A Stochastic ENSO Model for State Estimation, Uncertainty Quantification and Prediction
Abstract
El Niño Southern Oscillation (ENSO) is the largest inter-annual variability phenomenon in the tropical Pacific and its influence goes beyond tropics to higher latitudes via atmospheric and oceanic teleconnections; therefore, it has a significant impact on global climate predictions. Inspired by the recharge-discharge model, we have developed a stochastic ENSO model including Westerly Wind Bursting (WWB) through a stochastic differential equation that includes multiplicative noise, which reproduces the main observed non-Gaussian features from intra-seasonal to inter-annual time-scales. The model embodies the underlying physics of tropical air-sea interactions that drives inter-annual variability and thus describes the large-scale ENSO dynamical and statistical characteristics. Moreover, by introducing time-dependent coefficients (non-autonomous structure), we generate the seasonality that results in an accurate recovery of the "phase locking" of ENSO to the seasonal cycle and the "spring predictability barrier", and hence is of importance for reliable forecasts. We examine the sea surface temperature data using a Bayesian technique and thereby have a state and path-wise estimation procedure for the thermocline depth and WWB, which are not easily observed but are nonetheless strongly coupled to the sea surface temperature and hence play a crucial role in its future evolution. The special structure of our model facilitates an efficient data assimilation scheme that advances seasonal and inter-annual ENSO forecasts. Finally, the unique feature of the non-Gaussianity allows for uncertainty quantification and examination of the probability of the occurrence of extreme and rare ENSO events.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMNG0020005G
- Keywords:
-
- 3315 Data assimilation;
- ATMOSPHERIC PROCESSES;
- 3238 Prediction;
- MATHEMATICAL GEOPHYSICS;
- 3260 Inverse theory;
- MATHEMATICAL GEOPHYSICS;
- 3275 Uncertainty quantification;
- MATHEMATICAL GEOPHYSICS