Realistic Madden–Julian Oscillation Research Articles

Sources of MJO teleconnection errors in the ECMWF extended‐range forecasts

AbstractThe European Centre for Medium‐range Weather Forecasts (ECMWF) extended‐range forecasts display large errors in the representations of Madden–Julian Oscillation (MJO) teleconnections over the North Atlantic with a strong underestimation of the impact of the MJO on the North Atlantic Oscillation (NAO) following an MJO in Phase 2–3. The origin of this error was investigated using a large set of re‐forecasts covering 20 years where part of the atmosphere was relaxed towards the ECMWF Reanalysis v5 (ERA5) reanalysis. These relaxation experiments show that relaxing the Tropics significantly improves the MJO teleconnections associated with Phase 2–3, with a reduction of about 50% in the amplitude error of the teleconnections. However, model errors outside the Tropics also play an important role. Results also suggest that errors in local processes reduce by about 20% the amplitude of the MJO teleconnections over the North Atlantic. Another example of a source of MJO teleconnection errors is the near surface over the Tibetan and Mongolian Plateaux. Nudging this area towards ERA5 improves the representation of the Pacific subtropical jet and the amplitude of MJO teleconnections associated with Phase 2–3 in the extended‐range re‐forecasts. Nudging the stratosphere exerts a comparatively weaker impact on the MJO teleconnections 11–15 days after an MJO in Phase 2–3. Overall, these experiments indicate that there are multiple sources of MJO teleconnection errors, making the representation of realistic MJO teleconnection by dynamical models a particularly challenging task.

Projected future changes in equatorial wave spectrum in CMIP6

The simulation of the Madden–Julian Oscillation (MJO) and convectively coupled equatorial waves (CCEWs) is considered in 13 state-of-the-art models from phase 6 of the Coupled Model Intercomparison Project (CMIP6). We use frequency–wavenumber power spectra of the models and observations for Outgoing Longwave Radiation (OLR) and zonal winds at 250 hPa (U250), and consider the historical simulations and end of twenty-first century projections for the SSP245 and SSP585 scenarios. The models simulate a spectrum quantitatively resembling that observed, though systematic biases exist. MJO and Kelvin waves (KW) are mostly underestimated, while equatorial Rossby waves (ER) are overestimated. Most models project a future increase in power spectra for the MJO, while nearly all project a robust increase for KW and weaker power values for most other wavenumber–frequency combinations, including higher wavenumber ER. In addition to strengthening, KW also shift toward higher phase speeds (or equivalent depths). Models with a more realistic MJO in their control climate tend to simulate a stronger future intensification.

Intraseasonal Surface Salinity Variability and the MJO in a Climate Model

AbstractDue to the sparsity of ocean observations, understanding inconsistencies exist even in qualitative respect about the mechanisms of intraseasonal sea surface salinity (SSS) over the Indo‐Pacific basin. In this study, a quantitative diagnostic of salinity/freshwater budgets related to the Madden‐Julian oscillation (MJO) was conducted based on model simulations. The model applied is a modified version of the NCEP Climate Forecast System, version 2. The application of a coupled model other than ocean‐only models is one of the unique aspects of this study. The model modifications resulted in a more realistic MJO simulation and a better simulation of intraseasonal SSS variability as validated against the Aquarius estimates. The mechanisms of the MJO‐related SSS variability are further explored by conducting a salinity budget analysis over the mixed layer. The analysis indicates a strong spatial variation of dominant physical processes controlling the local SSS variability. For example, while the role of surface fluxes dominates in the western Indian basin, oceanic advection plays an important role in the eastern Indian and western Pacific oceans.

The Effect of the Madden–Julian Oscillation on the North Atlantic Oscillation Using Idealized Numerical Experiments

AbstractThe aim of the paper is to investigate the influence of the Madden–Julian oscillation (MJO) on the North Atlantic Oscillation (NAO) using a quasigeostrophic model on the sphere. A simplified forcing based on potential vorticity anomalies in the tropics is used to mimic the MJO. The idealized nature of our setup allows us to determine the distinct roles played by stationary and synoptic waves. This is done by means of several series of almost 10 000 short runs of 30 days. Ensemble averages and a streamfunction budget analysis are used to study the modifications of the flow induced by the MJO. We find that a stationary Rossby wave is excited in the tropics during MJO phase 3. The western part of the Pacific jet is displaced poleward, which modifies the transient eddy activity in that basin. These changes create a ridge south of Alaska, which favors equatorward propagation of synoptic waves and larger poleward eddy momentum fluxes from the eastern Pacific toward the Atlantic, increasing the frequency of occurrence of the positive NAO events. The situation is essentially reversed following phase 6 of the MJO and conducive to the negative phase of the NAO. For a realistic MJO forcing amplitude, we find increases in both NAO phases to be around 30%, in reasonable agreement with the observations given the model simplicity. Finally, we present a series of experiments to assess the relative importance of linear versus nonlinear effects.

Prospects for Erratic and Intensifying Madden-Julian Oscillations

The Madden–Julian Oscillation (MJO) is a planetary-scale convective disturbance that typically forms in the equatorial Indian Ocean, propagates slowly eastward, and dissipates near the date line. This study examines how the MJO changes in response to a changing radiative forcing in a fully-Lagrangian coupled model (LCM) that is shown to simulate robust and realistic MJOs. After the LCM is spun up for 160 years to reproduce the late 20th century climate, non-water-vapor longwave optical depth is increased over 70 years to model the effects of increasing concentrations of greenhouse gases. The model is then run for another 30 years without additional changes to the radiative forcing. After the radiative forcing is modified, the MJO generally becomes more frequent and intense, but it is also more variable from one year to the next. Not only do composite MJO rainfall perturbations increase, but wind, temperature, and moisture perturbations also become stronger. The aspect of the MJO’s structure that changes the most is the largely dry equatorial Kelvin wave circulation that circumnavigates the globe between moist phases of the MJO. Potential impacts of these changes included alterations to the way in which the MJO modulates tropical cyclones, monsoon disturbances, and El Niño.

Sensitivity of the Madden Julian Oscillation to Ocean Warming in a Lagrangian Atmospheric Model

The Madden Julian Oscillation (MJO) is the largest contributor to intraseasonal weather variations in the tropics. It is associated with a broad region of enhanced rainfall that moves slowly eastward over the Indian and western Pacific Oceans, which has global impacts on atmospheric circulations. A number of recent observational and modeling studies have suggested that the MJO is becoming stronger as the oceans warm. In this study, the author explores the sensitivity of the MJO to ocean warming in a recently developed Lagrangian Atmospheric Model (LAM), which has been shown to simulate robust and realistic MJOs in previous work. Numerical simulations suggest that ocean warming leads to more frequent and intense MJOs that propagate more rapidly and cover a larger region of the tropics. The strengthening of the MJO is attributed to enhanced surface fluxes of moisture coming from the warmer ocean waters. While the LAM simulations have a number of limitations owing to idealized physical parameterizations and the use of prescribed sea surface temperatures, they provide additional evidence that the MJO will strengthen if the oceans continue to warm, and they also shed light on the mechanism of this strengthening.

MJO Propagation Shaped by Zonal Asymmetric Structures: Results from 24 GCM Simulations

Abstract Eastward propagation is an essential characteristic of the Madden–Julian oscillation (MJO). Yet, simulation of MJO propagation in general circulation models (GCMs) remains a major challenge and understanding the causes of propagation remains controversial. The present study explores why the GCMs have diverse performances in MJO simulation by diagnosis of 24 GCM simulations. An intrinsic linkage is found between MJO propagation and the zonal structural asymmetry with respect to the MJO convective center. The observed and realistically simulated MJO eastward propagations are characterized by stronger Kelvin easterly waves than Rossby westerly waves in the lower troposphere, which is opposite to the Gill pattern where Rossby westerly waves are 2 times stronger than Kelvin easterly waves. The GCMs simulating stronger Rossby westerly waves tend to show a stationary MJO. MJO propagation performances are robustly correlated with the quality of simulated zonal asymmetries in the 850-hPa equatorial zonal winds, 700-hPa diabatic heating, 1000–700-hPa equivalent potential temperature, and convective instability. The models simulating realistic MJO propagation are exemplified by an eastward propagation of boundary layer moisture convergence (BLMC) that leads precipitation propagation by about 5 days. The BLMC stimulates MJO eastward propagation by preconditioning and predestabilizing the atmosphere, and by generating lower-tropospheric heating and available potential energy to the east of precipitation center. The MJO structural asymmetry is generated by the three-way interaction among convective heating, moisture, and equatorial wave and boundary layer dynamics. In GCMs, differing convective heating representation could produce different MJO structural asymmetry, and thus different propagations. Diagnosis of structural asymmetry may help revealing the models’ deficiency in representing the complex three-way interaction processes, which involves various parameterized processes.

Winter mean lower tropospheric moisture over the Maritime Continent as a climate model diagnostic metric for the propagation of the Madden‐Julian oscillation

AbstractDespite its widespread influences on the atmosphere, the Madden‐Julian oscillation (MJO) remains poorly represented in state‐of‐the‐art general circulation models (GCMs). Motivated by recent findings that the horizontal advection of the mean low‐tropospheric moist static energy or moisture by the MJO winds plays a crucial role in the eastward propagation of the MJO, we investigate the relationship between lower tropospheric moisture patterns over the Indo‐Pacific and MJO eastward propagation in a suite of 23 GCM simulations. Model capability of reproducing the observed November–April mean lower tropospheric moisture pattern over the Indo‐Pacific, especially near the Maritime Continent (MC), is highly correlated with model skill in simulating MJO eastward propagation. In GCMs with difficulty capturing realistic MJO propagation, the amplitude of the mean low‐level moisture over the MC is greatly underestimated, leading to weak horizontal moisture gradients and thus discrepancies in moisture advection, significantly affecting MJO propagation. This study suggests that the mean lower tropospheric moisture pattern over the MC can serve as an important diagnostic metric for MJO propagation in climate models.

Simulations of MJO Propagation across the Maritime Continent: Impacts of SST Feedback

Abstract The observed Madden–Julian oscillation (MJO) tends to propagate eastward across the Maritime Continent from the eastern equatorial Indian Ocean to the western Pacific. However, numerical simulations present different levels of fidelity in representing the propagation, especially for the tropical convection associated with the MJO. This study conducts a series of coupled simulations using the NCEP CFSv2 to explore the impacts of SST feedback and convection parameterization on the propagation simulations. First, two simulations differing in the model horizontal resolutions are conducted. The MJO propagation in these two simulations is found generally insensitive to the resolution change. Further, based on the CFSv2 with a lower resolution, two additional experiments are performed with model SSTs nudged to climatologies with different time scales representing different air–sea coupling strength. It is demonstrated that weakening the air–sea coupling strength significantly degrades the MJO propagation simulation, suggesting the critical role of SST feedback in maintaining MJO propagation. Last, the sensitivity to convection parameterization is explored by comparing two simulations with different convection parameterization schemes. Analyses of these simulations indicate that including air–sea coupling alone in a dynamical model does not result in realistic maintenance of the MJO eastward propagation without the development of favorable SST conditions in the western Pacific. In both observations and one simulation with realistic MJO propagations, the preconditioning of SSTs is strongly affected by surface latent heat fluxes that are modulated by surface wind anomalies in both zonal and meridional directions. The diagnostics highlight the critical contribution from meridional winds in wind speed variations, which has been neglected in most MJO studies.

MJO simulation in a cloud‐system‐resolving global ocean‐atmosphere coupled model

AbstractAn observed Madden‐Julian Oscillation (MJO) propagating from the central Indian Ocean to the western Pacific from 15 December 2006 to 10 January 2007 was successfully simulated by a cloud‐system‐resolving global ocean‐atmosphere coupled model without parameterization of cumulus convection. We found that the ocean coupling has significant impacts on the MJO simulation, e.g., strength of the moisture convergence, and the timing and strength of the westerly wind burst over the Maritime Continent. The model also generally well simulated the decay of the MJO in the western Pacific, as well as the changes in sea surface temperature. These results demonstrate that the cloud‐system‐resolving global ocean‐atmosphere coupled model can be used for realistic MJO simulation.

Boreal Winter MJO Teleconnection in the Community Atmosphere Model Version 5 with the Unified Convection Parameterization

Abstract The Madden–Julian oscillation (MJO), the dominant mode of tropical intraseasonal variability, influences weather and climate in the extratropics through atmospheric teleconnection. In this study, two simulations using the Community Atmosphere Model version 5 (CAM5)—one with the default shallow and deep convection schemes and the other with the unified convection scheme (UNICON)—are employed to examine the impacts of cumulus parameterizations on the simulation of the boreal wintertime MJO teleconnection in the Northern Hemisphere. It is demonstrated that the UNICON substantially improves the MJO teleconnection. When the UNICON is employed, the simulated circulation anomalies associated with the MJO better resemble the observed counterpart, compared to the simulation with the default convection schemes. Quantitatively, the pattern correlation for the 300-hPa geopotential height anomalies between the simulations and observation increases from 0.07 for the default schemes to 0.54 for the UNICON. These circulation anomalies associated with the MJO further help to enhance the surface air temperature and precipitation anomalies over North America, although room for improvement is still evident. Initial value calculations suggest that the realistic MJO teleconnection with the UNICON is not due to the changes in the background wind, but rather primarily to the improved tropical convective heating associated with the MJO.

Vertical structure and physical processes of the Madden‐Julian oscillation: Exploring key model physics in climate simulations

AbstractAimed at reducing deficiencies in representing the Madden‐Julian oscillation (MJO) in general circulation models (GCMs), a global model evaluation project on vertical structure and physical processes of the MJO was coordinated. In this paper, results from the climate simulation component of this project are reported. It is shown that the MJO remains a great challenge in these latest generation GCMs. The systematic eastward propagation of the MJO is only well simulated in about one fourth of the total participating models. The observed vertical westward tilt with altitude of the MJO is well simulated in good MJO models but not in the poor ones. Damped Kelvin wave responses to the east of convection in the lower troposphere could be responsible for the missing MJO preconditioning process in these poor MJO models. Several process‐oriented diagnostics were conducted to discriminate key processes for realistic MJO simulations. While large‐scale rainfall partition and low‐level mean zonal winds over the Indo‐Pacific in a model are not found to be closely associated with its MJO skill, two metrics, including the low‐level relative humidity difference between high‐ and low‐rain events and seasonal mean gross moist stability, exhibit statistically significant correlations with the MJO performance. It is further indicated that increased cloud‐radiative feedback tends to be associated with reduced amplitude of intraseasonal variability, which is incompatible with the radiative instability theory previously proposed for the MJO. Results in this study confirm that inclusion of air‐sea interaction can lead to significant improvement in simulating the MJO.

Propagation of the Madden–Julian Oscillation and scale interaction with the diurnal cycle in a high-resolution GCM

The Madden–Julian Oscillation (MJO) is the chief source of tropical intra-seasonal variability, but is simulated poorly by most state-of-the-art GCMs. Common errors include a lack of eastward propagation at the correct frequency and zonal extent, and too small a ratio of eastward- to westward-propagating variability. Here it is shown that HiGEM, a high-resolution GCM, simulates a very realistic MJO with approximately the correct spatial and temporal scale. Many MJO studies in GCMs are limited to diagnostics which average over a latitude band around the equator, allowing an analysis of the MJO’s structure in time and longitude only. In this study a wider range of diagnostics is applied. It is argued that such an approach is necessary for a comprehensive analysis of a model’s MJO. The standard analysis of Wheeler and Hendon (Mon Wea Rev 132(8):1917–1932, 2004; WH04) is applied to produce composites, which show a realistic spatial structure in the MJO envelopes but for the timing of the peak precipitation in the inter-tropical convergence zone, which bifurcates the MJO signal. Further diagnostics are developed to analyse the MJO’s episodic nature and the “MJO inertia” (the tendency to remain in the same WH04 phase from one day to the next). HiGEM favours phases 2, 3, 6 and 7; has too much MJO inertia; and dies out too frequently in phase 3. Recent research has shown that a key feature of the MJO is its interaction with the diurnal cycle over the Maritime Continent. This interaction is present in HiGEM but is unrealistically weak.

Role of large‐scale moisture advection for simulation of the MJO with increased entrainment

The intraseasonal moisture budget is analyzed in simulations with the UK Met Office climate model, which produces a reasonable representation of the Madden–Julian Oscillation (MJO) when the entrainment rate in the convective parametrization is increased by 50%. Analysis of the moisture budget shows that parametrized convection tends to dry the troposphere and that large‐scale vertical advection moistens the troposphere. These two tendencies mostly balance each other. However, the total moisture tendency is asymmetric relative to the maximum precipitation, corresponding to the recharge and discharge process of organized convection in the Tropics. This moistening before and drying after the maximum precipitation is largely due to large‐scale horizontal advection of moisture. By comparing to the control run that does not simulate a realistic MJO, we find that increasing the entrainment acts to reduce deep convection in the relatively dry environment by increasing the mixing of the dry air. As a result, large‐scale convection and large‐scale advective processes play a stronger role during the development stage, which implies that convection is necessarily better organized.

Resolving the upper-ocean warm layer improves the simulation of the Madden–Julian oscillation

Here we show that coupling a high-resolution one-column ocean model to an atmospheric general circulation model dramatically improves simulation of the Madden–Julian oscillation (MJO) to have realistic strength, period, and propagation speed. The mechanism for the simulated MJO involves both frictional wave-convective conditional instability of the second kind (Frictional wave-CISK) and air–sea convective intraseasonal interaction (ASCII). In particular, better resolving the fine structure of upper ocean temperature, especially the warm layer, produces more vigorous atmosphere–ocean interaction and strengthens intraseasonal variations in both SST and atmospheric circulation. This helps organize and strengthen deep convection, inducing a stronger Kelvin-wave like perturbation and frictional near-surface convergence to the east. In addition, the warmer SST ahead of the MJO also acts to destabilize the boundary layer and enhance frictional convergence. These lead to a more realistic eastward-propagating MJO. A suite of sensitivity experiments were performed to show the robustness of the mechanisms and to demonstrate: (1) that mean state differences are not the main contributors to the improved simulation of our coupled model; (2) the role of SST variability in enhancing frictional convergence and intraseasonal variations in precipitation, and (3) that the simulation is significantly degraded when the first ocean model layer is thicker than 10 m. Our coupled model results are consistent with observations and demonstrate a simple but effective means to significantly improve MJO simulation and potentially also forecasts.

The moist static energy budget in NCAR CAM5 hindcasts during DYNAMO

AbstractThe Dynamics of the MJO (DYNAMO) field campaign took place in the Indian Ocean during boreal fall and winter of 2011–2012 to collect observations of Madden‐Julian Oscillation (MJO) initiation. Hindcast experiments are conducted with an atmospheric general circulation model with varying values of a dilute CAPE entrainment rate parameter for the first two MJO events of DYNAMO from 1 October 2011 to 15 December 2011. Higher entrainment rates better reproduce MJO precipitation and zonal wind, with RMM skill up to 20 days. Simulations with lower entrainment rapidly diverge from observations with no coherent MJO convective signal after 5 days, and no MJO predictive skill beyond 12 days. Analysis of the tropical Indian Ocean column moist static energy (MSE) budget reveals that the simulations with superior MJO performance exhibit a mean positive MSE tendency by vertical advection; inconsistent with reanalysis that indicates a weak negative tendency. All simulations have weaker mean MSE source tendency and significantly weaker cloud‐radiative feedbacks. The vertical gross moist stability (VGMS) is used to interpret these MSE budget results in a normalized framework relevant to moisture mode theory. VGMS in the high entrainment runs is far too low compared to ERAi, indicating that it cannot be used in isolation as a measure of model success in producing a realistic MJO hindcast, contrary to previous studies. However, effective VGMS that includes radiative feedbacks is similar among the high entrainment runs and ERAi. We conclude that the MJO is erroneously improved by increasing the entrainment parameter because moistening by vertical MSE advection compensates for the overly weak cloud‐radiative feedbacks.

Causal Evidence that Rotational Moisture Advection is Critical to the Superparameterized Madden–Julian Oscillation

AbstractThe authors investigate the hypothesis that horizontal moisture advection is critical to the eastward propagation of the Madden–Julian oscillation (MJO). Consistent diagnostic evidence has been found in recent MJO-permitting global models viewed from the moisture-mode dynamical paradigm. To test this idea in a causal sense, tropical moisture advection by vorticity anomalies is artificially modulated in a superparameterized global model known to produce a realistic MJO signal. Boosting horizontal moisture advection by tropical vorticity anomalies accelerates and amplifies the simulated MJO in tandem with reduced environmental gross moist stability. Limiting rotational horizontal moisture advection shuts the MJO down. These sensitivities are robust in that they are nearly monotonic with respect to the control parameter and emerge despite basic-state sensitivities favoring the opposite response. Speedup confirms what several diagnostic lines of evidence already suggest—that anomalous moisture advection is fundamental to MJO propagation. The rotational component is shown to be especially critical. Amplification further suggests it may play a role in adiabatically maintaining the MJO.

Evolution of ECMWF sub-seasonal forecast skill scores

Sub-seasonal forecasts have been routinely produced at ECMWF since 2002 with reforecasts produced ‘on the fly’ to calibrate the real-time sub-seasonal forecasts. In this study, the skill of the reforecasts from April 2002 to March 2012 and covering a common set of years (1995 to 2001) has been evaluated. Results indicate that the skill of the ECMWF reforecasts to predict the Madden–Julian Oscillation (MJO) has improved significantly since 2002, with an average gain of about 1 day of prediction skill per year. The amplitude of the MJO has also become more realistic, although the model still tends to produce MJOs which are weaker than in the ECMWF re-analysis. As a consequence, the ability of the ECMWF model to simulate realistic MJO teleconnections over the Northern and Southern Extratropics has improved dramatically over the 10-year period. Forecast skill scores have also improved in the Extratropics. For instance, weekly mean forecasts of the North Atlantic Oscillation Index are more skilful in recent years than 10 years ago. A large part of this improvement seems to be linked to the improvements in the representation of the MJO. Skill to predict 2 m temperature anomalies over the Northern Extratropics has also improved almost continuously since 2002. Changes in the horizontal and vertical resolutions of the atmospheric model had only a small impact on the skill scores, suggesting that most of the improvements in the ECMWF sub-seasonal forecasts were due to changes in model physics which were primarily designed to improve the model climate and medium-range forecasts. The impact of changes in the data assimilation system and in the observing data has not been considered in this study, since all the reforecasts used for this study were initialized from the same re-analysis over a common set of years.

Lagrangian overturning and the Madden–Julian Oscillation

AbstractThe Madden–Julian Oscillation (MJO), a planetary‐scale disturbance in zonal winds and equatorial convection that dominates intraseasonal variability in the Tropics, is a challenge to explain and notoriously difficult to simulate with conventional climate models. This study discusses numerical experiments conducted with a novel Lagrangian atmospheric model (LAM) that produce surprisingly robust and realistic MJOs, even at very low resolution. The LAM represents an atmosphere as a collection of conforming air parcels with motions that are predicted using Newtonian mechanics. The model employs a unique convective parametrization, referred to as Lagrangian overturning (LO), in which air parcels exchange vertical positions in convectively unstable regions. A key model parameter for simulating MJOs is the mixing between adjacent ascending and descending parcels, with more frequent and stronger MJOs occurring when greater mixing is prescribed. Sensitivity tests suggest that MJOs simulated with the LAM are not particularly sensitive to model resolution, but their structure and propagation speed do depend on sea‐surface temperatures, large‐scale precipitation patterns and surface fluxes. An important conclusion of this article is that the most fundamental dynamics of the MJO are captured by the LO convective parametrization coupled with large‐scale atmospheric circulations.

The Madden–Julian Oscillation in ECHAM6 and the Introduction of an Objective MJO Metric

Abstract This study presents a quantitative evaluation of the simulated Madden–Julian oscillation (MJO) in an ensemble of 42 experiments performed with ECHAM6 and previous ECHAM versions. The ECHAM6 experiments differ in their parameter settings, resolution, and whether the atmosphere is coupled to an ocean or not. The analysis concentrates on a few basic features of the MJO, namely, the signatures of convection/precipitation coupled with the circulation system and the eastward propagation strength of outgoing longwave radiation (OLR) and 850- and 200-hPa zonal winds within the MJO-related frequency–wavenumber range. It also examines whether precipitation and OLR show similar signatures in the MJO as simulated by ECHAM. The experiments reveal an MJO, however, to different degrees and in different aspects, so that a sound assessment requires a multivariate approach. In particular, the convective rainfall signatures are decoupled from the dynamic signature of the MJO in the simulations herein, which eventually leads to the introduction of a new MJO diagram and metric that incorporate OLR and the zonal winds in 850 and 200 hPa. The analysis here confirms the importance of the convection scheme: only with the Nordeng modifications to the Tiedtke scheme can realistic MJO features be simulated. High-resolution coupled experiments better represent the MJO as compared to low-resolution AMIP experiments. This is shown to follow from two more general findings, namely, that 1) air–sea interaction mainly increases the convective signature and 2) increased resolution enhances eastward propagation.