The impact of Madden-Julian Oscillation on polar surface air temperature
Open Access
- Author:
- Yoo, Changhyun
- Graduate Program:
- Meteorology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- October 19, 2011
- Committee Members:
- Sukyoung Lee, Dissertation Advisor/Co-Advisor
Sukyoung Lee, Committee Chair/Co-Chair
Steven B Feldstein, Committee Member
Peter R Bannon, Committee Member
Diane Marie Henderson, Committee Member - Keywords:
- polar amplification
Rossby wave
MJO - Abstract:
- This dissertation investigates the impact of the Madden-Julian Oscillation (MJO) on the extratropical surface air temperature (SAT). The underlying idea of this study is that tropical convective heating is linked to the extratropical circulation and SAT through poleward propagating Rossby waves. Given this perspective, two questions are addressed. One is the extent to which the MJO contribute to interdecadal time scale polar amplification of SAT (Chapters 2 and 3) and the other is the mechanism by which the MJO alters the intraseasonal time scale extratropical SAT change (Chapters 4 and 5). Polar amplification, i.e., interdecadal time-scale SAT increase being greatest at high latitudes, is one of prominent features of current climate change. Numerous observational and modeling studies have documented this phenomenon. However, its cause remains uncertain. The surface albedo feedback, which is associated with retreats in snow and ice cover and hence increases in surface albedo with a warmer climate, is the most prominent explanation. However, one of important characteristics of polar amplification is that its maximum amplitude is retained during the winter season when incoming solar radiation is minimal. In Chapters 2 and 3, we show that polar amplification during the 1979-2008 winters is tropically excited. More specifically, we show evidence that polar amplification is linked to interdecadal time scale change in the MJO phase frequency of occurrence. We present both the extended boreal winter (November to March) and austral winter (May to September), with our focus being on winter Hemisphere, where the MJO has strong influence. First, during the 30-year boreal winter, MJO phases 4-6 have occurred with an increased frequency of occurrence while phases 1 and 2 have showed a moderate decrease in their frequency of occurrence. Using lagged composites of the SAT, we show that Arctic warming takes place 1-2 weeks after MJO passes its phases 4-6. Similarly, MJO phases 1 and 2 are 1-2 weeks later followed by Arctic cooling. By incorporating both the trend in MJO phase and the intraseasonal SAT anomaly associated with the MJO, it was found that the MJO-induced SAT trend accounts for 10-20% of the observed Arctic amplification over the Arctic Ocean. It is also presented that the Arctic SAT change is closely related to the spatial structure of the tropical heating associated with the MJO. Phases 4-6 (1-2) are associated with more zonally localized (uniform) tropical convection. The relationship between the Arctic SAT and the spatial structure of the tropical heating is consistent with Lee et al. (2011a; 2011b), who suggest that more localized (uniform) convection enhances (reduces) excitation of poleward propagating Rossby waves that transport heat poleward. A similar relationship between the MJO and polar amplification is found in the Southern Hemisphere. During the 30-year austral winter, MJO phases 7-1, which is associated with high latitude warming, exhibited a large increase in their frequency of occurrence, while phase 5, which is followed by high latitude cooling, showed a decrease. The interdecadal SAT trend induced by the MJO, which is obtained using the trend in MJO phase and the intraseasonal SAT anomaly associated with the MJO, explains 10-20% of the interdecadal warming over the Antarctica. It is also shown that the spatial structure of the tropical heating associated with MJO phases 7-1 (phase 5) is more zonally localized (uniform). To investigate the physical mechanism that accounts for the linkage between tropical convection, Rossby wave propagation, and polar amplification, we use lagged composites and projections with the thermodynamic energy equation. We show that the Arctic SAT increase (decrease) associated with MJO phase 5 (phase 1) takes place through following a three-step processes; adiabatic warming (cooling) is followed by eddy heat flux convergence (divergence) and then by enhanced (reduced) downward infrared radiation (IR). These three above steps are closely related to the enhanced (reduced) poleward Rossby wave activity associated with more zonally localized (uniform) tropical heating. For instance, the enhanced poleward Rossby wave propagation associated with MJO phase 5 leads to an increased eddy momentum flux convergence at the equator and divergence over the Arctic. At high latitudes, this eddy momentum flux divergence results in a deceleration of the zonal-mean zonal wind. Through thermal wind adjustment, these zonal wind changes further cause inducement of a thermally direct mean meridional circulation that warms the Arctic. It is the opposite for MJO phase 1. Then, an increased (decreased) poleward heat flux associated with the enhanced (reduced) poleward activity propagation contributes to Arctic warming (cooling). The eddy heat flux is dominated by zonal wavenumbers 1-3. The Arctic SAT change is further amplified by changes in downward IR. To further ascertain whether the above processes are driven by the tropical convective heating anomalies associated with a particular MJO phase, we employ an initial value approach with an idealized atmospheric general circulation model. It is shown that initial value calculations with observed background flows and MJO-based tropical heating successfully reproduce the above processes. Analogous to the observation, the more zonally localized (uniform) tropical heating associated with MJO phase 5 (phase 1) strengthens (weakens) the excitation of poleward propagating Rossby waves. It is shown that the poleward propagating Rossby waves contribute to alteration of extratropical circulation and, hence, SAT, through adiabatic warming/cooling and planetary-scale eddy heat flux. In addition, it is presented that poleward (equatorward) passive tracer transport takes place in the upper polar troposphere, in response to more zonally local (uniform) tropical heating, which contributes to the observed changes in downward IR associated with MJO phase 5 (phase 1). We lastly present sensitivity test to the initial background flow, which further verifies the causal relationship between tropical heating and planetary-scale response, and also indicates considerable variability due to interaction with pre-existing transient eddies.