1. Introduction
As a dominant intraseasonal mode in the tropics, the Madden–Julian oscillation (MJO) is characterized by a planetary-scale atmospheric circulation coupled with deep convection and slow eastward propagation along the equator (Madden and Julian 1972; Lau and Chan 1986; Li and Zhou 2009). The initiation of MJO convection often occurs in the western equatorial Indian Ocean, triggered by tropical and midlatitude processes (e.g., Zhao et al. 2013). The MJO exhibits a first baroclinic mode vertical structure (Wang and Rui 1990) with a phase leading of boundary layer convergence (Wang and Li 1994; Li and Wang 1994) and moisture (Sperber 2003; Hsu and Li 2012).
Although the MJO main convection is confined in the tropics, it may exert widespread impacts on the global climate and weather. For example, the MJO may exert a great impact on the onset and demise of the monsoons (e.g., Hendon and Liebmann 1990; Jiang et al. 2004; Lorenz and Hartmann 2006; Wheeler et al. 2009), the genesis and tracks of tropical cyclones (e.g., Nakazawa 1988; Liebmann et al. 1994; Maloney and Hartmann 2000; Bi et al. 2015; Zhao and Li 2019), and the onsets of El Niño events (e.g., McPhaden 1999; Kessler and Kleeman 2000). The MJO can also exert a remote impact on extreme precipitation over the U.S. West Coast via an upper-tropospheric wave train (Mo and Higgins 1998; Higgins et al. 2000; Donald et al. 2006). Atmospheric river (AR) activity over U.S. West Coast could be greatly modulated by the MJO phase (Guan et al. 2012; Baggett et al. 2017). A significant influence of the MJO on the subseasonal variability of surface air temperature and precipitation in North America was observed (e.g., Bond and Vecchi 2003; Lin and Brunet 2009; Zhou et al. 2012; Baxter et al. 2014; Zheng et al. 2018). The surface air temperature over East Asia and the Arctic region also could be significantly influenced by the MJO (Vecchi and Bond 2004; Jeong et al. 2005; Park et al. 2010; Yoo et al. 2012).
Understanding the local and remote impacts of the MJO can provide a basis for the forecast of severe weather extremes and climate events. However, the impacts and teleconnection patterns strongly depend on the MJO characteristics and the background state. A long-term climate change due to global warming would result in significant changes in the MJO characteristics and its teleconnection patterns. Therefore, to what extent and how MJO-associated influences and teleconnections change with global warming has become an interesting scientific issue.
So far most such studies have been based on climate model projections. Such an approach, however, involves large uncertainties due to poor simulations of the MJO by current climate models (e.g., Jiang et al. 2015; Wang et al. 2017). Some studies projected an increase of MJO precipitation variance in a warmer climate (e.g., Arnold et al. 2015; Rushley et al. 2019; Maloney et al. 2019; Cui and Li 2019) while other studies showed a decrease in MJO precipitation amplitude using some CMIP3 models (Takahashi et al. 2011) or in the MJO wind field (Bui and Maloney 2018). An alternative approach is to use long-term observational data. Benefitting from modern reanalysis systems, two century-long reanalysis datasets covering the whole twentieth century have been produced. One is the Twentieth Century Reanalysis version V2c (20CR; Compo et al. 2011) from the National Oceanic and Atmospheric Administration (NOAA), and the other is the European Centre for Medium-Range Weather Forecasts (ECMWF) first atmospheric reanalysis of the twentieth century (ERA-20C; Poli et al. 2016). Because the two long-term reanalysis datasets assimilated only the surface observations (i.e., sea level pressure fields) into the model, a key question is whether or not the two reanalysis products can reproduce the observed MJO amplitude and phase evolution characteristics. This issue was resolved recently by a comparison of the two datasets with observed outgoing longwave radiation (OLR) since 1979 and modern reanalysis products [ERA-Interim (Dee et al. 2011) and the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) Reanalysis-II (NCEP-II; Kanamitsu et al. 2002)]. Cui et al. (2020) noted that the MJO amplitude, vertical structure, and evolution derived from the NOAA-20CR and ERA-20C were reasonably well captured, compared to those derived from the observational data. This motivates us to further investigate the long-term change of the MJO characteristics and climate impacts in the past century (1900–2010). Given that the global mean surface temperature during the period increased by 1°C, the results derived from this study would provide a basis for assessing the future changes of the MJO intensity and evolution characteristics and associated climate impacts under global warming.
A positive trend in MJO amplitude over the last few decades of the twentieth century was noticed by previous studies. For example, Jones and Carvalho (2006) using the NCEP reanalysis data illustrated a positive linear trend in the amplitude of MJO zonal wind for 1958–2004, while Pohl and Matthews (2007) pointed out that the MJO amplitude during 1976–2005 increased by 16% compared to that during 1950–75 period. By extending the period into the early twentieth century, Oliver and Thompson (2011) showed a weak positive trend in the MJO intensity using a reconstructed MJO index, and this result was verified by Dasgupta et al. (2020) using a machine learning method. Fu et al. (2020) detected pronounced multidecadal variations along with an increasing trend in MJO amplitude during the twentieth century using two century-long reanalysis products. Caution is needed in interpreting the long-term trend of the MJO amplitude, due to uncertainty in observational data in the early twentieth century. The trend would change with the selected locations and the bias can be as large as 30% (Oliver 2016). Roxy et al. (2019) suggested a longer persistence of MJO in the late twentieth century compared to that in the early twentieth century, which had far-reaching effects on rainfall over Southeast Asia, northern Australia, and the west coast of the United States. Therefore, the long-term change of MJO activity may have a significant impact on climate worldwide.
The objective of the current study is to reveal the long-term change of the MJO and its teleconnection patterns in the past century using the NOAA-20CR. In section 2, the data and methods are introduced. The changes of the MJO characteristics, as well as their remote impact on tropical and midlatitude precipitation, are described in section 3. The MJO impact on the precipitation extremes and the high-frequency variability over the Northern and Southern Hemispheres are further discussed in section 4. Finally, a summary and discussion are provided in the last section.
2. Data, methods, and model
a. Datasets
The primary dataset used in the current study is the NOAA-20CR, which covers from the beginning of the twentieth century to the early twenty-first century (1901–2010). It provides long and gap-free three-dimensional variables, such as horizontal (u and υ) wind components, vertical pressure velocity ω, specific humidity q fields, and variables related to convective activity, such as OLR and precipitation.
To address the uncertainties in the early twentieth century of the long-term reanalysis products, the ERA-20C dataset was also used for comparison. To verify the reliability of the twentieth century reanalysis dataset in the description of the global MJO-related precipitation features, the NOAA interpolated OLR product (Liebmann and Smith 1996) and the ERA-Interim dataset were employed, which were generated by the assimilation of improved observational data after 1979 and are good at describing the true climate conditions (e.g., Tian et al. 2010; Jiang et al. 2011; Kim et al. 2014). In addition, observational precipitation provided by the Tropical Rainfall Measuring Mission (TRMM) 3B42 version 7 (Huffman et al. 2007) covering the tropical and subtropical regions during 1998–2015 was also employed for verification. Because they just cover the recent decades, the verification only can be carried out for the late twentieth century. All the data were archived on 2.5° × 2.5° grid points, with a daily interval.
A time slice method is used to assess the long-term change of the MJO and its climate impact. Two 20-yr periods (1901–20 and 1991–2010) are used to represent the change of MJO characteristics and climate impacts in the early twentieth century (E20C) and the late twentieth century (L20C). Because tropical intraseasonal oscillations exhibit pronounced seasonality (Wang and Rui 1990), we only focus on the boreal winter season (November–February) when the MJO signal is most evident.
b. Methods
To isolate the eastward-propagating MJO signal, a space–time filter from the U.S. CLIVAR MJO working group (Waliser et al. 2009) was employed. All variables are subject to a temporal and spatial filtering to retain 20–80-day-period and zonal wavenumber-1–4 signals. Hereafter, the term “MJO-related” will refer to the space–time-filtered variables after removing the annual cycle at first. MJO phase composites were further employed, based on the Real-Time Multivariate MJO (RMM) index (Wheeler and Hendon 2004) defined by the leading pair of principal components from an empirical orthogonal function analysis of OLR and 850- and 200-hPa zonal wind fields averaged over 15°N–15°S. Here, we extended the meridional range of all fields into 20°S–20°N to involve the MJO signal as much as possible. The composite MJO anomaly fields for each phase are constructed from days when the normalized RMM amplitude is greater than one. In addition, a wave frequency spectrum analysis, a lagged correlation analysis, and a linear fitting method were also used to illustrate the MJO structure and propagation characteristics.
In Eq. (2), the variable with an overbar denotes the background mean state, the variable with a prime denotes the temporal–spatial (20–80 days and wavenumbers 1–4) filtered field. The terms related to the double-primed variables are omitted due to their small magnitudes.
c. Model description
To validate the MJO climate impacts, an anomaly atmospheric general circulation model (AGCM) (Jiang and Li 2005; Li 2006) was used. This anomaly model was constructed based on a global spectrum dry AGCM developed in the Geophysical Fluid Dynamics Laboratory (Held and Suarez 1994). The governing equations for momentum, temperature, and surface pressure tendency were written under a specified background three-dimensional mean state. In this study, the observed climatological winter [November–February (NDJF)] mean state is prescribed as the model basic state, which is derived from a long time mean of the NCEP-II reanalysis by linearly interpolating the original pressure surface data to the five equally distributed sigma levels. An anomalous heating field with the horizontal pattern of the MJO-scale OLR field is specified in the model to examine the atmospheric response to a prescribed MJO heating. The model has a horizontal resolution of T42. The model is integrated for 20 days to get a steady-state response. A Newtonian type damping with an e-folding time scale of 5 days is applied in the anomalous temperature equation.
3. Changes of MJO characteristics and impacts on precipitation
a. Changes of MJO intensity and propagation characteristics
First, we examine the changes in the MJO amplitude and propagation features in the past century. Figures 1a and 1b show the differences in the standard deviation of MJO-related OLR and U850 fields between the L20C and E20C periods. It is found that the amplitudes of both the OLR and U850 variabilities increase in the tropical Indo-western Pacific warm pool, the most active region of the MJO. The two fields show a similar percentage increase (about 50%) in the green box–averaged region. The result demonstrates that the MJO amplitude was strengthened in the past century. The MJO main activity region, on the other hand, has little change from E20C to L20C (figure not shown).
Differences of the standard deviation of space–time-filtered (a) OLR (W m2) and (b) 850-hPa zonal wind (U850; m s−1) fields between L20C and E20C. Power spectra of space–time-filtered (c) OLR and (d) U850 fields averaged for zonal wavenumbers 1–4 in E20C (green line) and L20C (orange line). The green box in (a) and (b) denotes the MJO active region (20°S–0°, 80°E–180°). The stippling denotes the difference exceeding the 99% confidence level based on a t test.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
Additionally, a power spectrum analysis was performed. Figures 1c and 1d show the power spectra of MJO-related OLR and U850 fields for the two periods. A consistent increase of the spectrum peaks from E20C to L20C occurs in the OLR and wind fields, confirming a robust and significant increase of the MJO amplitude in the past century.
Next, we examine the changes in the MJO eastward propagation speed. Figure 2 displays the longitude–time section of the MJO-related OLR field regressed onto a reference time series of the OLR anomaly averaged over 15°S–0°, 120°–140°E. The contour shows the eastward phase propagation of the MJO in E20C, whereas the shading shows the characteristic MJO phase propagation in L20C. The two fields approximately overlap with each other, indicating that there is no noticeable change in the overall MJO phase speed over the MJO active region. To examine whether or not the change of MJO propagation speed is significant between the L20C and E20C, the average phase speed for each winter season (19 winters for each period) was calculated and then a t test was applied for the two sets of sample data. The result showed an insignificant difference between the L20C and E20C. Then the phase propagation was further separated into the western Pacific (east of 130°E) and the Indian Ocean (west of 110°E) regions to examine the regional difference. It turns out that the changes of the MJO phase speed over the two regions are also statistically insignificant. Sensitivity tests for the reference location (say, the tropical Indian Ocean vs the western Pacific) were also conducted, and no significant differences were found.
Longitude–time section of correlation coefficients of space–time filtered OLR averaged over 20°S–20°N regressed onto the OLR time series averaged over 15°S–0°, 120°–140°E for the E20C (contour) and L20C (shading) periods.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
To confirm the result above, we also examine the RMM-based phase evolution patterns for the two periods (result not shown). It is noted that the horizontal patterns of the anomalous OLR and wind fields are similar between E20C and L20C at each of the eight MJO phases, and the main difference lies in the amplitude of these fields. The result further supports the notion that in the past century with the mean global warming of 1°C, the MJO activity area, frequency, and phase propagation had no obvious changes while its amplitude increased markedly.
b. Changes of MJO-induced precipitation in selected regions
In this subsection, we investigate to what extent MJO-induced precipitation changed in the past century. Figures 3a and 3b show the horizontal distributions of the standard deviation of MJO-induced precipitation anomalies in E20C and L20C respectively. The amplitude of the MJO-induced precipitation variability had an obvious increase from E20C to L20C in the tropical Indian Ocean and western Pacific, which is consistent with the strengthening of the MJO intensity in the past century. Moreover, the intraseasonal precipitation variability was also strengthened in the extratropics and higher latitudes, in particular over the Northern Hemispheric storm track, the South Pacific convergence zone (SPCZ), South China, and the U.S. West Coast. Figure 3c shows the horizontal distribution of percentage increase (ratio between the difference of L20C and E20C and E20C) of MJO-induced precipitation variability. While large values appear in the open ocean (such as over the tropical Indian Ocean, tropical western Pacific, and North Pacific), additional large percentage increase centers appear in densely populated coastal regions such as South China (SC), the Maritime Continent/northern Australia (AU), and California (CA). Among the three selected coastal regions, the average percentage increase in AU is the largest (67%), followed by SC (40%) and CA (14%).
Horizontal patterns of the standard deviation of space–time filtered precipitation anomaly field (mm day−1) in (a) E20C and (b) L20C, and (c) the percentage increase (%) of the precipitation standard deviation from E20C to L20C. Orange boxes in (c) denote the South China (SC; 22°–34°N, 105°–120°E), northern Australia (AU; 18°–5°S, 120°–150°E), and California (CA; 30°–45°N, 115°–130°W) regions respectively. The stippling denotes the difference exceeding the 99% confidence level based on a t test. Purple boxes are regions with large high-frequency precipitation variability discussed in section 4.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
To illustrate how the intraseasonal precipitation anomaly in the three regions (SC, AU, and CA) evolved with the MJO phase, we rely on a composite analysis, with the MJO phase being determined by the RMM indices. A systematic error was identified in the analysis of the precipitation evolution over SC in E20C compared to L20C, whereas the rainfall phase evolutions in AU and CA are consistent between E20C and L20C. It was further noted that the OLR field is a more reliable variable in SC (see section 5b for details). Thus, for a moisture budget analysis in SC, the OLR field is used to replace the precipitation field, based on a statistical relationship between the OLR and precipitation fields in L20C. Figure 4 is a scatterplot between MJO-scale OLR and precipitation fields in L20C in the tropics (25°S–35°N, 40°–160°E) at a 10° × 10° grid. A significant linear relationship is derived.
Scatterplot of MJO-scale precipitation (mm day−1) vs OLR (W m2) anomaly at a 10° × 10° grid in the tropics (25°S–35°N, 40°–160°E) during L20C. The green line denotes a linear fitting result with a regression coefficient of −0.12.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
Figure 5 shows the MJO-phase-dependent evolutions of composite precipitation anomalies in SC, AU, and CA. In SC the reconstructed precipitation field based on the OLR has been used for E20C. To verify the NOAA-20CR derived phase evolutions, the observed NOAA OLR field, together with the modern reanalysis product (ERA-Interim), during the same L20C period are used and the result is shown as the green curves. Note that in all three regions the orange curves match well the green curves in both the amplitude and phase, suggesting that the NOAA-20CR is reliable in capturing the observed MJO-induced precipitation evolutions in L20C in these regions, even though only surface observations were assimilated in the NOAA-20CR product.
Evolutions of anomalous precipitation (mm day−1) with MJO phases in (a) SC, (b) AU, and (c) CA. Blue (orange) lines are results from NOAA-20CR for the E20C (L20C) period. Green lines are results derived from the observed NOAA Interpolated OLR and ERA-Interim during the L20C period.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
Figure 5 illustrates that MJO-induced rainfall variability increased markedly from E20C to L20C in the three regions. The rainfall anomaly in SC reaches a peak at phases 2–3 when the MJO convective center is located over the eastern Indian Ocean. For AU, the intraseasonal rainfall anomaly reaches a peak at phase 5 when the MJO convective center is located in the Maritime Continent/western Pacific. In CA, there is a phase shift in the precipitation evolution between E20C and L20C. The minimum is at phases 3–4 in E20C and shifts to phase 5 in L20C. It is unclear whether this shift is attributed to low data quality in E20C or reflects a real change.
To explore how the MJO convection affects the precipitation anomaly in the three regions, a composite analysis and a moisture budget diagnosis were employed. We focused on the analysis of wet phases at each of the regions in L20C, as the circulation patterns in the two periods are similar but signals are stronger and data quality is better in L20C. Here, the wet phases in each region are listed below: phases 2–3 for SC, phases 4–6 for AU, and phases 8–1 for CA.
Figure 6a shows the composite precipitation and 850-hPa wind anomaly fields in the wet phases of SC. During the period, the MJO convection is located over the Indian Ocean while a suppressed convective center is located in the western Pacific. The low-level wind response to the positive heating anomaly, according to Gill (1980), is anomalous westerlies to the west and anomalous easterlies to the east of the heat source. The negative heating anomaly in the western Pacific induces a Rossby wave response with an anomalous anticyclone to the northwest of the negative heat source. Anomalous southerlies associated with the anticyclone over the western North Pacific transport high mean moisture from the ocean into the SC region, leading to a positive precipitation anomaly there.
Composite patterns of anomalous precipitation (shading; mm day−1) and 850-hPa wind (vectors; m s−1) fields for the wet phases of the (a) SC, (b) AU, and (c) CA regions (shown as orange boxes). Pink vectors in (c) show the wave activity flux (m2 s−2) at 200 hPa.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
In contrast, the precipitation anomaly in AU is a direct response to MJO. Here, the “direct” response means that the MJO center passes through the region. It differs from the other regions where the MJO impact is through remote teleconnection (e.g., Rossby wave train). As the main convective branch of the MJO moves to the Maritime Continent, there are pronounced westerlies (easterlies) west (east) of the Maritime Continent (Fig. 6b). The convergence of the anomalous winds leads to positive precipitation anomalies over the AU. The enhanced rainfall anomaly shifts south of the equator because boreal winter is the Australian monsoon season and the mean precipitation shifts to the south.
Accompanied to the wet phases of CA is a cyclonic flow anomaly to the west of the region (Fig. 6c). This anomalous cyclone is part of the upper-tropospheric wave train emitted from Southeast Asia, as seen in the 200-hPa wave activity flux (pink vectors in Fig. 6c). A negative geopotential height anomaly appears over Southeast Asia at 200 hPa (figure not shown), and it is a direct response to the negative heating in situ. Anomalous southwesterlies to the south of the cyclone west of CA transport the moisture into the U.S. West Coast region, leading to a positive precipitation anomaly in the region.
It is worth mentioning that the MJO-induced wave train shown in Fig. 6c differs from the Pacific–North America (PNA) pattern, which is often triggered by anomalous heating over the central equatorial Pacific in association with ENSO. The different wave train patterns are possibly attributed to the difference in the tropical precipitation patterns associated with the MJO and ENSO.
Next, we diagnose a column-integrated moisture budget to understand the relative roles of dynamic and thermodynamic processes in causing the precipitation anomaly during the wet phases. Take the SC result as an example. Figure 7a shows the contributions of individual budget terms in Eq. (1) to the precipitation difference between L20C and E20C. A positive Pr' value denotes a larger precipitation anomaly in L20C compared to E20C. As shown in Fig. 7a, the rainfall difference is primarily attributed to the vertically integrated horizontal moisture convergence term—that is,
Differences (L20C minus E20C) of vertically integrated (1000–200 hPa) moisture budget terms (mm day−1) during the wet phases of SC. (b) Decomposition of the anomalous horizontal moisture convergence term [i.e., the third term in (a)]. (c) Relative contributions of MJO-scale meridional wind change vs the mean moisture change, with “1” (“2”) indicating the variable for the E20C (L20C) period.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
To further explore the relative role of the mean moisture difference versus the anomalous meridional wind difference, we conduct the following analysis. The first bar (referred to as “bar1” hereinafter) in Fig. 7c represents the meridional moisture convergence term in E20C. The second bar denotes the change due to the mean moisture difference, which shows a 13% increase compared to bar1. The third bar denotes the contribution by the anomalous wind change, which has a 66% increase compared to bar1. The last bar represents the contribution from both the mean moisture and the anomalous wind changes. Therefore, the analysis result indicates that the change of MJO-scale meridional wind anomaly plays a dominant role in causing a greater intraseasonal precipitation variability in L20C compared to E20C.
The same diagnostic approach was applied to the AU and CA regions. The results show that the increase in the precipitation anomaly in the two regions is mainly caused by the anomalous convergence of the mean moisture, in a way similar to that in the SC region. The difference arises from the relative role of the anomalous zonal and meridional wind components. In AU and CA, the contribution of the anomalous zonal wind component is greater.
To demonstrate that the increase of the anomalous circulation in the three regions is a result of the strengthened MJO, we conducted the anomaly AGCM experiments with a specified MJO heating in the tropics. The details of the model are described in section 2c.
First, we elaborate on SC simulation results. An idealized heating distribution is specified over the tropical Indian Ocean and western Pacific (Fig. 8a). This heating pattern is to mimic the anomalous precipitation pattern in MJO phases 2–3 when a positive rainfall anomaly occurs in the SC (Fig. 6a). The idealized heating has a dipole pattern with a positive center at 80°E and a negative center at 150°E. The vertical profile of the heating is based on the observed vertical heating structure for the MJO (e.g., Wang et al. 2017), with a maximum at σ = 0.3 (about 300 hPa). The amplitude of the anomalous heating field is 6 K day−1, which is proportional to the MJO precipitation anomaly in E20C. This experiment is referred to as EXP1.
The horizontal patterns of geopotential height (shading; gpm) and wind (vectors; m s−1) anomalies at (a) 850 hPa and (c) 200 hPa in response to the prescribed dipole heating (shaded circles, with green denoting a positive heating anomaly) in EXP1 and EXP3 respectively. The difference of 850-hPa geopotential height and wind anomaly fields (b) between EXP2 and EXP1 and (d) between EXP4 and EXP3. The amplitude of the specified heating rate in (a) and (c) is 6 K day−1. Green boxes denote the SC and CA regions.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
Figure 8a shows the low-level circulation response to the specified heating in EXP1. The simulated wind field resembles that shown in Fig. 6a to a large extent. The circulation in the tropics closely follows the classical Gill-type solution. The most noted feature in the extratropics is anomalous southerlies to the south of SC, which is part of Rossby wave-induced low-level anticyclone in response to the negative heating in the equatorial western Pacific.
To mimic the strengthening of the observed MJO intensity from E20C to L20C, we conducted a parallel experiment in which the equatorial heating amplitude is increased by 50%, while other settings remain unchanged. This experiment is named EXP2. Figure 8b illustrates the difference in the simulated circulation fields between EXP2 and EXP1. The anomalous southerly wind south of SC exhibits a significant increase with a rate of 67%, which is close to the MJO heating increase rate. This indicates that the increase of the anomalous wind over SC from E20C to L20C is primarily caused by the strengthening of the MJO intensity.
In the additional sensitivity experiments, we keep only the positive or negative heat source on the equator. The results show that the anomalous southerly over SC is primarily caused by the negative heating source in the western Pacific.
Parallel sets of experiments were conducted for the wet phases of AU and CA respectively. The anomalous circulation pattern associated with the AU wet phases (figure not shown) is similar to that in Fig. 6b as expected because it is a direct response to the MJO heating. For CA, an idealized dipole heating pattern with a positive center located at 60°E and a negative center at 120°E is specified (Fig. 8c) while other settings are same as those in EXP1. This experiment is referred to as EXP3. A Rossby wave train pattern appears in the upper troposphere (Fig. 8c). This wave train originates from the tropical western Pacific and the Indian Ocean, as seen from a negative geopotential height anomaly over Southeast Asia and a positive geopotential height anomaly over the northern Indian Ocean. Off the coast of CA, a low-level cyclonic anomaly with pronounced southwesterly flow appears, which can be inferred from Fig. 8d. The enhanced low-level westerly over CA as shown in Fig. 8d points out again the importance of the MJO intensification in causing the remote circulation change.
Additional sensitivity experiments were further carried out to reveal the relative importance of the heating anomalies over the Maritime Continent and the Indian Ocean. It was found that for the remote circulation change in CA, the negative heating anomaly over the Maritime Continent plays a more important role, while the effect of the positive heating over the Indian Ocean is relatively weaker.
4. Changes of precipitation extremes and high-frequency activity
As a low-frequency mode, MJO may modulate high-frequency activities including weather extremes. In this section, we will pay special attention to the changes in characteristics of precipitation extremes and high-frequency variability.
First, we define the extreme precipitation threshold as 95% of the winter season climatology in E20C and L20C respectively. Additional two indices are further defined. One is the number of the extreme day that exceed the E20C threshold, and the other is the cumulative extreme precipitation amount during the extreme days. Figures 9a–c show the differences of the extreme precipitation threshold, the cumulative extreme precipitation amount, and the number of the extreme day between L20C and E20C. It can be seen that the three indices all show a marked increase in most of the tropical and midlatitude oceans as well as coastal regions. Table 1 lists the percentage increase in the SC, AU, and CA regions. Results show that the maximum percentage increase appears in SC for all three indices, followed by AU and CA. The percentage increase of the cumulative extreme precipitation amount is around 140%–150% in SC and AU and 100% in CA. The percentage increase of the number of the extreme day is in the range of 70%–118%.
Horizontal patterns of differences (L20C minus E20C) of (a) extreme precipitation threshold (P95; mm day−1), (b) cumulative extreme precipitation amount (CEPA; 102 mm), and (c) the number of the extreme day (nDays; day). Orange boxes denote the SC, AU, and CA regions.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
Percentage increase of the extreme precipitation threshold (P95), the cumulative extreme precipitation amount (CEPA), and the number of the extreme day (nDays) from E20C to L20C in SC, AU, and CA.
The increase of the precipitation extremes can be further detected from the probability distribution function (PDF) maps shown in Fig. 10. The two ends of the PDF curves in Figs. 10a–c represent the probability of the extreme precipitation occurrence. The PDF maps show a clear increase in the probability from E20C to L20C in all three analysis regions. The result further demonstrates consistency with those shown in Figs. 9a–c and Table 1.
Probability distributions of precipitation anomaly in the (a) SC, (b) AU, and (c) CA regions respectively. Orange (blue) lines denote results for the L20C (E20C) period. Green dashed lines denote observation results derived from the TRMM precipitation product for the period of 1998–2015.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
The increase of the probability for extreme precipitation is likely attributed to the strengthened MJO amplitude and associated remote and local circulation anomalies. For example, a strengthened southerly during the wet phases over SC may promote a great moist enthalpy (i.e., temperature and moisture) advection, which increases precipitation extremes.
Next, we examine the change characteristics of high-frequency (HF; with a period less than 10 days) precipitation variability. Figures 11a and 11b show the standard deviation of the HF precipitation anomaly field for E20C and L20C respectively. As one can see, compared with E20C, the HF precipitation variability is much stronger in L20C, especially over the tropics and the Northern Hemispheric storm track regions. The percentage increase map (Fig. 11c) shows a maximum increase of 200% along the SPCZ and an average increase of 70%–100% in most of the tropical and midlatitude regions.
As in Fig. 3, but for the high-frequency precipitation anomaly (mm day−1). Purple boxes denote regions with large high-frequency precipitation variability: North Pacific storm track (NPST), North Atlantic storm track (NAST), and the southern Indian Ocean and Pacific (SIP). The stippling denotes the difference exceeding the 99% confidence level based on a t test.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
An interesting question is to what extent the MJO-scale precipitation increase may contribute to the increase of the HF variability. Note that the percentage increase pattern of MJO-scale precipitation shown in Fig. 3c is quite similar to that shown in Fig. 11c, with a pattern correlation coefficient of 0.91. This indicates that the HF rainfall variability tends to increase more in the region where the MJO-scale precipitation variability increases. The result implies a large-scale control of the HF rainfall variability by the MJO.
Table 2 lists the percentage increases of MJO-scale versus HF precipitation in the three selected coastal regions (SC, AU, and CA). To add more samples, we also show the results in three open ocean regions where the HF variability is large: North Pacific storm track region (NPST), North Atlantic storm track region (NAST), and the southern Indian Ocean and Pacific region (SIP) (indicated by the purple boxes in Figs. 11b and 3c). It is interesting to note that compared to the MJO-scale, the HF precipitation variability increases at a much faster rate. Such a difference is more evident in the midlatitude regions such as CA and NPST.
Percentage increase of HF precipitation standard deviation and MJO-scale precipitation standard deviation from E20C to L20C in six selected regions.
Figures 12a–d display the phase evolution of the RGR distribution. Note that the area with the largest RGR of the HF precipitation variability moves from the tropical Indian Ocean at phases 1–2 to the Maritime Continent and the western Pacific at phases 3–4 and 5–6. The eastward propagation of the maximum RGR is consistent with the phase evolution of the maximum MJO-scale precipitation, both of which move eastward along the equator. The phase-dependent RGR patterns clearly demonstrate the large-scale control of the MJO on the high-frequency rainfall variability. While such a large-scale control primarily occurs in the tropical warm pool where MJO is most active, it is also observed in CA and SC regions. The evolutions of RGRs of HF precipitation variability over CA and SC were particularly examined, and it was found that a stronger HF variability appears during the local active precipitation phase (figure not shown).
Horizontal patterns of the relative growth rate (RGR) of HF precipitation variability during different MJO phases. The dashed line indicates the eastward movement of the RGR center.
Citation: Journal of Climate 35, 2; 10.1175/JCLI-D-21-0306.1
Physically, it is argued that the MJO-scale motion may modulate the HF variability through the change of the background mean state. For example, an increase of the background moisture associated with the active phase of the MJO (include by vertical moisture advection and boundary layer convergence) may provide a favorable condition for the intensification of the HF precipitation variability.
5. Summary and discussion
a. Summary
In this paper, we investigate how and to what extent the MJO characteristics and its teleconnection patterns change in the past century using the NOAA-20CR dataset. The MJO features examined include amplitude and phase propagation. The teleconnection patterns include MJO-induced global precipitation, precipitation extremes, and high-frequency variabilities. The analysis primarily focused on the difference between the early-twentieth-century (E20C; 1901–20) and the late-twentieth-century (L20C; 1991–2010) periods.
The MJO amplitude was strengthened about 50% from E20C to L20C, according to the standard deviation of both the anomalous OLR and U850 fields. A power spectrum analysis also shows a marked increase of the two fields on the intraseasonal band. On the other hand, the eastward phase speed of the MJO and the main MJO activity region hardly change during the past century.
The anomalous precipitation pattern reveals a far-reaching impact of the MJO, and the amplitude of the precipitation pattern increases worldwide from E20C to L20C. A maximum increase of the precipitation anomaly appears in the tropical Indo-western Pacific warm pool. For the three selected coastal regions (SC, AU, and CA), the percentage increase of the anomalous precipitation ranges from 67% in AU to 14% in CA (Fig. 3c).
The processes through which the MJO affects the regional precipitation anomaly in SC, AU, and CA were examined. For SC, a positive precipitation anomaly occurs when a suppressed (enhanced) MJO convective center is located in the western Pacific (Indian) Ocean. An anomalous southerly as a part of the Rossby wave response to the negative heat source in the western Pacific is a primary contributor that impacts the intraseasonal rainfall anomaly in SC. For AU, the main circulation that affects the precipitation anomaly is the zonal wind anomaly directly related to MJO circulation. For CA, the remote MJO impact is through an upper-tropospheric Rossby wave train emanating from the western North Pacific. A cyclonic circulation anomaly with equivalent barotropic structure occurs southwest of CA. Anomalous southwesterlies to the south of the cyclone transport moisture from the ocean to the CA coastal region, leading to enhanced precipitation there.
An anomalous AGCM was employed to test the sensitivity of the remote circulation response to the MJO heating strength. The result indicates that the circulation teleconnection pattern is enhanced when the prescribed MJO heating is strengthened. Thus the numerical model experiments confirm a moisture budget diagnosis that the strengthened MJO intensity is responsible for enhanced global precipitation anomalies on the intraseasonal time scale.
Precipitation extremes and high-frequency (with a period less than 10 days) variability were examined. It is found that the extreme precipitation threshold, the number of extreme days, and the cumulative extreme precipitation amount all increase significantly from E20C to L20C. In particular, the number of extreme days increases by 110% in SC and AU and 70% in CA, and the cumulative extreme precipitation amount increases by 140%–150% in SC and AU and 100% in CA. Such increases are consistent with the strengthening of the high-frequency precipitation variability. The pattern of the percentage increase of the HF variability resembles well (with a pattern correlation coefficient of 0.91) the pattern of the MJO-scale precipitation increase (Figs. 11c and 3c), implying that the HF rainfall variability is greatly modulated by the MJO. Besides, the maximum HF variability center moves eastward, closely following the MJO phase propagation. The rate of increase of the HF variability from E20C to L20C is much greater than its intraseasonal counterpart. This implies that HF variability is the main contributor to extreme precipitation changes.
The close link between the changes of the HF variability and the MJO-scale rainfall variability suggests that the MJO may exert a large-scale control on the HF activity. Physically it is argued that the large-scale modulation is through the change of the background moisture as a positive intraseasonal precipitation anomaly implies anomalous ascent, which increases local specific humidity through vertical advection. A moister background convection provides a favorable condition for the intensification of extreme precipitation and HF activities.
b. Discussion
When analyzing the MJO-induced precipitation evolution in SC (Fig. 5a), a mismatch was found in the OLR and precipitation fields in E20C, while such a mismatch was not found in AU and CA. Through the comparison with local precipitation–OLR phase evolution in L20C when more reliable observations were assimilated, and considering the fact that MJO propagation characteristics have little change throughout the twentieth century, the OLR field in E20C appears a better choice.
An interesting question is what caused the systematic bias and mismatch between the reanalysis-generated precipitation and OLR field. We speculate that the mismatch may arise from different physical parameterization schemes that control the OLR and precipitation fields in the general circulation models (GCMs). Rain in GCMs comes from two sources: the convective parameterization and stratiform (or large-scale) parameterization. OLR, on the other hand, is primarily determined by the latter. In addition, OLR may also depend on the model cloud and radiative schemes. Therefore, biases in the model parameterization schemes can cause a mismatch between precipitation and OLR.
Caution is needed in examining the long-term trend using reanalysis products as they are generated by numerical models with various observational data being assimilated into the models. The scarcity of observations during the early period or a significant change of observational systems would affect the realism of the reanalysis product. Oliver (2016) found that the estimated centennial-scale trend of MJO amplitude could range from nearly zero to an increase of 30% by just changing the choice of predictor locations. In particular, the use of surface pressure data from locations that have either been poorly observed or have experienced significant changes in the observing system over the twentieth century would change the trend of the estimated MJO amplitude. Therefore, a comparison of results from different reanalysis products is needed in order to obtain a robust signal. To address this issue, we have carried out a parallel calculation with the use of the ERA-20C product. The result confirms that the MJO intensity experienced a significant increase throughout the twentieth century, even though the increase rate is about half of that derived from NOAA-20CR.
Given the large uncertainty in the NOAA-20CR precipitation fields, an additional sensitivity test with a velocity potential MJO (VPM) index (Ventrice et al. 2013) was conducted. The VPM index is a pure circulation index that consists of low-level and upper-level zonal wind fields as well as a velocity potential field to mimic the tropical convection. Again, a significant increase of MJO intensity is found using the VPM index, confirming the RMM-based results. A further sensitivity test using a precipitation-based index for tropical intraseasonal oscillation (PII; Wang 2020) was employed. The result shows a consistent, significant increase of the MJO variability from E20C to L20C.
Acknowledgments
This work was jointly supported by NSFC Grant 42088101, China National Key R&D Program 2017YFA0603802, 2018YFA0605604, NOAA NA18OAR4310298, NSF AGS-2006553, and the Postgraduate Research and Practice Innovation Program of Jiangsu Province KYCX20_0915. This is SOEST contribution number 11417 and IPRC contribution number 1543.
Data availability statement.
The datasets used in this study are openly available. The NOAA-20CR dataset and interpolated OLR data product are downloaded from NOAA Physical Sciences Laboratory at https://psl.noaa.gov/data/. The ERA-20C and ERA-Interim datasets from ECMWF are available at https://www.ecmwf.int/en/orecasts/datasets/reanalysis-datasets/era-20c and https://www.ecmwfint/en/forecasts/datasets/reanalysis-datasets/era-interim. The TRMM 3B42 data are download from NASA Goddard Earth Science Data and Information Services Center at https://disc.gsfc.nasa.gov/TRMM.
REFERENCES
Arnold, N. P., M. Branson, Z. Kuang, D. A. Randall, and E. Tziperman, 2015: MJO intensification with warming in the superparameterized CESM. J. Climate, 28, 2706–2724, https://doi.org/10.1175/JCLI-D-14-00494.1.
Baggett, C. F., E. A. Barnes, E. D. Maloney, and B. D. Mundhenk, 2017: Advancing atmospheric river forecasts into subseasonal-to-seasonal time scales. Geophys. Res. Lett., 44, 7528–7536, https://doi.org/10.1002/2017GL074434.
Banacos, P. C., and D. M. Schultz, 2005: The use of moisture flux convergence in forecasting convective initiation: Historical and operational perspectives. Wea. Forecasting, 20, 351–366, https://doi.org/10.1175/WAF858.1.
Baxter, S., S. Weaver, J. Gottschalck, and Y. Xue, 2014: Pentad evolution of wintertime impacts of the Madden–Julian oscillation over the contiguous United States. J. Climate, 27, 7356–7367, https://doi.org/10.1175/JCLI-D-14-00105.1.
Bi, M., T. Li, X. Shen, and M. Peng, 2015: To what extent the presence of real-strength tropical cyclones influences the estimation of atmospheric intraseasonal oscillation intensity?. Atmos. Sci. Lett., 16, 438–444, https://doi.org/10.1002/asl.579.
Bond, N. A., and G. A. Vecchi, 2003: The influence of the Madden–Julian oscillation on precipitation in Oregon and Washington. Wea. Forecasting, 18, 600–613, https://doi.org/10.1175/1520-0434(2003)018<0600:TIOTMO>2.0.CO;2.
Bui, H. X., and E. D. Maloney, 2018: Changes in Madden–Julian oscillation precipitation and wind variance under global warming. Geophys. Res. Lett., 45, 7148–7155, https://doi.org/10.1029/2018GL078504.
Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137 (654), 1–28, https://doi.org/10.1002/qj.776.
Cui, J., and T. Li, 2019: Changes of MJO propagation characteristics under global warming. Climate Dyn., 53, 5311–5327, https://doi.org/10.1007/s00382-019-04864-4.
Cui, J., L. Wang, T. Li, and B. Wu, 2020: Can reanalysis products with only surface variables assimilated capture Madden–Julian Oscillation characteristics?. Int. J. Climatol., 40, 1279–1293, https://doi.org/10.1002/joc.6270.
Dasgupta, P., A. Metya, C. V. Naidu, M. Singh, and M. K. Roxy, 2020: Exploring the long-term changes in the Madden–Julian Oscillation using machine learning. Sci. Rep., 10, 18567, https://doi.org/10.1038/s41598-020-75508-5.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828.
Donald, A., H. Meinke, B. Power, A. D. N. Maia, M. C. Wheeler, N. White, R. C. Stone, and J. Ribbe, 2006: Near-global impact of the Madden–Julian Oscillation on rainfall. Geophys. Res. Lett., 33, L09704, https://doi.org/10.1029/2005GL025155.
Fu, Z., P.-C. Hsu, and F. Liu, 2020: Factors regulating the multidecadal changes in MJO amplitude over the twentieth century. J. Climate, 33, 9513–9529, https://doi.org/10.1175/JCLI-D-20-0111.1.
Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447–462, https://doi.org/10.1002/qj.49710644905.
Guan, B., D. E. Waliser, N. P. Molotch, E. J. Fetzer, and P. J. Neiman, 2012: Does the Madden–Julian oscillation influence wintertime atmospheric rivers and snowpack in the Sierra Nevada?. Mon. Wea. Rev., 140, 325–342, https://doi.org/10.1175/MWR-D-11-00087.1.
Held, I. M., and M. J. Suarez, 1994: A proposal for the intercomparison of the dynamical cores of atmospheric general circulation models. Bull. Amer. Meteor. Soc., 75, 1825–1830, https://doi.org/10.1175/1520-0477(1994)075<1825:APFTIO>2.0.CO;2.
Hendon, H. H., and B. Liebmann, 1990: The intraseasonal (30–50 day) oscillation of the Australian summer monsoon. J. Atmos. Sci., 47, 2909–2924, https://doi.org/10.1175/1520-0469(1990)047<2909:TIDOOT>2.0.CO;2.
Higgins, R. W., J.-K. E. Schemm, W. Shi, and A. Leetmaa, 2000: Extreme precipitation events in the western United States related to tropical forcing. J. Climate, 13, 793–820, https://doi.org/10.1175/1520-0442(2000)013<0793:EPEITW>2.0.CO;2.
Hsu, P., and T. Li, 2012: Role of the boundary layer moisture asymmetry in causing the eastward propagation of the Madden–Julian oscillation. J. Climate, 25, 4914–4931, https://doi.org/10.1175/JCLI-D-11-00310.1.
Hsu, P., T. Li, and C.-H. Tsou, 2011: Interactions between boreal summer intraseasonal oscillations and synoptic-scale disturbances over the western North Pacific. Part I: Energetics diagnosis. J. Climate, 24, 927–941, https://doi.org/10.1175/2010JCLI3833.1.
Huffman, G. J., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 38–55, https://doi.org/10.1175/JHM560.1.
Jeong, J.-H., C.-H. Ho, B.-M. Kim, and W.-T. Kwon, 2005: Influence of the Madden–Julian Oscillation on wintertime surface air temperature and cold surges in East Asia. J. Geophys. Res., 110, D11104, https://doi.org/10.1029/2004JD005408.
Jiang, X., and T. Li, 2005: Reinitiation of the boreal summer intraseasonal oscillation in the tropical Indian Ocean. J. Climate, 18, 3777–3795, https://doi.org/10.1175/JCLI3516.1.
Jiang, X., T. Li, and B. Wang, 2004: Structures and mechanisms of the northward propagating boreal summer intraseasonal oscillation. J. Climate, 17, 1022–1039, https://doi.org/10.1175/1520-0442(2004)017<1022:SAMOTN>2.0.CO;2.
Jiang, X., and Coauthors, 2011: Vertical diabatic heating structure of the MJO: Intercomparison between recent reanalyses and TRMM estimates. Mon. Wea. Rev., 139, 3208–3223, https://doi.org/10.1175/2011MWR3636.1.
Jiang, X., and Coauthors, 2015: Vertical structure and physical processes of the Madden–Julian Oscillation: Exploring key model physics in climate simulations. J. Geophys. Res. Atmos., 120, 4718–4748, https://doi.org/10.1002/2014JD022375.
Jones, C., and L. M. V. Carvalho, 2006: Changes in the activity of the Madden–Julian oscillation during 1958–2004. J. Climate, 19, 6353–6370, https://doi.org/10.1175/JCLI3972.1.
Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter, 2002: NCEP–DOE AMIP-II reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631–1644, https://doi.org/10.1175/BAMS-83-11-1631.
Kessler, W. S., and R. Kleeman, 2000: Rectification of the Madden–Julian oscillation into the ENSO cycle. J. Climate, 13, 3560–3575, https://doi.org/10.1175/1520-0442(2000)013<3560:ROTMJO>2.0.CO;2.
Kim, D., M.-I. Lee, D. Kim, S. D. Schubert, D. E. Waliser, and B. Tian, 2014: Representation of tropical subseasonal variability of precipitation in global reanalyses. Climate Dyn., 43, 517–534, https://doi.org/10.1007/s00382-013-1890-x.
Lau, K.-M., and P. H. Chan, 1986: Aspects of the 40–50 day oscillation during the northern summer as inferred from outgoing longwave radiation. Mon. Wea. Rev., 114, 1354–1367, https://doi.org/10.1175/1520-0493(1986)114<1354:AOTDOD>2.0.CO;2.
Li, T., 2006: Origin of the summertime synoptic-scale wave train in the western North Pacific. J. Atmos. Sci., 63, 1093–1102, https://doi.org/10.1175/JAS3676.1.
Li, T., and B. Wang, 1994: A thermodynamic equilibrium climate model for monthly mean surface winds and precipitation over the tropical Pacific. J. Atmos. Sci., 51, 1372–1385, https://doi.org/10.1175/1520-0469(1994)051<1372:ATECMF>2.0.CO;2.
Li, T., and C. Zhou, 2009: Planetary scale selection of the Madden–Julian oscillation. J. Atmos. Sci., 66, 2429–2443, https://doi.org/10.1175/2009JAS2968.1.
Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77, 1275–1277, https://doi.org/10.1175/1520-0477-77.6.1274.
Liebmann, B., H. H. Hendon, and J. D. Glick, 1994: The relationship between tropical cyclones of the western Pacific and Indian Oceans and the Madden–Julian Oscillation. J. Meteor. Soc. Japan, 72, 401–412, https://doi.org/10.2151/jmsj1965.72.3_401.
Lin, H., and G. Brunet, 2009: The influence of the Madden–Julian oscillation on Canadian wintertime surface air temperature. Mon. Wea. Rev., 137, 2250–2262, https://doi.org/10.1175/2009MWR2831.1.
Lorenz, D. J., and D. L. Hartmann, 2006: The effect of the MJO on the North American monsoon. J. Climate, 19, 333–343, https://doi.org/10.1175/JCLI3684.1.
Madden, R. A., and P. R. Julian, 1972: Description of global-scale circulation cells in the tropics with a 40–50 day period. J. Atmos. Sci., 29, 1109–1123, https://doi.org/10.1175/1520-0469(1972)029<1109:DOGSCC>2.0.CO;2.
Maloney, E. D., and D. L. Hartmann, 2000: Modulation of eastern North Pacific hurricanes by the Madden–Julian oscillation. J. Climate, 13, 1451–1460, https://doi.org/10.1175/1520-0442(2000)013<1451:MOENPH>2.0.CO;2.
Maloney, E. D., Á. F. Adames, and H. X. Bui, 2019: Madden–Julian Oscillation changes under anthropogenic warming. Nat. Climate Change, 9, 26–33, https://doi.org/10.1038/s41558-018-0331-6.
McPhaden, M. J., 1999: Genesis and evolution of the 1997-98 El Niño. Science, 283, 950–954, https://doi.org/10.1126/science.283.5404.950.
Mo, K. C., and R. W. Higgins, 1998: Tropical convection and precipitation regimes in the western United States. J. Climate, 11, 2404–2423, https://doi.org/10.1175/1520-0442(1998)011<2404:TCAPRI>2.0.CO;2.
Nakazawa, T., 1988: Tropical super clusters within intraseasonal variations over the western Pacific. J. Meteor. Soc. Japan, 66, 823–839, https://doi.org/10.2151/jmsj1965.66.6_823.
Oliver, E. C. J., 2016: Blind use of reanalysis data: Apparent trends in Madden–Julian Oscillation activity driven by observational changes. Int. J. Climatol., 36, 3458–3468, https://doi.org/10.1002/joc.4568.
Oliver, E. C. J., and K. R. Thompson, 2011: A reconstruction of Madden–Julian oscillation variability from 1905 to 2008. J. Climate, 25, 1996–2019, https://doi.org/10.1175/JCLI-D-11-00154.1.
Park, T.-W., C.-H. Ho, S. Yang, and J.-H. Jeong, 2010: Influences of Arctic Oscillation and Madden–Julian Oscillation on cold surges and heavy snowfalls over Korea: A case study for the winter of 2009–2010. J. Geophys. Res., 115, D23122, https://doi.org/10.1029/2010JD014794.
Pohl, B., and A. J. Matthews, 2007: Observed changes in the lifetime and amplitude of the Madden–Julian oscillation associated with interannual ENSO sea surface temperature anomalies. J. Climate, 20, 2659–2674, https://doi.org/10.1175/JCLI4230.1.
Poli, P., and Coauthors, 2016: ERA-20C: An atmospheric reanalysis of the twentieth century. J. Climate, 29, 4083–4097, https://doi.org/10.1175/JCLI-D-15-0556.1.
Roxy, M. K., P. Dasgupta, M. J. McPhaden, T. Suematsu, C. Zhang, and D. Kim, 2019: Twofold expansion of the Indo-Pacific warm pool warps the MJO life cycle. Nature, 575, 647–651, https://doi.org/10.1038/s41586-019-1764-4.
Rushley, S. S., D. Kim, and Á. F. Adames, 2019: Changes in the MJO under greenhouse gas–induced warming in CMIP5 models. J. Climate, 32, 803–821, https://doi.org/10.1175/JCLI-D-18-0437.1.
Sperber, K. R., 2003: Propagation and the vertical structure of the Madden–Julian Oscillation. Mon. Wea. Rev., 131, 3018–3037, https://doi.org/10.1175/1520-0493(2003)131<3018:PATVSO>2.0.CO;2.
Takahashi, C., N. Sato, A. Seiki, K. Yoneyama, and R. Shirooka, 2011: Projected future change of MJO and its extratropical teleconnection in East Asia during the northern winter simulated in IPCC AR4 models. SOLA, 7, 201–204, https://doi.org/10.2151/sola.2011-051.
Tian, B., D. E. Waliser, E. J. Fetzer, and Y. L. Yung, 2010: Vertical moist thermodynamic structure of the Madden–Julian oscillation in atmospheric infrared sounder retrievals: An update and a comparison to ECMWF Interim re-analysis. Mon. Wea. Rev., 138, 4576–4582, https://doi.org/10.1175/2010MWR3486.1.
Vecchi, G. A., and N. A. Bond, 2004: The Madden–Julian Oscillation (MJO) and northern high latitude wintertime surface air temperatures. Geophys. Res. Lett., 31, L04104, https://doi.org/10.1029/2003GL018645.
Ventrice, M. J., M. C. Wheeler, H. H. Hendon, C. J. Schreck, C. D. Thorncroft, and G. N. Kiladis, 2013: A modified multivariate Madden–Julian oscillation index using velocity potential. Mon. Wea. Rev., 141, 4197–4210, https://doi.org/10.1175/MWR-D-12-00327.1.
Waliser, D., and Coauthors, 2009: MJO simulation diagnostics. J. Climate, 22, 3006–3030, https://doi.org/10.1175/2008JCLI2731.1.
Wang, B., and H. Rui, 1990: Dynamics of the coupled moist Kelvin–Rossby wave on an equatorial β-plane. J. Atmos. Sci., 47, 397–413, https://doi.org/10.1175/1520-0469(1990)047<0397:DOTCMK>2.0.CO;2.
Wang, B., and T. Li, 1994: Convective interaction with boundary-layer dynamics in the development of a tropical intraseasonal system. J. Atmos. Sci., 51, 1386–1400, https://doi.org/10.1175/1520-0469(1994)051<1386:CIWBLD>2.0.CO;2.
Wang, L., T. Li, E. Maloney, and B. Wang, 2017: Fundamental causes of propagating and nonpropagating MJOs in MJOTF/GASS models. J. Climate, 30, 3743–3769, https://doi.org/10.1175/JCLI-D-16-0765.1.
Wang, S., 2020: A precipitation-based index for tropical intraseasonal oscillations. J. Climate, 33, 805–823, https://doi.org/10.1175/JCLI-D-19-0019.1.
Wheeler, M. C., and H. H. Hendon, 2004: An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132, 1917–1932, https://doi.org/10.1175/1520-0493(2004)132<1917:AARMMI>2.0.CO;2.
Wheeler, M. C., H. H. Hendon, S. Cleland, H. Meinke, and A. Donald, 2009: Impacts of the Madden–Julian oscillation on Australian rainfall and circulation. J. Climate, 22, 1482–1498, https://doi.org/10.1175/2008JCLI2595.1.
Yoo, C., S. Lee, and S. B. Feldstein, 2012: Mechanisms of Arctic surface air temperature change in response to the Madden–Julian oscillation. J. Climate, 25, 5777–5790, https://doi.org/10.1175/JCLI-D-11-00566.1.
Zhao, C., and T. Li, 2019: Basin dependence of the MJO modulating tropical cyclone genesis. Climate Dyn., 52, 6081–6096, https://doi.org/10.1007/s00382-018-4502-y.
Zhao, C., T. Li, and T. Zhou, 2013: Precursor signals and processes associated with MJO initiation over the tropical Indian Ocean. J. Climate, 26, 291–307, https://doi.org/10.1175/JCLI-D-12-00113.1.
Zheng, C., E. K.-M. Chang, H.-M. Kim, M. Zhang, and W. Wang, 2018: Impacts of the Madden–Julian oscillation on storm-track activity, surface air temperature, and precipitation over North America. J. Climate, 31, 6113–6134, https://doi.org/10.1175/JCLI-D-17-0534.1.
Zhou, S., M. L’Heureux, S. Weaver, and A. Kumar, 2012: A composite study of the MJO influence on the surface air temperature and precipitation over the continental United States. Climate Dyn., 38, 1459–1471, https://doi.org/10.1007/s00382-011-1001-9.