Editorial Type:
Article
Article Type:
Research Article

Madden–Julian Oscillation: Bridging Weather and Climate

Chidong Zhang RSMAS, University of Miami, Miami, Florida

Search for other papers by Chidong Zhang in
Current site
Google Scholar
PubMed Close
Online Publication:
01 Dec 2013
Print Publication:
01 Dec 2013
DOI:
https://doi.org/10.1175/BAMS-D-12-00026.1
Page(s):
1849–1870
Full access

The Madden–Julian oscillation exerts broad influences on global weather and climate as its center of convection moves from the tropical Indian Ocean into the Pacific. Weather events under the influence of the MJO include precipitation, surface temperature, tropical cyclones, tornadoes, flood, wildfire, and lightning, among others. Several climate phenomena are also affected by the MJO. They are the monsoons, El Niño–Southern Oscillation, the North Atlantic Oscillation, the Pacific and North American pattern, the Arctic and Antarctic Oscillations or northern and southern annual modes, the Indian Ocean dipole, the Wyrtki jets, and the Indonesian Through-flow. This article provides a brief summary of the connections between the MJO and these weather and climate phenomena. These connections demonstrate the critical role of the MJO in the weather–climate continuum and its prediction.

CORRESPONDING AUTHOR: Chidong Zhang, RSMAS/MPO, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33133, E-mail: czhang@rsmas.miami.edu

The Madden–Julian oscillation exerts broad influences on global weather and climate as its center of convection moves from the tropical Indian Ocean into the Pacific. Weather events under the influence of the MJO include precipitation, surface temperature, tropical cyclones, tornadoes, flood, wildfire, and lightning, among others. Several climate phenomena are also affected by the MJO. They are the monsoons, El Niño–Southern Oscillation, the North Atlantic Oscillation, the Pacific and North American pattern, the Arctic and Antarctic Oscillations or northern and southern annual modes, the Indian Ocean dipole, the Wyrtki jets, and the Indonesian Through-flow. This article provides a brief summary of the connections between the MJO and these weather and climate phenomena. These connections demonstrate the critical role of the MJO in the weather–climate continuum and its prediction.

CORRESPONDING AUTHOR: Chidong Zhang, RSMAS/MPO, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33133, E-mail: czhang@rsmas.miami.edu

As a prime example of intraseasonal variability, the Madden–Julian Oscillation affects— and is pivotal to predicting—both weather and climate.

The conceptual separation of weather and climate is deeply rooted in our daily experience, as Herbertson (1901) put it: “Climate is what on an average we may expect, weather is what actually we get.”1 Translated into a scientific language, weather is a state of the atmosphere at a particular instance and climate is a set of statistics of an ensemble of many different states (Lorenz 1975). The weather–climate separation had its scientific basis in numerical prediction. It has been perceived that weather predictability comes from initiation conditions, while climate predictability from boundary conditions (Charney and Shukla 1977). This distinction would cease to exist in the modern practice of “seamless prediction” for weather and climate using “unified prediction models” (Hurrell et al. 2009; Brown et al. 2012). In such models, all components of the Earth system are coupled to each other, the only boundary condition needed is at the top of the atmosphere, and the source of predictability comes from, in addition to initiation conditions, the “memory” of slowly varying subsystems (the ocean, soil moisture, land, and sea ice), quasiperiodic phenomena, and known external forcing (Lorenz 1975). Yet, the weather–climate separation has penetrated so deep in our thinking that their traditional definitions are still often used in scientific and official documents, leaving a gaping vacancy in between. This vacancy is occupied by intraseasonal (20–90 days) variability.

Intraseasonal variability is by no means merely red noise filling the gap between synoptic and seasonal variability. Intraseasonal phenomena are distinct from higher- and lower-frequency variability by their significant spectral peaks and coherent spatial patterns. The Madden–Julian oscillation (MJO; Madden and Julian 1971, 1972) is the best example. Its large-scale signals in the atmospheric circulation, deep convection, and other variables propagating eastward slowly (~5 m s−1) from the Indian to Pacific Oceans are the dominant component of the tropical intraseasonal variability. They are so robust that they can be discerned from raw data without statistical manipulation (Zhang 2005).

The MJO plays a critical role in connecting or bridging weather and climate. This bridging role can be appreciated from different perspectives. The MJO affects many weather and climate phenomena. Its effects on weather depend on the state or phase of certain climate phenomena (e.g., ENSO), and their combined effects may lead to extreme weather events. Climate modes under the influence of the MJO in turn modulate weather in many parts of the world. The MJO is involved in scale interactions across a wide range of spectrum from the diurnal cycle to interannual variability (Moncrieff et al. 2012). Forecast of the Earth system to serve the society requires seamless prediction that covers daily, intraseasonal, seasonal, interannual, and longer variabilities (Dole 2008; Brunet et al. 2010; Shapiro et al. 2010; Chang et al. 2011). Improved MJO forecasting benefits prediction of tropical cyclones (Vitart 2009; Vitart et al. 2010), extratropical weather regimes (Marshall et al. 2010; Vitart and Molteni 2010), and ENSO (Shi et al. 2009); serves users from many sectors of the society (Gottschalck et al. 2010); and helps close the gap between traditional weather and short-term climate prediction (Waliser et al. 2006).

This article provides a brief summary of MJO effects on certain types of weather and climate events. The author hopes to convince its readers that weather and climate must be treated as a continuum by including the MJO and intraseasonal variability in general and reinforce the notion that the societal need for weather and climate prediction must be met with improved understanding and forecast of the MJO (Waliser et al. 2003a).

PRECIPITATION.

The sidebar “Detecting MJO influences on weather and climate” (Fig. SB2) illustrates rainfall variability in the tropics associated with the MJO during boreal winter (November–April). MJO influences on precipitation are not limited to the tropics and this season. A global map of precipitation anomalies associated with MJO in austral winter is given in Fig. 1. Anomalies in precipitation change signs between MJO phases in many places of the world.

DETECTING MJO INFLUENCES ON WEATHER AND CLIMATE

When discussing possible effects of the MJO on a particular type of weather or climate events, we must be mindful that all MJO episodes do not cause those events and all those events are not related to the MJO. The issue is whether and how the MJO may modulate the chances of occurrence, strengths, or spatial patterns/distributions of those events, as illustrated by examples given in this article.

MJO influences on weather events are commonly described as how those events vary with its phases. MJO phases can be defined in terms of the timing and locations of its center of convection (maximum rainfall anomalies) and associated wind fields. Most commonly used MJO phases are based on the real-time multivariate MJO (RMM) index of Wheeler and Hendon (2004). The RMM index is derived from a combined EOF analysis of daily anomalies in upper-and lower-level zonal wind and outgoing longwave radiation (OLR). MJO phases are defined by the principle components of the first two leading EOFs, normalized by their standard deviation (Fig. SB1). Each day, represented by a dot on the phase diagram, belongs to a particular phase. The distance of the dot from the center measures the amplitude of the MJO on that day. Composites of rainfall or any other field for each phase illustrate the canonical behavior of the MJO. In the boreal winter composite (Fig. SB2), the convection center of the MJO, represented by the maximum of positive anomalies in rainfall, starts over the Indian Ocean in phases 1–3, passes through the Maritime Continent in phases 4 and 5 and into the western Pacific in phases 6 and 7, and may continue their circumnavigating journey into the western hemisphere in phases 8 and 1 and thus complete its full cycle. During boreal summer, the zonal movement of the MJO convection center is accompanied by an additional northward movement associated with the Asian summer monsoon (Lawrence and Webster 2002). When the amplitude is less than 1 (within the circle on the phase diagram), the MJO is considered very weak or not existing (no MJO) and can be assigned as phase 0.

Fig. SB1
Fig. SB1

(above). Phase diagram of the RMM index. Each point represents a day. Eight phases and corresponding approximate locations of enhanced convective signals of the MJO are labeled. Points within the circle represent weak or no MJO (from Wheeler and Hendon 2004).

Citation: Bulletin of the American Meteorological Society 94, 12; 10.1175/BAMS-D-12-00026.1

Fig. SB2
Fig. SB2

(right). Composites of intraseasonal (30– 90 days) anomalies in TRMM precipitation (mm day−1) during November–April of 1998–2012 based on the RMM index.

Citation: Bulletin of the American Meteorological Society 94, 12; 10.1175/BAMS-D-12-00026.1

MJO influences on climate may also depend on its phases. Some climate events are more likely to start, amplify, or change sign in certain MJO phases than others. Some other climate phenomena are related to activities of a group of MJO events over a period (e.g., a season), instead of phases of individual MJO events. There can be a time lag between the group MJO activities and the climate phenomena they affect.

Fig. 1.
Fig. 1.

Rainfall anomalies measured by surface rain gauges during MJO phases (a) 2, (b) 4, (c) 6, and (d) 8 for austral winter. Anomalies are expressed as maximum vertical distance between the unconditional cumulative distribution function (CDF) and the corresponding conditional CDF for a particular MJO phase (vertical differences are measured at the point of maximum divergence in dimensionless units of “percent change in probability”). Positive (negative) distances indicate evidence of enhanced (suppressed) rainfall during the respective phase (from Donald et al. 2006).

Citation: Bulletin of the American Meteorological Society 94, 12; 10.1175/BAMS-D-12-00026.1