【作者】 胡宗达；

【导师】 刘世荣； 刘兴良； 史作民；

【作者基本信息】 中国林业科学研究院 ， 生态学， 2013， 博士

【摘要】 土壤呼吸是陆地生态系统向大气圈释放CO2的主要源，指示着土壤中异养微生物活性和土壤有机碳矿化速率，其时空变化直接影响到陆地生态系统的碳循环。土壤呼吸涉及碳循环中复杂的生物学和生态学过程，对环境变化程度响应各异，至今还存在很多不确定因素。因此，揭示昼、夜间土壤呼吸的时空动态变化特征及其驱动机制有助于加深我们对土壤呼吸变化机理的了解，为准确估算土壤碳排放以及碳循环机制对全球变化的响应提供参考。目前，在亚热带亚高山地区森林和草甸土壤呼吸的研究工作相对欠缺，尤其是夜间土壤呼吸更为少见。为此，本研究选择青藏高原东南缘的亚高山川滇高山栎灌丛和草甸为研究对象，旨在为该区陆地生态系统碳循环提供理论基础。川滇高山栎灌丛群落是该类型中面积最大且分布于高海拔高山峡谷地带的典型地带性植被，具有重要的生态水文功能。亚高山草甸是高寒草甸的一种类型，也是川西地区的一种主要植被类型。本研究在川西巴郎山，沿海拔梯度(2551、3091和3549m)布设了土壤呼吸实验样地。于2010年9月至2011年12月用LI-8100A土壤碳通量自动测量系统测定了土壤呼吸速率及其环境因子，并于2010年12月在相应样地土壤5cm深处布设一套土壤温度和水分的自动记录仪。同时在2010年11月、2011年4月、8月和11月采集土壤并分析土壤活性有机碳和氮素含量。旨在探明川滇高山栎群落昼夜土壤呼吸的时空格局及其影响因素，结果表明：(1)巴郎山川滇高山栎群落原始林地土壤异养(RH)和根呼吸(RR)速率均具有明显季节变化趋势。整个观测期，昼、夜间RH的均值分别为1.57±1.06和1.13±0.96μmol m-2s-1，RR分别为0.60±0.61和0.47±0.41μmol m-2s-1；川滇高山栎群落昼间RH和RR的年均累积量显著高于夜间；土壤温度、土壤含水量是昼间和夜间的RH和RR的主要限制因子。累积土壤总呼吸和RH与0~30cm土壤微生物生物量氮、易氧化有机碳和轻组有机碳显著相关；RR与易氧化有机碳显著相关，但与细根生物量(根径<0.5cm)相关性不显著。(2)观测期内各原始群落林地中，异养呼吸和土壤总呼吸的温度敏感性(Q10)均表现为昼间温度敏感性低于夜间。整个群落昼间异养呼吸和土壤总呼吸的Q10值分别为3.77±0.25(均值±标准误差)和4.51±0.41，夜间分别为4.96±0.45和5.28±0.79。(3)不同海拔川滇高山栎原始群落昼间土壤表面CO2通量总体趋势与地下5cm处土壤温度的季节变化一致，均出现昼高夜低现象，且生长季节显著高于非生长季节；土壤温度、土壤含水量及其交互作用是影响土壤表面CO2通量的主要环境因子，其构建的模型解释了川滇高山栎群落昼、夜间土壤表面CO2通量变异的83.2%和96.6%；温度敏感性指标(Q10)的分析显示，该区生长季节的Q10值明显低于非生长季节，且夜间高于昼间。各海拔梯度川滇高山栎群落全年的Q10值昼间为4.35~4.48，夜间为5.37~6.42。累积CO2通量与土壤微生物生物量氮和易氧化有机碳显著相关(P<0.05)，而与轻组有机碳呈弱相关。土壤表面CO2通量的年累积量为371.08~951.55Cm-2yr-1(平均为684.75Cm-2yr-1)，其中非生长季节的累积通量是生长季节的0.244倍，而夜间CO2通量对年积累量的贡献率为39.82%~44.60%(平均为42.46%)。如果仅用昼间数据拟合的模型估测年累积CO2通量，在非生长季节会高估12.96%，生长季节高估3.19%。综合分析表明，在不同时间尺度上，生物和非生物因素对川滇高山栎群落土壤表面CO2通量的影响程度不同，土壤表面CO2通量的季节变化主要受温度和水分变化的调控外，还可能受到底物供给状况的影响。为了较为客观地推算土壤表面CO2通量的日通量和年通量，在构建土壤呼吸模型时，除了考虑对该地区土壤水热条件的综合效应及土壤活性有机碳组分的库容量外，还应考虑其它非生物因子的影响，并需加强夜间土壤表面CO2通量监测。(4)火烧迹地川滇高山栎群落与其对照群落中土壤总呼吸(RS)和异养呼吸(RH)呈现显著的相似“抛物线”变化趋势，表现为生长季节最高，非生长季节最低的变化格局，且火烧迹地群落在非生长季节的RS和RH均显著高于对照群落，而生长季节则低于对照群落；两种群落RS和RH在生长和非生长季节以及整个观测期内均与土壤5cm温度呈显著指数相关，而与土壤水分含量则表现为在非生长季节和整个观测期内呈显著的幂函数相关，但土壤温度和水分含量构建的双因素模型可更好地解释RS和RH的时间变异。不同季节火烧迹地RS和RH的Q10值均低于对照群落，整个观测期内火烧迹地群落RS和RH的Q10分别为3.56和3.00，对照群落分别为5.33和4.23；两群落土壤轻组有机碳(FLOC)、颗粒有机碳(POC)和硝态氮(NO3-N)与RS和RH的时间变异呈显著线性相关；非生长季节火烧迹地群落RS和RH显著高于对照群落，生长季节低于对照群落但未表现出显著水平；火烧迹地群落RS和RH年累积通量分别为768.23和655.60gCm-2yr-1，分别比对照群落低13.57%和高20.59%。总的来说，火干扰改变了土壤地表温湿度及土壤底物有机碳氮素的供给水平等因素，进而影响到群落RS和RH。(5)环割造成川滇高山栎群落的土壤总呼吸在不同时间尺度上均不同程度地低于未环割群落，即在生长季节和非生长季节比未环割群落分别约降低了10.6%和16.5%；整个观测期内，平均降低了11.96%。(6)川西亚高山草甸昼、夜间土壤呼吸的变化格局不同，昼间呈双峰型，夜间呈抛物线型；整个观测期(4~11月)内，夜间土壤呼吸占总土壤呼吸的46.79%，其中草盛期(6~9月)的昼、夜土壤呼吸占有较大比例，与同海拔的川滇高山栎群落比较，草甸昼、夜间土壤呼吸速率分别高4.7%和20.5%；整个观测期草甸昼、夜间的Q10值分别为3.90和3.74。研究结果说明，采用昼、夜间土壤呼吸的瞬时值来推算土壤呼吸的季节通量和年通量时，需加大夜间土壤呼吸的测定。此外，观测发现，巴郎山阳坡(东南坡)川滇高山栎群落林地5cm处土壤月平均温度，海拔升高100m的递减速率为0.22℃。但在非生长季节发现，海拔3091m处原始林地5cm土壤温度显著最低(比与之相邻的火烧迹地群落林地土壤温度低2.41℃，比海拔3549m和2551m处林地土壤温度地分别低2.28℃和2.74℃)；生长季节则呈现出随海拔升高而降低。这说明在进行构建土壤呼吸模型时，应当考虑植被覆盖差异引发土壤温度变化的因素。综上所述，为较为精确的预测亚高山川滇高山栎群落和草甸土壤表面CO2通量，需加强非生长季节和夜间土壤呼吸测定。同时还应考虑除了昼、夜间土壤温度和土壤水分外，还应考虑土壤底物(主要是土壤活性有机碳、氮库)的供给状况及地形条件和它生态因子的综合影响。更多还原

【Abstract】 Soil respiration is the main source of carbon efflux from terrestrial ecosystems to theatmosphere, which indicates the rates of soil microbial activity and soil organic carbonmineralization. As a complicate ecological process, soil respiration is affected not only bybiological factors, but also by environmental factors. However considerable uncertainties existabout the mechanisms underlying the soil respiration in the present studies. So a goodunderstanding of temporal-spatial dynamic variation and driving factors of day and nocturnalsoil respiration will help us to better comprehend the mechanism of soil respiration variation.At present, studies on soil respiration were more concentrated in the subtropical mid-tolow-elevation areas and less in the subtropical high-altitude subalpine forest and meadow, andwithout knowledge of nocturnal soil respiration. Therefore, this study was designed to examinesoil respiration along with its controlling factors in sclerophyllous evergreen broad forest andmeadow at the southeastern margin of the Qinghai-Tibet Plateau in order to assess the regionalcarbon budget.The Quercus aquifolioides shrub with the eco-hydrological functions distributes in highaltitude areas. Subalpine meadow is one of important vegetation types in the western Sichuan.In this study we measured RS, soil temperature and water content with a LI-8100A SoilRespiration System in three altitudinal plots (3549m，3091m and2551m respectively) inBanlang Mountain in Wolong Nature Reserve from Sept.2010to Dec.2011and sampled soilthere and analyzed their contents of soil active organic carbon and soil nitrogen in order toexplore the seasonal variation of day-and night-time soil respiration in relation to theenvironmental factors in the Q.aquifolioides stand. The results indicate that:(1) Both RHand RRsignificantly and similarly varied with the shifts of seasons. Duringthe study period the mean rates of day-and night-time RHwere1.57±1.06(mean±Std.) and1.13±0.96μmol m-2s-1respectively, and the RRvalues were0.60±0.61and0.47±0.41μmolm-2s-1. The annual cumulative RHand RR(295.02and200.46gCm-2yr-1) in the day were obviously higher than those (125.43and121.20gCm-2yr-1) at night in the primary stand. Soiltemperature and soil water content are the main factors affecting RHand RRof the day and thenight. The annual cumulative RHand soil total respiration highly correlated with soil microbialbiomass nitrogen (MBN), soil labile organic carbon (LOC) and light fraction organic carbon(FLOC) at0~30cm soil depth; RRhighly correlated with FLOC but there was little correlationbetween RRand the fine root biomass (root diameter<0.5cm).(2) The temperature sensitivity values (Q10) of soil heterotrophic respiration and total soilrespiration in the daytime were lower than those at night. The annual Q10values of soilheterotrophic respiration and total soil respiration in the daytime measured by T5were3.77±0.25(mean±standard error) and4.96±0.45, and those at night were4.51±0.41and5.28±0.79, respectively.(3) Soil surface CO2effluxes in the different stands showed similar variation patternsbetween day and night. Namely, soil surface CO2effluxes were greater in the daytime and inthe growing seasons. Soil temperature, soil water content and their interaction were the majorenvironmental factors affecting soil surface CO2effluxes. The model constructed by thesefactors explained the variance of soil surface CO2effluxes (83.2%in the daytime and96.6%at night). The temperature sensitivity values (Q10) in the growing seasons were higher thanthose in the non-growing seasons and those at night were higher than those in the daytime.During the whole measurement period their day Q10ranged from4.35to4.48(mean4.44)and their nocturnal Q10ranged from5.37to6.42(mean4.44). However, we found their Q10values increased with the decrease of soil temperature, and reached their maximums in thenon-growing seasons and minimum in the growing seasons. The cumulative CO2effluxhighly correlated with MBN and LOC stocks (P<0.05) and slightly correlated with LFOC.The estimated annual soil surface CO2efflux values for the Q.aquifolioides stand ranged from371.08to951.55Cm-2yr-1(mean684.75Cm-2yr-1) in2011, and the cumulative flux in thegrowing seasons was as4.1times as that in the non-growing seasons, and the mean ratio ofnocturnal soil surface CO2efflux to annual flux was42.46%(ranging from39.82to44.60%). So if only daytime data are used to estimate the annual soil surface CO2efflux, the efflux willbe overestimated by12.96%in the non-growing seasons and by3.19%the growing seasonsrespectively. We conclude that biotic and abiotic factors influence soil respiration differentlyat the different time scales. Seasonal variations of soil CO2efflux strongly depends on soiltemperature, soil moisture and their interactions. The variances of soil CO2efflux betweenday and night maybe mainly result from quantity and quality of soil substrates. Therefore,apart from soil temperature, soil moisture and their synthetic effect and components of soilactive organic carbon, we should also take into consideration other abiotic factors andhighlight the measurement of the soil nocturnal CO2efflux so as to estimate daily and annualsoil surface CO2effluxes more accurately.(4) The RSand RHof the burned stand and the control Q.aquifolioides stand showedsimilar parabolic curve patterns, which were highest in the growing seasons and lowest in thenon-growing seasons. The RSand RHof the burned stand were obviously higher than those ofthe controlled stand. There was a significant exponential correlation between the soiltemperature and the RSand RHof the burned stand and the controlled stand during the wholemeasurement period whether in the growing seasons or in the non-growing seasons. However,there was a significant power function correlation between the soil water content and the RSand RHin the non-growing seasons and the whole measurement period, but no correlation inthe growing seasons. So the double-factor model of soil temperature and soil water contentcan be used to explain the variance of rhizospheric and heterotrophic respiration as timebeing. In the whole study period, the RSand RHvalues (Q10) of the burned stand were3.00and3.56, respectively, which were lower than those (4.23and5.33, respectively) of thecontrolled stand. Linear regression analyses suggested that soil light fraction organic carbon(LFOC), particulate organic carbon (POC) and nitrate nitrogen (NO3-N) were importantparameters in soil respiration studies. The annual RSand RHof the burned area were768.23and655.60gCm-2yr-1respectively, lower by13.57%and higher by20.59%than those of thecontrolled stand correspondingly. Hence, we conclude that it was the fire which interfered with the soil temperature and the supply of soil organic carbon fractions and soil nitrogen thataffected the RSand RH.(5) Soil rhizospheric respiration was reduced by plot girdling. In specific, it was loweredby about10.6%and16.5%in the growing seasons and the non-growing seasons, and byabout11.96%annually.(6) Soil respiration rate (RS) in the alpine meadow showed large seasonal variations inthe daytime and the nighttime in the measurements period, with a two-peak curve in thedaytime and parabola curve at the nighttime. In the whole period (April to November), the RSof the nighttime accounted for46.7%of the total; the RSin the growing seasons was muchgreater than that in the non-growing seasons. Compared with the soil respiration rates in theQ.aquifolioides stand at the same altitude, those in the alpine meadow were higher by4.7%and20.5%respectively in the daytime and at night. The average Q10values in the daytimeand night-time were3.90and3.74respectively. The results showed that in estimating theseasonal and annual effluxes of soil respiration we should intensify the measurement ofnocturnal soil respiration.Further, during experimental time it was observed that the monthly mean soiltemperature at5cm soil depth linearly dropped by0.22℃with the elevation of each100meters in the Q.aquifolioides stand of sunny slope at the altitudes between2551m and3549mfrom January to December2011. But it is very interesting that the soil temperature at5cm inthe Q.aquifolioides primary stand was the significantly lowest at the altitude of3091m in thenon-growing seasons, which was2.41℃,2.28℃and2.74℃lower than those in contiguousburned stand, at3549m and2551m primary stands respectively. However, the soiltemperature increased with the increase of altitude in the growing seasons (May throughOctober2011). The result implies that in constructing a general soil respiration model weshould take the forest type into account for vegetation differences can lead to the timedifferences.By all accounts, for an accurate estimation of seasonal and annual soil carbon fluxes in Q.aquifolioides stands and in subalpine meadows, measurements of the soil nocturnal CO2efflux and in the non-growing seasons were essential. At the same time, soil physicalenvironments and substrate availability should be taken into consideration in analyzing andinterpreting measurements of soil surface respiration and its components.更多还原