华北平原是我国重要的粮食产区，有限的水资源量和大量灌溉用水导致水资源供需矛盾突出。气候变化倍受人们关注，定量把握水分及碳循环的特点是水资源合理配置和CO2减排的基础。陆气间的水分-能量-碳通量交换是该地区农田生态系统中生态水文的主导过程。因此，本论文从通量观测入手，分析通量交换过程的内在机理，并以此建立农田生态水文模型，用于生态水文过程的重建和预测。论文以位山引黄灌区作为典型区域，并选择典型农田开展了以“水分-能量-碳”通量为核心的农田生态水文过程观测。不同仪器测量结果的对比表明，通量观测具有较高精度。数据分析表明，蒸散发（ET）年际变异性的主导因素是平衡蒸发，而季节变化的主导因素有平衡蒸发和冠层导度。小麦和玉米生长盛期内，潜热通量占净辐射的比例分别为83%和57%，这一比例的全年平均值为59%。生态系统呼吸（Reco）对温度的响应以及净生态系统交换量（NEE）对太阳辐射的响应均表现出与作物物候密切相关的季节变化特点。Reco的季节变化主要与总初级生产力（GPP）及气温相关，而Reco和GPP的年际变化主要受温度控制。全年来看，该农田为较强的碳汇；考虑作物收割所引起的CO2释放后，可转变成弱碳源。基于陆面过程模型SiB2构建了田间尺度水文强化陆面过程模型（HELP）。将HELP与作物生长模型耦合，构建了田间尺度生态水文模型（HELP-C），实现了对作物生长过程的模拟。验证表明，HELP和HELP-C是水分、能量、碳通量模拟的有效工具。将分布式水文模型与HELP（或HELP-C）耦合，构建了灌区尺度分布式生态水文模型。基于遥感归一化植被指数，该模型可用于灌区过去水分及碳循环过程的模拟；基于未来气候情景，模型可用于预测灌区未来水分及碳循环的可能变化。针对1984~2006年的模拟结果表明，尽管位山灌区降雨的年际变异性较大，但因灌溉的保障，ET的年际变异性较小，且无显著的长期变化趋势。小麦和玉米生长季内，多年平均ET分别为284和314 mm，分别占降雨和灌溉总水量的99%和71%。灌区农田年净初级生产力（NPP）、土壤呼吸（Rs）及NEE均无显著的长期变化趋势，但净生物群系生产量显著升高，且小麦NPP显著增长，玉米NPP显著下降。基于未来气候情景的预测结果表明，未来50年内，在充分灌溉条件下，小麦和玉米的ET将分别下降10%和7%，但灌溉需水量变化分别是+1%和-22%；NPP分别上升11%和3%，但Rs变化较小；小麦和玉米产量变化分别为+21%和-3%。
The North China Plain is one of the main crop-producing regions in China. However, limited water resources and large amount of irrigation water use have led to serious water shortage. As the climate change is being much more concerned, quantitative investigation of the water and carbon cycles is essential for efficient water management and CO2 sequestration in this region. The coupled water-energy-carbon flux exchanges between land and atmosphere are the dominant ecohydrological processes in agroecosystems of this region. Therefore, this thesis focused on the variability of these dominant ecohydrological processes though a long-term flux observation, and then attempted to develop an ecohydrological model for exploring variations in the ecohydrological processes in the past and projecting changes of these ecohydroogical processes in the future.The Weishan Irrigation District, a typical irrigated area in the North China Plain, was selected as the study area. A flux tower was set up in a typical cropland in this irrigated area, aiming to observe the water, energy, and carbon fluxes and other associated hydro-meteorological elements. The inter-comparison of fluxes observed by different instruments showed that the accuracy of the flux observations was acceptably high. Analysis of the observed data showed that, the interannual variability of evapotranspiration (ET) was controlled by the equilibrium evaporation, while the intra-annual variability of ET was controlled by both of the equilibrium evaporation and the canopy conductance. During the peak growth seasons of winter wheat and summer maize, the ratio of latent heat flux to net radiation was 83% and 57%, respectively. Annual average value of this ratio was 59%. In seasonality, the temperature response of ecosystem respiration (Reco) and the light response of net ecosystem exchange (NEE) corresponded closely to the crop phenology. The seasonal variations in Reco were primarily controlled by both of the gross primary production (GPP) and air temperature, while the interannual variations in Reco and GPP were primarily controlled by air temperature. Annually, this cropland was a strong carbon sink, but could be possibly converted into a weak carbon source when the grain harvest was taken into account.A Hydrologically-Enhanced Land Process model (HELP) was developed based on the Simple Biosphere Model 2 (SiB2). A field-scale ecohydrological model (HELP coupled with Crop growth model, HELP-C) was then developed through coupling the HELP with a crop growth model, which realized the modeling of crop growth. Validations by the observed data showed that the HELP and HELP-C were effective tools for simulating water, energy, and carbon fluxes. Finally, a regional-scale ecohydrological model was developed through coupling the HELP (or HELP-C) with a distributed hydrological model. This regional-scale ecohydrological model can be used for simulating the historical variations in water and carbon cycles, based on remotely sensed vegetation index. Optionally, this model can be used for projecting the possible changes of water and carbon cycles based on the future climate scenarios.The model simulations during 1984~2006 showed that the interannual variability of ET was small due to the sufficient irrigation even though the interannual variability of precipitation was high. Likewise, the annual ET had no significant long-term trend. The accumulated ET in the wheat and maize seasons were 284 and 314 mm, respectively, corresponding to 99% and 71% of the total amount of precipitation and irrigation, respectively. The annual net primary production (NPP), soil respiration (Rs), and NEE showed no significant long-term trend. However, the net biome production showed significant increasing trend, and the accumulated NPP in the wheat and maize seasons showed significant increasing and decreasing trends, respectively. Projection under the future climate scenarios (2010~2049) and the sufficient irrigation condition showed that the ET in the wheat and maize seasons would decrease by 10% and 7%, respectively. However, the irrigation water requirement of wheat and maize would change by +1% and -22%, respectively. The NPP in the wheat and maize seasons would increase by 11% and 3%, respectively, but Rs in both crop seasons would change very slightly. The wheat and maize yields would change by +21% and -3%, respectively.
