The Role of Coupled Versus Uncoupled Reanalyses for Initialization of Decadal Climate Predictions

各种社会经济部门对可靠气候预测的需求日益增长。巧合的是，研究气候系统在年际到年代际时间尺度上的可预测性的研究表明，利用现有的海洋和大气观测对地球系统模型进行初始化可以提高模型预测未来几年区域表面温度的能力。然而，十年预测被证明受到模型误差和初始冲击的影响，这限制了可预测性。目前尚不清楚这一影响的程度。为了消除初始冲击和改进十年预报，我们需要能够以动态一致的方式估计耦合初始条件的ESM。在各种耦合数据同化(CDA)方法中，伴随方法是最有前途的方法之一，因为它的结果与ESM的方程是动态一致的。因此，由伴随项产生的气候状态的动态自洽时空演变适合于年代际预报的初始化。为此，我们将使用为中等复杂性的CESAM(Centrum für Erdsystem forschung und Nachhaltigkeit Erdsystem Convermations-MODELL)的ESM开发的耦合伴随模式来产生海洋-大气耦合再分析，这将作为追溯年代际预报集合的初始条件来源。然后将这些模拟与现有的观测结果进行比较，以估计CESAM的预测技能。将在CESAM内实施的十年预测框架将侧重于耦合气候系统内作用的大范围过程和反馈的多年变异性。所提出的研究的优势在于伴随和正向CESAM的可用性和计算效率。由于为ESMS编制伴随代码不是一件容易的事情，它目前阻碍了基于伴随的同化方法在十年预报初始化中的广泛应用。因此，伴随CDA用于年代际预测的优点仍有待理解和论证；这是拟议研究项目的主要目标。因此，在模式一致的方法中，该项目将比较基于海洋-大气耦合再分析和基于推动非耦合再分析的十年预测研究中的普遍战略的初始化。该项目的结果将指导今后综合环境监测系统的初始化发展。

The need for reliable climate prediction is growing in demand for various socio-economic sectors. Coincidently, studies that investigate predictability of the climate system at interannual to decadal timescales demonstrated that the initialization of Earth System Models (ESMs) with the available oceanic and atmospheric observations can improve the ability of models to predict regional surface temperature up to several years ahead. However, the decadal predictions are shown to suffer from model errors and initialization shocks which limit predictability. The magnitude of this effect remains unclear. To eliminate initialization shocks and improve decadal predictions, we need ESMs that are able to estimate coupled initial conditions in a dynamically consistent way. The current project will be concerned with quantifying the impact of dynamically balanced initial conditions on the decadal prediction skill.Among a variety of coupled data assimilation (CDA) methods, the adjoint method is one of the most promising because its result is dynamically consistent with the ESM’s equations. Thus, a dynamically self-consistent space-time evolution of the climate state produced with the adjoint is suitable for initialization of decadal predictions. To this end, we will use the coupled adjoint model developed for the ESM of intermediate complexity CESAM (Centrum für Erdsystemforschung und Nachhaltigkeit Erdsystem Assimilations-Modell) to produce a coupled ocean-atmosphere reanalysis, which will serve as a source of initial conditions for ensembles of retrospective decadal predictions. These simulations will then be compared against the available observations to estimate the prediction skill of CESAM. A decadal prediction framework, which will be implemented within CESAM, will focus on the multiannual variability of large-scale processes and feedbacks acting within the coupled climate system. The advantage of the proposed research is the availability and the computational efficiency of the adjoint and forward CESAM. As programming adjoint codes for ESMs is not a trivial task, it is currently hindering widespread application of the adjoint-based assimilation for initialization of decadal predictions. Therefore, the merits of the adjoint CDA for decadal predictions are still to be understood and demonstrated; this is the major objective of the proposed research project. Thus, in a model-consistent approach, the project will compare initializations based on the coupled ocean-atmosphere reanalysis and based on the widespread strategy in decadal prediction studies of nudging toward the un-coupled reanalyses. Results of the project will serve to guideline future initialization developments for comprehensive ESMs.