Mechanisms of Climate Variability in the North Atlantic Ocean

0136327/Hain为了理解年代际气候变率中可能涉及的一些海洋过程，本项目将对N.大西洋环流研究控制海面温度（SST）的机制。 1度模式的行为也将与1/3度和1/6度模式的前向积分进行比较，以研究不可预测的内部过程的重要性，如中尺度涡旋的输送。 伴随研究将用于确定SST对先前表面强迫的敏感性。 这项工作的基本假设之一是，表面强迫的变化可能导致温跃层的深度变化，随后通过斜压时间尺度上的海洋传播，并可能影响在随后的几年中，他们上面的SST。 这项工作将基于麻省理工学院开发的z坐标海洋环流模型。 该方法首先使用1度分辨率的模型进行1950-2000年的50年后投实验。 将通过比较模型SST与COADS SST来评估其保真度。 三组20年的伴随灵敏度计算，然后将使用的模型系统的线性化的基准轨迹给出的后投运行。 三个伴随计算将对应于NAO的不同阶段-低-60年代初，高-90年代初，接近零-70年代中期。 将检验其敏感性的变量是在海洋表面O（10，000平方公里）区域上空间平均的冬末SST。 模型计算将在JHU的Beowulf集群和一个国家超级计算中心进行。

0136327/HaineIn an effort to understand some of the oceanic processes that may be involved in decadal climate variability, this project will apply adjoint sensitivity analysis to 1-degree models of the N. Atlantic Ocean circulation to look at mechanisms that control sea-surface temperature (SST). The behavior of the 1-degree model will also be compared to forward integrations with 1/3-degree and 1/6-degree models in order to investigate the importance of unpredictable internal processes such as transport by mesoscale eddies. The adjoint study will be used to determine the sensitivity of SST to prior surface forcing. One of the hypotheses underlying this work is that changes to surface forcing can lead to changes in the depth of the thermocline which subsequently propagate through the ocean on baroclinic time scales and which may affect the SST above them in subsequent years. The work will be based on a z-coordinate ocean circulation model developed at MIT. The approach involves first conducting a 50-year hind-cast experiment for 1950-2000 with a 1-degree resolution version of the model. The fidelity of this will be assessed by comparing model SST with COADS SST. Three sets of 20-year long adjoint sensitivity calculations will then be made using the a linearization of the model system around the fiducial trajectory given by the hind-cast run. The three adjoint calculations will correspond to during different phases of the NAO -- low - early 60s, high - early 90s, near zero - mid-70s. The variable whose sensitivity will be examined will be the end-of-winter SST spatially averaged over O(10,000 Km^2) regions of the ocean surface. Model calculations will be performed on a Beowulf cluster at JHU and at one of the national supercomputing centers.