Highly-Scalable Algorithms for Ensemble Data Assimilation
نویسندگان
چکیده
A complete ensemble data assimilation algorithm can be composed into three basic computational tasks: advancing an ensemble of model forecasts, computing ensembles of forward operators for available observations, and assimilating the observations to modify the ensemble of model states. Each of these tasks requires a fundamentally different pattern of communication when implemented on current generation supercomputers.
منابع مشابه
A Hybrid Particle-Ensemble Kalman Filter for Lagrangian Data Assimilation
Lagrangian measurements from passive ocean instruments provide a useful source of data for estimating and forecasting the ocean’s state (velocity field, salinity field, etc). However, trajectories from these instruments are often highly nonlinear, leading to difficulties with widely-used data assimilation algorithms such as the ensemble Kalman filter (EnKF). Additionally, the velocity field is ...
متن کاملSoftware for ensemble-based data assimilation systems - Implementation strategies and scalability
Data assimilation algorithms combine a numerical model with observations in a quantitative way. For an optimal combination either variational minimization algorithms or ensemble-based estimation methods are applied. The computations of a data assimilation application are usually far more costly than a pure model integration. To cope with the large computational costs, a good scalability of the ...
متن کاملEnhanced Predictions of Tides and Surges through Data Assimilation (TECHNICAL NOTE)
The regional waters in Singapore Strait are characterized by complex hydrodynamic phenomena as a result of the combined effect of three large water bodies viz. the South China Sea, the Andaman Sea, and the Java Sea. This leads to anomalies in water levels and generates residual currents. Numerical hydrodynamic models are generally used for predicting water levels in the ocean and seas. But thei...
متن کاملAn Ensemble-Based Smoother with Retrospectively Updated Weights for Highly Nonlinear Systems
Monte Carlo computational methods have been introduced into data assimilation for nonlinear systems in order to alleviate the computational burden of updating and propagating the full probability distribution. By propagating an ensemble of representative states, algorithms like the ensemble Kalman filter (EnKF) and the resampled particle filter (RPF) rely on the existing modeling infrastructure...
متن کاملParallel Climate Data Assimilation PSAS Package Achieves 18 GFLOPS on 512-node Intel Paragon
We have designed and implemented a set of highly efficient and highly scalable algorithms for an unstructured computational package, the PSAS data assimilation package, as demonstrated by detailed performance analysis of systematic runs on up to 512-node Intel Paragon. The preconditioned Conjugate Gradient solver achieves a sustained 18 Gflops performance. As the results, we achieved an unprece...
متن کامل