Efficient Simulation of Tropical Cyclone Pathways with Stochastic Perturbations
Abstract
Global Climate Models (GCMs) are known to statistically underpredict intense tropical cyclones (TCs) because they fail to capture the rapid intensification and high wind speeds characteristic of the most destructive TCs. Stochastic parametrization schemes have the potential to improve the accuracy of GCMs. However, current analysis of these schemes through direct sampling is limited by the computational expense of simulating a rare weather event at fine spatial gridding. The present work introduces a stochastically perturbed parametrization tendency (SPPT) scheme to increase simulated intensity of TCs. We adapt the Weighted Ensemble algorithm to simulate the distribution of TCs at a fraction of the computational effort required in direct sampling. We illustrate the efficiency of the SPPT scheme by comparing simulations at different spatial resolutions and stochastic parameter regimes. Stochastic parametrization and rare event sampling strategies have great potential to improve TC prediction and aid understanding of tropical cyclogenesis. Since rising sea surface temperatures are postulated to increase the intensity of TCs, these strategies can also improve predictions about climate change-related weather patterns. The rare event sampling strategies used in the current work are not only a novel tool for studying TCs, but they may also be applied to sampling any range of extreme weather events.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2017
- Bibcode:
- 2017AGUFMNG34A..07W
- Keywords:
-
- 3337 Global climate models;
- ATMOSPHERIC PROCESSES;
- 3367 Theoretical modeling;
- ATMOSPHERIC PROCESSES;
- 3265 Stochastic processes;
- MATHEMATICAL GEOPHYSICS;
- 4532 General circulation;
- OCEANOGRAPHY: PHYSICAL