Westerly wind bursts (WWBs) are associated with the onset of major El Nino events. They occur more frequently and energetically and extend farther eastward prior to and during significant El Nino events than in non-El Nino conditions. While these wind events were initially thought to be completely stochastic, more recent analyses indicate that their occurrence and characteristics are at least partially modulated by the large-scale sea-surface temperature (SST). Model studies show that this modulation affects the characteristics of El Ninos as compared to the case of purely stochastic WWBs.
This project is divided into two parts. The first, more applied, component is the implementation of an empirical WWB parameterization into a state-of-the-art El Nino-Southern Oscillation (ENSO) prediction model, in an effort to improve the skill of ENSO prediction and to assess the impact of parameterized WWBs on predictability. A previously developed empirical semi-stochastic observationally-motivated WWB model will be applied in the NCAR CCSM3.0 (National Center for Atmospheric Research Community Climate System Model 3.0) coupled ocean-atmosphere model, and this model will be used as an ENSO prediction model. The mutual effects of the WWB model and the coupled ENSO model on each other will be studied, along with the effects of the WWBs on the ENSO simulation, on the predictability of ENSO, and on the spread of predictions in the model.
The second part is an effort to understand the basic mechanisms of WWBs and the dynamics of their dependence on the large-scale SST field. A cloud-resolving atmospheric model capable of producing WWBs will be run in a near-global configuration using the diabatic acceleration and rescaling (DARE) approach. The DARE model will be supplemented with analyses of observed wind and outgoing longwave radiation and with several simple models. It is intended that this strategy will result in an understanding of how the SST controls WWBs and will expose the mechanisms for the initiation and termination of the convective activity that creates WWBs.
The broader impacts of this project are in contributing to improved skill in predicting ENSO, with the resulting social and economic benefits. Two graduate students will be trained: one at Harvard and one at Miami.
We have developed an empirical-statistical Westerly Wind Burst (WWB) model and tested it against scatterometer observations (Gebbie and Tziperman, 2009a). We found the empirical-statistical WWB model was very successful in hindcasting observed the WWB events of the past 20 years or so given the observed SST. The results of the effect of this WWB model on ENSO prediction was mixed, with some successes in improving the prediction of the large 1997 event. We also tested the WWB model in an intermediate ENSO prediction model (Gebbie and Tziperman, 2009b) and in a state-of-the-art coupled ocean-atmosphere ENSO prediction (Lopez, Kirtman, Tziperman, Gebbie, 2012). We completed the study of WISHE dynamics in a simple shallow water model, as outlined in the proposal, obtaining some very interesting findings as described below (MA thesis by Solodoch 2010, Solodoch et al, 2001). The analysis of the WISHE-driven shallow water model resulted in an interesting slowly eastward-propagating signal somewhat reminiscent of the MJO. The analysis revealed that the signal is due to the forcing of slow kelvin wave by atmospheric heating caused by the WISHE effects of the wind induced by Yanai waves. We plan to next examine the relevance of these idealized results to MJO in more realistic models (MA thesis by Solodoch 2010, Solodoch et al 2011, JAS). In Collaboration with Doug MacMynowski (MacMynowski and Tziperman, 2010, 2011), we examined ENSO dynamics using TAO array observations, GFDL coupled model, and the Cane-Zebiak intermediate complexity model, using tools from control engineering, in particular "transfer functions". With graduate student Nathan Arnold we have been looking at the excitation of a "permanent El Nino" during the Pliocene by MJO/WWB convective noise at the equator, via the excitation of poleward propagating Rossby waves and the forcing of superrotation (westerlies) at the equator. This work uses a hierarchy of modeling tools, from the NCAR AGCM (CAM) in idealized aqua planet configuration, the same model in a realistic configuration, and all the way to the super parameterized CAM (SPCAM). (Arnold, Farrell and Tziperman, 2012). we find that it is easy to excite superrotation in the upper atmosphere, but more difficult to weaken the surface easterlies, a step that's needed to explain the Pliocene permanent El Nino. We identified an interesting positive feedback/ resonance mechanism between excited poleward propagating Rossby waves and the developing superrotation jet which leads to an abrupt transition to superrotation above some threshold of the imposed stochastic forcing representing convective variability (Arnold, Farrell, Tziperman 2012). We are now in the process of using SPCAM to examine whether the convective noise is expected to change amplitude or form as a result of warmer SST expected during a Pliocene permanent El Nino. We also studied the large-scale organization of convection that gives rise to the WWBs using state-of-the-art models. We have analyzed the moist static energy (MSE) budget of MJO-like disturbances in idealized SPCAM simulations (Andersen and Kuang 2012). we have found that MJO-like modes can emerge and reduced longwave cooling in enhanced convection region is the dominant term that sustains this mode and modulation of transient eddy activity by the MJO flow is important for its eastward propagation. We have also systematically simplified the setup in SPCAM to bridge the full SPCAM model simulation and the simple model of Andersen and Kuang (2008), and extended the simple model to include frictional convergence, column MSE sources, horizontal advection moisture. We have also elucidated the cause of the red noise background spectra of tropical convection. Two papers are being prepared for publication based on the above results. We have also used idealized simulations with a cloud system resolving model (CSRM) to show that the gross moist stability, a key concept in tropical meteorology, has a wavelength dependence, which can potentially explain the planetary scale selection of the Madden-Julian Oscillation (Kuang 2011). We have also done comparisons of the different approaches of using a limited domain model with parameterized large-scale dynamics to study the interaction between convection and the large-scale flow, and found that a commonly approach has inherent biases in modeling planetary scale motions because it neglects the varying wave speeds for the different vertical wavenumbers, while the approach of Kuang (2008) works considerably better (Kuang 2012). This grant has also supported the study of the mechanisms of poleward-propagating, intraseasonal convective anomalies (Boos and Kuang 2011), a way of describing convective transport using the transilient matrix (Romps and Kuang, 2011), a refined study of the effect the Tibetan Plateau on the South Asian Monsoon (Ma and Kuang, in preparation), and further development of a stochastic tracer model and a Lagrangian Particle Dispersion Model to study cumulus entrainment (Nie and Kuang, in press). The project also helped to train four graduate students and two research associates.
