1. Introduction
During the Asian summer monsoon (ASM) season (June–September) there are two preferred locations of convection over the Indian longitudes, over the continent and over the equatorial Indian Ocean (Fig. 1a), often referred to as a bimodal structure in convection (Sikka and Gadgil 1980). These regions exhibit maxima in the variance of convection (Fig. 1b). The transition of convection from its oceanic to continental regime often involves intraseasonal northward propagation over India and the Bay of Bengal (Yasunari 1979, 1981; Gadgil 1990). Similar north-northwestward migration of convection is also observed over the western Pacific (e.g., Murakami et al. 1984; Lau and Chan 1986; Wang and Rui 1990). These northward and northwestward propagating components coexist with the eastward propagating Madden–Julian oscillation (MJO) that is weaker in summer than in winter (Julian and Madden 1981; Hendon and Salby 1994). These three propagating components dominate the 30–50-day intraseasonal variability during the ASM season (Lau and Chan 1986; Webster et al. 1998; Annamalai and Slingo 2001, hereafter AS01). In the present paper we refer to these collectively as the boreal summer intraseasonal variability (BSISV).
The subseasonal variability of the ASM is also modulated by the 10–20-day mode and higher frequency synoptic systems (Gadgil and Asha 1992). However, in the present study our analysis is concentrated on the 30–50-day time scales because it is associated with large-scale convective and circulation anomalies (AS01) and it explains more variance than the 10–20-day mode. Importantly, many investigations indicate that the BSISV can influence the seasonal mean rainfall over India (e.g., Fennessy and Shukla 1994; Ferranti et al. 1997; Sperber et al. 2000; Krishnamurthy and Shukla 2000) and also the statistics of synoptic systems over the ASM domain (Liebmann et al. 1994; Goswami et al. 2003). Yet, no study has attempted to understand how the regional heat sources associated with the three components of the BSISV influence each other and what implications such an interaction might have on the simulation and prediction of wet and dry spells of the monsoon.
Several mechanisms have been proposed for the northward propagation of convection over the Indian and west Pacific longitudes. They include (i) surface hydrology, (ii) air–sea interaction, and (iii) equatorial wave dynamics. While Webster (1983) emphasized the role of land surface heat fluxes in destabilizing the atmosphere ahead of the region of ascent, Goswami and Shukla (1984) stressed the importance of a convection–thermal relaxation feedback. Implementing some modifications to Webster’s model and incorporating the space–time variations in surface pressure, the model realistically simulated the persistence of the active phase over the Indian subcontinent (Nanjundiah et al. 1992). From an atmospheric general circulation model (AGCM) study, Ferranti et al. (1999) showed that feedbacks from the land surface are not necessary for poleward propagation, but the variance associated with the BSISV is enhanced when the hydrology is interactive.
Based on the fact that poleward propagation is more coherent over the ocean than over the land, both observational studies (e.g., Krishnamurti et al. 1988; Hendon and Glick 1997; Sperber et al. 1997: Sengupta et al. 2001; Kemball-Cook and Wang 2001; Vecchi and Harrison 2002; Sperber 2003) and coupled modeling studies (e.g., Waliser et al. 1999; Kemball-Cook et al. 2002) find that at intraseasonal time scales positive SST anomalies lead the convection by about 10 days. This lead/lag relationship appears to be a factor in the organization and poleward migration of convection over the Indian longitudes (Fu et al. 2003). In AGCMs forced by SST, such a lead/lag relationship is absent and may account for the poor simulation of the BSISV (Waliser et al. 2004).
Lau and Peng (1990) show that, as a result of the interaction of the large-scale monsoon flow with equatorial Kelvin waves, unstable baroclinic Rossby modes are generated over the monsoon region at about 15°–20°N. Krishnan et al. (2000) suggest that forcing by suppressed convection anomalies over the Bay of Bengal leads to the development of low-level anticyclonic circulation anomalies as a Rossby wave response, which then propagate northwestward to initiate the monsoon break over India. Lawrence and Webster (2002) show that the northward propagating convection is forced by surface frictional convergence into the low pressure center of the Rossby cell that is excited by equatorial convection. Wang and Xie (1997), Kemball-Cook and Wang (2001), and Jiang et al. (2004) point out that the emanation of Rossby waves is largely responsible for the northward migration of convection in the western Pacific and over India. On the other hand, Hsu and Weng (2001) suggest that the combination of ocean–atmosphere interaction and circulation–convection interactions lead to the northwestward propagation in the western Pacific.
Despite these advancements in our understanding of the BSISV, both AGCMs and fully coupled models are limited in their ability to simulate the space–time evolution of the BSISV over the entire ASM domain. For example, an AGCM forced by observed SST simulates the intraseasonal activity over the tropical west Pacific but not over the Indian longitudes; however, when the same AGCM is coupled to an ocean model the intraseasonal variability over the Indian longitudes is reasonably captured, but that over the tropical west Pacific is virtually lost (Kemball-Cook et al. 2002; Fu et al. 2003).
a. Quadrapole structure in convection
Figure 1c shows the variance explained by outgoing longwave radiation (OLR) at intraseasonal time scales during boreal summer. High variability is noticeable over (i) the equatorial Indian Ocean, (ii) the northern Indian Ocean including the Arabian Sea and the Bay of Bengal, (iii) the northern tropical west Pacific, and (iv) to a lesser extent over the equatorial west Pacific. Of the four regions, the northern tropical west Pacific exhibits the largest variance, although that over the Arabian Sea is largest when expressed as a percent of the total daily variance (Fig. 1d). Consistent with the variance distribution, during its life cycle the convection associated with the BSISV depicts a “quadrapole” structure over the ASM domain (e.g., Lau and Chan 1986; Krishnan et al. 2000; AS01); see also Fig. 2. Lau and Chan (1986) applied extended empirical orthogonal functions (EEOF), and AS01 used principal oscillation pattern (POP) analysis on filtered OLR and obtained the quadrapole structure, a north–south dipole over the Indian longitudes and a complementary pattern over the tropical west Pacific longitudes. AS01 suggested that the quadrapole arises due to the coexistence of the three propagating components of the BSISV. Sperber et al. (2000) identified a similar pattern that was found to control the intraseasonal and interannual variations of the Indian summer monsoon. The quadrapole structure in convection is a robust feature, as will be further demonstrated in section 3a.
Despite the recognition of the quadrapole structure, virtually all of the studies mentioned above examined the BSISV either over the Indian or over the west Pacific longitudes in “isolation.” For example, it is known that the convective activity over the equatorial Indian Ocean modulates the active–break cycles over India (e.g., Sikka and Gadgil 1980), but it is unknown if the former influences the active–break cycles over the west Pacific. Additionally, the importance of the suppressed phase of convection to the life cycle of the BSISV is unclear.
b. Present study
Here, no attempt is made to simulate the transient evolution of the BSISV nor to understand the possible mechanisms for the poleward propagation of convection. Rather, we seek a broader understanding of the three propagating components of the BSISV and investigate whether they influence each other. The null hypothesis is that they are mutually independent. Specifically, we examine if the convective and associated circulation anomalies over the west Pacific longitudes exert considerable influence on the convective and circulation anomalies over the Indian longitudes and vice versa. This premise is valid over the Tropics since convective processes generating the anomalous heat source are tightly coupled with the circulation. If the circulation anomaly excited by such a convective anomaly has a larger spatial scale than the heating, an additional convective anomaly may develop in response to the circulation anomaly and in turn force the atmospheric circulation. To examine our hypothesis, we use reanalysis products, satellite data, observed precipitation, a new diagnostic tool to capture the life cycle of the BSISV, and identify the relative impacts of local heating anomalies to the overall flow in an idealized model.
One approach to investigate the proposed hypothesis is to use a fully coupled ocean–atmosphere model, but it is difficult to control these models to perform conclusive sensitivity studies. Here, we use a simple linear model to identify and diagnose the wave dynamics due to the regional heat sources, first by forcing the model with prescribed heatings. To account for the strong interaction between convection and the circulation, model solutions are also sought for interactive heatings. These experimental setups provide the simplest approach to investigate the complex problem posed above. An important implication of the present study is that, if a mutual influence among the components exists, then a realistic simulation of the regional heat sources associated with them is a prerequisite for the successful simulation of the BSISV in numerical models.
The paper is organized as follows: In section 2 the data, the method used, a brief description of the simple atmospheric model, and the experimental designs are given. Section 3 describes the observed characteristics of the BSISV in terms of its evolution, air–sea interaction and Rossby wave dynamics, and the inferences from the model solutions. The major results and their implications for the BSISV, along with the limitations of our study, are summarized in section 5.
2. Data, method, model, and experimental design
a. Atmospheric data
The bulk of the atmospheric data used in our study is from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis for the period 1979–95 (Kalnay et al. 1996). [This period was chosen to correspond with our previous intercomparison of summer monsoon in the NCEP–NCAR reanalysis and the 15-yr European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-15; Annamalai et al. 1999). Additionally, this is the base simulation period of the second Atmospheric Model Intercomparison Project (AMIP II) for which we will analyze summer monsoon variability in a subsequent paper.] The data assimilation and forecast model was implemented operationally at NCEP in January 1995. The model is run at a horizontal resolution of T62 with 28 vertical levels. Upper-air data on standard pressure surfaces have been supplied on a 2.5° latitude–longitude grid. Surface and 24-h forecast fields (e.g., fluxes) are on the equivalent T62 Gaussian grid. Optimally interpolated SST of Reynolds and Smith (1994) were linearly interpolated to daily values.
The Advanced Very High Resolution Radiometer (AVHRR) OLR on the National Oceanic and Atmospheric Administration (NOAA) polar orbiting satellites has been used to identify the convective signature of the BSISV. These data have been daily averaged and processed on to a 2.5° latitude–longitude grid with missing values filled by interpolation (Liebmann and Smith 1996). Intraseasonal variations of rainfall are characterized using the pentad-based Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP). This dataset uses essentially the same algorithm and data sources as the monthly CMAP dataset described by Xie and Arkin (1996).
In addition, we use two observed daily rainfall datasets over India to demonstrate how faithfully the cyclostationary EOF (CsEOF) analysis constructed from OLR captures the active–break spells of the monsoon over India. These data are 1) the area-weighted daily all-India rainfall (AIR) time series for the period 1979–95, produced by the Indian Institute of Tropical Meteorology, and 2) the daily gridded rainfall over the Indian subcontinent for the period 1979–90. The daily AIR is constructed from the station rainfall data supplied by the Indian Meteorological Department (K. Rupakumar 2004, personal communication). The gridded dataset, originally compiled by Singh et al. (1992) based on daily rainfall at 365 stations were regridded to 1.25° latitude–longitude grid by M. Fennessy (2004, personal communication) of the Center for Ocean–Land–Atmosphere Studies (COLA).
b. Cyclostationary EOFs
Some disadvantages in using POP analysis include fitting a sinusoidal function to the POP spatial patterns to obtain the life cycle of the BSISV (AS01), and in the case of EEOF analysis more than one EEOF and their corresponding principal component (PC) time series are needed to obtain the evolution of the BSISV (e.g., Lau and Chan 1986; Waliser et al. 2004). Additionally, the EEOF matrix subjected to analysis is overly large when using daily data since padding of lagged information is required. Since the CsEOF operates on the cyclicity of the covariance function, the space–time evolution is directly derived from the analysis (Kim and North 1997).
Kim and Wu (1999) compared CsEOF analysis with other techniques commonly employed to extract space–time characteristics from geophysical time series such as EEOF, complex EOF, and POP. They found that CsEOF analysis captures the propagative features more realistically, in particular when the covariance of the given time series is cyclic in nature. For example, this technique has been successfully used to understand the propagative features of the ENSO (Kim 2003).
CsEOF differs from standard EOF techniques in that it incorporates the “cyclicity” in the covariance function. Consider a time series in which the first two moment statistics, mean and covariance, are periodic with period “d1” and “d2,” respectively. The cyclicity in the mean represents, for example, the annual cycle (d1 = 1 yr) and can be removed from the data. The cyclicity in the covariance function indicates that the given time series is cyclostationary rather than stationary. The life cycle of the BSISV is approximately 40 days (Lau and Chan 1986; Lawrence and Webster 2002; AS01) and therefore taken as the cyclicity in the covariance function (d2 = 40 days). In CsEOF analysis d2 is called the “nested period.” Similar to standard EOF computations, CsEOFs are obtained as eigenfunctions of the cyclic covariance function, and the periodic time dependence of the Bloch functions (eigenfunctions) is due to the periodic time dependence of the covariance function. Therefore, instead of one spatial pattern as in EOF analysis, a Bloch function represents a series of spatial patterns for the nested period, d2.
The procedure for calculating CsEOFs involves three steps. First, a set of EOFs is constructed from the state space (original data) that retains more than 90% of the variance. Second, for the nested period of 40 days, the Bloch functions are constructed from the PC time series in temporal space. Finally, the Bloch functions are scaled by the EOFs to project the intended phenomenon in physical space. To preserve the cyclic stationarity in the covariance, the Bloch functions or CsEOFs repeat every nested period, and the PCs associated with them represent the low-frequency modulation of the Bloch functions and are known as the “stochastic undulation” (Kim 2003). Hence, the PCs of the CsEOFs are not the appropriate base time series to extract the coherent evolution of other variables. A different procedure is adopted here and explained in section 3b. However, it is important to stress that the CsEOF spatial patterns that characterize the BSISV are important so that we may ascertain the validity of our alternate approach.
c. Linear baroclinic model and experimental design
Linear models have been widely used as a diagnostic tool to elucidate the atmospheric response to idealized forcing (e.g., Matsuno 1966; Webster 1972; Gill 1980; Rodwell and Hoskins 1996; Watanabe and Jin 2002). In most cases the forcing, the anomalous diabatic heating, is prescribed and the model’s steady state response is sought. Using this approach, which we call the “dry case,” Hendon and Salby (1996) reproduced many “salient characteristics” of the observed MJO. Importantly, Matthews et al. (2004) used a more complex model to “assess the cause of the circulation anomalies” in response to tropical heating associated with the MJO, finding that the “tropical anomalies are consistent with a forced Rossby–Kelvin wave response to tropical MJO heating.” Additionally, they noted that “the tropical and extratropical MJO anomalies are approximately linear . . . even though the response to the heating anomalies included the nonlinear dynamical terms.” These aforementioned results support our use of a linear model with prescribed heating to investigate the complex observed circulation patterns associated with the BSISV (section 3c).
On the other hand, feedback between convection and dynamics has been recognized as an important factor in the simulation and propagation of MJO (e.g., Emanuel 1987; Salby et al. 1994). One way to incorporate this positive feedback is to force a linear model with SST anomalies (e.g., Su et al. 2001; Watanabe and Jin 2003). In this approach, the model-generated heating anomaly and the circulation anomaly interact, potentially creating a new heating anomaly elsewhere. To overcome this shortcoming of the dry case, in the “moist case” we force the model with intraseasonal SST anomaly patterns derived from observations. An added benefit of the moist case is that validation of the moisture field with reanalysis provides for an additional point of comparison in determining how well the linear solutions support our hypothesis. The merits and caveats of our approach are discussed below.
Linear models employed by previous studies to understand the dynamics associated with the BSISV (e.g., Wang and Xie 1997; Krishnan et al. 2000) have only two levels in the vertical (850 and 200 hPa) where the imposed heating is entirely projected on the first baroclinic mode. Wu et al. (2000) demonstrated that a single vertical mode approximation is incapable of approximating the qualitative 3D structure of the response to tropical heating. Further, Mapes and Houze (1995) showed that the truncation of vertical modes greatly affects the amplitude of the simulated low-level winds. Therefore, we employ a multilevel linear model.
The model used here has been explained in detail by Watanabe and Kimoto (2000, 2001). Our configuration is similar to that used by Watanabe and Jin (2002; 2003). It is a global, time-dependent, primitive equation model linearized about the boreal summer climatology derived from NCEP–NCAR reanalysis for 1958–97. The horizontal resolution is triangular truncation T21 and there are 20 vertical levels using the sigma (σ) coordinate system. Since the model has multiple levels, it is referred to as the linear baroclinic model (LBM), and it has the advantage of extracting all the vertical modes. The model employs diffusion, Rayleigh friction, and Newtonian damping with a time scale of 1 day−1 for σ levels greater than or equal to 0.9 and those less than or equal to 0.03; 5 and 15 day−1 for the fourth and fifth levels; while 30 day−1 is used elsewhere. Rayleigh friction coefficients represent not only friction but also other damping such as high-frequency transient Reynolds stress (Wang 2005). The relatively weaker damping in the free atmosphere mimics nonlinearity in linear models (e.g., Ting and Yu 1998). To absorb vertically propagating waves a common approach is to incorporate strong damping at the top of the models (e.g., Hendon and Salby 1996; Wu et al. 2000). In linear models, the equatorial wave responses to prescribed heating depends on the selection of the damping coefficients since the zonal distance that the signals can reach is characterized by the product of the damping time scale and the zonal group speed (Wu et al. 2000). The LBM results to be discussed are not sensitive qualitatively to the damping rate unless the damping is unrealistically small. The values that we use at the model boundary layer are comparable or even stronger than those used by Wang and Xie (1997) and Krishnan et al. (2000).
In the dry case, the prescribed forcing is the anomalous diabatic heating proportional to the observed OLR (or precipitation) anomalies at various stages in the life cycle of the BSISV (Figs. 2 and 4). Lau and Peng (1987) and Sui and Lau (1989) showed that the zonal propagation speed of the forced signal is sensitive to the vertical structure of the prescribed heating. The vertical heating profile used is similar to that of Reed and Recker (1971) and has been employed in many previous studies (e.g., Rodwell and Hoskins 1996). To represent the top heavy heating associated with the BSISV (Mapes 2000), the maximum prescribed heating is located at 400 hPa (σ = 0.45). To maintain the divergence that promotes the baroclinic structure associated with MJO, Wang (2005) suggests that a minimum heating from ∼2–3 mm day−1 rainfall is necessary. In our case, for example, to mimic the positive heating over the equatorial Indian Ocean (Fig. 2d), the maximum heating imposed at 400 hPa was 3 K day−1 (18–20 W m−2 of OLR) or ∼5 mm day−1 of rainfall (Fig. 4d). The horizontal shape of the heating is elliptical (e.g., Fig. 6a) and the heating is imposed on the summer (June–September) mean climatology derived from NCEP–NCAR reanalysis. Further, the heating amplitude is not crucial because numerical experiments show that the 3D structure of the response is not sensitive to the amplitude of the heating as long as the maximum heating rate corresponds to less than 20 mm day−1 precipitation (Wu et al. 2000).
In the moist case, the forcing is the anomalous SST patterns derived from observations whose horizontal shape is assumed to be elliptical as in the dry case. The surface heat fluxes generated by this forcing are parameterized as in Betts and Miller (1986). A linearized moisture equation for the perturbation specific humidity is incorporated in the model. Heat and moisture sources associated with cumulus convection are also parameterized. In the moist case, the mean fields of SST and ground wetness are also included in the basic state. More details can be found in Watanabe and Jin (2003). In both dry and moist cases, linearization about the zonally varying ambient flow is considered. The LBM was integrated for 30 days with fixed forcing. With the dissipation terms adopted, the tropical response approaches a steady state after day 10 and the response at day 20 is analyzed.
As in many simple modeling studies, the present experimental designs have clear advantages with some limitations as well. For example, solutions from simple idealized models reveal the essential dynamics and suggest the relative contributions from the individual heating patterns to the overall flow. Further, one can diagnose the effect of one regional heating pattern on the circulation and on the heating patterns in other regions in the model. One caveat in our approach is that in the moist cases the prescribed SST anomalies are stationary in contrast to observations where the SST anomalies propagate and lead the convection by about 10 days. The suppression of the lead–lag relationship between SST and convection does not alter our conclusions since our focus here is not to simulate the transient evolution of the BSISV.
3. Life cycle of the BSISV
a. Cyclostationary EOFs
To isolate the variations related to the BSISV, the daily OLR has been bandpass filtered with a 20–100-day 100-point Lanczos filter. The CsEOF analysis is then applied to the filtered OLR over the ASM domain (30°S–30°N, 40°E–180°) with a nested period of 40 days. It has been found that the leading CsEOFs are insensitive to the domain chosen.
Figure 2 shows the dominant CsEOF mode of OLR explaining 11.2% variance. Since the chosen nested period is 40 days, each panel in Fig. 2 is separated by 5 days. This CsEOF mode describes the life cycle of the BSISV and is consistent with the results presented in Lau and Chan (1986) and AS01. The space–time pattern clearly indicates that this mode is associated with northward propagation over the Indian longitudes and northwestward migration of convection over the tropical west Pacific. One major difference is that CsEOF analysis captures the west-northwestward migration of the “convective maxima” associated with the Rossby waves over the Indian and west Pacific longitudes with more clarity than Wang and Xie (1997) and Lawrence and Webster (2002). The northward propagation occurs in conjunction with equatorial eastward propagation from the Indian Ocean to the west Pacific, characteristic of the MJO. One can easily delineate the coexistence of the three propagating components, and as such the intraseasonal modes are more complex during northern summer compared to northern winter. From observations, Lawrence and Webster (2002) found that 78% of the northward propagating events are accompanied with an eastward propagating MJO. Although the domain-averaged variance explained by this CsEOF is modest, the amplitude of the OLR anomalies over the eastern Indian Ocean at day 0 (Fig. 2d) is ∼18– 20 W m−2.
b. Linear regression and daily AIR variability
The CsEOF patterns in Fig. 2 give the life cycle of convection based on the specification of an inherent 40-day periodicity. As discussed in section 2b, each CsEOF pattern is associated with a low-frequency time series not appropriate for extracting covarying signals. To characterize the intraseasonal fluctuations, 20–100-day bandpass-filtered OLR is projected onto the CsEOF pattern in Fig. 2d. This pattern, hereafter referred to as day 0, is chosen since it is also recovered as the dominant mode in a conventional EOF analysis (not shown). The resulting principal component time series, hereafter referred to as PC-4, is used for lagged linear regression. The regressions are calculated for lags of ±25 days. The regressed fields are then scaled by a one standard deviation perturbation of PC-4, and plotted where the regression attains at least 5% significance assuming each pentad is independent as in Sperber et al. (1997) and Sperber (2003).
To understand how faithfully PC-4 captures the temporal and spatial patterns associated with active–break phases over India, we calculated the lag correlations (±25 days) of PC-4 with observed daily AIR series and the gridded rainfall data over India, respectively. Prior to the calculation, both datasets were smoothed by a 9-day running mean to suppress high frequency synoptic variability. For the temporal pattern, for each year we checked the maximum positive correlation (R+) and the minimum negative correlation (R−), and have plotted the larger of the square of either R+ or R− in Fig. 3a. We consider both positive and negative correlations since during years when the BSISV is active PC-4 and AIR exhibit a characteristic lead–lag relationship (not shown) that oscillates between positive and negative extremes due to the cyclicity of the BSISV. Assuming 12 degrees of freedom (R > = |0.532|) the relationship is significant at the 5% level, particularly in the years when the BSISV is very active (e.g., 1979). For the spatial pattern, the PCs corresponding to the active phase at day 15 (Fig. 2g) and the break phase at day −5 (Fig. 2c) were regressed onto the observed gridded India rainfall, and the difference is shown in Fig. 3b. Typically, during active phases rainfall is increased over the west coast, central, and northern India while it is decreased over southeast and northeast India (e.g., Ramamurthy 1969; Krishnan et al. 2000). Additionally, this pattern is consistent with those of Krishnamurthy and Shukla (2000, their Fig. 9a), Sperber et al. (2000, their Figs. 8c, 9, and their Table 1), and Goswami and Ajayamohan (2001, their Fig. 5d) using different methods and indices. That the difference pattern in Fig. 3b looks identical to the classical rainfall departures during active phases reassures us that our CsEOF analysis captures both the large-scale and regional-scale BSISV rainfall variations.
c. Linear regression and LBM solutions
Figure 4 shows the regressions of PC-4 against 850-hPa winds and CMAP rainfall. Overall, the 850-hPa wind anomalies are dynamically consistent with the precipitation anomalies. Briefly, the active phase of the Indian monsoon develops as the anomalous cyclonic circulation, in the form of Rossby waves, moves north-northwestward. The equatorial rainfall enhancement moves into the western Pacific from day 0 to day 10 as the active monsoon over India matures (Figs. 4d,f). The active monsoon over India (Figs. 4f,g) is characterized by stronger cross-equatorial flow and cyclonic vorticity over the northern Indian Ocean. The extension of convection into the western Pacific has been preceded by above-normal SST and enhanced convergence and 850-hPa moisture (not shown). As over the Indian monsoon region, cyclonic vorticity over the tropical west Pacific dominates during the active phase. A near mirror image in convection and circulation anomalies exists during the break monsoon phases. Using filtered OLR, Lawrence and Webster (2002) constructed a regional time series by averaging the data over (0°–5°N, 85°–90°E). This regional index was used to regress against other variables, but did not reveal the covarying signals over the tropical west Pacific. Other studies (e.g., Kemball-Cook et al. 2002) show latitude–time Hovmöller plots averaged over certain longitudes to infer the poleward propagation. A benefit of the present analysis is that PC-4 represents the convective activity over the entire ASM domain.
Despite this close association between convection and circulation anomalies, it is rather difficult to test the null hypothesis from observations alone. As in the model of Gill (1980), a basic assumption in the dry case is that the circulation arises solely due to atmospheric heating. In the traditional Gill solution for symmetric equatorial heating, the steady-state response consists of westerlies at and to the west of the heating, with easterlies at the leading edge extending to the east. This is accompanied by cyclonic flow to the northwest and southwest of the heating. While the system that we are investigating is not in steady-state equilibrium, examination of the rainfall and 850-hPa wind anomalies for day −10 through day 0 in Figs. 4b–d suggests that heating due to convection is modifying the flow. At day −10 easterly anomalies prevail along the equatorial Indian Ocean. The westerly anomalies are observed only after the convection strengthens and persists through day 0, and then the system evolves to resemble the steady-state Gill response, similar to the modeling result of Jin and Hoskins (1995). From observations, Hendon and Salby (1996) noted a similar lagged dynamical response during the life cycle of the winter MJO and, they found “In the Eastern Hemisphere, the dynamical response involves equatorial Kelvin structure and Rossby gyres that straddle anomalous heating as it migrates eastward along the equator.” Here, the solutions from the prescribed heatings are interpreted as the “forced response.” Even so, this is an idealized view of an otherwise convective–dynamical interactive system (Emanuel 1987) in which dynamically induced heating is absent (Salby et al. 1994). We will demonstrate that the major results from the dry cases are reproducible using the moist cases in which the atmospheric heating is model generated and from which the moisture distributions further support our hypothesis.
Though the amplitude of the wind anomalies from the dry cases is stronger than those from the regressions, this does not influence our interpretation since the magnitude of the response, but not the spatial pattern, is proportional to the strength of the imposed heating. The eight panels in Figs. 2 and 4 describe the entire life cycle of the BSISV, and we obtained solutions for each of the heating anomalies in those panels. For brevity, we show the results at selected time lags that reveal new findings such as (i) initiation of convection in the western Indian Ocean, (ii) circulation anomalies forced by west Pacific convection on the active/break phases over India, and (iii) how the convective anomalies over the equatorial Indian Ocean influence the active–break phases over the tropical west Pacific.
1) Initiation of convection over the western Indian Ocean
At day −15 the initiation of convection occurs over the western Indian Ocean, although suppressed convection over the equatorial Indian Ocean is the prominent feature (Figs. 2a and 4a). In conjunction with the suppressed convection low-level twin anticyclones are present on either side of the equator (Figs. 4a and 5a). The Southern Hemispheric Rossby component is weaker than its Northern Hemispheric counterpart due to the asymmetry in the mean flow during the boreal summer (e.g., Krishnan et al. 2000).
Over the near-equatorial Indian Ocean (poleward of 2.5°S) the mean climatological winds are westerly (e.g., Annamalai et al. 1999) and the SST over the equatorial central–eastern Indian Ocean is >28°C, the threshold required for deep convection (Gadgil et al. 1984). The easterly wind anomalies there (Figs. 4a and 5a,b) act against the mean westerly flow and reduce evaporation (Fig. 5b). Similarly, the alongshore northerly anomalies reduce evaporation (Fig. 5b). This, in conjunction with increased net surface shortwave radiation (not shown), contributes to the development of positive SST anomalies over much of the Indian Ocean (Fig. 5a).1 Consistent with the observations of Maloney and Hartmann (1998, their Figs. 3 and 4) and Sperber (2003, his Fig. 13) for boreal winter, low-level convergence anomalies and increased lower-tropospheric water vapor occur near East Africa at intraseasonal time scales when suppressed convective anomalies and below-normal moisture occur over the equatorial Indian Ocean (Figs. 5c,d). The increased moisture near Africa is confined to the lower–middle troposphere and is likely associated with shallow convection, while the suppressed convection to the east is associated with below-normal moisture through the depth of the troposphere (Fig. 5e).
We now use the LBM to examine how well these features can be represented by linear dynamics. Based on the rainfall anomalies in Fig. 4a, we show the vertical integral of the heatings used for the “dry” simulation (Fig. 6a). As seen in Fig. 6b, the solution using the negative heating reproduces many of the observed features over the Indian Ocean seen in reanalysis (Fig. 5). Notably, these include 1) the twin anticyclones over the western/central Indian Ocean (cf. Figs. 6b and 4a), 2) enhanced convergence adjacent to the east coast of Africa, and 3) divergence anomalies from the southern Arabian Sea to the Malaysian Peninsula (cf. Figs. 6b and 5f). With the inclusion of the positive heating over the west Pacific in Fig. 6c, the solution of all heatings, Fig. 6d, shows improved consistency with reanalysis, especially over the west Pacific.
Figure 7 shows the results from the LBM when it is run in the moist configuration. To generate negative heating rates over the central–eastern Indian Ocean (Fig. 7a) consistent with the observed below-normal rainfall, negative SST anomalies over this region need to be imposed in the model. Importantly, an examination of the vertical profiles of the model-generated heatings over the regions of interest have maximum amplitude at about 400 hPa with minima near the surface and at the model’s top level (not shown), consistent with the top-heavy profile imposed in the dry simulations. This result, and those discussed below, validate the heating profile used in the dry model.
Remote responses to the local SST forcing are also apparent. Near the coast of Africa positive heating anomalies are generated (Fig. 7a), accompanied by convergence anomalies at 850 hPa (Fig. 7b). This result is consistent with the aquaplanet modeling experiments of Neale and Hoskins (2000) who showed that a localized negative (positive) SST anomaly will lead to enhancement (suppression) of the equatorial convection to the west as a result of the dynamical response to the heating associated with the suppressed (enhanced) convection. The large-scale patterns in Fig. 7b are consistent with the dry model solution (Figs. 6b,d) and they resemble the wind and divergence anomalies in the reanalysis (Figs. 4a and 5f) during the initiation of convection over the western Indian Ocean. Importantly, over the Indian Ocean the 850-hPa specific humidity (Fig. 7c) and its vertical structure from 10°S to 5°N (Fig. 7d) compare favorably to reanalysis (Figs. 5d,e, respectively). This result indicates that the dynamical response over the Indian Ocean is similar irrespective of whether an anomalous heat source is imposed or in the case where the heat source and dynamics are interactive. Our model results bear close similarity to Matthews (2004) who found that the Rossby wave response to suppressed convection was an important component in the onset of intraseasonal convection over Africa. Similarly, during boreal winter Sperber (2003) implicated a Rossby wave signal associated with suppressed heating over the equatorial Indian Ocean in the boreal winter onset of MJO convection in the western Indian Ocean. Thus, Rossby waves associated with tropical intraseasonal heating are a ubiquitous component of the climate system, and they appear to play a role in the development of enhanced convection to the west of suppressed convection.
While not germane to the initiation of convection over the western Indian Ocean, the positive heating anomalies generated over the South China Sea provide additional evidence of a mutually interactive system (Fig. 7a). However, the dry model suggests that the teleconnectivity of the negative heating over the equatorial Indian Ocean is greater than that for the positive heating over the tropical west Pacific (cf. Figs. 6b,c).
2) Active (dry) phase over the tropical west Pacific (equatorial Indian Ocean) and break initiation over India
Figure 8 shows the regressions of PC-4 with observations and reanalysis, and Figs. 9 and 10 show the model solutions for days 10 and 20 with prescribed heatings, respectively. At day 10 observations show the enhanced rainfall to have a tilted structure (Fig. 8a). In the vicinity of India this is associated with 850-hPa wind anomalies over the Bay of Bengal that recurve over northern India. In the LBM the positive heating over the Bay of Bengal gives rise to the cyclonic anomalies over India (Fig. 9a), though weak easterly and southeasterly wind anomalies over the Bay of Bengal are contributed from the positive heating over the equatorial west Pacific (Fig. 9b) and the negative heating over the tropical west Pacific (Fig. 9c), respectively. At this time, enhanced moisture over the Arabian Sea dominates the inflow into India in the reanalysis (Fig. 8b).
By day 20 the cross-equatorial flow adjacent to Africa weakens and the available moisture at 850 hPa over the Arabian Sea has diminished (Figs. 8c,d). The anomalous moisture signals over the Arabian Sea and southern Indian Ocean between 10° and 20°S at day 10 and day 20 are in agreement with the results of Cadet and Greco (1987) who investigated the moisture budgets during the active/break phases of the 30–50-day mode during the 1979 monsoon season. With the transition to westerly anomalies over India and the Bay of Bengal by day 20, the monsoon trough over India collapses (Fig. 8c), and enhanced moisture shifts eastward (Fig. 8d). Raghavan (1973) also noted the presence of westerlies over the plains of northern India during the initiation of monsoon breaks. From day 10 to day 20, suppressed convection persists and intensifies in the equatorial central/eastern Indian Ocean (Fig. 8c). In the LBM, it was shown previously that with negative heating over the equatorial Indian Ocean twin anticyclones are present to the west (Fig. 6b). These wind anomalies are consistent with the reduction of the cross-equatorial flow and the penetration of westerly anomalies over the Bay of Bengal and the western Pacific, seen in the reanalysis. Additionally, the LBM suggests two important contributions to the initiation of break over India from the positive heating over the tropical west Pacific. Firstly, it is accompanied by cyclonic vorticity (Figs. 10a,b), which relative to day 10 is now displaced farther east, contributing to the breakdown of the monsoon trough over the Bay of Bengal. Secondly, the positive heating over the tropical west Pacific forces descent anomalies over India (Fig. 10c) that would impede convective activity.
The moist solution in Fig. 11 confirms the result from the dry model version, and additionally depicts a mutually interactive system. Warm SST anomalies over the tropical west Pacific generate in situ positive heating (Fig. 11a) and low-level cyclonic vorticity anomalies (Fig. 11b). Remotely, negative heating anomalies near India and Southeast Asia and over the near-equatorial Indian Ocean are generated. This negative heating would be a positive feedback to suppressed convection that might otherwise be present (such as during the initiation of the next active phase of convection over the western Indian Ocean, Figs. 5 –7). The vertical velocity over India (Fig. 11c) shows strong descending motion through the depth of the troposphere over India, which is consistent, but much stronger than that produced by the dry model (Fig. 10c).
Bhat et al. (2001) and Rao et al. (2004) used the Bay of Bengal Monsoon Experiment (BOBMEX) data and satellite measurements, respectively, to show that the upper troposphere dries a few days before the commencement of monsoon breaks over India. This feature is also seen in the reanalysis, Fig. 12, which shows the vertical cross section of the lead–lag regressions of PC-4 with specific humidity and vertical velocity at 92.5°E averaged between 15° and 20°N. Apart from the fact that active (break) phases are associated with deep moist (dry) layer, consistent with other measurements, the upper troposphere dries well in advance of the break as the vertical velocity anomalies transition from upward to downward. Despite uncertainties in the humidity data from NCEP–NCAR reanalysis (e.g., Trenberth and Guillemot 1998) the results presented in Fig. 12 deserve attention. Raghavan (1973) and Bhat et al. (2001) speculated that the drying of the upper troposphere due to forced descent anomalies is the primary reason for the inhibition of convection during break phases. Our model results suggest that Rossby waves forced by active convection over the tropical west Pacific is a possible source for the descent anomalies over India. The negative heating over the tropical Indian Ocean contributes to descent in the upper troposphere over India (Fig. 10c), suggesting that it may help cap off deep convection over India in addition to reducing the cross-equatorial flow.
Due to the consistency in the solutions, to further illustrate the mechanisms in the LBM and to assess the relative role of west Pacific versus equatorial Indian Ocean in generating subsidence over India we show the evolution of the geopotential anomalies at 700 hPa from the dry model for the positive heating over the tropical west Pacific and the negative heating over the equatorial Indian Ocean (Figs. 13 and 14, respectively). In the tropical west Pacific solution, within 3 days of the imposed heating in situ negative height anomalies are well developed and are compensated by positive height anomalies to the east and west. During the course of the LBM integration, the Kelvin wave component radiates eastward, while the positive height anomalies associated with the forced Rossby waves propagate westward and cover the entire Indian subcontinent and Arabian Sea by day 6 of the simulation. As the negative height anomalies intensify over the tropical west Pacific, the Rossby waves continue to amplify attaining local height anomalies >2 m over the Arabian Sea, and the pattern attains a steady state by simulation day 12. The interpretations presented here have close similarity to those of Rodwell and Hoskins (1996).
In response to the equatorial Indian Ocean heating anomalies (Fig. 14), the in situ Kelvin–Rossby wave packet develops with the positive height anomalies over the equatorial Indian Ocean compensated by negative height anomalies over the Maritime Continent. By simulation day 6, the symmetric Rossby waves over the Indian Ocean begin their westward journey, while the Kelvin wave generates positive height anomalies over the Indonesian islands. During the evolution, the Northern Hemisphere Rossby wave response intensifies and attains a maximum amplitude of 1.5 m over the Arabian Sea by day 12. The Southern Hemisphere response weakens, primarily due to the asymmetry in the mean flow. The positive height anomalies in the Northern Hemisphere are strongest over southern peninsular India and the Arabian Sea.
Krishnan et al. (2000) showed that northwest propagating Rossby waves, triggered by suppressed convective anomalies over the Bay of Bengal, initiate a break over continental India. However, their mechanism becomes effective once suppressed convective anomalies get organized and spread over the Bay of Bengal (e.g., Fig. 4a at day −15, or Figs. 4d,e of their paper). The LBM solutions presented here suggest that, if one extends the analysis domain into the west Pacific and the effect of suppressed convective anomalies over the equatorial Indian Ocean at day 20, the initiation of breaks over India can be diagnosed 5–10 days earlier. In summary, the coexistence of forced descent anomalies (Figs. 10c and 11c), drying from above (Fig. 12), and reduced moisture due to weakened cross-equatorial flow (Fig. 8d) provides a plausible mechanism for initiating break conditions over India.
3) Role of equatorial mode in the active–break phases over the tropical west Pacific
The LBM also suggests that the equatorial Indian Ocean heating plays an important role for the amplification and reduction of convection over the tropical west Pacific. The day −15 model solution in Fig. 6b indicates that westerly wind anomalies forced by equatorial Indian Ocean suppressed convective anomalies extend zonally into the west Pacific and strengthen the convergence and therefore the ongoing active phase there. When compared with the regression map in Fig. 4a, the model solutions suggest that the low-level westerly anomalies over the Bay of Bengal and north Arabian Sea are associated with equatorial Indian Ocean suppressed convective anomalies and are not forced by west Pacific convective anomalies (cf. Figs. 6b,c). To understand the linear dynamics we return to the model solution for suppressed connection over the equatorial Indian Ocean (Fig. 14). By day 3 of the simulation, the below-normal height anomalies over the Maritime Continent begin to be perturbed by an eastward-moving Kelvin wave as westward propagating symmetric Rossby waves develop. As the solution approaches steady state, the Kelvin wave signal continues to radiate eastward while the Rossby wave propagates northwestward over the tropical west Pacific. In a theoretical study, Lau and Peng (1990) showed that during northern summer, when the equatorial Kelvin waves reach the monsoon region in the vicinity of 140°–150°E, unstable baroclinic Rossby modes are initiated at about 15°–20°N. The geopotenial perturbations in Fig. 14 are consistent with their results. In addition, under the presence of mean vertical easterly shear, these Rossby waves are confined to the lower troposphere and influence the low-level convergence (Figs. 5f, 5c, 6b and 6d) (Wang and Xie 1996), and hence the local convection (Fig. 7a).
All of the model results presented here occur when a zonally varying mean flow is prescribed in the model. When zonal mean flow is prescribed the growth of Rossby waves over the monsoon region is inhibited (not shown) due to the absence of easterly shear, consistent with the conclusions of Lau and Peng (1990) and Wang and Xie (1997).
4. Summary and discussion
The BSISV, associated with the 30–50-day mode, is represented by the coexistence of three components, the poleward propagation of convection over the Indian and tropical west Pacific regions and eastward propagation along the equatorial Indian and west Pacific Oceans (Fig. 2). In the present study we investigate whether the three propagating components mutually influence each other using observational data and solutions from an idealized linear model. Our null hypothesis is that the three components are independent.
The space–time evolution of the BSISV was extracted using CsEOF analysis on filtered OLR (Fig. 2), and the base PC time series, used for linear regressions against ocean–atmosphere variables, is significantly related to the daily observed rainfall over India (Fig. 3) and also captures the space–time evolution of convection over the ASM domain (Fig. 4). The dry model solutions, forced with heating proportional to the observed life cycle of OLR, capture many of the observed characteristics and have enabled us to suggest the relative contributions of the regional heat sources to the overall anomalous circulation. In particular, the low-level circulation anomalies forced by the individual heat sources result in convergence (divergence) anomalies in remote regions but largely within the ASM domain (Figs. 6, 9, 10). However, due to the strong interactive nature between convection and circulation in the Tropics, additional experiments were conducted where the model was forced by observed SST anomalies (moist case) and the solutions (Figs. 7, 11) are nearly identical to those from the dry cases. An added benefit of the latter solutions is the validation of the moisture field with the reanalysis that further supports our hypothesis.
Due to their consistency with reanalysis, the model solutions suggest three major findings. First, the Rossby wave response to suppressed convection over the central/eastern Indian Ocean (i.e., from the previous break phase of the BSISV) aids in the initiation of convection near the African coast as the next active cycle begins (Figs. 5 –7). This result is consistent with observations of the boreal winter MJO (e.g., Sperber 2003). Second, descent anomalies forced by active convection over the tropical west Pacific and the reduction of cross-equatorial flow due to the developing suppressed convection over the equatorial Indian Ocean trigger the break monsoon conditions over India (Figs. 9 –11). Finally, the circulation anomalies forced by equatorial Indian Ocean convective anomalies play a role in the active and break phases over the tropical west Pacific and vice versa (Fig. 11). In summary, based on the LBM results, the null hypothesis that the three components of the BSISV are independent is rejected. Our model solutions suggest that such an influence will be a positive feedback to the active or suppressed convection that might otherwise be present, but do not imply that the influence is responsible for the space–time evolution of the BSISV.
However, there are caveats that temper our results. First, we have assumed that linear interactions are sufficient to address the BSISV and, second, the role of air–sea interaction has not been considered in the model, nor has transient forcing. Thus, while the major features of the large-scale flow are captured by the linear model in both dry and moist configurations, the relative importance of the regional heat sources suggested by the LBM may have limited applicability to the observed system. With the aforementioned caveats in mind, the results presented herein suggest that the regional heating anomalies and their teleconnections need to be understood and that they are a basic element for a realistic simulation of the BSISV. We are currently examining the BSISV in a suite of coupled models and find that the features reported here are highly relevant for the simulation of BSISV.
Acknowledgments
The authors express their gratitude to Dr. Masahiro Watanabe for providing the linear model and offering many useful suggestions on its use. Dr. Yoo Yin Kim is thanked for the cyclostationary EOF routine. Dr. Rupakumar, Indian Institute of Tropical Meteorology, provided the daily AIR time series and Dr. Mike Fennessy (Center for Ocean–Land–Atmosphere Studies) provided the gridded rainfall data. Drs. Brian Mapes, Ragu Murtugudde, P. V. Joseph, Masahiro Watanabe, and Xianan Jiang are acknowledged for comments on the draft version of the manuscript. This research is partly funded by the Japan Agency for Marine–Earth Science and Technology (JAMSTEC) through its sponsorship of the International Pacific Research Center (IPRC). K. R. Sperber was supported by the U.S. Department of Energy, Office of Science, Climate Change Prediction Program at the University of California Lawrence Livermore National Laboratory under Contract W-7405-ENG-48. Comments from Dr. Kerry Emanuel, the editor, and the anonymous reviewers helped improve the manuscript.
REFERENCES
Annamalai, H., and J M. Slingo, 2001: Active/break cycles: Diagnosis of the intraseasonal variability over the Asian summer monsoon. Climate Dyn., 18 , 85–102.
Annamalai, H., J M. Slingo, K R. Sperber, and K. Hodges, 1999: The mean evolution and variability of the Asian summer monsoon: Comparison of ECMWF and NCEP–NCAR reanalyses. Mon. Wea. Rev., 127 , 1157–1186.
Betts, A., and M. Miller, 1986: A new convective adjustment scheme. Part II: Single column test using GATE wave, BOMEX, ATEX, and arctic air-mass data sets. Quart. J. Roy. Meteor. Soc., 112 , 693–709.
Bhat, G S., and Coauthors, 2001: BOBMEX: The Bay of Bengal Monsoon Experiment. Bull. Amer. Meteor. Soc., 82 , 2217–2243.
Cadet, D L., and S. Greco, 1987: Water vapor transport over the Indian Ocean during the 1979 summer monsoon. Part I: Water vapor fluxes. Mon. Wea. Rev., 115 , 653–663.
Emanuel, K A., 1987: An air–sea interaction model of intraseasonal oscillations in the Tropics. J. Atmos. Sci., 44 , 2324–2340.
Fennessy, M J., and J. Shukla, 1994: GCM simulations of active and break monsoon periods. Proc. Int. Conf. on Monsoon Variability and Prediction, WCRP-84, WMO/TD-619, Vol. 2, Trieste, Italy, WMO, 576–585.
Ferranti, L., J M. Slingo, T N. Palmer, and B J. Hoskins, 1997: Relations between interannual and intraseasonal monsoon variability as diagnosed from AMIP integrations. Quart. J. Roy. Meteor. Soc., 123 , 1323–1357.
Ferranti, L., J M. Slingo, T N. Palmer, and B J. Hoskins, 1999: The effect of land surface feedbacks on the monsoon circulation. Quart. J. Roy. Meteor. Soc., 125 , 1527–1550.
Fu, X., B. Wang, T. Li, and J P. McCreary, 2003: Coupling between northward-propagating, intraseasonal oscillations and sea surface temperature in the Indian Ocean. J. Atmos. Sci., 60 , 1733–1753.
Gadgil, S., 1990: Poleward propagation of the ITCZ: Observations and theory. Mausam, 41 , 285–290.
Gadgil, S., and G. Asha, 1992: Intraseasonal variation of the summer monsoon. I: Observational aspects. J. Meteor. Soc. Japan, 70 , 517–527.
Gadgil, S., P V. Joseph, and N V. Joshi, 1984: Ocean–atmosphere coupling over monsoon regions. Nature, 312 , 141–143.
Gill, A E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106 , 447–462.
Goswami, B N., and J. Shukla, 1984: Quasiperiodic oscillations in a symmetric general circulation model. J. Atmos. Sci., 41 , 20–37.
Goswami, B N., and R S. Ajayamohan, 2001: Intraseasonal oscillation and interannual variability of the Indian summer monsoon. J. Climate, 14 , 1180–1198.
Goswami, B. N., R. S. Ajayamohan, P. K. Xavier, and D. Sengupta, 2003: Clustering of synoptic activity by Indian summer monsoon intraseasonal oscillations. Geophys. Res. Lett., 30 .1431, doi:10.1029/2002GL016734.
Hendon, H., and M L. Salby, 1994: The life cycle of the Madden–Julian oscillation. J. Atmos. Sci., 51 , 2207–2219.
Hendon, H., and M L. Salby, 1996: Planetary-scale circulations forced by intraseasonal variations of observed convection. J. Atmos. Sci., 53 , 1751–1758.
Hendon, H., and J. Glick, 1997: Intraseasonal air–sea interaction in the tropical Indian and Pacific Oceans. J. Climate, 10 , 647–661.
Hsu, H H., and C H. Weng, 2001: Northwestward propagation of the intraseasonal oscillation in the western North Pacific during the boreal summer: Structure and mechanism. J. Climate, 14 , 3834–3850.
Jiang, X., T. Li, and B. Wang, 2004: Structures and mechanisms of the northward-propagating boreal summer intraseasonal oscillation. J. Climate, 17 , 1022–1039.
Jin, F-F., and B J. Hoskins, 1995: The direct response to tropical heating in a baroclinic atmosphere. J. Atmos. Sci., 52 , 307–319.
Julian, P R., and R A. Madden, 1981: Comments on a paper by T. Yasunari, “A quasi-stationary appearance of a 30 to 40 day period in the cloudiness fluctuations during the summer monsoon over India. J. Meteor. Soc. Japan, 59 , 435–437.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Kemball-Cook, S., and B. Wang, 2001: Equatorial waves and air–sea interaction in the boreal summer intraseasonal oscillation. J. Climate, 14 , 2923–2942.
Kemball-Cook, S., B. Wang, and X. Fu, 2002: Simulation of the intraseasonal oscillation in the ECHAM-4 model: The impact of coupling with an ocean mode. J. Atmos. Sci., 59 , 1433–1453.
Kim, K Y., 2003: Investigation of ENSO variability using cyclostationary EOFs of observational data. Meteor. Atmos. Phys., 81 , 149–168.
Kim, K Y., and G R. North, 1997: EOFs of harmonizable cyclostationary processes. J. Atmos. Sci., 54 , 2416–2427.
Kim, K Y., and Q. Wu, 1999: A comparison study of EOF techniques: Analysis of nonstationary data with periodic statistics. J. Climate, 12 , 185–199.
Krishnamurthy, V., and J. Shukla, 2000: Intraseasonal and interannual variability of rainfall over India. J. Climate, 13 , 4366–4377.
Krishnamurti, T N., D. Dosterhof, and A. Mehta, 1988: Air–sea interaction on the time scale of 30–50 days. J. Atmos. Sci., 45 , 1304–1322.
Krishnan, R., C. Zhang, and M. Sugi, 2000: Dynamics of breaks in the Indian summer monsoon. J. Atmos. Sci., 57 , 1354–1372.
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.
Lau, K M., and L. Peng, 1987: Origin of low frequency (intraseasonal) oscillations in the tropical atmosphere. Part I: Basic theory. J. Atmos. Sci., 44 , 950–972.
Lau, K M., and L. Peng, 1990: Origin of low frequency (intraseasonal) oscillations in the tropical atmosphere. Part III: Monsoon dynamics. J. Atmos. Sci., 47 , 1443–1462.
Lawrence, D M., and P J. Webster, 2002: The boreal summer intraseasonal oscillation: Relationship between northward and eastward movement of convection. J. Atmos. Sci., 59 , 1593–1606.
Liebmann, B., and C A. Smith, 1996: Description of a complete (interpolated) OLR dataset. Bull. Amer. Meteor. Soc., 77 , 1275–1277.
Liebmann, B., H H. Hendon, and J D. Glick, 1994: The relationship between tropical cyclones of the western Pacific and Indian Ocean and Madden–Julian oscillation. J. Meteor. Soc. Japan, 72 , 401–412.
Maloney, E D., and D L. Hartmann, 1998: Frictional moisture convergence in a composite life cycle of the Madden–Julian oscillation. J. Climate, 11 , 2387–2403.
Mapes, B E., 2000: Convective inhibition, subgrid-scale triggering energy, and stratiform instability in a simple tropical wave model. J. Atmos. Sci., 57 , 1515–1535.
Mapes, B E., and R A. Houze, 1995: Diabatic divergence profiles in western Pacific mesoscale convective systems. J. Atmos. Sci., 52 , 1807–1828.
Matsuno, T., 1966: Quasi-geostrophic motions in the equatorial area. J. Meteor. Soc. Japan, 44 , 25–42.
Matthews, A J., 2004: Intraseasonal variability over tropical Africa during northern summer. J. Climate, 17 , 2427–2440.
Matthews, A J., B J. Hoskins, and M. Masutani, 2004: The global response to tropical heating in the Madden–Julian oscillation during northern winter. Quart. J. Roy. Meteor. Soc., 130 , 1991–2011.
Murakami, T., T. Nakazawa, and J. He, 1984: On the 40–50 day oscillations during the 1979 Northern Hemisphere summer. I: Phase propagation. J. Meteor. Soc. Japan, 62 , 440–468.
Nanjundiah, R S., J. Srinivasan, and S. Gadgil, 1992: Intraseasonal variation of the Indian summer monsoon. II: Theoretical aspects. J. Meteor. Soc. Japan, 70 , 529–550.
Neale, R B., and B J. Hoskins, 2000: A standard test for AGCMs including their physical parametrizations. II: Results from the Met. Office Model. Atmos. Sci. Lett, 1 .doi:10.1006/asle.2000.0020.
Raghavan, K., 1973: Break-monsoon over India. Mon. Wea. Rev., 101 , 33–43.
Ramamurthy, K., 1969: Some aspects of “break” in the Indian southwest monsoon during July and August. Indian Meteorological Department Publication, Forecasting Manual, FMU, Rep. 4, 18.3, 57 pp. [Available from India Meteorological Department, Mausam Bhavan, Lodi Road, New Delhi 110 003, India.].
Rao, K G., M. Desbois, and R. Roca, 2004: Upper tropospheric drying and the transition to break in the Indian summer monsoon during 1999. Geophys. Res. Lett., 31 .L03206, doi:10.1029/2003GL018269.
Reed, R J., and E E. Recker, 1971: Structure and properties of synoptic-scale wave disturbances in the equatorial western Pacific. J. Atmos. Sci., 28 , 1117–1133.
Reynolds, R W., and T M. Smith, 1994: Improved global sea surface temperature analyses using optimum interpolation. J. Climate, 7 , 929–948.
Rodwell, M J., and B J. Hoskins, 1996: Monsoons and the dynamics of deserts. Quart. J. Roy. Meteor. Soc., 122 , 1385–1404.
Salby, M L., R R. Garcia, and H H. Hendon, 1994: Planetary-scale circulations in the presence of climatological and wave-induced heating. J. Atmos. Sci., 51 , 2344–2367.
Sengupta, D., B N. Goswami, and R. Senan, 2001: Coherent intraseasonal oscillations of the ocean and atmosphere during the Asian summer monsoon. Geophys. Res. Lett., 28 , 4127–4130.
Sikka, D R., and S. Gadgil, 1980: On the maximum cloud zone and the ITCZ over Indian longitudes during the southwest monsoon. Mon. Wea. Rev., 108 , 1840–1853.
Singh, S V., R H. Kripalani, and D R. Sikka, 1992: Interannual variability of the Madden–Julian oscillation in the Indian summer monsoon rainfall. J. Climate, 5 , 973–978.
Sperber, K R., 2003: Propagation and the vertical structure of the Madden–Julian oscillation. Mon. Wea. Rev., 131 , 3018–3037.
Sperber, K R., J M. Slingo, P M. Inness, and W-K. M. Lau, 1997: On the maintenance and initiation of the intraseasonal oscillation in the NCEP/NCAR reanalysis and in the GLA and UKMO AMIP simulations. Climate Dyn., 13 , 769–795.
Sperber, K R., J M. Slingo, P M. Inness, and H. Annamalai, 2000: Predictability and the relationship between subseasonal and interannual variability during the Asian summer monsoon. Quart. J. Roy. Meteor. Soc., 126 , 2545–2574.
Su, H., J D. Neelin, and C. Chou, 2001: Tropical teleconnection and local response to SST anomalies during the 1997–98 El Nino. J. Geophys. Res., 106 , 20025–20043.
Sui, C H., and K M. Lau, 1989: Origin of low-frequency (intraseasonal) oscillations in the tropical atmosphere. Part II: Structure and propagation of mobile wave-CISK modes and their modification by lower boundary forcing. J. Atmos. Sci., 46 , 38–56.
Ting, M., and L. Yu, 1998: Steady response to tropical heating in wave linear and nonlinear baroclinic model. J. Atmos. Sci., 55 , 2509–2527.
Trenberth, K E., and C J. Guillemot, 1998: Evaluation of the atmospheric moisture and hydrological cycle in the NCEP/NCAR reanalyses. Climate Dyn., 14 , 213–231.
Vecchi, G A., and D E. Harrison, 2002: Monsoon breaks and subseasonal sea surface temperature variability in the Bay of Bengal. J. Climate, 15 , 1485–1493.
Waliser, D E., K M. Lau, and J H. Kim, 1999: The influence of coupled sea surface temperatures on the Madden–Julian oscillation: A model perturbation experiment. J. Atmos. Sci., 56 , 333–358.
Waliser, D E., and Coauthors, 2004: AGCM simulations of intraseasonal variability associated with the Asian summer monsoon. Climate Dyn., 21 , 423–446.
Wang, B., 2005: Basic theories. Intraseasonal Variability of the Ocean and Atmosphere, K. M. Lau and D. E. Waliser, Eds., Praxis, 307–360.
Wang, B., and H. Rui, 1990: Synoptic climatology of transient tropical intraseasonal convection anomalies: 1975–1985. Meteor. Atmos. Phys., 44 , 43–61.
Wang, B., and X. Xie, 1996: Low-frequency equatorial waves in vertically sheared zonal flow. Part II: Unstable waves. J. Atmos. Sci., 53 , 3589–3605.
Wang, B., and X. Xie, 1997: A model for the boreal summer intraseasonal oscillation. J. Atmos. Sci., 54 , 72–86.
Watanabe, M., and M. Kimoto, 2000: Atmosphere–ocean thermal coupling in the North Atlantic: A positive feedback. Quart. J. Roy. Meteor. Soc., 126 , 3343–3369.
Watanabe, M., and M. Kimoto, 2001: Corrigendum. Quart. J. Roy. Meteor. Soc., 127 , 733–734.
Watanabe, M., and F F. Jin, 2002: Role of Indian Ocean warming in the development of Philippine Sea anticyclone during ENSO. Geophys. Res. Lett, 29 .1478, doi:10.1029/2001GL014318.
Watanabe, M., and F F. Jin, 2003: A moist linear baroclinic model: Coupled dynamical-convective response to El Niño. J. Climate, 16 , 1121–1139.
Webster, P J., 1972: Response of the tropical atmosphere to local steady forcing. Mon. Wea. Rev., 100 , 518–541.
Webster, P J., 1983: Mechanisms of monsoon low-frequency variability: Surface hydrological effects. J. Atmos. Sci., 40 , 2110–2124.
Webster, P J., V O. Magana, T N. Palmer, J. Shukla, R A. Thomas, M. Yanai, and T. Yasunari, 1998: The monsoon: Processes, predictability and prediction. J. Geophys. Res., 103 , 14451–14510.
Wu, Z., E S. Sarachik, and D. Battisti, 2000: Vertical structure of convective heating and the three-dimensional structure of the forced circulation on an equatorial beta plane. J. Atmos. Sci., 57 , 2169–2187.
Xie, P., and P. Arkin, 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate, 9 , 840–858.
Yasunari, T., 1979: Cloudiness fluctuations associated with the Northern Hemisphere summer monsoon. J. Meteor. Soc. Japan, 57 , 227–242.
Yasunari, T., 1981: Structure of an Indian monsoon system with around 40-day period. J. Meteor. Soc. Japan, 59 , 336–354.
Jun–Sep 1979–95 AVHRR OLR (a) climatological mean with contour interval of 10 W m−2 , (b) daily variance with a contour interval of 250 (W m−2)2, (c) 20–100-day bandpass-filtered variance with contour interval of 100 (W m−2)2, and (d) percent of daily variance explained by 20–100-day filtered data with a contour interval of 5%.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Jun–Sep 1979–95 AVHRR OLR (a) climatological mean with contour interval of 10 W m−2 , (b) daily variance with a contour interval of 250 (W m−2)2, (c) 20–100-day bandpass-filtered variance with contour interval of 100 (W m−2)2, and (d) percent of daily variance explained by 20–100-day filtered data with a contour interval of 5%.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Jun–Sep 1979–95 AVHRR OLR (a) climatological mean with contour interval of 10 W m−2 , (b) daily variance with a contour interval of 250 (W m−2)2, (c) 20–100-day bandpass-filtered variance with contour interval of 100 (W m−2)2, and (d) percent of daily variance explained by 20–100-day filtered data with a contour interval of 5%.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Cyclostationary EOFs of 20–100-day bandpass-filtered AVHRR OLR for Jun–Sep 1979–95. The EOFs have been scaled by one standard deviation of the PCs to give units of W m−2: (a) day −15, (b) day −10, (c) day −5, (d) day 0, (e) day 5, (f) day 10, (g) day 15, and (h) day 20.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Cyclostationary EOFs of 20–100-day bandpass-filtered AVHRR OLR for Jun–Sep 1979–95. The EOFs have been scaled by one standard deviation of the PCs to give units of W m−2: (a) day −15, (b) day −10, (c) day −5, (d) day 0, (e) day 5, (f) day 10, (g) day 15, and (h) day 20.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Cyclostationary EOFs of 20–100-day bandpass-filtered AVHRR OLR for Jun–Sep 1979–95. The EOFs have been scaled by one standard deviation of the PCs to give units of W m−2: (a) day −15, (b) day −10, (c) day −5, (d) day 0, (e) day 5, (f) day 10, (g) day 15, and (h) day 20.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) For each summer (Jun–Sep 1979–95) the percent variance of observed all India rainfall (AIR) departures explained by PC-4 as determined from lagged correlation analysis. The daily observed AIR departures were smoothed with a 9-day running mean prior to the analysis. The horizontal line corresponds to the 5% significance level assuming 12 degrees of freedom. (b) Active minus break phase daily AIR (smoothed with a 9-day running mean prior to the analysis) for Jun–Sep 1979–90 (the period for which the gridded data are available). The active [break] phase AIR was calculated from regression with the PC associated with day 15 (Fig. 2g) [day −5 (Fig. 2c)]. The units are mm day−1.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) For each summer (Jun–Sep 1979–95) the percent variance of observed all India rainfall (AIR) departures explained by PC-4 as determined from lagged correlation analysis. The daily observed AIR departures were smoothed with a 9-day running mean prior to the analysis. The horizontal line corresponds to the 5% significance level assuming 12 degrees of freedom. (b) Active minus break phase daily AIR (smoothed with a 9-day running mean prior to the analysis) for Jun–Sep 1979–90 (the period for which the gridded data are available). The active [break] phase AIR was calculated from regression with the PC associated with day 15 (Fig. 2g) [day −5 (Fig. 2c)]. The units are mm day−1.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) For each summer (Jun–Sep 1979–95) the percent variance of observed all India rainfall (AIR) departures explained by PC-4 as determined from lagged correlation analysis. The daily observed AIR departures were smoothed with a 9-day running mean prior to the analysis. The horizontal line corresponds to the 5% significance level assuming 12 degrees of freedom. (b) Active minus break phase daily AIR (smoothed with a 9-day running mean prior to the analysis) for Jun–Sep 1979–90 (the period for which the gridded data are available). The active [break] phase AIR was calculated from regression with the PC associated with day 15 (Fig. 2g) [day −5 (Fig. 2c)]. The units are mm day−1.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Linear regressions of PC-4 with 20–100-day bandpass-filtered CMAP rainfall (mm day−1) and NCEP–NCAR reanalysis 850-hPa wind (m s−1) (a) day −15, (b) day −10, (c) day −5, (d) day 0, (e) day 5, (f) day 10, (g) day 15, and (h) day 20. Data are plotted where the regression is significant at the 5% level or better assuming each pentad is independent. All regressions have been scaled by a one standard deviation perturbation of PC-4 to give the aforementioned units.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Linear regressions of PC-4 with 20–100-day bandpass-filtered CMAP rainfall (mm day−1) and NCEP–NCAR reanalysis 850-hPa wind (m s−1) (a) day −15, (b) day −10, (c) day −5, (d) day 0, (e) day 5, (f) day 10, (g) day 15, and (h) day 20. Data are plotted where the regression is significant at the 5% level or better assuming each pentad is independent. All regressions have been scaled by a one standard deviation perturbation of PC-4 to give the aforementioned units.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Linear regressions of PC-4 with 20–100-day bandpass-filtered CMAP rainfall (mm day−1) and NCEP–NCAR reanalysis 850-hPa wind (m s−1) (a) day −15, (b) day −10, (c) day −5, (d) day 0, (e) day 5, (f) day 10, (g) day 15, and (h) day 20. Data are plotted where the regression is significant at the 5% level or better assuming each pentad is independent. All regressions have been scaled by a one standard deviation perturbation of PC-4 to give the aforementioned units.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day −15 linear regressions of PC-4 with 20–100-day bandpass-filtered (a) SST and surface temperature (°C) and surface wind (m s−1), (b) latent heat flux (W m−2) and surface wind stress (N m−2), (c) 1000-hPa divergence (s−1), (d) 850-hPa specific humidity (kg kg−1), (e) specific humidity (10°S–5°N; kg kg−1), and (f) 850-hPa divergence (s−1). Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day −15 linear regressions of PC-4 with 20–100-day bandpass-filtered (a) SST and surface temperature (°C) and surface wind (m s−1), (b) latent heat flux (W m−2) and surface wind stress (N m−2), (c) 1000-hPa divergence (s−1), (d) 850-hPa specific humidity (kg kg−1), (e) specific humidity (10°S–5°N; kg kg−1), and (f) 850-hPa divergence (s−1). Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day −15 linear regressions of PC-4 with 20–100-day bandpass-filtered (a) SST and surface temperature (°C) and surface wind (m s−1), (b) latent heat flux (W m−2) and surface wind stress (N m−2), (c) 1000-hPa divergence (s−1), (d) 850-hPa specific humidity (kg kg−1), (e) specific humidity (10°S–5°N; kg kg−1), and (f) 850-hPa divergence (s−1). Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day −15 (a) column integrated heating anomalies (K day−1), steady-state response of 850-hPa wind (m s−1), and divergence (s−1) to day −15 heating: (b) negative heating over the equatorial Indian Ocean, (c) positive heating over the tropical west Pacific, and (d) total response [sum of (b) and (c)].
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day −15 (a) column integrated heating anomalies (K day−1), steady-state response of 850-hPa wind (m s−1), and divergence (s−1) to day −15 heating: (b) negative heating over the equatorial Indian Ocean, (c) positive heating over the tropical west Pacific, and (d) total response [sum of (b) and (c)].
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day −15 (a) column integrated heating anomalies (K day−1), steady-state response of 850-hPa wind (m s−1), and divergence (s−1) to day −15 heating: (b) negative heating over the equatorial Indian Ocean, (c) positive heating over the tropical west Pacific, and (d) total response [sum of (b) and (c)].
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) Model-generated heating anomalies (K day−1) for prescribed cold SST anomalies over the equatorial Indian Ocean, (b) steady-state response of 850-hPa wind (m s−1) and divergence (s−1), (c) 850-hPa specific humidity (kg kg−1), and (d) specific humidity averaged over 10°S–5°N from 1000 to 100 hPa. The negative (positive) contour interval is 0.2 (0.1) K day−1 in (a) and is dashed (solid). The negative contour interval in (b) is 1.0 e−07 and the positive values are shaded progressively. The contour interval for positive (negative) values is 2 × 10−5 (4 × 10−5) in (c)–(d).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) Model-generated heating anomalies (K day−1) for prescribed cold SST anomalies over the equatorial Indian Ocean, (b) steady-state response of 850-hPa wind (m s−1) and divergence (s−1), (c) 850-hPa specific humidity (kg kg−1), and (d) specific humidity averaged over 10°S–5°N from 1000 to 100 hPa. The negative (positive) contour interval is 0.2 (0.1) K day−1 in (a) and is dashed (solid). The negative contour interval in (b) is 1.0 e−07 and the positive values are shaded progressively. The contour interval for positive (negative) values is 2 × 10−5 (4 × 10−5) in (c)–(d).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) Model-generated heating anomalies (K day−1) for prescribed cold SST anomalies over the equatorial Indian Ocean, (b) steady-state response of 850-hPa wind (m s−1) and divergence (s−1), (c) 850-hPa specific humidity (kg kg−1), and (d) specific humidity averaged over 10°S–5°N from 1000 to 100 hPa. The negative (positive) contour interval is 0.2 (0.1) K day−1 in (a) and is dashed (solid). The negative contour interval in (b) is 1.0 e−07 and the positive values are shaded progressively. The contour interval for positive (negative) values is 2 × 10−5 (4 × 10−5) in (c)–(d).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day 10 linear regressions of PC-4 with 20–100-day bandpass-filtered (a) rainfall (mm day−1) and 850-hPa wind (m s−1) and (b) 850-hPa specific humidity (kg kg−1); (c), (d) as in (a), (b) but for day 20. Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day 10 linear regressions of PC-4 with 20–100-day bandpass-filtered (a) rainfall (mm day−1) and 850-hPa wind (m s−1) and (b) 850-hPa specific humidity (kg kg−1); (c), (d) as in (a), (b) but for day 20. Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Day 10 linear regressions of PC-4 with 20–100-day bandpass-filtered (a) rainfall (mm day−1) and 850-hPa wind (m s−1) and (b) 850-hPa specific humidity (kg kg−1); (c), (d) as in (a), (b) but for day 20. Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1) to day 10 heating: (a) positive heating over the Bay of Bengal, (b) positive heating over the equatorial west Pacific, (c) negative heating over the tropical west Pacific, and (d) total response [sum of (a), (b), and (c)]. The negative contour interval in (a)–(d) is 2.0 × 107 and the positive values are shaded progressively.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1) to day 10 heating: (a) positive heating over the Bay of Bengal, (b) positive heating over the equatorial west Pacific, (c) negative heating over the tropical west Pacific, and (d) total response [sum of (a), (b), and (c)]. The negative contour interval in (a)–(d) is 2.0 × 107 and the positive values are shaded progressively.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1) to day 10 heating: (a) positive heating over the Bay of Bengal, (b) positive heating over the equatorial west Pacific, (c) negative heating over the tropical west Pacific, and (d) total response [sum of (a), (b), and (c)]. The negative contour interval in (a)–(d) is 2.0 × 107 and the positive values are shaded progressively.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1) to day 20 (a) positive heating over the tropical west Pacific, (b) all heatings, and (c) vertical profile of anomalous vertical velocity (ω, hPa s−1) averaged over India (10°–25°N, 70°–100°E) based on northern tropical west Pacific heating (closed circles) and negative heating over the equatorial Indian Ocean (open circles). Positive (negative) values correspond to descent (ascent).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1) to day 20 (a) positive heating over the tropical west Pacific, (b) all heatings, and (c) vertical profile of anomalous vertical velocity (ω, hPa s−1) averaged over India (10°–25°N, 70°–100°E) based on northern tropical west Pacific heating (closed circles) and negative heating over the equatorial Indian Ocean (open circles). Positive (negative) values correspond to descent (ascent).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1) to day 20 (a) positive heating over the tropical west Pacific, (b) all heatings, and (c) vertical profile of anomalous vertical velocity (ω, hPa s−1) averaged over India (10°–25°N, 70°–100°E) based on northern tropical west Pacific heating (closed circles) and negative heating over the equatorial Indian Ocean (open circles). Positive (negative) values correspond to descent (ascent).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) Model-generated heating anomalies (K day−1) for prescribed warm SST anomalies over the tropical west Pacific, (b) steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1), and (c) vertical profile of anomalous vertical velocity (ω, hPa s−1) averaged over India (10°–25°N, 70°–100°E). Positive values correspond to descent. The negative (positive) contour interval in (a) is 0.1 (0.2) K day−1. The negative contour interval in (b) is 1.0 × 10−6, and the positive values are shaded progressively.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) Model-generated heating anomalies (K day−1) for prescribed warm SST anomalies over the tropical west Pacific, (b) steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1), and (c) vertical profile of anomalous vertical velocity (ω, hPa s−1) averaged over India (10°–25°N, 70°–100°E). Positive values correspond to descent. The negative (positive) contour interval in (a) is 0.1 (0.2) K day−1. The negative contour interval in (b) is 1.0 × 10−6, and the positive values are shaded progressively.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
(a) Model-generated heating anomalies (K day−1) for prescribed warm SST anomalies over the tropical west Pacific, (b) steady-state response of 850-hPa wind (m s−1) and relative vorticity (s−1), and (c) vertical profile of anomalous vertical velocity (ω, hPa s−1) averaged over India (10°–25°N, 70°–100°E). Positive values correspond to descent. The negative (positive) contour interval in (a) is 0.1 (0.2) K day−1. The negative contour interval in (b) is 1.0 × 10−6, and the positive values are shaded progressively.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Linear regressions of PC-4 with 20–100-day bandpass-filtered specific humidity (kg kg−1) and vertical velocity (positive downward with isoline increment of 0.002 Pa s−1) at 92.5°E, averaged between 15° and 20°N plotted from 1000 to 300 hPa as a function of time lag (−25 to 25 days). Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Linear regressions of PC-4 with 20–100-day bandpass-filtered specific humidity (kg kg−1) and vertical velocity (positive downward with isoline increment of 0.002 Pa s−1) at 92.5°E, averaged between 15° and 20°N plotted from 1000 to 300 hPa as a function of time lag (−25 to 25 days). Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Linear regressions of PC-4 with 20–100-day bandpass-filtered specific humidity (kg kg−1) and vertical velocity (positive downward with isoline increment of 0.002 Pa s−1) at 92.5°E, averaged between 15° and 20°N plotted from 1000 to 300 hPa as a function of time lag (−25 to 25 days). Significance testing and scaling as in Fig. 4.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Temporal evolution of 700-hPa geopotential height anomalies (m) for day 20 positive heating over the tropical west Pacific: (a) day 3, (b) day 6, (c) day 9, and (d) day 12. The contour interval is 0.5 m and the positive (negative) contours are solid (dashed).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Temporal evolution of 700-hPa geopotential height anomalies (m) for day 20 positive heating over the tropical west Pacific: (a) day 3, (b) day 6, (c) day 9, and (d) day 12. The contour interval is 0.5 m and the positive (negative) contours are solid (dashed).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
Temporal evolution of 700-hPa geopotential height anomalies (m) for day 20 positive heating over the tropical west Pacific: (a) day 3, (b) day 6, (c) day 9, and (d) day 12. The contour interval is 0.5 m and the positive (negative) contours are solid (dashed).
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
As in Fig. 10 but for day 20 negative heating over the equatorial Indian Ocean.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
As in Fig. 10 but for day 20 negative heating over the equatorial Indian Ocean.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
As in Fig. 10 but for day 20 negative heating over the equatorial Indian Ocean.
Citation: Journal of the Atmospheric Sciences 62, 8; 10.1175/JAS3504.1
* International Pacific Research Center Contribution Number 307 and School of Ocean and Earth Science and Technology Contribution Number 6525.
The SST regressions presented throughout are an underestimate of the actual anomalies. In situ measurements during BOBMEX (Bhat et al. 2001) and SST from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (Sengupta et al. 2001) indicate that during strong BSISV events the peak to peak SST amplitude is as large as 2°C. The underestimate arises due to 1) the regression over a large number of intraseasonal events and 2) observed monthly (weekly) averaged SSTs were used as the boundary condition in the reanalaysis prior (and subsequent) to December 1981.
