃左右时，土壤碳通量与土壤温度呈指数关系，温度敏感性指数Q10为3.3；当土壤温度超过30℃之后，土壤含水量会明显降低，进而使得土壤碳通量减小。要特别注意的是在估算区域土壤呼吸时要考虑土壤分布的空间不连续性，否则会导致土壤呼吸的严重高估。同时，土壤呼吸产生的CO2也有可能向下扩散至裂隙洞穴等再经过洞口集中排放，其占比还需进一步研究。 （3）样地内洞穴（沙湾洞）CO2浓度呈现短期至季节以及年际变化，变化范围约为1000至22000 ppm。季节尺度上，以夏天最低、秋天升至最高、冬天降低、春天小幅升高为主要特征；年际尺度上，峰值从2015年的16000 ppm 增加至2018年的22000 ppm。洞穴二氧化碳和洞穴氡气变化规律基本一致，本论文通过对全球有连续监测的35个洞穴氡气季节变化和洞穴结构形态进行归纳整理，将其分为3种季节变化类型和5种通风模式。对于沙湾洞而言，当洞穴温度和大气温度接近时，洞穴通风弱，导致二氧化碳和氡气的累计；当洞穴温度低于大气温度时，受密度差驱动，导致洞穴内部气流向下运移，从而使得大气通过洞口进入洞穴，混合作用使得洞穴二氧化碳浓度和氡气浓度降低；当洞穴温度高于大气温度时，洞穴内气流向上运移，从而带来包气带高浓度二氧化碳和氡气，使得洞穴二氧化碳和氡气浓度升高。初步估算，沙湾洞通过上部洞口向大气碳排放量约为每年2.3吨。（4）综上所述，喀斯特地区坡耕地退耕后，自然恢复过程可以显著增加生物量，进而形成植被碳汇。另外，由于喀斯特地区地下空间发育，其二氧化碳浓度远高于大气背景值，形成次级碳库，在与大气交换过程中形成碳源，其本质是喀斯特地区土壤二氧化碳向下扩散以及土壤有机质下渗至包气带后分解所形成的地下次级碳库通过洞口和裂隙向大气的再排放过程，体现了喀斯特地区碳循环过程的复杂性和独特性。涡度相关技术本身在复杂下垫面情况下已有很多成功的应用，但在数据解释时需要因地制宜的考虑所观测对象的特殊性，特别是在喀斯特地区，不能忽视地下通风过程的影响。

A comprehensive observation system for CO2 flux through karst critical zone was established on an abandoned farmland undergo natural restoration in Puding Karst Ecosystem Research Station. The observation system mainly consist of: the eddy covariance observation system between canopy and atmosphere interface, automatic multi-chamber observation system between soil and atmosphere interface, CO2 gradient observation system between cave and atmosphere. In addition, standard weather station can provide relevant auxiliary meteorological parameters in the station. Also, temperature and humidity sensors, air pressure sensors, drip rates sensors and wind speed analyzer were installed inside the cave to record the micro-meteorological changes. In addition, radon gas analyzer (RAD7) was installed to record the changes of the radon concentration inside the cave to track the ventilation process. Following are main conclusions after continuous observation from 2015 to 2018:(1) The study site was abandoned in 2010, the vegetation recovery was obvious during natural restoration process. During the observation period from March 2015 to December 2018, the study site was almost completely covered by vegetation. The annual precipitation change is large, and experienced extremely high rainfall of 1637 mm during 2015, but extremely low rainfall of 963 mm during 2016, and then increased to 1166 and 1454 mm during 2017 and 2018 respectively. The annual average temperature show a slight decline trend which changes from 16.5 to 16.1℃. The photosynthetic effective radiation changes associated with precipitation, the relationship between annual photosynthetic effective radiation (PAR) and rainfall shows a single peak curve, and the PAR reaches its peak when the rainfall is about 1500 mm. The average NEE measured by eddy covariance system is about 544±77 g C m-2 a-1 from 2016 to 2018, the rainy season (April to September) accounts for 68 % to 82 % of the annual NEE. The NEE increases with the increase of the annual photosynthetic effective radiation and decreases with the increase of soil temperature. Vegetation survey showed that carbon sinks induced by biomass increase is about 95 g C m-2 a-1, biomass increase during restoration should be the main part of NEE, but the value is far less than the result measured by eddy covariance. On the one hand, this may be due to the sampling representativeness as well as it is hard to accurately survey root biomass increment. On the other hand, the NEE measured by eddy covariance should also include carbonate weathering process and SOC change process. In addition, the overestimated result may be mainly related to the ventilation process. Part of soil respired CO2 can diffused down to the caves and fissures which will be emitted to the atmosphere as point source later, the eddy covariance system cannot capture this part of ecosystem respiration because of the limited footprint, which lead to overstimated NEE. (2) Soil CO2 flux shows large heterogeneity, in the case of NO. 2 observation spot, annual average carbon emission is about 973 ±45 g C m-2 a-1, and rainy season accounts for about 70% of the whole year value. Soil carbon flux is mainly controlled by the coupling change of soil temperature and soil water content. In general, when soil temperature at 5 cm below 30 ℃, soil carbon flux shows exponential relationship with soil temperature, and soil carbon flux temperature sensitivity index Q10 is about 3.3. When the soil temperature exceeds 30 ℃, soil carbon flux began to decreases because of reduced soil water content. Great attention should be paid to the spatial discontinuity of soil distribution in the estimation of regional soil carbon emission, otherwise it will lead to serious overestimation. In addition, soil respired CO2 not only diffused to the atmosphere through the soil surface, but also can diffused downward to underground spaces such as caves and fissures to form a temporary pool and become a second emission path later. As a result, the flux observation only based on the soil surface will underestimate the soil carbon emissions in karst areas, but the specific contribution of this process needs to be further quantified. (3) CO2 concentration inside Shawan cave is much higher than atmospheric background and variation ranges from 1000 to 22000 ppm. The maximum concentration appears at autumn, the minimum concentration appears at summer, the concentration decrease at winter, the concentration increase at spring. The peak CO2 concentration shows a rising trend which increased from 16000 to 22000 ppm during the study period. The variations of cave CO2 and cave radon gas show same pattern. We summarized 35 caves with continuously monitoring of radon gas and divided them into three types of seasonal variation pattern and five types of ventilation mode. For Shawan cave, when the cave temperature and atmospheric temperature are close, the cave ventilation is weak, leading to the accumulation of CO2 and radon gas. When the cave temperature is lower than the atmospheric temperature, the air inside the cave will move downwardly driven by the density difference, as a result, the atmosphere air will enter the cave through the entrance, and the mixing will reduce the concentration of CO2 and radon gas in the cave. When the cave temperature is higher than the atmospheric temperature, the cave air will moves upwardly, which bring the high concentration CO2 and radon gas air of the vadose zone, which lead to high concentrations in the cave. Preliminary estimates show that the carbon emissions through the upper entrance of the cave is about 2.3 tons per year.(4) In conclusion, the natural restoration process can significantly increase the biomass after the abandonment of farmland and contribute to mitigate global warming. Especially, the underground space are fully developed in karst region and the inside CO2 concentration is much higher than atmospheric background, which forms the secondary carbon pool in karst ecosystem. This pool becomes a source of atmosphere CO2 during the process of ventilation. Essentially, the source of this carbon pool is come from the downward diffusion of soil respiration and the decay of infiltrated organic matter inside the vadose zone, but emitted through the cave entrance and fissures instead. This characteristic reflects the complex and uniqueness of carbon cycle process in karst ecosystems. Although eddy covariance method itself has been successfully applied in complex underlying surface, but attention should be paid according to local conditions when interpret the results. Especially, the influence of underground ventilation process cannot be ignored for carbon cycle in karst ecosystem.