【作者】 王进；

【导师】 王绍明；

【作者基本信息】 石河子大学 ， 作物栽培学与耕作学， 2017， 博士

【摘要】 人类活动排放的以CO2为主的温室气体增加导致全球变暖,全球公认,事实确凿,减少温室气体排放是减缓气候变暖的关键途径。海洋是公认的碳汇系统,但陆地生态系统的碳源汇功能十分复杂,存在很大的不确定,这其中农田生态系统扮演重要角色;从研究的热点和重点分析,对干旱区农业生态系统的碳源汇功能重视程度不够:新疆是干旱区的典型代表,自1950s以来,进行了大规模的土地开发,绿洲农业规模不断扩大,先进的绿洲灌溉技术普及以及其他绿洲农业管理集约化程度不断提高,显著改变了绿洲系统碳源汇功能,但这方面的研究还较为薄弱。本文以新疆绿洲农业的典型代表--天山北坡玛纳斯河流域绿洲为研究区,通过长期定点观测和生物地球化学循环模型,定量评估绿洲农田生态系统碳动态近40年来的时空格局及其对绿洲扩张、绿洲管理和气候波动的响应旨在探索干旱区绿洲农田生态系统碳收支时空格局和动态特征及其对农田管理措施的响应机制。本文首先在站点尺度上,基于安装在农田生态系统的涡度相关系统的观测数据、土壤呼吸和农气观测数据（2009-2013年）,分析覆膜滴灌棉田的碳通量变化特征和不同种植模式下棉田产量和耗水量的时间变化特征,揭示棉田碳通量与气象因子的响应关系及种植模式优化对农田水分利用效率和灌溉效率的影响。然后基于这些站点数据校验生物地球化学循环模型BIOME-BGC,在站点尺度上利用农业气象观测数据和产量统计数据验证模型,模拟流域绿洲农业集约化管理,对农田生态系统碳收支的影响。最后利用基于空间化的绿洲农业BIOME-BGC模型,在1 km×1 km空间分辨率网格尺度上,估算玛河流域1970-2013年绿洲农田碳收支时空分布变化,并对气候变化、绿洲扩张和人类集约化管理（施肥、灌溉）,开展不同情景的模拟和预测分析,取得的主要结果和认识包括以下几方面。1.基于涡度相关系统,研究了覆膜沟灌（MF）和膜下滴灌（MD）种植模式下绿洲棉田碳循环过程和水分利用特征MF和MD的播种、出苗和现蕾期较裸地沟灌（NF）种植时期提前了13-20 d,且整个生育期较NF时期分别延长了12-16 d。MD棉田水分利用效率和灌溉利用效率较NF（裸地沟灌）时期的分别提高了65.2%和62.3%。覆膜滴灌棉田生育期内（5-10月）GPP和NEE的昼夜变化都呈现显著的正弦曲线变化特征,峰值出现在14:00;GPP和NEE的季节变化与LAI变化较为一致,峰值出现在7月;月积累的GPP在6-9月主要分配给NEE,而其余月份则分配给R_es;整个生育期内,总GPP为816.2 gC m^-2,其中约58.6%分配给了NEE(总NEE为-478.6 gC m^-2),表现为强碳汇。2.利用土壤呼吸系统,识别了不同秸秆管理和施肥方式下农田土壤呼吸速率变化,其在作物生长季表现出明显的季节变化规律,季节变化峰值出现在7月中旬,最小值出现在冬季（1-2月）。秸秆还田下棉田土壤呼吸速率显著大于秸秆不还田处理;有机无机肥配施棉田土壤呼吸速率最大,单施有机肥土壤呼吸速率次之,单施化肥处理土壤呼吸速率最小。土壤呼吸速率与10 cm地温呈极显著正相关。说明干旱区农田土壤呼吸速率受土壤温度影响很大。由于干旱区棉田采用膜下滴灌,土壤含水量变化不大,对土壤呼吸速率没有显著影响。3.基于耦合了施肥和灌溉模块的BIOME-BGC模型,分析和模拟了乌兰乌苏农业气象试验站的LAI、NPP和NEE数据。冬小麦、棉花和玉米LAI的模拟值与观测值的相关系数R²分别为0.84、0.84和0.73,均方根误差（RMSE）分别为0.75、1.05和1.33;冬小麦和玉米净初级生产力（NPP）的模拟值与观测值的相关系数R²分别为0.76和0.80,均方根误差（RMSE）分别为39.46和25.33。棉田净碳交换量（NEE）的模拟值与观测值的相关系数R²和均方根误差（RMSE）分别为0.75和2.78。这表明改进后的BIOME-BGC模型可较为准确的模拟玛河流域绿洲区冬小麦、棉花和玉米作物生长过程和净初级生产力。4.利用空间化BIOME-BGC模型,对玛河流域绿洲区农田生态系统净初级生产力进行模拟和估算。1971-2013年玛河流域绿洲区农作物NPP总量呈显著增加趋势,从1970年的6.72 PgC m^-1增加到2010年的12.36 PgC m^-1,其中棉花的贡献比例由1970年的42%增加到2010年的89%。绿洲扩张有效增加了玛河流域耕地面积,棉花播种面积也由1970年的72.09 Kha增加到2010年的152.25 Kha,因此该区域棉花对CO₂的固定比冬小麦和玉米生态系统要大,由1970年的2.88 PgC m^-1增加到2010年的12.36 PgC m^-1,增加了4倍。5.基于BIOME-BGC模型多情景模拟,玛河流域绿洲农田NPP的增加主要来源于农业植被面积的增加（贡献率为49%）,其次为农业管理措施（贡献率分别为30%）,而气候变化贡献较低（贡献率为5%）。因此,近40年来,玛河流域绿洲扩张和农业管理是NPP大幅度增加的主要因素。在当前禁止大规模水土开发的情况下,改善农田管理措施是提高绿洲NPP和增加农田生态系统固碳能力,和减少农田生态系统CO₂的排放量主要途径。更多还原

【Abstract】 This Global warming is the result of increased CO2 emissions from anthropogenic greenhouse gases,and the global consensus is that reducing greenhouse gas emissions is a key way to mitigate climate change.The ocean is recognized as a carbon sink system,but the carbon sink function of terrestrial ecosystem is very complex,and there is a large uncertainty.The farmland ecosystem plays an important role in carbon sink function.From the research hotspot and the analysis,there is not enough emphasis on the carbon sink function of agricultural ecosystem in arid area.Xinjiang is a typical representative of the arid areas,and there was carried on the massive land development,oasis agriculture scale expands unceasingly,an oasis of advanced irrigation technology popularization and other oasis agricultural management intensive degree enhances unceasingly since 1950s,which significantly changed the carbon sink function of oasis system.This study aimed to explore tempo-spatial patterns of carbon budgets and the mechanisms of its responses to the change of agricultural management.By using real-time monitoring and biogeochemical cycle model,we studied the change of tempo-spatial patterns of carbon budgets and its responses to the oasis expansion and climate fluctuations in field ecosystem in arid areas.Firstly,to reveal the response relationship between cotton carbon flux and the meteorological factors and planting pattern optimization on the farmland water use efficiency and irrigation efficiency,we studied the carbon flux variation under drip irrigation cotton field and cotton yield and water consumption under different planting patterns,which based on the eddy covariance and soil respiration data（2009-2013）.Secondly,using these site data validation Biome-BGC,and simulate the effect of agricultural management on the farmland ecosystem carbon balance.Finally,estimated the tempo-spatial patterns of carbon balance from 1970 to 2013 by BIOME-BGC model on 1km×1km grid scale.And carried out the simulation and prediction analysis of different scenarios which aimed on climate change,oasis expansion and agricultural management（fertilization,irrigation）.The results suggested that:1.Based on the eddy correlation system,we studied the carbon and water cycles under different cropping pattern.The results showed that:sowing,emergence and squaring dates were 13-20 d in advance in furrow irrigation under mulching（MF）and drip irrigation under mulching（MD）compared to furrow irrigation without mulching（NF）,and the whole life dates was extended for 12-16 d.The water use efficiency and irrigation water use efficiency were increased 65.2%and 62.3%in MD compared to NF.The gross primary production（GPP）and net ecosystem exchange（NEE）showed an obviously sinusoidal dynamic between daytime and nighttime,and the peak value occurred in 14:00.The seasonal variations of GPP and NEE were followed changes in LAI,and the peak value occurred in July.The GPP was assigned primarily to NEE from June to September and R_es in other months.The cumulative GPP averaged 816.2 gC m^-2 during the growing season,and about 58.6%of GPP was partitioned into NEE(meanly-478.6 gC m^-2)which showed a carbon sink.2.The soil respiration rate was measured by using soil respiration system under different straw management and fertilization treatments.The results showed that soil respiration rate has a significant seasonal characteristic,and the peak value occurred in mid-July and the minimum in January-February.The residue incorporation significantly increased soil respiration.The soil respiration rate under different fertilization treatments were shown as NPK+OM>OM>NPK>CK.There was a significant correlation between soil respiration and soil temperature,indicated that soil respiration was effected primarily by soil temperature.There was no significant seasonal variation under drip irrigation with mulching,resulted in there was no significant correlation between soil respiration and soil moisture.3.After irrigation and fertilization sub-model of BIOME-BGC model.It used the LAI,NPP and NEE data in Wulanwusu agrometeorological station.The results showed that the correlation coefficients between observed and simulated data of winter wheat,cotton and corn were 0.84,0.84 and 0.73,and root mean square errors were 0.75,1.05 and 1.33,respectively.The correlation coefficients between observed and simulated data of winter wheat and corn were 0.76 and 0.80,and root mean square errors were 39.46 and 25.33,respectively.The correlation coefficients and root mean square error between observed and simulated NEE data of cotton were 0.75 and 2.78,respectively.The results showed that the BIOME-BGC sub-model could simulated growth and NPP of winter wheat,cotton and corn.4.Simulated and estimated the NPP of oasis field ecosystem by using spatial Biome-BGC model.The total NPP in Manas River during 1970-2010 showed a trend of a significant increase,from 6.72 PgC m^-1 in 1970 increased to 12.36 PgC m^-1 in 2010,in which the contribution of cotton increased from 42%in 1970 to 89%in 2010.Oasis expansion effectively increased arable land,the cotton area was increased from 72.09 Kha in 1970 to152.25 Kha in 2010.Therefore,the CO₂ fixation of cotton was increased from 2.88 PgC m^-1in 1970 to 12.36 PgC m^-1 in 2010（about 4 times）,which was bigger than wheat or maize in this region.5.Comprehensive analysis the simulate NPP data by BIOME-BGC.The results showed that the contribution of the increased arable land was first,the agricultural management was second and climate change was third to NPP.Therefore,oasis expansion is the major factor for increasing NPP in nearly 40 years.Now,the large land reclamation is banned in Xinjiang,but improving the agricultural management can effectively increase the oasis farmland ecosystem carbon sequestration and reduce CO₂ emission.更多还原