1. Introduction
The tropical western Pacific Ocean (TWP) is well known for high sea surface temperatures and an accompanying large amount of precipitation. The diabatic heating from convective clouds over this region is an important factor that drives global atmospheric circulation.
One of the major scales to regulate convection over TWP is intraseasonal variation (ISV). Following the pioneering studies by Madden and Julian (1972) that identified the 40–50-day oscillation propagating eastward along the equator [Madden–Julian oscillation (MJO)], a northward-propagating boreal summer ISV was found to occur particularly during the Asian summer monsoon (ASM) season (e.g., Yasunari 1979; Lau and Chen 1986). We refer to this as the boreal summer intraseasonal variation (BSISV). Wang and Rui (1990) demonstrated that BSISV is independent of the eastward-propagating component. They also showed that BSISV is confined to a limited region of TWP and the Indian Ocean where active convection prevails during ASM (e.g., Murakami and Matsumoto 1994; Wang and LinHo 2002); that is, convective activity during ASM over these regions is regulated by ISV. The seasonal march of the onset of ASM accompanies ISVs that are phase locked in the early boreal summer (e.g., Nakazawa 1992; LinHo and Wang 2002). The activity of typhoons, another factor that regulates convective activity in TWP (e.g., Kubota and Wang 2009), is also associated with ISV (e.g., Nakazawa 1986; Nitta 1987; Hartmann et al. 1992). These studies indicate that the nature and the mechanism of the ISV are essential for understanding the weather and climate over the Asian continent, TWP, and the Indian Ocean.
To explain the mechanism of ISV with its characteristic time scale (tens of days), direction of propagation (eastward or northward), and slow propagation speed (from one to several meters per second), various hypotheses have been proposed that emphasize key processes such as wave dynamics, latent heating, moisture, radiation, and air–sea interaction (see review by Wang 2005). Efforts have been made to obtain detailed observational evidence in support of these hypotheses by utilizing satellite observations (e.g., Masunaga et al. 2006; Jiang et al. 2011) and by running field experiments [e.g., Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE; Webster and Lukas 1992); the Joint Air–Sea Monsoon Interaction Experiment (JASMINE; Webster et al. 2002)]. For the MJO, the observational studies revealed the preexisting conditions that favor convection (low-level convergence, ascending motion, positive humidity anomalies, etc.) or “preconditioning” (e.g., Johnson et al. 1999) as proposed in theoretical studies [see review by Zhang (2005)]. In the Indian Ocean, JASMINE revealed some similarities between BSISV and the MJO during TOGA COARE (Webster et al. 2002). Over TWP, however, in situ evidence for the mechanism of BSISV or for the similarity between BSISV and the MJO remains insufficient, particularly in regard to the details of the evolution of the atmospheric vertical profile and the convective features.
To fill the existing gap in the in situ observational data for BSISV over TWP, the Japan Agency for Marine-Earth Science and Technology (JAMSTEC) conducted a field project named the Pacific Area Long-Term Atmospheric Observation for Understanding of Climate Change (PALAU2008).1 In this project, a sounding and radar network was deployed over the western Pacific during the early summer of 2008 to capture the ISV around the onset of ASM. The project successfully captured the variation of the convective systems during ASM, as described in Geng et al. (2011, hereafter G11). In the present study, we focused on the ISV-scale evolution of the atmospheric profile and the accompanying convective systems by utilizing the observational networks in PALAU2008. The sounding network, in particular, was utilized to estimate the budget of heat and moisture to evaluate how the large- and convective-scale systems modify the atmosphere’s thermodynamic conditions before and during the convectively active phase of the BSISV.
The paper is organized as follows. Section 2 describes the observational network and the methods used to process the dataset obtained by the network. Section 3 presents an outline of the situation during PALAU2008. Section 4 describes the variations during PALAU2008, as revealed primarily by in situ atmospheric observations. Section 5 discusses the observed atmospheric variations, particularly during the first half of the period, which includes the convectively inactive period, and some additional observed features. Section 6 summarizes the results of the study.
2. Observations and data
a. Radiosonde observations and budget analyses
During the PALAU2008 field project, special soundings were launched during June and July at four sites: Koror (7.3°N, 134.5°E), Yap (9.5°N, 138.1°E), Woleai (7.4°N, 143.9°E), and R/V Mirai. The R/V Mirai stayed at (12°N, 135°E) for 22 days during the project to form a sounding array as shown in Fig. 1. We refer hereafter to these 22 days, from 0000 UTC 6 June to 0000 UTC 28 June, as the arrayed sounding period (ASP).
Map of the PALAU2008 observational area, with the monthly-averaged TB(IR) for June 2008. The filled squares represent the upper-air stations used in the present study. The two circles at stations R/V Mirai and Koror indicate the observational ranges of the radar (in a volume scan to obtain three-dimensional echo distribution) used. The triangle connecting three sites (R/V Mirai, Koror, and Yap) indicates the area used for budget analyses (and some other corresponding parameters as described in the texts and figure captions). The quadrilateral connecting four points (R/V Mirai, Koror, Woleai, and Guam) indicates the area used for Fig. 13. The vertical dashed white lines indicate the zonal range to average the data used in Figs. 2 and 3.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
During ASP, the radiosondes were launched every 3 h on R/V Mirai and every 6 h at Koror and Yap. At Woleai, observations were made every 12 h until 0000 UTC 11 June and every 6 h thereafter. This study utilizes all of these sounding datasets obtained during ASP and operational 12-hourly soundings obtained at Guam to construct a 6-hourly dataset. When necessary, the sounding data were interpolated or averaged to facilitate a 6-h analysis frequency throughout ASP at all the sites. A dry bias of the humidity sensors of Vaisala RS92, used at Mirai and Woleai, is corrected as described in Yoneyama et al. (2008) for all of the individual soundings. The diurnal dry bias of Vaisala RS80, used at Yap and Koror, is confirmed and corrected as described in appendix A.
The analysis area in this study is the inside of the triangle with the vertices formed by R/V Mirai, Koror, and Yap (see Fig. 1). This area was used 1) to confine the mesoscale array in order to better capture the effect of the convective systems, 2) to compare the results in the array with the radar data at two of the vertices (R/V Mirai and Koror), and 3) to utilize only the soundings with shorter intervals (6- or 3-hourly for the entire period). The resulting values of Q1 and Q2 from the triangular area qualitatively resemble those from the larger rectangular area (with the vertices formed by R/V Mirai, Koror, Woleai, and Guam; see Fig. 1), although the amplitude is larger in the triangular area, reflecting the smaller size of this area. Hereafter, “sounding array” or simply “array” refers to the triangular sounding array consisting of R/V Mirai, Koror, and Yap.
The reliability of the budget analyses is evaluated by comparing the budget-estimated rainfall values with those from individual sources such as the Tropical Rainfall Measuring Mission (TRMM) 3B42 (Huffman et al. 2007), Global Satellite Mapping of Precipitation (GSMaP) near-real-time (NRT) product (Kubota et al. 2007), Special Sensor Microwave Imager (SSM/I) estimates (Wentz and Spencer 1998), and surface rain gauges, as described in appendix B. The budget-estimated rainfall values are fairly consistent with those from the individual sources except around 17 June, when a tropical depression passed over the array. We therefore omitted the 2-day period of 16–17 June from the analyses in the present study.
b. Radar data
During ASP, two Doppler scanning precipitation radars operated continuously in the array. A C-band Doppler radar was installed on the R/V Mirai, and an X-band radar was installed at Aimeliik (approximately 10 km north of Koror, Republic of Palau). These two radars collected volume-scan data every 10 and 7.5 min, respectively. To equalize the data sampling, we utilized data taken every hour for both radars. The reflectivity data from both radars were quality controlled (appendix C) essentially as described by Katsumata et al. (2008). After correction, these data were interpolated to a three-dimensional Cartesian grid with a volume of 200 km × 200 km × 20 km and a horizontal and vertical grid spacing of 1 and 0.5 km, respectively.
The echo-top height was calculated using a 10-dBZ threshold. The threshold was determined to be well above the minimum detectable reflectivity within the Cartesian grid for both the R/V Mirai and Aimeliik radars. The echo-top height was determined by counting upward from the bottom to detect the first echo top and to exclude the overlying multilayer clouds.
3. Outline of the analyzed period
As described in G11, the convective activity during PALAU2008 ASP changed from the inactive phase to the active phase. G11 concluded that this change represented the onset of ASM in TWP during this year. Actually, the area of the low brightness temperature in the satellite infrared channels (hereafter TB(IR); Janowiak et al. 2001) with the surface westerly detected by the Quick Scatterometer (QuikSCAT), a satellite-borne scatterometer (data and detailed descriptions are provided by Remote Sensing Systems; http://www.remss.com), extended northward to approximately 20°N after the middle of June (Fig. 2). During the same time, one can also see embedded northward-propagating signals with a period of approximately 15–30 days, similar to observations in previous studies (e.g., Yasunari 1979). A higher magnification of the image in Fig. 2 for ASP (Figs. 3a,b) indicates that one of BSISVs passed over the sounding array and proceeded northward with a speed of approximately 1.8 m s−1. This speed is slightly faster than (but comparable to) those recorded in previous studies (e.g., Yasunari 1979; Jiang et al. 2004, 2011). Based on these observed features, we regard the observed transition in ASP from the convectively inactive to active phase as BSISV.
Time–latitude cross section of the (a) TB(IR) and (b) QuikSCAT-derived surface zonal wind from 1 Mar to 31 Dec 2008. The horizontal dashed black lines indicate the meridional coverage of the triangle sounding array. The vertical dashed black lines indicate the beginning and the end of ASP.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
Time–latitude cross section for ASP: (a) TB(IR), (b) QuikSCAT-derived surface zonal wind, (c) QuikSCAT-derived surface divergence, and (d) precipitable water (Wentz and Spencer 1998). The horizontal dashed black lines indicate the meridional coverage of the triangle sounding array. The slanted dashed black line indicates northward propagation with the speed of 1.8 m s−1 starting at the equator on 10 Jun. The center positions of Typhoon Fengshen are indicated as filled circles (when between 130° and 140°E) and open circles (when between 120° and 130°E), every 12 h.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
The horizontal map of the 5-day-averaged TB(IR) and QuikSCAT-derived surface wind during ASP are displayed in Fig. 4 to confirm the horizontal structure of BSISV. Before the appearance of BSISV (Fig. 4a for 4–8 June), a zonally elongated ITCZ above the prevailing surface easterlies was apparent. This situation was altered in the next 5 days (Fig. 4b for 9–13 June) with a scattering of the cloudy area and weakening of the easterlies. At the third 5-day period (Fig. 4c for 14 to 18 June), a westerly area appeared in the southern part of the array, with the northward procession of the cloudy area. Finally, on 19–23 June (Fig. 4d), the cloudy area with the westerly component propagated farther to the north of the array.
Five-day-averaged horizontal distributions of the (shaded) TB(IR) and (arrows) QuikSCAT-derived surface wind for (a) 4–8, (b) 9–13, (c) 14–18, and (d) 19–23 Jun, respectively. The area, triangle, and vertical dashed white lines have the same meaning as in Fig. 1. The positions of Typhoon Fengshen in each 5-day period are shown as filled circles (when between 130° and 140°E) and open circles (when between 120° and 130°E), every 12 h.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
The change of the zonal wind component from easterly to westerly at the convectively active region of BSISV has been observed in previous studies (e.g., Kemball-Cook and Wang 2001). Close inspection of the time–latitude cross section (Fig. 3), however, reveals that the low TB(IR) region propagated at a slower speed than the westerly region (and also than the southerly region; not shown). This discrepancy indicates that the simple schematic used to represent BSISV as the shear zone between easterlies and westerlies (or northerlies and southerlies) is not always realistic, at least in the present case. In Fig. 3c, the horizontal convergence of the surface wind, derived from the QuikSCAT wind vector fields, is located around the leading edge of the low TB(IR) zone (shown as a dashed slanted line in Fig. 3), although a clear divergence signal appeared in the middle and south of the low TB(IR) zone after 20 June. The horizontal convergence near the leading edge of the BSISV is consistent with the results in Kemball-Cook and Wang (2001).
The characteristic parameter that follows the northward-propagating low TB(IR) region is the precipitable water (PW). In the time–latitude cross section of the SSM/I-derived PW (Fig. 3d), the high PW region (>52 mm) appeared almost simultaneously with the low TB(IR) region. A striking feature is that north of 10°N, the region with medium PW (>48 mm) appeared in advance of the low TB(IR) and high PW region and propagated northward at almost the same speed as the low TB(IR) region. This medium PW appeared north of 10°N where the northern edge of the period-averaged ITCZ and the sounding array were located (Fig. 1). The medium PW region became unclear north of 20°N where the northward-propagating low TB(IR) region was also unclear. The coexistence of the medium PW region and the low TB(IR) region suggests that the medium PW region is a component of BSISV.
In addition to the northward-propagating signal described above, we noted a relationship between BSISV and the Tropical Depression (TD) Fengshen that was observed during ASP. The best-track dataset from the Joint Typhoon Warning Center indicates that Fengshen appeared around the array on 17 June and then moved toward the northwest (Fig. 4). When Fengshen was located east of 130°E (as indicated by filled circles), its position corresponded well to those of low TB(IR), surface convergence, and meridional shear of zonal wind (Fig. 3). These findings indicate that the observed parameters over the array were under the strong influence of Fengshen around 17 June. After Fengshen moved west of 130°E (as indicated by open circles in Figs. 3 and 4), the low TB(IR) region corresponds well to the surface southerly region (Fig. 4d). The southerly zone located between the cyclonic circulation near Fengshen and the anticyclonic circulation in the east had a half wavelength of approximately 1500 km. This resembles the case in Dickinson and Molinari (2002) in which the pattern appeared in the 6–10-day filtered field. These suggest that the disturbance after 17 June was not directly influenced by Fengshen. Furthermore, the medium PW extended toward the north of the array approximately 4 days before the appearance of Fengshen and then extended approximately 1000 km north of the initial position of Fengshen.
These results indicate that the observed BSISV is not clearly associated with Fengshen except when Fengshen was located near the array (16–17 June). The period of 16–17 June was omitted from the subsequent analyses because of the possible large error of the budget analyses, as mentioned in section 3. To simplify our discussion of the results of the present study, the effects of Fengshen were not considered further in the following sections. Our detailed investigation of the effects of Fengshen will be described in future reports.
4. Evolution of the atmospheric profiles over the sounding array
One BSISV passed over the sounding array during PALAU2008 ASP, as described in the previous section. In this section, we describe the in situ observed atmospheric profiles over the array and their temporal variations in detail.
The temporal variation of the radar-echo area during ASP with the averaged TB(IR) over the sounding array is shown in Fig. 5 to illustrate the changes in the precipitating systems under the BSISV cloud. A large peak of the low TB(IR) over the array is found from 17 to 22 June when the northward-propagating low TB(IR) region (i.e., BSISV) was over the array (Fig. 3a). Prior to the 5-day period from 17 to 22 June, the radar-echo coverage is continuously small at Mirai, while peaks with large coverage are observed at Aimeliik, where the low TB(IR) was continuous (Fig. 4). In contrast, the radars at both Mirai and Aimellik detected widespread echo from 17 to 22 June. In terms of these temporal variations in convective activities, we regard the periods before and after 17 June as differing in their atmospheric conditions. Bearing in mind the exclusion of 16 and 17 June from further analyses as explained in the two preceding sections, we hereafter refer to the 10-day periods from 6 to 15 June and from 18 to 27 June as the “preactive period” and the “active period,” respectively.
Time series during ASP for (a) the brightness temperature at the channel IR1 [TB(IR); black line with gauge at left axis] and the difference between the brightness temperatures at channels IR1 and IR2 [δTB = TB (IR1) − TB (IR2); dashed gray line with gauge at right axis], and (b) the areal coverage of the radar echo (>10 dBZ) at a 3-km height within the 200 km × 200 km square centered at R/V Mirai (black line) and Aimeliik (gray line).
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
The time–height cross sections of the winds over the array, corresponding to the QuikSCAT-derived surface winds, are shown in Fig. 6. This figure illustrates the drastic change that occurred around 17 June. The vertical wind shear in the meridional component changed from northerly in the preactive period to southerly in the active period. Considering the northward procession of BSISV, this temporal change of the meridional shear could be explained by the northward passage of the lower-tropospheric convergence and upper-tropospheric divergence, in combination. The zonal wind also changes drastically around 17 June, but the barotropic westerly lasts only a few days when the most active convection appeared on the sounding array. This temporal variation in the zonal wind corresponds well to the conceptual model of Kemball-Cook and Wang (2001), at least over the sounding array, although not always to the north of the sounding array (as described in the previous section).
Time–height cross section of (a) zonal wind and (b) meridional wind over the triangle array, derived by averaging the profiles obtained at three stations at the vertices (R/V Mirai, Koror, and Yap). A 1-day running mean is applied.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
Corresponding to the SSM/I-derived PW field shown in Fig. 3d, the variation of moisture over the array, obtained by the radiosonde, is shown as PW and the layer-averaged water vapor mixing ratio (hereafter simply “mixing ratio”) in Fig. 7a. As in the SSM/I image, PW over the sounding array exceeds 52 mm during the active period. In the preactive period, PW over the sounding array gradually increased toward the beginning of the active phase. The period of medium PW (48 < PW < 52 mm), observed by SSM/I (Fig. 3d), is also well captured by the radiosonde from 10 to 16 June. The temporal variation of the PW is dominated by the averaged mixing ratio in the layer between 900 and 500 hPa rather than that of the boundary layer (below 950 hPa), as shown by Yoneyama (2003).
(a) Temporal variation of the precipitable water (thick black line), and the layer-averaged water vapor mixing ratio between 500 and 900 hPa (thin gray line with gauge at right axis and gray numbers) and between 950 and 1000 hPa (thin black line with gauge at right axis and black numbers). (b) Time–height cross section of the MSE over the sounding array. All values are averages of the three sounding sites (R/V Mirai, Koror, and Yap), as described for Fig. 6. A 1-day running mean is applied.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
To better elucidate the moisture variability detail, the time–height cross section of the moisture is displayed in Fig. 7b as the moist static energy (MSE) by the difference from the ASP-averaged value at each height (hereafter anomaly). The largest variation corresponding to the convective activity appears in the middle of the troposphere (between 500 and 700 hPa), with the dry anomaly in the preactive period and the moist anomaly in the active period. The positive anomaly is confined to the lower troposphere in the preactive period but extended to the upper troposphere in the active period. However, close inspection of the preactive period reveals the gradual thickening of the moist layer in the lower troposphere, which corresponds well to the gradual increase of the PW and the mixing ratio in the lower troposphere (Fig. 7a).
To reveal the processes underlying the temporal variations described above in detail, we utilized the budget analyses. The time–height cross sections of Q1 and Q2 are shown in Fig. 8. The Q1 profiles are “top heavy” from the end of the preactive period to the end of the active period, in contrast to the relatively “bottom heavy” profiles in the first half of the preactive period. The Q2 profiles indicate bottom-heavy heating (drying) profiles in the preactive period and top-heavy ones in the active period. These changes correspond well to the temporal variation of the convective activity and the dominant type of convection over R/V Mirai as in G11; that is, the dominant cloud type changes from the isolated, shallow, and convective clouds in the preactive period to more widespread, deep, and stratiform clouds. These findings indicate that the results of the budget analyses are reasonable, not only for the vertically integrated results (in appendix B) but also for the vertical profile.
Time–height cross section of (a) the apparent heat source (Q1) and (b) the apparent moisture sink (Q2) obtained for the triangle sounding array. A 1-day running mean is applied.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
Detailed inspection of Fig. 8 reveals that, in the preactive period, the positive Q2 layer capped by a negative Q2 layer gradually thickens. This gradual growth appeared during the period of medium PW, that is, when PW and the mixing ratio in the middle troposphere increase (Fig. 7a) and the moist layer gradually thickens (Fig. 7b). This observation suggests that the convective-scale moisture transport worked to moisten the middle troposphere, resulting in the gradual growth of the moist layer.
To quantitatively evaluate the effect of convections and large-scale situations, the results of the budget analyses are averaged temporally for each term in the budget analyses (Fig. 9). The plotted terms are the three from the rhs of Eqs. (1) and (2) in units of daily rate of heating (K day−1). All profiles are averaged for two 10-day periods (i.e., the preactive and active periods). The averaging period was set to be as long as possible to minimize the sampling error as recommended by Mapes et al. (2003).
Vertical profiles of the derived parameters from the budget analyses over the array. (a),(c) Each term of the heat and moisture budgets [Eqs. (1) and (2)]. The terms “hadv” and “vadv” in the caption refer to the first and second terms in Eqs. (1) and (2), respectively. (b),(d) The horizontal divergence over the array. (a),(b) Preactive period. (c),(d) Active period. Note that the abscissa of (a) and (c) differ.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
The vertical profiles in the preactive and active periods clearly differ. On Q1, the positive peak is at approximately 700 hPa in the preactive period but approximately 400–500 hPa in the active period. These two vertical profiles resemble the second and first modes of the EOF analyses presented by Zhang and Hagos (2009). The order of appearance; that is, an earlier bottom-heavy heating followed by a middle-heavy heating [i.e., a peak at approximately 400 hPa without significant cooling below as in Zhang and Hagos (2009)], is also consistent with their results derived from the previous field projects for the tropical and subtropical regions.
A close inspection of Fig. 9 reveals the detailed processes in both the preactive and active periods. In the active period, not only Q1 but also Q2 have a peak at approximately 400–500 hPa, with positive values below. This finding indicates that the diabatic heating is maintained primarily by the release of latent heat in the middle and upper troposphere. The vertical profile of the horizontal divergence (Fig. 9d) indicates a stratiform-type circulation, with significant convergence in the middle troposphere and significant divergence in the upper troposphere. The effects of horizontal advections are small for both heat and moisture.
In the preactive period, all of the profiles are of the bottom-heavy type in which the significant values appear only below a 400-hPa height. Large positive values of the large-scale horizontal advection of moisture indicate the horizontal advection of very dry air onto the array. This advection could be caused by the easterly below 400 hPa (Fig. 6) from the cloud-free area under a subtropical high (Fig. 4). The factors that cause moistening of the middle troposphere (Fig. 7) are the large-scale vertical advection of moisture and Q2. The graph indicates that convective-scale processes cause more moistening around a 500–600-hPa height, while large-scale vertical advection causes more moistening below a 700-hPa height. The vertical profile of the large-scale horizontal divergence (Fig. 9b) indicates a large-scale upward motion below a 500-hPa height to transport moisture from the lower to the middle troposphere. In terms of the convective-scale process, the large difference between Q1 and Q2 indicates the active vertical transport of the moist static energy in the convective scale [as estimated from the equation Q1 − Q2 − QR = −∂(ω′h′)/∂p], which may reflect an abundance of shallow convection types such as cumulus congestus (Johnson et al. 1999). The combination of a large negative Q2 at approximately 500–600 hPa, and a positive Q2 below a 900-hPa height indicates vertical moisture transport by the convective-scale features from the lowest to the middle troposphere.
To confirm the relationship between the budget-derived convective-scale effects (Q1 and Q2) and actual convections, we utilized the three-dimensional radar echo distribution at two radar sites: Aimeliik (7.3°N, 134.5°E) near Koror at the southern tip of the array and R/V Mirai (12.0°N, 135.0°E) at the northern tip. The frequency distributions of the echo-top height as the areal ratio to all of the observed echo areas for the preactive and the active periods at the two radars are shown in Fig. 10. The vertical distribution of rainfall-normalized Q1 and Q2 are also displayed in the figure by comparing as the relative frequency distribution.
(a) Vertical profiles of the 10-day average of Q1 and Q2 for the preactive period (6–15 Jun) and for the active period (18–27 Jun). The values are normalized by the 10-day averages of the budget-derived rainfall amounts for each period. The thin lines indicate the original results with a 25-hPa interval; the thick lines are smoothed by averaging vertically within ±100 hPa. The vertical profiles of the relative frequency distribution of the radar echo-top height, displayed as the ratio to the total echo area, at (b) R/V Mirai (12°N, 135°E) and (c) Aimeliik (7.3°N, 134.5°E). The black and gray lines represent the preactive period and the active period, respectively.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
At Aimeliik (Koror), at the southern tip of the array, the profiles are almost identical for the preactive versus active periods. This similarity is reasonable considering that a low TB(IR) was observed in both the preactive and active periods at the southern tip of the sounding array (Fig. 3a). The single clear mode at approximately 400–500 hPa (i.e., just above the melting layer) is also a reasonable finding for the convectively active days considering that the radar reflectivity is generally weak for ice particles above the melting layer. These results are consistent with those of Kubota et al. (2005), who found that the ISV in June was generally not clear over the Palau islands (where Koror and Aimeliik are located).
At R/V Mirai, at the northern tip of the array, the distributions differ clearly between the preactive and active periods. In the active period, the vertical profile is similar to those at Aimeliik. In contrast, in the preactive period, the spectrum of the radar echo-top height ranges more widely, from 700 to 400 hPa with a peak at approximately 700 hPa. This distribution indicates that precipitating clouds with the cloud top below and around the melting level were dominant. The difference in the radar echo-top height statistics clearly shows the transition of the dominating convection from the shallow cumulus and congestus type in the preactive period to the deep convection in the active period. This temporal variation is consistent with the results of the budget analyses.
Unfortunately, the available radars for the present study are at the vertices of the triangle, whereas the budget analyses estimate the effects of convections inside the array. To investigate the convections inside the array, we utilized the data from the Multifunctional Transport Satellite-1R (MTSAT-1R) infrared “split window” channels [infrared channels 1 (10.8 μm) and 2 (12.0 μm)] by the method of Inoue (1987). The difference between the TB values from two split-window channels (hereafter δTB) were averaged for the triangle array and plotted in Fig. 5a, along with the TB at infrared channel 1 [hereafter TB(IR)]. These two parameters, TB(IR) and δTB, show a negative correlation in the preactive period but a positive correlation in the active period. The variations in the active period can reasonably be assumed to result from the frequent appearance of tall clouds with an optically thick top. In contrast, the variations in the preactive period can be assumed to result from the appearance of low clouds (and/or sparse clouds) with an optically thin top. In view of the results of the budget analyses, it is reasonable to assume that the shallow nonorganized clouds (e.g., cumulus and/or congestus) were overcast by remnant cloud particles that evaporated to moisten the top of the cloud layer. These results from split-window channels indicate that the convection observed at R/V Mirai well represented the characteristics of BSISV that propagated northward through the array.
The results presented in this section, derived primarily from in situ observations, indicate that the gradual moistening of the lower and middle troposphere in the preactive period was driven by both shallow convection and large-scale shallow vertical advection. The preactive period corresponds well to the satellite-observed “medium PW” period (Fig. 3d) over the sounding array. Considering that the PW is controlled largely by the amount of water vapor in the middle troposphere (Fig. 7a), the satellite-observed medium PW period can be regarded as the period during which the gradual moistening occurred. In the present study, BSISV can therefore be recognized as the northward propagation of a packet that consists of a leading “gradual moistening” period and following deep convection.
A similar gradual moistening period was suggested by previous observational studies of the MJO (e.g., Kemball-Cook and Weare 2001; Agudelo et al. 2006; Masunaga et al. 2006; Katsumata et al. 2009). These studies described gradual moistening as a preconditioning for deep convection during the convectively active period. For BSISV, similar processes were observed using satellite data and objective reanalyses (e.g., Jiang et al. 2004; Fu et al. 2006; Jiang et al. 2011) over the Indian Ocean. The in situ data in the present study reveal details of the preconditioning process in TWP.
5. Factors that regulate the preconditioning
The results presented in the previous sections demonstrated that the period of gradual moistening of the lower troposphere (preconditioning) occurred prior to the deep convection of BSISV. In this section, we discuss the possible factors that regulate the preconditioning for BSISV in our study. As described in the preceding section, large-scale and convective-scale processes worked together to cause moistening of the lower and middle troposphere.
On the convective scale, the CAPE is one parameter that promotes convection. McBride and Frank (1999) observed high CAPE before the convectively active phase of the ISV over Australia. Agudelo et al. (2006) identified adequate buildup of CAPE as a factor required to promote deep convection in the MJO. Numerical simulations for preconditioning by Waite and Khouider (2010) indicated that rising CAPE during preconditioning is a factor that promotes deepening of the convection during the preactive period.
The time series of CAPE in the present study is shown in Fig. 11. During the preactive period, the CAPE rose to a relatively high value (>2 kJ kg−1). This tendency is consistent with results reported in the previous studies mentioned above. Generally, an increase in CAPE is caused primarily by a rise in the equivalent potential temperature in the boundary layer and/or a cooling of the free troposphere above. In the present study, the mixing ratio in the boundary layer (indicated by thin black line in Fig. 7a) increased continuously during the period of increasing CAPE. A simultaneous decrease of temperature was also observed in the free troposphere above the boundary layer (not shown). To examine the contributions of these factors, the following simple experiment was conducted. The CAPE for the 3-day-averaged profile from 8 to 10 June is 1.65 kJ kg−1. When we replaced the moisture in the boundary layer (assuming a height up to 950 hPa) with that of the 3-day average from 12 to 14 June, the calculated CAPE became 2.10 kJ kg−1. This value is comparable to that from the 3-day-averaged profile from 12 to 14 June (2.31 kJ kg−1). In contrast, replacing the temperature profile with that from 12 to 14 June resulted in a calculated CAPE of 1.76 kJ kg−1, which is only slightly higher than the original value (1.65 kJ kg−1). These experiments suggest that the increase of moisture in the boundary layer (0.5 g kg−1) was the key factor in the preconditioning for the observed BSISV in the convective-scale process. The increase of the mixing ratio in the boundary layer (0.5 g kg−1) corresponds to approximately 2% of the relative humidity, while the actual relative humidity increased only 1% because of the rise in temperature at the same time. Therefore, rise of the temperature could also contribute to moisten the boundary layer.
Time series of the CAPE (black) and CIN (gray) over the triangle array. A 1-day running mean is applied.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
To investigate the possible factors that increase the moisture in the boundary layer, we first considered the effect of surface flux. Roxy and Tanimoto (2011) demonstrated recently that the active phase of BSISV over the South China Sea follows the period with a high SST, high latent heat flux, and relatively weakened zonal wind. The in situ observed surface parameters at R/V Mirai in the present study are shown in Fig. 12. The increase of SST and sensible heat flux in the preactive period corresponds well to the increased temperature in the boundary layer. The high SST in the preactive period is consistent with the results in Roxy and Tanimoto (2011). However, the latent heat flux has no clear peak in the preactive period. This finding reflects the decreased wind speed, particularly in the zonal component. The discrepancy between our results and those of Roxy and Tanimoto (2011) indicates that their proposed scenario cannot be applied in a straightforward manner in the present case.
Time series of the parameters obtained at R/V Mirai: (a) sea surface temperature, (b) latent heat flux (solid line) and sensible heat flux (dashed line), and (c) wind speed (thick black line) and its zonal (thin black line) and meridional (thin gray line) components. The thin and thick lines for each parameter represent hourly and daily averaged values, respectively.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
To investigate other factors that cause moistening of the boundary layer, the contributions of each term of the budget analyses were again evaluated. The large-scale vertical moisture advection worked only to moisten the boundary layer, while Q2 and the large-scale horizontal moisture advection worked to dry the layer (Fig. 9a). Related to the large-scale vertical moisture transport, there was a large-scale horizontal convergence near the surface (Fig. 9b). It is reasonable to assume that the horizontal convergence at the lowermost layer accompanied the large-scale updraft near the surface and transported the more humid air to the rest of the boundary layer. This hypothesis also explains the large-scale process for transporting moisture to the free troposphere above the boundary layer. The large-scale convergence near the surface is therefore considered to be a key factor for the preconditioning in both the large-scale and the convective-scale processes.
As the mechanism to produce the surface convergence off the center of the active convection, frictional convergence is suggested by Hsu and Weng (2001). In the present study, the zonal component of the surface wind (Fig. 3b) increases (i.e., easterly to westerly) in the leading edge of the northward-propagating low TB(IR) region where the horizontal convergence was observed by QuikSCAT (Fig. 3c) in the comparable order to that obtained by the budget analyses (Fig. 9b). The increase of the zonal component of the surface wind during the preactive period is also shown in Fig. 12c. These observed features may reflect the coexistence of a cyclonic circulation and convergence zone ahead of the deep convection of BSISV, in which the frictional convergence may occur.
The length of the preconditioning period is also of interest in terms of regulating the time scale of BSISV. Previous studies using cloud-resolving models (Tompkins 2001; Waite and Khouider 2010) demonstrated that the preconditioning requires only a few days to attain deep organized convection. This period is shorter than those observed in the present or previous studies (e.g., Jiang et al. 2011; i.e., several to 10 days), although the simulations did not include a large-scale forcing as observed in the present study. Waite and Khouider (2010) suggested that large-scale forcing (i.e., subsidence) could enable the preconditioning for a longer period. Agudelo et al. (2006) also concluded that large-scale subsidence is an important factor that causes moistening of the lower troposphere by inhibiting the moisture-consuming deep convection. In the present study, large-scale vertical motion is estimated from the large quadrilateral array (with R/V Mirai, Koror, Woleai, and Guam as vertices; see Fig. 1) as shown in Fig. 13. This figure indicates that the subsidence is found in most of the troposphere before 14 June. This corresponds to the period during which both CAPE and moisture in the boundary layer increased (Figs. 7a and 11). The magnitude of the estimated subsidence found in the present study is comparable to that found by Agudelo et al. (2006). Considering that the moistening of the boundary layer is maintained by large-scale upward advection, this subsidence would tend to confine the moist layer in the lower troposphere. These findings suggest that the large-scale subsidence also prolonged the period of the preconditioning in the present study.
Time–height cross section of the vertical motion (in pressure coordinate) estimated from the budget analyses and averaged for the area of the quadrilateral sounding array with the vertices formed by R/V Mirai, Koror, Woleai, and Guam (see Fig. 1 for locations).
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
6. Conclusions
This study investigated the temporal evolution of the atmospheric profiles for a case of northward-propagating intraseasonal variation (ISV) over the tropical western Pacific in the boreal early summer. The data were obtained during the field experiment PALAU2008, which deployed a sounding network together with a few Doppler radars for approximately one month. During the 22-day period with enhanced sounding [the arrayed sounding period (ASP)] at three points [R/V Mirai (at 12°N, 135°E), Koror, and Yap], the northward-propagating ISV [the boreal summer intraseasonal variation (BSISV)] was successfully captured.
The cloudy period of the observed BSISV appeared during the second half of ASP. Therefore, BSISV can be divided into two periods: that preceding the convectively active period (“preactive period”) in the first half and the convectively active period (“active period”) in the second half.
The active period was characterized by positive MSE anomalies in the middle and upper troposphere and negative MSE anomalies in the lower troposphere. Both the apparent heat source and the moisture sink (Q1 and Q2; as described in Yanai et al. 1973), derived from the networked upper-air soundings, are top heavy, with a peak at approximately 400–500 hPa. This height corresponds well to the peak of the frequency distribution of the radar-derived echo-top height. These characteristics represent the period with the deep convection.
The observed features in the preactive period contrast with those of the active period. The MSE had positive anomalies in the lower troposphere, and the positive layer was gradually deepened toward the active phase. At the same time, the positive Q2 layer capped by the negative Q2 layer also deepened. The period-averaged Q2 was positive in the lowest kilometer but negative above, with the negative peak at approximately 500 hPa in height. The Q1 profile was bottom heavy, with the positive layer below 400 hPa. These observed features indicate that the convective-scale transport by shallow convection contributed to the gradual moistening of the lower and middle troposphere. This concept is supported by the frequency distribution of the radar echo top at R/V Mirai, in which the peak appeared at approximately 600–700 hPa. The large-scale updraft also transported the moisture from the boundary layer to the lower and middle troposphere. In summary, the moistening by both the convective and large-scale processes served as the “preconditioning” for the deep convection in the following active period.
To investigate the processes that affect the preconditioning, the other observed parameters are inspected. The observed increase in CAPE during the preactive period is consistent with the results of previous studies using numerical models, in which a rise of CAPE was essential for the preconditioning. The present data also indicate that the observed increase in CAPE was caused largely by the increase of moisture in the boundary layer. In contrast, the latent heat flux decreased during this period. The budget analyses indicated that large-scale vertical transport is the only source of moistening for the boundary layer. The in situ and satellite data suggest that frictional convergence is the mechanism that promotes large-scale updraft, as suggested by Hsu and Weng (2001). The large-scale subsidence is also suggested to prolong the preactive phase.
Our results represent the detailed evolution of atmospheric profiles and the roles of convection in BSISV. Among the notable observations are the gradual moistening of the lower troposphere along with shallow convection during the preactive period and the preconditioning for the convectively active period. Similar process were observed for the Madden–Julian oscillation (e.g., Kemball-Cook and Weare 2001; Kikuchi and Takayabu 2004). This similarity implies the commonality of the mechanisms in BSISV and the MJO.
Our results demonstrated the effectiveness of in situ observations using a multiplatform network to investigate BSISV. However, these results are for only one case. We do not yet have comprehensive understanding of the mechanism of BSISV. One of the major unresolved questions is the reason for its very slow northward propagation speed. Previous studies suggested that air–sea interaction is the key factor that promotes the very slow propagation speed (e.g., Sengupta et al. 2001; Kemball-Cook and Wang 2001; Roxy and Tanimoto 2011). However, the present study did not provide clear evidence for a role of air–sea interaction. The effects of higher-frequency variations (e.g., a 3–4-day cycle; as implied by Figs. 5a and 7) or typhoons (such as Fengshen in the present case), should be investigated further to clarify the mechanism of the BSISV. Investigation of the details of convection at the interface between ITCZ and the subtropical region (e.g., Katsumata and Yoneyama 2004) is also important for understanding the details of the transition from preactive to active period. The effects of the landmass around the basin (e.g., Moteki et al. 2008) should also be investigated.
To investigate these unresolved factors in BSISV, future studies will use in situ observations, particularly enhanced sounding to enable more precise budget analyses, radar networks to capture convections, and oceanic observations to help elucidate the air–sea interactions.
Acknowledgments
The authors acknowledge Drs. T. Nasuno, P. E. Ciesielski, R. H. Johnson, and W. H. Schubert for valuable suggestions. The comments from three anonymous reviewers were invaluable to improve the manuscript. All of the participants and supporters of the PALAU2008 observations are acknowledged for their efforts to obtain and provide the dataset. GPS precipitable water data were kindly provided by Dr. M. Fujita. QuikSCAT and SSM/I datasets were provided by Remote Sensing Systems Co. (http://www.remss.com). Global-IR data were generated by NOAA and provided by NASA. MTSAT data on the split-window channels were provided by Kochi University. TRMM data were provided by NASA and JAXA. GSMaP data were processed and provided by JAXA.
APPENDIX A
Diurnal Dry Bias of the Precipitable Water Measured by RS80 Radiosonde at Koror
Previous studies reported that the Vaisala RS80-type radiosonde has a dry bias (e.g., Ciesielski et al. 2003). To confirm and to reduce the effect of the dry bias in the observed data used in the present study, we utilized the GPS-measured precipitable water (GPS-PW) as a true value and compared it with the radiosonde-derived precipitable water (RS-PW). The GPS sensor was installed in the vicinity of the Koror upper-air station (approximately 500 m in distance). The GPS-PW was obtained by the method of Fujita et al. (2008).
The results of comparisons between RS-PW and GPS-PW are shown in Fig. A1a. The two snapshots shown by a broken circle were excluded from the analyses because the soundings were suspected to penetrate the clouds where the PW was high locally, while GPS measured the averaged value within a volume of approximately several tens of kilometers in diameter.
(a),(b) Scatterplot of precipitable water measured by radiosonde (ordinate) and GPS at the same time as radiosonde (abscissa). The radiosonde-derived plot is corrected in (b) using the method of Cady-Pereira et al. (2008) and is uncorrected in (a). The dot and cross indicate the daytime (0000 and 0600 UTC) and nighttime (1200 and 1800 UTC) data, respectively. See the text regarding the circled points.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
The results indicate an average dry bias of 1.33 mm on the RS-PW. When we grouped the snapshots into daytime (0000 and 0600 UTC; i.e., namely, 0900 and 1500 LST) and nighttime (1200 and 1800 UTC; i.e., namely, 2100 and 0300 LST), the dry bias was large in daytime (2.46 mm) and very small in nighttime (0.15 mm). This diurnal cycle of the dry bias is consistent with results in previous studies (e.g., Turner et al. 2003). To correct this dry bias, the variation of sunshine should be considered.
Based on the above results, we chose to apply the correction of Cady-Pereira et al. (2008), which utilized the solar zenith angle as the parameter to determine how much humidity should be corrected. The result is shown in Fig. A1b. This figure indicates that the PW in the daytime was corrected properly. The statistical parameters before and after the correction are shown in Table A1. The bias error and RMS error in the daytime become comparable to those in the nighttime after applying the correction.
The statistical parameters derived by comparing precipitable waters derived from radiosonde with those derived from GPS at Koror in June 2008. All values are expressed in millimeters.
These results indicate that the method of Cady-Pereira et al. (2008) works properly to correct the diurnal bias of the precipitable water measured by the RS80-type radiosonde. In the present study, the corrected dataset in this method was utilized for both Koror and Yap, where the RS80-type radiosonde was used.
APPENDIX B
Evaluation of the Budget-Derived Rainfall
To evaluate the reliability of the budget analyses, we compare the estimated rainfall from the budget analyses to that from other individual sources. The dataset and method are similar to those described by Katsumata et al. (2011). The estimated rainfall from the budget analyses is calculated as in Yanai et al. (1973) as 
We adopted four rainfall datasets from individual datasets. Three are satellite estimates and the other is from the surface rain gauges. The satellite estimates are the TRMM 3B42 product (Huffman et al. 2007), GSMaP near-real-time (NRT) product (Kubota et al. 2007), and SSM/I (Wentz and Spencer 1998). SSM/I utilizes only two satellite-borne passive microwave sensors, while the other two utilize all available passive microwave sensors, combined with the infrared data. TRMM 3B42 also adjusts the monthly rainfall to that from the rain gauge data. The surface rain gauge dataset is produced by averaging the data from three rain gauges at the vertices of the sounding triangle: R/V Mirai, Koror, and Yap.
The time series of the estimated rainfalls are shown in Fig. B1. The temporal variations generally correspond well with each other. Both the budget-estimate rainfall and observed rainfall captured the near-zero period and the following small peak (around 14 June) in the preactive phase and two peaks (around 21 and 26 June) in the active phase. In contrast, the budget analyses indicated near-zero rainfall around 17 June while the other indicated a large amount of rainfall.
Time series of the observed and estimated rainfall during PALAU2008 ASP. A two-day running mean is applied to the data.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
These relationships are summarized more clearly by the temporally averaged rainfall for each period shown in Table B1. The differences between budget analyses and the other sources are within 15% except for the rain gauge values. The differences are the same as in the active period.
Observed and estimated rainfall amounts averaged for various periods in PALAU2008 ASP.
The largest differences are found for 16 and 17 June (“TDdays” in Table B1 and hereafter). During these days, the Tropical Depression Fengshen was located near the array (see Fig. 4). As in the studies by Ciesielski et al. (1999) and Katsumata et al. (2011), the discrepancy in the present case seems to result from falsely sampled rotational wind by the triangle sounding array. Considering that the developing TD typically produced large amounts of rainfall nearby, it is reasonable to assume that the budget analyses underestimated the rainfall on TDdays.
The other large discrepancy is in the rain gauge data. The large differences from the other estimates are found in TDdays (16–17 June) and in the preactive period (6–15 June). In the preactive period, the large difference between the data from rain gauges and budget analyses is found around 11 June (Fig. 5b). For these two periods (11 June and TDdays), most of the rainfall was measured at Koror (at the southern tip of the array). This finding suggests that the large rainfall amount measured by the rain gauges resulted from the local heavy rain near the tip of the array, which was not well captured by the budget analyses and the satellite estimates.
The comparison between the results from the budget analyses and other satellite estimates indicates the possibility of overestimation by the budget analyses. However, it is interesting that the satellite estimates with more sources resulted in values that were closer to those of the budget analyses. This matter requires further investigation.
In summary, it is reasonable to assume that the results of the budget analyses are acceptable for purposes of the analyses in the present study, except when the tropical depression was located over the array.
APPENDIX C
Correction of Radar Reflectivity
First we apply only the corrections in section 3 of K08 (i.e., without adjusting to the reflectivity of TRMM/PR). The resulting histogram of the difference between TRMM/PR and the ground-based radars is shown in Fig. C1. This figure indicates that the distributions of the differences between the Aimeliik X-band radar and the Mirai C-band radar are comparable. Based on these findings, we decided to match the reflectivity from the ground-based radar to that from TRMM/PR only by adding a constant value to the all observed reflectivities. For the Aimeliik X-band radar, we add 0.5 dB to all of the values. For the Mirai C-band radar, no adjustment is required to match the reflectivity to that from TRMM/PR.
Histogram for frequency distribution of the reflectivity difference between TRMM/PR and the ground-based radars. The thick gray line and the thin black line represent the Aimeliik X-band radar without adjustment and with +0.5-dB adjustment, respectively. The dashed black line represents the Mirai C-band radar without adjustment. The averaged values for a 5.0-dB bin width are plotted.
Citation: Monthly Weather Review 141, 2; 10.1175/MWR-D-12-00011.1
REFERENCES
Agudelo, P. A., J. A. Curry, C. D. Hoyos, and P. J. Webster, 2006: Transition between suppressed and active phases of intraseasonal oscillations in the Indo-Pacific warm pool. J. Climate, 19, 5519–5530.
Cady-Pereira, K. E., M. W. Shephard, D. D. Turner, E. J. Mlawer, S. A. Clough, and T. J. Wagner, 2008: Improved daytime column-integrated precipitable water vapor from Vaisala radiosonde humidity sensors. J. Atmos. Oceanic Technol., 25, 873–883.
Ciesielski, P. E., W. H. Schubert, and R. H. Johnson, 1999: Large-scale heat and moisture budgets over the ASTEX region. J. Atmos. Sci., 56, 3241–3261.
Ciesielski, P. E., R. H. Johnson, P. T. Haertel, and J. Wang, 2003: Corrected TOGA COARE sounding humidity data: Impact on diagnosed properties of convection and climate over the warm pool. J. Climate, 16, 2370–2384.
Dickinson, M., and J. Molinari, 2002: Mixed Rossby-gravity waves and western Pacific tropical cyclogenesis. Part I: Synoptic evolution. J. Atmos. Sci., 59, 2183–2196.
Fu, X., B. Wang, and L. Tao, 2006: Satellite data reveal the 3-D moisture structure of Tropical Intraseasonal Oscillation and its coupling with underlying ocean. Geophys. Res. Lett., 33, L03705, doi:10.1029/2005GL025074.
Fujita, M., F. Kimura, K. Yoneyama, and M. Yoshizaki, 2008: Verification of precipitable water vapor estimated from shipborne GPS measurements. Geophys. Res. Lett., 35, L13803, doi:10.1029/2008GL033764.
Geng, B., K. Yoneyama, R. Shirooka, and M. Yoshizaki, 2011: Characteristics of precipitation systems and their environment during the onset of the western North Pacific summer monsoon in 2008. J. Meteor. Soc. Japan, 89A, 1–25.
Hartmann, D. L., M. L. Michelsen, and S. A. Klein, 1992: Seasonal variations of tropical intraseasonal oscillations: A 20–25 day oscillation in the western Pacific. J. Atmos. Sci., 49, 1277–1289.
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.
Huffman, G., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 38–55.
Inoue, T., 1987: A cloud type classification with NOAA 7 split-window measurements. J. Geophys. Res., 92 (D4), 3991–4000.
Janowiak, J. E., R. J. Joyce, and Y. Yarosh, 2001: A real–time global half–hourly pixel–resolution infrared dataset and its applications. Bull. Amer. Meteor. Soc., 82, 205–217.
Jiang, X., T. Li, and B. Wang, 2004: Structures and mechanisms of the northward propagating boreal summer intraseasonal oscillation. J. Climate, 17, 1022–1039.
Jiang, X., D. E. Waliser, J.-L. Li, and C. Woods, 2011: Vertical cloud structures of the boreal summer intraseasonal variability based on CloudSat observations and ERA-interim reanalysis. Climate Dyn., 36, 2219–2232.
Johnson, R. H., T. M. Rickenbach, S. A. Rutledge, P. E. Ciesielski, and W. H. Schubert, 1999: Trimodal characteristics of tropical convection. J. Climate, 12, 2397–2418.
Joss, J., and A. Waldvogel, 1967: A raindrop spectrograph with automatic analysis. Pure Appl. Geophys., 68, 240–246.
Katsumata, M., and K. Yoneyama, 2004: Internal structure of ITCZ mesoscale convective systems and related environmental factors in the tropical western Pacific: An observational case study. J. Meteor. Soc. Japan, 82, 1035–1056.
Katsumata, M., T. Ushiyama, K. Yoneyama, and Y. Fujiyoshi, 2008: Combined use of TRMM/PR and disdrometer data to correct reflectivity of ground-based radars. SOLA, 4, 101–104, doi:10.2151/sola.2008-026.
Katsumata, M., R. H. Johnson, and P. E. Ciesielski, 2009: Observed synoptic-scale variability during the developing phase of an ISO over the Indian Ocean during MISMO. J. Atmos. Sci., 66, 3434–3448.
Katsumata, M., P. E. Ciesielski, and R. H. Johnson, 2011: Evaluation of budget analyses during MISMO. J. Appl. Meteor. Climatol., 50, 241–254.
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., and B. C. Weare, 2001: The onset of convection in the Madden–Julian oscillation. J. Climate, 14, 780–793.
Kikuchi, K., and Y. N. Takayabu, 2004: The development of organized convection associated with the MJO during TOGA COARE IOP: Trimodal characteristics. Geophys. Res. Lett., 31, L10101, doi:10.1029/2004GL019601.
Kubota, H., and B. Wang, 2009: How much do tropical cyclones affect seasonal and interannual rainfall variability over the western North Pacific? J. Climate, 22, 5495–5510.
Kubota, H., R. Shirooka, T. Ushiyama, T. Chuda, S. Iwasaki, and K. Takeuchi, 2005: Seasonal variations of precipitation properties associated with the monsoon over Palau in the western Pacific. J. Hydrometeor., 6, 518–531.
Kubota, T., and Coauthors, 2007: Global precipitation map using satellite-borne microwave radiometers by the GSMaP project: Production and validation. IEEE Trans. Geosci. Remote Sens., 45, 2259–2275.
Lau, K.-M., and P. H. Chen, 1986: Aspects of the 40–50 day oscillation during the northern summer as inferred from outgoing longwave radiation. Mon. Wea. Rev., 114, 1354–1367.
LinHo, and B. Wang, 2002: The time–space structure of the Asian–Pacific summer monsoon: A fast annual cycle view. J. Climate, 15, 2001–2019.
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.
Mapes, B. E., P. E. Ciesielski, and R. H. Johnson, 2003: Sampling errors in rawinsonde-array budgets. J. Atmos. Sci., 60, 2697–2714.
Masunaga, H., T. S. L’Ecuyer, and C. D. Kummerow, 2006: The Madden–Julian Oscillation recorded in early observations from the Tropical Rainfall Measurement Mission (TRMM). J. Atmos. Sci., 63, 2777–2794.
McBride, J. L., and W. M. Frank, 1999: Relationships between stability and monsoon convection. J. Atmos. Sci., 56, 24–36.
Moteki, Q., and Coauthors, 2008: Mechanism of the northward propagation of mesoscale convective systems observed on 15 June 2005 during PALAU2005. J. Geophys. Res., 113, D14126, doi:10.1029/2008JD009793.
Murakami, T., and J. Matsumoto, 1994: Summer monsoon over the Asian continent and western North Pacific. J. Meteor. Soc. Japan, 72, 719–745.
Nakazawa, T., 1986: Mean features of 30–60 day variations as inferred from 8-year OLR data. J. Meteor. Soc. Japan, 64, 777–786.
Nakazawa, T., 1992: Seasonal phase lock of intraseasonal variation during the Asian summer monsoon. J. Meteor. Soc. Japan, 70, 597–611.
Nitta, T., 1987: Convective activities in the tropical western Pacific and their impacts on the Northern Hemisphere summer circulation. J. Meteor. Soc. Japan, 65, 165–171.
Nuss, W. A., and D. W. Titley, 1994: Use of multiquadric interpolation for meteorological objective analysis. Mon. Wea. Rev., 122, 1611–1631.
O’Brien, J. J., 1970: Alternative solutions to the classical vertical velocity problem. J. Appl. Meteor., 9, 198–203.
Roxy, M., and Y. Tanimoto, 2011: Influence of sea surface temperature on the intraseasonal variability of the South China Sea summer monsoon. Climate Dyn., 37, 1209–1218, doi:10.1007/s00382-011-1118-x.
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.
Takahashi, S., F. Kondo, O. Tsukamoto, Y. Ito, S. Hirayama and H. Ishida, 2005: On-board automated eddy flux measurement system over open ocean. SOLA,1, 37–40, doi:10.2151/sola.2005-011.
Tompkins, A. M., 2001: Organization of tropical convection in low vertical wind shears: The role of water vapor. J. Atmos. Sci.,58, 529–545.
Turner, D. D., B. M. Lesht, S. A. Clough, J. C. Liljegren, H. E. Revercomb, and D. C. Tobin, 2003: Dry bias and variability in Vaisala RS80-H radiosondes: The ARM experience. J. Atmos. Oceanic Technol., 20, 117–132.
Waite, M. L., and B. Khouider, 2010: The deepening of tropical convection by congestus preconditioning. J. Atmos. Sci., 67, 2601–2615.
Wang, B., 2005: Theory. Intraseasonal Variability in the Atmosphere-Ocean Climate System, W. K.-M. Lau and D. E. Waliser, Eds., Springer, 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 LinHo, 2002: Rainy season of the Asian–Pacific summer monsoon. J. Climate, 15, 386–398.
Webster, P. J., and R. Lukas, 1992: TOGA COARE: The Coupled Ocean–Atmosphere Response Experiment. Bull. Amer. Meteor. Soc., 73, 1377–1416.
Webster, P. J., and Coauthors, 2002: The JASMINE pilot study. Bull. Amer. Meteor. Soc., 83, 1603–1630.
Wentz, F. J., and R. W. Spencer, 1998: SSM/I rain retrievals within a unified all-weather ocean algorithm. J. Atmos. Sci., 55, 1613–1627.
Yanai, M., S. Esbensen, and J.-H. Chu, 1973: Determination of bulk properties of tropical cloud clusters from large-scale heat and moisture budgets. J. Atmos. Sci., 30, 611–627.
Yasunari, T., 1979: Cloudiness fluctuations associated with the Northern Hemisphere summer monsoon. J. Meteor. Soc. Japan, 57, 227–242.
Yoneyama, K., 2003: Moisture variability over the tropical western Pacific Ocean. J. Meteor. Soc. Japan, 81, 317–337.
Yoneyama, K., M. Fujita, N. Sato, M. Fujiwara, Y. Inai, and F. Hasebe, 2008: Correction for radiation dry bias found in RS92 radiosonde data during the MISMO field experiment. SOLA, 4, 13–16.
Yu, L., X. Jin, and R. A. Weller, 2008: Multidecade global flux datasets from the objectively analyzed air-sea fluxes (OAFlux) project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. OAFlux Project Tech. Rep. OA-2008-01, Woods Hole Oceanographic Institution, 64 pp.
Zhang, C., 2005: Madden-Julian Oscillation. Rev. Geophys., 43, RG2003, doi:10.1029/2004RG000158.
Zhang, C., and S. Hagos, 2009: Bi-modal structure and variability of large-scale diabatic heating in the tropics. J. Atmos. Sci., 66, 3621–3640.
PALAU intensive observation in 2008.
