Download
1 / 1
10 likes | 148 Vues
This project, led by Dennis McLaughlin at MIT, explores methodological issues in earth sciences through five distinct research clusters. Each cluster addresses critical challenges like nonlinearity and uncertainty, fostering collaboration among researchers from various disciplines. Key applications include dynamic image segmentation, multiscale data assimilation, and advanced variational methods. With a strong emphasis on multidisciplinary training, the initiative aims to integrate innovative data assimilation techniques to enhance predictive capabilities in the earth sciences. Broader impacts involve partnerships with institutions like the Boston Museum of Science.
E N D
An Ensemble Approach to Data Assimilation in the Earth SciencesITR- 0121182 PI: Dennis McLaughlin Massachusetts Institute of Technology Focus: Methodological issues that cut across earth science disciplines.Nonlinearity Dimensionality Uncertainty Structure: Five research clusters, each deals with a particular issue, brings together researchers from different disciplines, focuses on one or more applications Broader Impact: Multidisciplinary training, Boston Museum of Science, Alliance for Computational Earth Sciences, and Earth Systems Initiative 1) Dynamic Image Segmentation 2) Multiscale Data Assimilation Gulf Stream “field and boundary” estimation from sparse tracer-field measurements Best Student Paper Award AGU 2003 Samples conditioned on real-time Infrared (GOES) data Replicates from unconditional Model GOES IR 3) Advanced Variational Methods Multiscale ensemble filtering, with replicates conditioned on satellite cloud observations, is both realistic and efficient 4) Field Alignment Adjoint methods for global state estimation merge diverse data sources Representative Articles(2003-2004) 5) Assimilation for Chaotic Systems • Buehner, M. and P. Malanotte-Rizzoli, “Reduced-rank Kalman filters applied to an idealized model of the wind-driven circulation”, (accepted) Journal of Geophysical Research. • Hansen J.A. and K. A. Emanuel, “Forecast 4d-Var: Exploiting Model Output Statistics”, (accepted) Quarterly Journal of the Royal Meteorological Society. • Hansen, J.A., “Accounting for model error in ensemble-based state estimation and forecasting”, (accepted) Monthly Weather Review. • Lawson, W. G. and Hansen, J. A, “Implications of stochastic and deterministic filters as ensemble-based data assimilation methods in varying regimes of error growth”, (in review) Monthly Weather Review. • Naumann, U. and P. Heimbach, “Coupling tangent-linear and adjoint models” , vol. 2668, part II, pp. 105-114, V. Kumar, M. Gavrilova, C.J.K. Tan, P. L'Ecuyer (Ed.) Lecture Notes in Computer Science, Springer-Verlag, 2003. • S. Ravela, K. Emanuel and D. McLaughlin, “Data Assimilation by Field Alignment”, (in review) Monthly Weather Review. • W. Sun, M. Çetin, W. C. Thacker, T. M. Chin, A. S. Willsky, “Localization of Oceanic Fronts & Feature Boundaries Using a Variational Technique”, AGU 2003 (Best Student Paper Award). • Zang, X. and P.Malanotte-Rizzoli, "Practical Implementation of the Ensemble Kalman filter for a realistic Primitive Equation model", (in review) Monthly Weather Review. Variational adjustment of displacements at grid nodes compensates for position errors in hurricane forecats Ensemble Kalman filtering with realistic ocean models. Topex/Poseidon altimetry reduces errors in sea-surface height anomalies by over 40%
